Anesthesia & Analgesia Vol 109 Issue 01 2009

Editorial OpenAnesthesia.org: Graduate Medical Education on the World Wide Web “Education is not the filling of a pail,...

Author: International Anesthesia Research Society

38 downloads 1593 Views 7MB Size Report

This content was uploaded by our users and we assume good faith they have the permission to share this book. If you own the copyright to this book and it is wrongfully on our website, we offer a simple DMCA procedure to remove your content from our site. Start by pressing the button below!
Report copyright / DMCA form

DOWNLOAD PDF

Editorial

OpenAnesthesia.org: Graduate Medical Education on the World Wide Web “Education is not the filling of a pail, but the lighting of a fire.”

Edward C. Nemergut, MD

T

—William Butler Yeats

he International Anesthesia Research Society (IARS) was founded in 1922 “to foster progress and research in all phases of anesthesia.” Implicit in this statement is the understanding that the value of a medical society rests not in the number of printed journals that make their way around the globe each month, but in its ability to enhance the progress of medicine. Keeping with that broad mission, the IARS, in combination with Anesthesia & Analgesia, will be the first major medical journal to sponsor an online multimodal toolkit specifically designed to advance graduate medical education (GME) in anesthesia: OpenAnesthesia.org. It is the hope of the editorial board that OpenAnesthesia.org will not only enhance the traditional paradigm of scientific journalism, but also revolutionize GME itself. OpenAnesthesia.org creates a unique method for resident physicians to discover and appreciate the primary medical literature. It also allows residents to cooperate and collaborate while providing Program Directors with a tool to document core competency activities for Accreditation Council for Graduate Medical Education (ACGME)-mandated learning portfolios. OpenAnesthesia.org can roughly be divided into 2 overlapping components, a Journal Section and a Wiki section.

JOURNAL SECTION

From the Department of Anesthesiology, University of Virginia Health System, Charlottsville, Virginia. Accepted for publication May 4, 2009. Reprints will not be available from the author. Address correspondence to Edward C. Nemergut, MD, Associate Professor of Anesthesiology and Neurosurgery, University of Virginia Health System, P.O. Box 800710, Charlottesville, VA 22908-0710. Address e-mail to [email protected]. Copyright © 2009 International Anesthesia Research Society DOI: 10.1213/ane.0b013e3181ace68d

Vol. 109, No. 1, July 2009

The Journal Section is intended to help resident physicians learn from the primary literature, with a focus on Anesthesia & Analgesia. The centerpiece of the Journal Section is the “Featured Article.” Each month, the editorial board will select an article from the current Anesthesia & Analgesia issue. Anesthesia & Analgesia is a comprehensive journal, encompassing anesthesia and analgesia (of course) as well as critical care, pain medicine, perioperative medicine, experimental neuroscience, medical economics, and occasional forays into the philosophy of science and medical ethics. Selecting high-quality articles from this potpourri of diverse papers will be a simple task. Residents will be invited to read the featured article and listen to an interview with one of the article’s authors. During the interview, the author will discuss the specifics of the article as well as general topics geared towards improving each resident’s appreciation of basic or clinical research. For example, the interview might include a discussion of why a given statistical test was employed by the authors, or why particular subjects were chosen or excluded from the study. Each month’s interview will be available as a podcast to which anyone may subscribe and listen. After listening to the podcast and reading the article, residents will be able to login to the Anesthesia & Analgesia GME page using their IARS/ Anesthesia & Analgesia password and answer 5 questions in order to demonstrate their mastery of the topics discussed (similar to the Anesthesia & Analgesia Continuing Medical Education (CME) section). Like the CME section, after demonstrating proficiency, a resident will receive a printable certificate that will specify which ACGME core competencies were addressed 1

in the article and interview. The certificates can be put in each resident’s ACGME-required learning portfolio. In addition, the journal section of OpenAnesthesia. org will feature an “Ask the Experts” podcast interview each month. During this interview, an “expert” on a specific topic (typically a member of the Anesthesia & Analgesia editorial board) will respond to presubmitted resident questions. These “Ask the Experts” podcasts will make international expertise on fundamental or timely topics directly available to IARS members every month.

ANESTHESIA & ANALGESIA WIKI SECTION The second component of OpenAnesthesia.org is a “wiki” that aims to cover all aspects of anesthesiology, critical care, pain medicine, and perioperative medicine. For the uninitiated, a wiki is a website designed to enable anyone who can access it to contribute or modify its content. The Anesthesia & Analgesia wiki facilitates worldwide interaction and thoughtful collaboration among residents, faculty, scientists, and regulators. In general terms, the OpenAnesthesia.org wiki can be considered an encyclopedic anesthesia textbook, which is continually updated by its own readers. Information is instantaneously integrated and incorporated into the knowledge base, permitting physicians to extrapolate and bridge data and knowledge from various medical topics. Controversial topics are actively discussed, and typically thoughtful threads are pursued while baseless speculation is discouraged by the collective wisdom. Most importantly, the OpenAnesthesia.org wiki allows anesthesia residents to take an active role in their own education by writing new content and actively sharing their knowledge with other physicians. The process compels residents to take charge of their own education, and by doing so, improve the education of others. This reminds

2

Editorial

residents that they are one of a community of physicians and that a community needs the citizenship of all its members to prosper. Can multiple unpaid authors and editors really create something that is both accurate and useful? The answer is an unequivocal “yes.” The skeptic reader is referred to Nature for an objective comparison of Wikipedia (the world’s most famous wiki, on which OpenAnesthesia.org was modeled) and Encyclopedia Britannica.1 The cost of these services is modest: an IARS membership. For United States residents that is $40/year, which roughly defrays the cost of mailing the printed journal. For non-residents, IARS membership increases to $140/year. IARS membership, of course, also includes the printed Anesthesia & Analgesia, on-line Anesthesia & Analgesia, the on-line (and excellent!) Anesthesia & Analgesia CME program offering 24 hours of category 1 credit every year. It’s not free, but it is as good a bargain as physicians get anywhere. The goal of OpenAnesthesia.org is identical to that of Anesthesia & Analgesia itself, improving patient care. Although the focus will be on resident education, we expect all IARS members will participate, and benefit, from this new initiative. Indeed, medical students, basic scientists, and physicians from other disciplines are likely to benefit from this program as well. OpenAnesthesia.org is not “done.” Indeed, we don’t intend to ever finish it. That is how a “wiki” works. It is forever dynamic and changing, just like the medicine itself. But you can go there today and probably learn something. Or better yet, help someone else learn something. We hope you are sufficiently intrigued to spend some time exploring OpenAnesthesia.org. We hope, even more, that it lights your fire. REFERENCE 1. Giles J. Internet encyclopaedias go head to head. Nature 2005; 438:900 –1

ANESTHESIA & ANALGESIA

Editorial

PRO: Manometry Should Routinely Be Used During Central Venous Catheterization Andrew B. Leibowitz, MD* Marc A. Rozner, PhD, MD†

From the *Mount Sinai Medical Center, New York, New York; and †University of Texas MD Anderson Cancer Center, Houston, Texas. Accepted for publication August 14, 2008. Address correspondence and reprint requests to Marc A. Rozner, PhD, MD, University of Texas MD Anderson Cancer Center, 1400 Holcombe Blvd, Mail Code 0409, Houston, TX 77030. Address e-mail to [email protected]. Copyright © 2009 International Anesthesia Research Society DOI: 10.1213/ane.0b013e31818e4347

Vol. 109, No. 1, July 2009

I

n the beginning, the specialty of Anesthesiology emphasized analgesia, amnesia, hypnosis, and optimization of operating conditions, usually through muscle relaxation to prevent unwanted muscle tone or movement. During the last half-century, Anesthesiologists added attention to intraoperative hemodynamic stability, active minimization of perioperative morbidity and mortality, patient satisfaction, operating room throughput, and length of postoperative stay. Underlying all of these issues is every Anesthesiologist’s desire to avoid any complication specifically attributable to the administration of Anesthesia. Two important results stand out: 1) the risk of significant Anesthesia-related morbidity and mortality typically is orders of magnitude lower than that of the surgical procedure for which the Anesthesia is provided and 2) our Anesthesiology devices and medications have an enviable efficacy and safety profile. This is why our specialty is held as the model for overall patient safety in medicine.1 In fact, Anesthesiologists have created a myriad of technologies, practice guidelines, and alternative techniques to identify and mitigate a wide range of problems that occur exceedingly rarely. Malignant hyperthermia has an incidence of 1/50,000,2 but it has resulted in tens of thousands of dantrolene carts and tomes of hospital policy (with attendant maintenance costs) dedicated to the purpose of treating this very rare event. The combination of “can’t intubate and can’t ventilate” occurs, at most, in only about 1/1000 –1/5000 patients,3 but it has spawned multiple iterations of difficult airway algorithms,4,5 nearly 800 journal articles that include the term “difficult airway” since 1980 and an entire industry devoted to performing ventilation and tracheal intubation in just those few patients. Intraoperative awareness during general anesthesia might occur in 1/500 –1/1000 patients (this is the subject of intense argument and investigation).6 Nevertheless, it has resulted in another industry that manufactures and sells highly touted (and expensive) monitors that might, at best, be marginally effective.7,8 Many other rare events (e.g., pulmonary aspiration, postoperative visual loss, epidural hematoma, and local anesthetic toxicity) also have resulted in routine drug prophylaxis, standards, guidelines, practice parameters, advisories, etc. Let’s face it, Anesthesiologists are amazingly risk averse, and preventing rare events and complication is, in large part, what our specialty is all about. So, now we turn to the subject of central venous catheterization and the investigation reported herein by Ezaru et al.9 This study, entitled “Eliminating arterial injury during central venous catheterization using manometry” reports the experience of a universityaffiliated Veterans Administration hospital that implemented “mandatory utilization of manometry to verify venous placement” in response to a sentinel event of arterial cannulation. This retrospective study encompassed 16 years with two different data collection schemes. For the first 15 yr, 9348 central venous catheters were inserted without any arterial cannulation with a large bore catheter (7F or larger). During the final year of reported cases, the database was refined and revealed that in the 511 central venous catheters placed, arterial puncture (with 18-gauge or 3

smaller needle) occurred in 28 patients (5%) and was recognized without manometry in only 24 of these occasions. Arterial puncture, however, was identified through the use of manometry in these remaining four cases, so there was no incident of arterial dilation or cannulation with a large bore catheter. Criticisms of this study will include the nature of a retrospective database review, which is subject to all of the problems suffered by every retrospective study. Next, the study took place in a Veteran’s Administration hospital, wherein attending physicians performed only two-thirds of the central line placements; perhaps this accounts for the 5% rate of arterial puncture noted during the final year of the study, or that recognition of initial arterial puncture could result only after applying manometry. However, this 5% is in line with other reports of unintended arterial (real or simulated) puncture based upon routine use of anatomic landmarks,10 –14 even with ultrasound guidance.11,15 On the other hand, the retrospective nature of this study should not dissuade one from accepting its results, because every arterial dilation and cannulation would likely have become a high-profile event and, therefore, not escaped the attention of the institution or these investigators. We do not know of any data suggesting this problem is more or less likely in veterans or Veterans Administration hospitals, so this experience likely represents the patient population at large. Considering the number of teaching hospitals’ residents and fellows, as well as the widespread use of physician extenders in nonteaching facilities, perhaps more than one-third of all central line placements are not performed by attending physicians. Also, perhaps many attending physicians will have relatively less experience than the group reporting the data in this study, and, thus, will have a greater risk of complications when performing the procedures themselves. Unfortunately, the study did not identify the primary operator in all of the cases involving arterial puncture, but, again, this fact might be unimportant. Several articles have been published regarding the incidence of arterial puncture and cannulation which confirm these authors’ experiences. Unintentional arterial puncture with something larger than the traditional 20 –22-gauge “finder needle” occurs in 0.5%–11.4% (mean 5.9%) of all central line attempts.13 Rarely will this event result in patient harm. Vessel dilation and cannulation with a 7F or larger catheter has been reported in 0.1%–1% of all central line attempts.16 The results of this error can be devastating and include hemothorax, pseudoaneurysm, stroke, and death.17 In fact, in their review of 14 years of hospital data, Shah et al.13 report repairing 11 arterial injuries after unrecognized arterial cannulation with a large bore instrument (their total number of central lines for this period was not shown). In three of their cases, infusions were started before the arterial position of the cannula was identified, and all the three patients developed neurological symptoms. 4

Editorial

So, if a simple, quick, and inexpensive method of risk prevention such as manometry was successful in even a fraction of cases, it is quite hard to understand any objection to incorporating it into practice. To put this into perspective, using the lower limit of incidence, if a hospital performs just 2000 central catheterizations per year, and the incidence of this problem is 0.1% (two cases), and manometry is only half as effective as reported, then one major morbidity, or perhaps mortality, will be avoided per year in that facility. Relative to many other rare events, which cause concern to us as Anesthesiologists, the payback here is tremendous. Since time and cost seem negligible, what could be the objection(s) to routine manometry? First, this methodology might create a break in sterility. To perform tubing manometry with most current central line kits, an extra step of locating a sterile tubing set to create the manometer becomes necessary. Because some tubing kits are nonsterile packages with only sterile fluid pathways, this step might create some confusion with a resultant break in sterility. However, some kits already contain manometry tubing and all kits with Raulerson syringes have a transducing adaptor. If tubing manometry becomes a standard, then commercial reconfiguration will rapidly follow. Second, many practitioners will state that there is a potential difficulty in performing routine manometry, especially if one’s practice is to use the metal 18-gauge needle for guidewire access instead of an 18-gauge angiocatheter. This technique would then entail delicate attachment of the manometry tubing to the metal needle (M.A.R.’s technique) or insertion of the guidewire through the metal needle and exchange (over the wire) of the metal needle to an 18-gauge angiocatheter (A.B.L.’s method). We attest to the ease of incorporating this extra step into practice having collectively performed several hundred central venous catheterizations in our practices since adapting manometry as a standard. We agree with Ezaru et al. and argue that use of tubing manometry for all elective central vein catheterizations to ensure entry into a vein, rather than artery, before vessel dilation, will prevent patient injury. Surely, anyone who insists that their personal safety record obviates the need for manometry likely has the dexterity to add manometry to their practice with less difficulty than it takes to argue against it. REFERENCES 1. Hallinan JT. Once seen as risky, one group of doctors changes its ways. Wall Street J 2005;A1. Available at: http://onlinewsj. com/article/0,,SB111931728319164845,00.html. Accessed 14 July, 2008 2. Rosenberg H, Fletcher JE. An update on the malignant hyperthermia syndrome [review]. Ann Acad Med Singapore 1994;23(6 suppl):84 –97 3. Vasdev GM, Harrison BA, Keegan MT, Burkle CM. Management of the difficult and failed airway in obstetric anesthesia [review]. J Anesthesia 2008;22:38 – 48 4. Henderson JJ, Popat MT, Latto IP, Pearce AC, Difficult Airway Society. Difficult Airway Society guidelines for management of the unanticipated difficult intubation [see comment]. Anaesthesia 2004;59:675–94


5. American Society of Anesthesiologists Task Force on Management of the Difficult Airway. Practice guidelines for management of the difficult airway: an updated report by the American Society of Anesthesiologists Task Force on Management of the Difficult Airway [erratum appears in Anesthesiology 2004;101:565]. Anesthesiology 2003;98:1269 –77 6. American Society of Anesthesiologists Task Force on Intraoperative Awareness. Practice advisory for intraoperative awareness and brain function monitoring: a report by the American Society of Anesthesiologists Task Force on intraoperative awareness [review]. Anesthesiology 2006;104:847– 64 7. Avidan MS, Zhang L, Burnside BA, Finkel KJ, Searleman AC, Selvidge JA, Saager L, Turner MS, Rao S, Bottros M, Hantler C, Jacobsohn E, Evers AS. Anesthesia awareness and the bispectral Index. N Engl J Med 2008;358:1097–108 8. Myles PS, Leslie K, McNeil J, Forbes A, Chan MTV. Bispectral index monitoring to prevent awareness during anaesthesia: the B-Aware randomised controlled trial. The Lancet 2004; 363:1757– 63 9. Ezaru CS, Mangione MP, Oravitz TM, Ibinson JW, Bjerke RJ. Eliminating arterial injury during central venous catheterization using manometry. Anesth Analg 2009;109:130 – 4 10. Schaffartzik W, Neu J. Injuries in anaesthesia—results of the Hannover arbitration procedure 2001–2005. Anaesthesist 2007;56:444 – 8

Vol. 109, No. 1, July 2009

11. Lamperti M, Cortellazzi P, D’Onofrio G, Subert M, Falcone C, Filippini G, Caldiroli D. An outcome study on complications using routine ultrasound assistance for internal jugular vein cannulation. Acta Anaesthesiol Scand 2007;51:1327–30 12. Bailey PL, Whitaker EE, Palmer LS, Glance LG. The accuracy of the central landmark used for central venous catheterization of the internal jugular vein. Anesth Analg 2006;102:1327–32 13. Shah PM, Babu SC, Goyal A, Mateo RB, Madden RE. Arterial misplacement of large-caliber cannulas during jugular vein catheterization: case for surgical management. J Am Coll Surg 2004;198:939 – 44 14. Reuber M, Dunkley LA, Turton EP, Bell MD, Bamford JM. Stroke after internal jugular venous cannulation. Acta Neurol Scand 2002;105:235–9 15. Augoustides JG, Diaz D, Weiner J, Clarke C, Jobes DR. Current practice of internal jugular venous cannulation in a university anesthesia department: influence of operator experience on success of cannulation and arterial injury. J Cardiothorac Vasc Anesth 2002;16:567–71 16. Kusminsky RE. Complications of central venous catheterization. J Am Coll Surg 2007;204:681–96 17. Domino KB, Bowdle TA, Posner KL, Spitellie PH, Lee LA, Cheney FW. Injuries and liability related to central vascular catheters: a closed claims analysis [review]. Anesthesiology 2004;100:1411–18

© 2009 International Anesthesia Research Society

5

Editorial

CON: Manometry During Internal Jugular Cannulation: Case Not Proven Bruce J. Leone, MD

From the Department of Anesthesiology, Mayo Clinic, Jacksonville, Florida. Accepted for publication August 27, 2008. Address correspondence and reprint requests to Bruce J. Leone, MD, Department of Anesthesiology, Mayo Clinic, Jacksonville, FL 32224. Address e-mail to leone. bruce@ mayo.edu. Copyright © 2009 International Anesthesia Research Society DOI: 10.1213/ane.0b013e31818e462c

6

I

n the current issue of Anesthesia & Analgesia, Ezaru et al.1 have presented data collected over a 15-year period regarding internal jugular vein cannulation and inadvertent carotid artery cannulation. After a single sentinel event, the authors’ group instituted a mandatory policy of examining the blood return in a piece of extension tubing to determine whether the cannulation needle was located in the carotid artery before placement of a large bore indwelling cannula. The authors report no adverse events in 15 years, comprising 9859 internal jugular vein cannulations, of inadvertent arterial cannulation and arterial injury. The authors have thus suggested that all internal jugular vein cannulae be placed after this simplistic method of manometry to eliminate arterial cannulation and injury. The literature is replete with unfortunate case reports of carotid arterial puncture and cannulation resulting in injury and morbidity. However, it is difficult to gain insight into the relative risk of mishap given the vast number of internal jugular vein cannulations and the unreported percentage of mishaps. Ezaru et al. report an arterial puncture (without cannulation) rate of 5% in 511 internal jugular vein cannulation attempts; a similar arterial puncture rate has been reported,2 as well as an almost identical incidence reported earlier by Jobes et al.3 Indeed, as Ezaru et al.1 emphasize, the incidence of arterial puncture is significantly higher than the incidence of carotid cannulation. However, can one make the case for adoption of simple manometry (without pressure transduction) as de rigueur practice before placement of an internal jugular large-bore cannula? The data are compelling from the aspect of avoidance of a potentially lethal complication, but is the risk-benefit ratio sufficient to change practice? The authors1 speculate that the absence of carotid injury mishaps in their 15 years of required manometry would suggest adoption of manometry with a high benefit, yet the incidence quoted of inadvertent arterial cannula placement is vanishingly small, varying from 0.0995% to 0.775%.4 Thus, assuming Ezaru et al. have an extremely low incidence of arterial cannula placement, one would expect to see 1 possible arterial cannulation in 10,000 patients. The fact that no arterial cannulation was seen could be the natural variability associated with random chance (or rare events), rather than an elimination of the event due to changes in practice. And what of the potential risks in adding an extra task to the already intricate performance of internal jugular cannulation? In manipulating the 18-gauge cannula to affix the extension tubing and then aspirating or manipulating the cannula tubing to obtain a sufficient column of blood, one could envision many other mishaps: air embolization, dislodgement of the cannula, infection and violation of the sterile field are very real possibilities not discussed in Ezaru et al.’s study. While these may be viewed as negligible risks or unlikely, an incidence of more than 1 in 10,000 cases would obviate some of the benefit gained by performing manometry. The outcomes may not be significant initially (e.g., blood loss), but may become severe and/or debilitating (e.g., infection or air embolization). Even if one considers the higher rate of reported incidence (0.775%) of inadvertent carotid cannulation, an incidence of manometry-induced Vol. 109, No. 1, July 2009

mishaps of once a year would be more than double the incidence of carotid cannulation injury with standard techniques. Ezaru et al. regrettably did not have data on the incidence of carotid cannulation before the inception of mandatory manometry. Data quoted above4 suggest that this singular catastrophic incident, which occasioned mandatory manometry, may have been the only incident at the authors’ institution in the preceding 15 years (assuming approximately 10,000 cannulations per 15 years). Thus, the results presented in their article concerning manometry may be statistically unrelated to the application of routine manometry. Although the prospective data of 4 in 511 unrecognized arterial punctures suggests that cannulae may have been placed without the performance of manometry, this incidence (8%) is extremely high for inadvertent large-bore carotid cannulation. The manometry Ezaru et al. describe is simple and fosters some sense of security, but the data presented do not make the case for routine application of this technique in practice. If one accepts the risks associated with the traditional methods of internal jugular cannulation, this maximum reported risk (7 in 10,000) is less than the mortality risk associated with anesthesia in ASA physical status 3 patients (8 in 10,877) or ASA physical status 4 patients (34 in 2939),5 groups in whom internal jugular cannulation is most likely to be performed. While it is unknown what role invasive monitoring mishaps played in the overall mortality figures presented by Lagasse,5 it is clear that internal jugular cannulation without mandatory manometry is as safe as the administration of anesthesia. Acceptance of status quo limits quality improvements in our specialty. Therefore, we should not shrug and accept that “What if this is as good as it gets?” (with apologies to Jack Nicholson) but rather look critically at our performance of internal jugular cannulation and improve our technique. Ezaru et al. mention en passant the use of ultrasound in assisting

Vol. 109, No. 1, July 2009

internal jugular cannulation but discount this as not user friendly when compared to simple manometry. Perhaps our improvements will come with improved availability of and facility with of ultrasound, which can readily identify structures and detect successful cannulation of the internal jugular vein. Manometry can only detect arterial placement of the cannula, whereas ultrasound can identify the internal jugular vein before placement of the finder needle as well as arterial placement of an 18-gauge catheter. Use of ultrasound in highly dangerous situations with the potential for serious complications, such as the liver failure patient or the fully anticoagulated patient arriving from unsuccessful coronary artery interventions, may be the entreé needed to propel ultrasound to the forefront of technological requirement for internal jugular cannulation. Caution is warranted here, however, as carotid puncture despite the use of ultrasound to detect apparent successful internal jugular puncture continue to be reported; however, like Ezaru et al., these authors also report zero arterial cannulations.6 Manometry, while providing a comforting affirmation before large-bore indwelling cannula placement, is not the answer. REFERENCES 1. Ezaru CS, Mangione MP, Oravitz TM, Ibinson JW, Bjerke RJ. Eliminating arterial injury during central venous catheterization using manometry. Anesth Analg 2009;109:130 – 4 2. McGee DC, Gould MK. Preventing complications of central venous catheterization. N Engl J Med 2003;348:1123–33 3. Jobes DR, Schwartz AJ, Greenhow DE, Stephenson LW, Ellison N. Safer jugular vein cannulation: recognition of arterial puncture and preferential use of the external jugular route. Anesthesiology 1983;59:353–5 4. Shah PM, Babu SC, Goyal A, Mateo RB, Madden RE. Arterial misplacement of large-caliber cannulas during jugular vein catheterization: case for surgical management. J Am Coll Surg 2004;198:939 – 44 5. Lagasse RS. Anesthesia safety: model or myth? A review of the published literature and analysis of current original data. Anesthesiology 2002;97:1609 –17 6. Augoustides JG, Jobes DR, Diaz D, Weiner J. Safe internal jugular vein cannulation. J Cardiothorac Vasc Anesth 2002;16:262–3


7

Editorial

Toward Improving the Safety of Transforaminal Injection James P. Rathmell, MD*†

From the *Division of Pain Medicine, Department of Anesthesia and Critical Care, and Pain Medicine, Massachusetts General Hospital; and †Department of Anesthesia, Harvard Medical School, Boston, Massachusetts. Accepted for publication March 9, 2009. Address correspondence and reprint requests to James P. Rathmell, MD, Department of Anesthesia and Critical Care, Massachusetts General Hospital, 55 Fruit St., GRB 444, Boston, MA 02114. Address e-mail to [email protected]. Copyright © 2009 International Anesthesia Research Society DOI: 10.1213/ane.0b013e3181a81ee1

8

T

ransforaminal injection of steroids has emerged as a common technique for the treatment of acute radicular pain associated with intervertebral disk herniation.1 In contrast to placing steroid in the epidural space using a standard, midline approach between adjacent laminae (the “interlaminar” approach), the transforaminal technique uses a needle that approaches the posterolateral aspect of the intervertebral foramen and places the steroid directly adjacent to the affected spinal nerve at the site of inflammation. The concept is to inject the steroid solution directly at the target site in high concentration. There are limited data suggesting that the efficacy of steroid placed via a transforaminal approach is superior to that of the same steroid placed via an interlaminar approach.2,3 In recent years, catastrophic neurologic complications have been reported after transforaminal injection of steroids. After cervical transforaminal injection: spinal cord infarction resulting in quadriplegia,4,5 cortical blindness, and fatal strokes in the territory of the posterior cerebral circulation;6,7 after lumbar transforaminal injection: paraplegia resulting from infarction of the conus medullaris.8 The clinical suspicion has been that the mechanism of injury is ischemia caused by end-arteriolar occlusion after the inadvertent intraarterial injection of particulate steroid,9 which has been recently confirmed in a convincing study performed in large animals.10 It follows that the only safe manner to perform transforaminal injection using the common particulate formulations of steroid that have been used for this application is to devise a means to detect intravascular needle position before particulate steroid is injected, allowing the needle position to be adjusted to avoid intraarterial injection of even a portion of the injected steroid. Experts have recommended the use of radiographic contrast injected under continuous or “live” fluoroscopy with or without digital subtraction to detect intravascular needle position.9 In this issue, Kim et al.11 prospectively examined the frequency of intravascular injection using radiographic contrast injected under continuous fluoroscopy. What is most striking is that the incidence of intravascular injection that Kim et al.11 report is frighteningly high: 56 of 182 injections (30.8%) overall, 45 of 71 (63%) cervical, 11 of 110 (10%) lumbar with no blood flashback through the 21-gauge needle in 45 of 56 (80%) these cases. The incidence is markedly higher during cervical injections when compared with lumbar injections. Static fluoroscopic injections performed during and immediately after injection do not reliably detect intravascular injection in more than half of the cases. Simultaneous perineural spread and intravascular uptake frequently occurred (23 of 45 [52%] in cervical and 1 of 11 [9%] in lumbar injections), whereas pure vascular uptake occurred in 11.3% of cervical and 0.9% of lumbar injections, underscoring the need for a continuous fluoroscopy. However, something is curious: the authors tell us that they cannot distinguish between IV (benign) and arterial (potentially catastrophic) uptake. A previous report also did not discern between IV and intraarterial injection, finding a somewhat lower incidence of intravascular injection during cervical transforaminal injection (19.4% of 504 injections)12 and a similar rate during lumbar transforaminal injection (11.2% in a series of 761 injections).13 Clinical practice and a few sporadic observations suggest that Vol. 109, No. 1, July 2009

Figure 1. Anterior-posterior view of the cervical spine during cervical transforaminal injection demonstrating intraarterial contrast injection. Left panel, image as seen on fluoroscopy. Right panel, image as seen with use of digital subtraction. The needle tip lies in the left C7/T1 intervertebral foramen. Contrast outlines the spinal nerve within and lateral to the foramen (arrowhead) and digital subtraction clearly reveals that contrast extends medially toward the center of the spinal canal via a spinal medullary artery (small arrow) to the anterior spinal artery. (Reproduced from Ref. 9, with permission.)

Figure 2. Anterior-posterior view of the cervical spine during cervical transforaminal injection demonstrating IV contrast injection. Left panel, image as seen on fluoroscopy. Right panel, image as seen with use of digital subtraction. The needle tip lies in the left C5/6 intervertebral foramen. Contrast outlines the spinal nerve within and lateral to the foramen (arrowhead) and digital subtraction clearly reveals that contrast extends laterally and inferiorly away from the spinal canal toward the central venous circulation (small arrow).

IV injection is common and intraarterial injection is rare, and that the two can be distinguished with imaging. The first hint is discussed by Kim et al.11 in this study: flow toward the midline. They tell us, “Eleven cases out of 71 cervical transforaminal epidural injections (TEI) (15%) showed vascular flow pattern on anterior-posterior view running to the vertebral column. However, there was a significantly different result in lumbar TEI; one case of intravascular uptake (0.9%) running toward the midline of vertebral column.” Flow toward the spinal canal is consistent with arterial or venous flow, suggesting that 15% of cervical transforaminal injections in this series were potentially intraarterial before the needle was repositioned and this is alarming. Are there better means to differentiate IV from intraarterial injection? The use of digital subtraction technology has been suggested.9 There is little doubt that digital subtraction technology can enhance the visualization of small vascular structures by “subtracting” the digital image that exists before contrast injection from subsequent images, leaving the practitioner to see only the pattern of contrast flow without Vol. 109, No. 1, July 2009

distracting overlying shadows. Most modern imaging equipment stores the digital subtraction runs as a series of individual frames and allows immediate playback of the images as a motion picture loop. In this way, the practitioner can reexamine the sequence of injection looking for subtle patterns of contrast flow. When continuous fluoroscopy is used without the capability of storing and reviewing a series of consecutive images, the practitioner must repeat the injection using additional x-ray exposure and contrast media. Most modern portable c-arms now have the capability to allow for routine use of digital subtraction. The difficulties described by Kim et al.11 in distinguishing IV from intraarterial injection may have been reduced or eliminated by the addition of digital subtraction. Indeed, the brisk arterial flow toward the spinal canal and the clear delineation of the arterial supply to the spinal cord itself (Fig. 1) would seem difficult to confuse with the sluggish venous flow away from the spinal canal and toward the central venous circulation, also well delineated with digital subtraction (Fig. 2). So, what have we learned from this study? The authors have confirmed that return of blood from the © 2009 International Anesthesia Research Society

9

needle during transforaminal injection, either passively or with aspiration, is not a reliable means to detect intravascular needle placement. They also confirm that intravascular injection is common and may be more common during cervical than lumbar transforaminal injection, appearing in as many as one third of these procedures. They clearly demonstrate that still radiographs, no matter when they are obtained during the course of the injection, cannot reliably detect intravascular needle placement. However, the current findings do not help us to understand how best to distinguish IV injection, which is unlikely to lead to serious sequelae, from intraarterial injection. Moreover, although no adverse events are reported in this small series, it is unclear what practitioners should do once they detect intravascular needle placement. Is it really safe to proceed with injection of particulate steroid after simply adjusting the position of the needle? It seems feasible that a portion of the steroid could be forced into the arterial system once the artery has been penetrated, even if the tip of the needle no longer resides within the arterial lumen. Finally, digital subtraction technology seems to offer some promise of improving the safety of transforaminal injection, but is in need of further study. Other critical questions also remain, foremost among them: should we move to the routine use of nonparticulate steroid or abandon the transforaminal technique altogether? For now, this new study reinforces the critical need for real-time imaging in detecting intravascular injection during transforaminal injection.

10

Editorial

REFERENCES 1. Young IA, Hyman GS, Packia-Raj LN, Cole AJ. The use of lumbar epidural/transforaminal steroids for managing spinal disease. J Am Acad Orthop Surg 2007;15:228 –38 2. Schaufele MK, Hatch L, Jones W. Interlaminar versus transforaminal epidural injections for the treatment of symptomatic lumbar intervertebral disc herniations. Pain Physician 2006;9:361– 6 3. Ackerman WE III, Ahmad M. The efficacy of lumbar epidural steroid injections in patients with lumbar discherniations. Anesth Analg 2007;104:1217–22 4. Baker R, Dreyfuss P, Mercer S, Bogduk N. Cervical transforaminal injection of corticosteroids into a radicular artery: a possible mechanism for spinal cord injury. Pain 2003;103:211–5 5. Muro K, O’Shaughnessy B, Ganju A. Infarction of the cervical spinal cord following multilevel transforaminal epidural steroid injection: case report and review of the literature. J Spinal Cord Med 2007;30:385– 8 6. Tiso RL, Cutler T, Catania JA, Whalen K. Adverse central nervous system sequelae after selective transforaminal block: the role of corticosteroids. Spine J 2004;4:468 –74 7. Rozin L, Rozin R, Koehler SA, ShakirA, Ladham S, Barmada M, Dominick J, Wecht CH. Death during transforaminal epidural steroid nerve root block (C7) due to perforation of the left vertebral artery. Am J Forensic Med Pathol 2003;24:351–5 8. Houten JK, Errico TJ. Paraplegia after lumbosacral nerve root block: report of three cases. Spine J 2002;2:70 –5 9. Rathmell JP, Aprill C, Bogduk N. Cervical transforaminal injection of steroids. Anesthesiology 2004;100:1595– 600 10. Okubadejo GO, Talcott MR, Schmidt RE, Sharma A, Patel AA, Mackey RB, Guarino AH, Moran CJ, Riew KD. Perils of intravascular methylprednisolone injection into the vertebral artery. An animal study. J Bone Joint Surg Am 2008;90:1932– 8 11. Kim DW, Han KR, Kim C, Chae YJ. Intravascular flow patterns in transforaminal epidural injections: a comparative study of the cervical and lumbar vertebral segments. Anesth Analg 2009;109:233–9 12. Furman MB, Giovanniello MT, O’Brien EM. Incidence of intravascular penetration in transforaminal cervical epidural steroid injections. Spine 2003;28:21–5 13. Furman MB, O’Brien EM, Zgleszewski TM. Incidence of intravascular penetration in transforaminal lumbosacral epidural steroid injections. Spine 2000;25:2628 –32


Cardiovascular Anesthesiology

Cardiovascular and Thoracic Education

Hemostasis and Transfusion Medicine

Section Editor: Charles W. Houge, Jr.

Section Editor: Martin J. London

Section Editor: Jerrold H. Levy

The Effect of Epsilon-Aminocaproic Acid and Aprotinin on Fibrinolysis and Blood Loss in Patients Undergoing Primary, Isolated Coronary Artery Bypass Surgery: A Randomized, Double-Blind, Placebo-Controlled, Noninferiority Trial Philip E. Greilich, MD, FAHA* Michael E. Jessen, MD† Neeraj Satyanarayana, BS* Charles W. Whitten, MD* Gregory A. Nuttall, MD‡ Joseph M. Beckham, MD* Michael H. Wall, MD§ John F. Butterworth, MD储

BACKGROUND: Until recently, aprotinin was the only antifibrinolytic drug with a licensed indication in cardiac surgery in the United States. The most popular alternative, ⑀-aminocaproic acid (EACA), has not been adequately compared with aprotinin. We undertook this study to test the hypothesis that EACA, when dosed appropriately, is not inferior to aprotinin at reducing fibrinolysis and blood loss. METHODS: Seventy-eight patients scheduled for primary, isolated coronary artery bypass graft surgery were randomly assigned to receive “full Hammersmith” dose aprotinin, high dose EACA (100 mg/kg initial loading dose, 5 g in the pump prime solution, 30 mg 䡠 kg⫺1 䡠 h⫺1 maintenance infusion) or equal volumes of a salineplacebo in a double-blind trial. Reductions in peak d-dimer formation (a measure of fibrinolysis) and 24-h chest tube drainage (CTD) were the primary end points by which noninferiority of EACA was tested. The noninferiority limit was set at a 30% increase in peak d-dimer formation (a difference of 250 ␮g/mL) and 24-h CTD (a difference of 350 mL) relative to aprotinin. RESULTS: The between-group differences (EACA versus aprotinin) in peak d-dimer formation (⫺3.58 ␮g/L, 95% CI ⫺203 to 195 ␮g/L) and 24-h CTD (67 mL, 95% CI ⫺90 to 230 mL) were within the predetermined noninferiority margins (250 ␮g/mL and 350 mL, respectively) and satisfied the criteria for noninferiority. Compared with saline, significant between-group reductions in peak d-dimer formation were observed using EACA (589 ␮g/L, 95% CI 399 –788 ␮g/L; P ⬍ 0.0001) and aprotinin (585 ␮g/L, 95% CI 393–778 ␮g/L; P ⬍ 0.0001). Similar reductions in 24 h CTD were also seen using EACA (239 mL, 95% CI 50 – 415 mL; P ⬍ 0.05) and aprotinin (323 mL, 95% CI 105– 485 mL; P ⬍ 0.05) compared with saline. Plasma EACA levels were maintained well above a target of 260 ␮g/mL. CONCLUSIONS: When dosed in a pharmacologically guided manner, EACA is not inferior to aprotinin in reducing fibrinolysis and blood loss in patients undergoing primary, isolated coronary artery bypass surgery. (Anesth Analg 2009;109:15–24)

A

ntifibrinolytic drugs are recommended in cardiac surgery to promote blood conservation (Class I; Level of evidence A).1 These drugs reduce blood loss by inhibiting fibrinolysis during and after cardiopulmonary bypass (CPB).2–5 Nevertheless, there is no drug From the Departments of *Anesthesiology and Pain Management, and †Cardiovascular and Thoracic Surgery, University of Texas Southwestern Medical Center at Dallas, Dallas, Texas; ‡Department of Anesthesiology, Mayo Clinic College of Medicine, Rochester, Minnesota; §Department of Anesthesiology, Washington University, St. Louis, Missouri; and 储Department of Anesthesia, Indiana University School of Medicine, Indianapolis, Indiana. Accepted for publication January 16, 2009. Supported by University of Texas Southwestern Department of Anesthesia and Pain Management and Department of Veterans’ Affairs Clinical Research Funds. Address correspondence and reprint requests to Philip E. Greilich, MD, 5323 Harry Hines Blvd., Dallas, TX 75390-8894. Address e-mail to [email protected]. Copyright © 2009 International Anesthesia Research Society DOI: 10.1213/ane.0b013e3181a40b5d

Vol. 109, No. 1, July 2009

currently licensed for this indication. Aprotinin was the only antifibrinolytic drug to receive approval by the United States Food and Drug Administration (FDA) for use in cardiac surgery but was recently removed from marketing for safety reasons.* Epsilonaminocaproic acid (EACA) remains the most commonly used antifibrinolytic in coronary artery bypass graft (CABG)/CPB surgery in the United States, yet it has not been subjected to the scrutiny (pharmacokinetics, dose ranging, safety, and efficacy) required of any newly licensed pharmaceutical.6 In May of 2008, aprotinin was withdrawn from marketing after a large randomized controlled trial (BART Trial) confirmed the link between aprotinin use and increased mortality and renal dysfunction (compared with lysine analogs) in high risk cardiac surgery *FDA News: Manufacturer Removes Remaining Stocks of Trasylol, FDA, 2008; http://www.fda.gov/bbs/topics/NEWS/2008/NEW01834. html.

15

patients.6 –10 Unfortunately, withdrawal of aprotinin occurred before the relative efficacy of EACA and aprotinin in CABG/CPB surgery could be established. Few head-to-head trials have compared aprotinin and EACA in cardiac surgery and these studies are illsuited to meta-analytic techniques because of significant heterogeneity, widely varying EACA dosing and lack of placebo-controls.11–14 This study was conceived and performed at a time when EACA and aprotinin were in common, seemingly competitive use, and avoids many of these limitations by providing: 1) a homogenous surgical group (primary, isolated CABG); 2) a pharmacokinectically guided high dose EACA regimen; and 3) a placebo-control group. The questions regarding EACA’s dosing and efficacy assume greater importance now that aprotinin has been withdrawn. This study tests whether EACA and aprotinin provide comparable degrees of inhibition of fibrinolysis and, more importantly, whether they provide similar efficacy at reducing blood loss in patients undergoing CABG/CPB surgery. Plasma d-dimer formation and chest tube drainage (CTD), our primary end points, are well-established outcome measures in studies involving aprotinin.15–17 In addition, we measured plasma levels of EACA to confirm that our high-dose regimen maintained drug levels well above the therapeutic target.

METHODS Study Design This was a prospective, randomized, double-blind, placebo-controlled, noninferiority trial. The noninferiority design permits the investigator to define the margin below which between-group (aprotinin, EACA) differences for the primary end point are considered medically unimportant.18 Noninferiority margins were based on conservative clinical judgment and statistical reasoning as described by the International Council on Harmonization.19 After a review of the literature and discussion with experienced cardiac anesthesiologists, surgeons, hematologists, and transfusion medicine specialists in this field, a consensus was established that a noninferiority margin of 30% in peak d-dimer formation (difference of 250 ␮g/L) and 24 h CTD (difference of 350 mL) should be the limit of tolerability for this study.15,16,20 Sample size calculations were then determined using PASS 2004 software (Kaysville, UT) to achieve a power (1⫺␤) of ⬎0.80. For peak d-dimer formation, a mean difference of 250 ␮g/L and standard deviation of 300 ␮g/L would require 25 subjects per group to achieve a power of 0.84. For 24-h CTD, a mean difference of 350 mL and a standard deviation of 400 mL, 25 subjects per group would achieve a power of 0.83. We hypothesized that the upper limit of the 95% confidence interval for peak d-dimer formation and 24-h CTD would be less than their respective noninferiority margins which would 16

EACA Versus Aprotinin for Primary CABG

allow us to reject the hypothesis that EACA is inferior to aprotinin in this population. After IRB review and approval, 81 patients at the Dallas Veterans’ Affairs Medical Center were consented and enrolled over a period of 28 mo (September 1998 to January 2001). Eligible patients undergoing primary, isolated CABG surgery were randomly assigned to receive placebo, EACA, or aprotinin. Exclusion criteria included emergency surgery, preexisting bleeding diathesis, symptomatic cerebrovascular disease, hepatic dysfunction, serum creatinine ⬎2.0 mg/dL, and left-ventricular ejection fraction ⬍30%. Aspirin use was allowed, but the use of adenosine diphosphate or glycoprotein IIb/IIIa receptor inhibitors or thrombolytic drugs were prohibited. Baseline patient characteristics were collected and EuroSCORE and Society of Thoracic Surgeons (STS) risk of 30-day mortality were calculated for each study subject.†21 Of note, this report contains data from patients included in articles previously published in Circulation, Anesthesiology, and Journal of Thoracic and Cardiovascular Surgery.22–24 These works focused on the effects of antifibrinolytic drugs on the inflammatory response. Although d-dimer formation and blood loss data were reported, these data were not used as primary end points nor were they subjected to noninferiority analyses. All consents were obtained before allocation to a treatment group. Allocation was performed by a research pharmacist using a simple random numbers table for three groups (1:1:1 ratio): placebo, EACA, and aprotinin. Colorless study material was dispensed by the pharmacist to the anesthesia team using three 250 mL bags (normal saline) on the morning of the surgery. All study personnel and participants were blinded to the treatment assignment for the duration of the study. The code was revealed to the investigators once recruitment, data collection, and laboratory analysis were complete.

Drug Regimens All patients received study drug from three 250 mL bags of clear solution labeled “Loading Dose,” “Pump Prime,” and “Infusion” as shown in Figure 1. All patients received a 1 mL test dose from the bag labeled “Loading Dose.” The EACA group received a loading dose of 100 mg/kg of EACA (Amicar, Xanodyne Pharmaceuticals, Newport, KY) over 15 min, plus 5 g added to the pump prime and an infusion of 30 mg 䡠 kg⫺1 䡠 h⫺1 until the patient arrived in the intensive care unit (ICU). The dosing regimen for the aprotinin group was based on the package insert and consisted of a loading dose of 2 ⫻ 106 kallikrein inhibiting units (KIU) of aprotinin (Trasylol, Bayer Pharmaceuticals Corp, West Haven, CT) over 15 min, plus 2 ⫻ 106 KIU added to the pump prime and an †STS: STS National Database Risk Calculator, The Society of Thoracic Surgeons, 2007; http://www.sts.org/sections/stsnationaldatabase/ riskcalculator/.


Figure 1. Protocol for administration of study material. Study material was administered from three (loading dose, pump prime, infusion) 250 mL bag of clear solution. The amount of ⑀-aminocaproic acid (EACA) in bags used for the loading dose and infusion were weight-adjusted. All loading doses were administered over 20 min. The pump prime was added to the cardiopulmonary bypass (CPB) circuit before initiating CPB. All infusions were initiated after the loading dose was complete and continued at a rate of 50 mL/h until arrival in the intensive care unit.

infusion of 5 ⫻ 105 KIU/h until the patient arrived in the ICU.‡ This regimen is commonly called the “full Hammersmith” dose. The placebo group received equal volumes of saline for the loading dose, pump prime, and infusion. All loading doses were administered immediately after anticoagulation with heparin.25

Sample Collection, D-Dimer Assay, and Plasma Drug Levels Arterial blood samples were collected after the removal of five dead space volumes at six time points: 1) before anesthesia induction or Preop; 2) 10 min after the initiation of CPB or CPB 10 min; 3) immediately before separation from CPB or CPB warm; 4) 10 min after the completion of protamine sulfate or End CPB; 5) 4 h after the conclusion of CPB or 4 h Post-CPB; and 6) 24 h after CPB or 24 h Post CPB. Hemoglobin levels, platelet counts, and fibrinogen concentrations were measured at all six time points. d-dimer levels and plasma EACA and aprotinin levels were collected at multiple time points until 4 h after CPB. d-dimer levels were determined using a turbidometric immunoassay (Beckman Coulter, Miami, FL) and plasma drug levels were determined as previously described.26 –28 Serum troponin I and creatinine phosphokinase-MB isoenzyme levels were determined at baseline, 4 and 24 h after CPB.

Surgical, CPB, and ICU Management Patients underwent surgery via median sternotomy. The CPB technique included a membrane oxygenator (model CEO 123; Gish, Irvine, CA), nonpulsatile flow using a centrifugal pump (Biomedicus, Eden Prairie, MN), and a 2.0L prime (1.8L lactated Ringer’s solution, 100 mL albumin 25%, 44.6 mEq sodium bicarbonate, 50 g mannitol, 10,000 IU bovine heparin). Myocardial protection was provided by moderate hypothermia (28°C–32°C) with antegrade/retrograde sanguineous (4:1 blood: crystalloid) cardioplegia infused every 10 –20 min. CPB flow rates were 2 L 䡠 min⫺1 䡠 m⫺2 during hypothermia and 2.5 L 䡠 min⫺1 䡠 m⫺2 during normothermia. ‡Bayer: Trasylol (aprotinin injection), Bayer HealthCare, 2006; http://www.univgraph.com/bayer/inserts/trasylol.pdf. Vol. 109, No. 1, July 2009

Anticoagulation using bovine heparin produced an initial kaolin activated clotting time of (ACT) ⱖ480 s. The Hepcon system (Medtronic, Minneapolis, MN) was used for heparin and protamine titration.29 Heparin concentration was checked every 30 min during CPB and additional heparin was added as needed. Blood aspirated by the cardiotomy suction was reinfused before chest closure. Thoracostomy drainage tubes were inserted before chest wound closure. A scaled1–5 assessment of the quality of hemostasis (poor, marginal, acceptable, good, and excellent) and prediction of the patient’s study assignment group were made by the blinded operating surgeon immediately before chest closure. CTD was recorded hourly. Shed mediastinal blood was not reinfused.

Transfusion Criteria The criteria for the transfusion of packed red blood cells were hemoglobin concentrations ⬍6.0 g/dL during CPB, ⬍7.0 g/dL at normothermia immediately before separation from CPB, and ⬍8.0 g/dL in the early postbypass period. The criteria for transfusion of platelet concentrates and fresh frozen plasma were excessive bleeding and a platelet count ⬍70,000/␮L or a prothrombin time and/or activated partial thromboplastin time of ⬎1.5 times the upper limit of normal (after heparin reversal), respectively. Additional protamine was administered if the ACT was prolonged.

Adverse Events Deaths occurring within the same hospital admission, within 30 days of surgery but after discharge from the hospital, or within the 12 mo after surgery were recorded. Stroke, myocardial infarction (MI), renal dysfunction, and renal failure requiring dialysis were defined as previously described.8,30 –32 Criteria for these serious adverse events included a clinical diagnosis of stroke or evidence of a focal or global defect on computed tomography or magnetic resonance imaging, peak troponin I and/or creatine phosphokinase-MB isoenzyme levels 30 times the upper limit of normal for MI and an increase in percent change in creatinine of more than 50% from baseline for renal dysfunction and a new requirement for © 2009 International Anesthesia Research Society

17

dialysis for renal failure. Excessive blood loss was defined as ⬎750 mL in the first 6 h after thoracostomy tube placement.29 Prolonged mechanical ventilation, ICU, and hospital length of stay (LOS) were defined as 48 h, 96 h, and 8 days, respectively.

Statistical Analysis Noninferiority was assessed by observing the upper limit of the 95% CI based on observed differences (using mean or median values as indicated). EACA was considered as effective as the aprotinin when the upper limit of the 95% confidence interval (based on the differences) is less than the noninferiority margin.18 Noninferiority analyses were based on parametric confidence interval calculations for data considered normally distributed (peak d-dimer formation) and by using Hedges-Leman confidence interval estimates of median differences when data were considered nonnormally distributed (24 h blood loss).33 Standard statistical comparisons between groups were performed using SAS 9.13 software (SAS Institute, Cary, NC). A two-way analysis of variance (ANOVA) with repeated measures was performed testing for interaction between time and groups. Post hoc comparison of groups was performed using Tukey procedure. Continuous data for patient and operative characteristics were analyzed using an ANOVA for parametric data and Kruskal–Wallis analysis for nonparametric data. Nominal data were analyzed using ␹2

or Fisher’s exact test as appropriate. Spearman correlation was used to determine if there was an association between drugs levels, peak d-dimer formation, and CTD. Multivariate linear regression analyses were performed to determine which independent variables were predictive of 24 h CTD and/or peak d-dimer formation. Baseline patient characteristics and operative data served as potential candidate variables. A hierarchical approach was used to identify potential independent variables in conjunction with automated model building procedures to arrive at a parsimonious model. A P value of ⬍0.05 was considered statistically significant. P values were not adjusted for multiple tests.

RESULTS Of the 81 patients consented for the study, one patient had his surgery cancelled after induction of anesthesia and two others did not receive study material because they were emergently placed on CPB. This provided 78 evaluable subjects for the on-treatment analysis. Table 1 summarizes the baseline (preoperative) characteristics of the patients. There were no significant differences in age, weight, comorbid conditions, preoperative medications, relevant baseline laboratory, EuroSCORE, or STS predicted risk among groups. All patients in this study received aspirin (ⱖ81 mg daily) for at least 72 h before surgery.

Table 1. Baseline Patient Characteristics General Age at surgery (yr) Weight (kg) Male no. (%) Cardiovascular risk no. (%) Hypertension Hypercholesterolemia Tobacco use Diabetes (any type) Risk stratification EuroSCOREa STS risk of death (%)b Current cardiac history no. (%) Unstable angina Previous MIc Ejection fraction ⬍50% Medications no. (%) Beta-blocker ACE inhibitor Statin therapy Salicylates Intravenous heparin Laboratory data no. (%) Hemoglobin ⬍12.5 g/dL Creatinine level ⬎1.4 mg/dL

Saline (n ⫽ 27)

Aprotinin (n ⫽ 26)

EACA (n ⫽ 25)

P

62 ⫾ 7 89 ⫾ 15 27 (100)

65 ⫾ 9 91 ⫾ 21 26 (100)

62 ⫾ 8 95 ⫾ 16 25 (100)

0.24 0.53 1.00

20 (74) 21 (78) 24 (89) 14 (52)

22 (85) 22 (85) 20 (77) 10 (38)

22 (88) 20 (80) 21 (84) 14 (56)

0.43 0.88 0.52 0.42

4.7 ⫾ 1.9 1.4 ⫾ 0.8

5.4 ⫾ 2.0 1.3 ⫾ 1.0

4.4 ⫾ 1.7 1.2 ⫾ 0.7

0.13 0.54

4 (15) 15 (56) 10 (37)

4 (15) 8 (31) 13 (50)

3 (12) 13 (52) 10 (40)

1.00 0.27 0.61

24 (87) 18 (67) 16 (59) 27 (100) 4 (15)

23 (89) 20 (77) 18 (69) 26 (100) 2 (8)

21 (84) 17 (68) 20 (80) 25 (100) 6 (24)

0.84 0.68 0.27 1.00 0.26

16 (59) 0 (0)

16 (62) 0 (0)

14 (56) 0 (0)

0.92 1.00

Data expressed as mean ⫾ SD unless otherwise stated. ACE ⫽ angiotensin converting enzyme; EACA ⫽ epsilon-aminocaproic acid. a European System for Cardiac Operative Risk Evaluation (EuroSCORE): Low risk (0 –2), Medium risk (3–5), High risk (6 –24). b American Society of Thoracic Surgeons (STS) predicted in-hospital mortality. c Previous history of Myocardial Infarction (MI).

18



Table 2. Operative Data Group

Saline (n ⫽ 27)


EACA (n ⫽ 25)

P

Median number of grafts Internal thoracic artery used no. (%) Duration of CPB (min) Duration of aortic cross-clamp (min) Time to skin closure (min) Total heparin dose (U/kg)a Protamine/heparin reversal ratiob Median hemostasis scorec Correct prediction of treatment no. (%)d OR chest tube drainage (mL)e

3 100% 119 ⫾ 30 73 ⫾ 21 46 ⫾ 18 536 ⫾ 129 0.8 ⫾ 0.3 3 54% 100 (120)

3 96% 143 ⫾ 64 82 ⫾ 30 44 ⫾ 18 557 ⫾ 160 0.8 ⫾ 0.3 4 38% 68 (50)

3 96% 131 ⫾ 36 74 ⫾ 19 50 ⫾ 20 482 ⫾ 111 0.8 ⫾ 0.3 3 41% 100 (78)

0.95 0.54 0.17 0.40 0.52 0.13 0.85 0.10 0.37 0.03

Data expressed as mean ⫾ SD unless otherwise stated. EACA ⫽ epsilon-aminocaproic acid; CPB ⫽ cardiopulmonary bypass. a Includes heparin in pump prime and in cardioplegic solutions. b Protamine/heparin ratio is based on 10 mg protamine per 100 IU heparin. c Hemostasis is scored immediately after chest closure and range from 1 to 5 (1-poor, 2-marginal, 3-adequate, 4-good, 5-excellent). d Surgeon’s prediction of treatment group before leaving operating room (OR). e Chest tube drainage expressed as median (interquartile range).

Figure 2. Mean hemoglobin concentration. Hemoglobin was measured before induction (preinduction), 10 min after cardiopulmonary bypass (CPB) was initiated (10 min CPB), immediately before separation from CPB (CPB warm), 10 min after administration of protamine sulfate (End CPB), 4 h after CPB (4 h Post-CPB), and 24 h after CPB (24 h Post-CPB). There were no significant between-group differences in hemoglobin concentration at any time point. Error bars were omitted given the absence of these differences.

Operative details are provided in Table 2. There were no differences among the groups in the number of grafts, use of internal thoracic artery grafts, heparin dose (IU/kg), duration of CPB or aortic crossclamping, protamine/heparin reversal ratios, time to skin closure, or hemostasis scores. There were also no differences in ACT values or heparin concentrations at any time between groups. The intraoperative CTD was lower in the aprotinin group (P ⫽ 0.03) compared with saline and EACA. The hemoglobin concentrations (Fig. 2) and platelet counts did not differ among the groups at any time. An average of 26.5 g (95% CI 24.5–28.5 g) of EACA and 6.2 ⫻ 106 KIU (95% CI 6.0 ⫻ 106– 6.4 ⫻ 106 KIU) of Vol. 109, No. 1, July 2009

Figure 3. Plasma ⑀-aminocaproic acid (EACA) concentrations. Plasma drug levels were measured in a subset of patients (n ⫽ 14) for EACA before induction of anesthesia (Preinduction), 10 min after cardiopulmonary bypass (CPB) was initiated (10 min CPB), immediately before separation from CPB (CPB-Warm), after heparin reversal (End CPB), and 4 h after the CPB (4 h Post-CPB). Infusion of the study material was stopped upon arrival in the intensive care unit. The therapeutic target for EACA is between 130 and 260 ␮g/mL.

aprotinin were administered to patients in the EACA and aprotinin groups, respectively. Based on the purchase price of these drugs at our institution, the direct pharmacy costs per patient was $1048.77 for aprotinin and $3.84 for EACA. The plasma drug levels for a subset of patients receiving EACA (n ⫽ 14) are illustrated in Figure 3. All of the EACA patients tested had plasma drug levels ⬎260 ␮g/mL shortly after protamine administration and ⬎130 ␮g/mL 4 h after CPB. There was no correlation found between plasma EACA levels and peak d-dimer formation or blood © 2009 International Anesthesia Research Society

19

Figure 4. Peak d-dimer level confidence intervals. The confidence intervals (CI) for the between-group differences for peak d-dimer are shown. The lines represent parametric 95% CIs for the difference in mean peak d-dimer formation (in ␮g/L) between the specified groups. loss at any time. Plasma aprotinin levels in similar subset of patients (n ⫽ 15) peaked shortly after CPB was initiated (202 ⫾ 80 KIU/mL) and decreased to 125 ⫾ 46 KIU/mL immediately before separation from CPB and further to 40 ⫾ 33 KIU/mL 4 h after CPB. There were negative correlations between aprotinin levels and peak d-dimer formation (r2 ⫽ ⫺0.304; P ⫽ 0.03) 4 h after CPB and between aprotinin levels and 6 h CTD (r2 ⫽ ⫺0.300; P ⫽ 0.03) 10 min after initiation of CPB. The confidence intervals and the estimated mean differences between treatment groups for peak d-dimer levels are shown in Figure 4. The observed

difference for EACA and aprotinin (⫺3.58 ␮g/L, 95% CI ⫺203 to 195 ␮g/L) was less than the predetermined noninferiority margin of 250 ␮g/L. d-dimer formation was increased in all treatment groups compared with baseline during and immediately after CPB, but was markedly attenuated in those receiving active drug (EACA and aprotinin; P ⬍ 0.0001). These differences were most pronounced during rewarming and in the early post-CPB period. The saline placebo group had significantly higher peak d-dimer levels (1197 ⫾ 279 ␮g/L) compared with aprotinin (608 ⫾ 279 ␮g/L; P ⬍ 0.0001) or EACA (612 ⫾ 335 ␮g/L; P ⬍ 0.0001). The multivariate linear regression model for peak d-dimer formation included (P ⬍ 0.2) treatment group, tobacco use, STS risk, statin therapy, and heparin:protamine reversal ratio based on univariate analysis. This model predicted 51% of the variance (r2 ⫽ 0.511; P ⬍ 0001) in peak d-dimer formation when all variables were included. Treatment group (P ⬍ 0.0001) and tobacco use (P ⫽ 0.006) were the only independent variables identified in this model. The confidence intervals and the estimated median differences between treatment groups for 6 and 24 h CTD are shown in Figure 5. The observed difference in 24 h CTD for EACA and aprotinin (67 mL, 95% CI ⫺90 to 230 mL) was less than the predetermined noninferiority margin of 350 mL. In addition, there was a significant overall treatment effect (ANOVA with repeated measures; P ⫽ 0.009) for EACA and aprotinin (compared with saline) on 24 h CTD. This effect was not present when EACA was compared with aprotinin. Significant point reductions in CTD at 6 and 24 h were also observed for both EACA and aprotinin groups compared with the saline control group.

Figure 5. Chest tube drainage (CTD) confidence intervals. The confidence intervals (CI) for the between-group differences for 6 and 24 h CTD are shown. The lines represent nonparametric 95% CIs for the difference in median chest tube drainage (in mL) between the specified groups.

20



Table 3. Blood Product Exposure Group Packed red blood cells no. (%) Platelets no. (%) Fresh frozen plasma no. (%) Any blood products no. (%)

Table 4. Adverse Events

Saline Aprotinin EACA (n ⴝ 27) (n ⴝ 26) (n ⴝ 25) 17 (63)

18 (69)

15 (60)

Group

0.78

Mortality 30-day no. (%) 1-yr no. (%) Stroke no. (%) Myocardial infarction no. (%)a Renal dysfunction no. (%)b Renal dysfunction requiring dialysis no. (%) Excessive chest tube drainage no. (%)c

8 (30) 6 (22)

9 (35) 9 (35)

4 (16) 5 (20)

0.31 0.47

18 (67)

18 (69)

15 (60)

0.77

EACA ⫽ epsilon-aminocaproic acid.

Again, no such differences were found when EACA was compared with aprotinin (P ⬎ 0.3). The median (interquartile) 24 h CTD for the saline, EACA, and aprotinin groups were 870 (671–1015), 640 (435–795), and 478 mL (409 – 879), respectively. The mean (⫾sd) 24 h CTD for the saline, EACA, and aprotinin groups were 1002 ⫾ 627, 715 ⫾ 394, and 685 ⫾ 505 mL. The multivariate linear regression model for predicting 24 h CTD included (P ⬍ 0.2) treatment group, diabetes, previous MI, and ejection fraction ⬍50% based on univariate analysis. This model predicted 16% of the variance (r2 ⫽ 0.160; P ⫽ 0.013) in 24 h CTD. Treatment group (P ⫽ 0.011) was the only independent variable identified in this model. Table 3 summarizes the patients exposed to blood transfusion before discharge from the ICU. This study was not powered to detect differences in blood transfusion. Overall, 65% of patients were transfused, and the likelihood of transfusion did not seem to differ significantly among groups (P ⫽ 0.77). The mean (⫾sd) number of units transfused per person in the saline, aprotinin, and EACA groups was 1.8 ⫾ 2.0, 2.5 ⫾ 2.7, 1.7 ⫾ 1.8, respectively, for packed red blood cells, 2.7 ⫾ 5.4, 2.6 ⫾ 4.3, 1.0 ⫾ 2.2, respectively, for platelets, and 0.8 ⫾ 1.7, 1.0 ⫾ 1.8, 0.4 ⫾ 0.8, respectively, for fresh frozen plasma. There were no significant differences in number of each blood product or total blood products between any of the treatment groups. Table 4 summarizes clinical adverse events. This study was not powered to detect differences in death, stroke, MI, renal dysfunction, or LOS. The predicted 30-day mortality in this study was 3.0% based on a mean EuroScore of 4.8 (intermediate risk class ⫽ 3–5), yet the observed 30-day mortality rate was 1.3%. There were no significant differences in 1-yr all-cause mortality, stroke, MI, renal dysfunction/failure, excessive blood loss, prolonged mechanical ventilation, ICU, or hospital LOS detected between groups. Four patients underwent mediastinal reexploration and two for excessive bleeding (both in the saline group). One patient had a surgical source for the bleeding and the other did not. One patient in the EACA group underwent a negative reexploration after cardiac arrest in the ICU. One aprotinin patient who suffered ventricular fibrillation cardiac arrest Vol. 109, No. 1, July 2009

Saline Aprotinin EACA (n ⫽ 27) (n ⫽ 26) (n ⫽ 25) P

P

0 (0) 0 (0) 1 (4) 7 (26)

1 (4) 2 (8) 2 (8) 6 (31)

0 (0) 0 (0) 1 (4) 6 (24)

0.65 0.33 0.77 0.85

9 (33)

8 (31)

3 (12)

0.16

0 (0)

1 (4)

1 (4)

0.54

4 (15)

1 (4)

1 (4)

0.23

EACA ⫽ epsilon-aminocaproic acid. a Myocardial infarction was defined as an increase in Troponin I (TnI) and/or creatinine phosphokinase-MB isoenzyme ⬎30⫻ upper limit normal, and/or an abnormal Q-Wave. b Renal dysfunction was defined as increase in ⌬ % creatinine ⬎50%. c Excessive chest tube drainage was defined as ⬎750 mL in the first 6 h after placement.

shortly after arrival in the ICU was returned to the operating room for regrafting of a thrombosed vein graft.

DISCUSSION This study demonstrates that EACA is not inferior to aprotinin at reducing fibrinolysis or blood loss in patients undergoing primary, isolated CABG/CPB surgery. The estimated difference in 24 h CTD between these two antifibrinolytic drugs was only 67 mL (95% CI ⫺90 to 230 mL), despite the use of aspirin in 100% of the patients studied. The maintenance of plasma EACA levels above ⬎260 ␮g/mL suggests the maximal therapeutic benefit was likely achieved. Multivariate linear regression analysis confirmed that treatment group was the only independent predictor of reductions in both peak d-dimer formation and 24 h blood loss. Our results suggest that patients undergoing primary, isolated CABG/CPB surgery are at negligible increased risk for excessive bleeding due to substitution of EACA for aprotinin. Aprotinin inhibits fibrinolytic activity by directly binding the active site of free plasmin, whereas EACA does so by preventing plasmin from binding to fibrin.2,4 In vitro studies have shown that plasmin activity can be reduced by up to 1000-fold when fibrin binding is inhibited with lysine analogs.3,34 d-dimer formation is the most commonly used marker for the putative effects of antifibrinolytic drugs in cardiac surgery. Multiple studies have associated reductions in d-dimer formation with decreases in CTD and blood transfusion.5,13,16,17,35 Although d-dimer levels were increased (compared with baseline) in every patient in this study, they were significantly attenuated in those receiving EACA or aprotinin. The absence of any between-group differences between EACA and aprotinin is consistent with their being equally effective for preventing blood loss. © 2009 International Anesthesia Research Society

21

This is the first randomized, placebo-controlled trial to demonstrate that EACA is not inferior to aprotinin at reducing blood loss in patients receiving aspirin before CABG/CPB. Preoperative aspirin use can significantly increase the risk of bleeding and had been the justification for using aprotinin in CABG/CPB surgery in some centers.1,36 In fact, a recent study in patients undergoing complex cardiac surgery (BART Trial) demonstrated that aprotinin provided its greatest benefit is those receiving preoperative aspirin.7 The lack of significant between-group (aprotinin versus EACA) differences in blood loss in this study was mostly likely due to our use of a much lower risk patient population, even so, aspirin use was common. Our inability to detect any subjective differences in hemostasis by the treatment-blinded surgeon (Table 2) or incidence of excessive postoperative bleeding further supports the clinical import of our findings. Two recent meta-analyses concluded that postoperative blood loss in patients receiving EACA is significantly greater (approximately 185 mL; 95% CI 110 –250 mL) than in those receiving aprotinin.30,37 Our estimated between-group (aprotinin, EACA) difference in 24 h CTD was lower (67 mL; 95% CI ⫺90 to 230 mL) and not significant by both standard methods of comparison and our primary noninferiority analyses. In contrast, the BART trial detected significant reductions in massive bleeding in high-risk patients receiving aprotinin compared with EACA.7 Our use of a low-risk patient population (primary, isolated CABG) provides the most plausible explanation for the difference found between these two studies. The use of blood products, in this study, was similar to that of Brown et al.30 who found no differences in red blood cell transfusions in the 18 head-to-head studies (2358 patients) comparing aprotinin to EACA or tranexamic acid. Like Brown et al., the upper 95% confidence limit for differences in CTD between saline and either aprotinin or EACA were much higher (485 and 415 mL, respectively, in this study) and likely explain the transfusion-sparing benefit found with these two antifibrinolytic drugs in cardiac surgery.1,27 The dosing regimens in the published trials using EACA in cardiac surgery have varied widely, undoubtedly leading to wide variations in EACA concentrations.11,12,26,38 Earlier in vitro studies demonstrated that the effective concentration for inhibiting 50% (EC50) of plasmin activity by EACA is 65–70 ␮g/mL.26 In 1999, Butterworth et al.26 reported the first multicompartmental pharmacokinetic analysis of EACA in patients undergoing cardiac surgery. Given the rapid changes in circulating blood volume in this patient population and normal patient-to-patient variation, they reasoned that plasma levels should be targeted at two to four times (130 –260 ␮g/mL) the EC50 to insure suppression of plasmin activity in all patients, although this has never been validated with clinical studies.26 22


Butterworth et al.26 and Bennett-Guerrero et al.39 used loading doses between 50 and 150 mg/kg and infusions between 25 and 30 mg 䡠 kg⫺1 䡠 h⫺1 and found that plasma EACA levels were maintained above 130 ␮g/mL until the time of protamine administration. The dosing regimen used in this trial (100 mg/kg loading dose, 5 g pump prime, and 30 mg 䡠 kg⫺1 䡠 h⫺1) exceeds that recommended by Butterworth et al. and resulted in plasma levels above 260 ␮g/mL at the time of protamine administration and above 130 ␮g/mL 4 h after CPB. Although this could have accounted for the lack of differences found (between aprotinin and EACA), optimal EACA blood concentration will remain speculative until dose-ranging outcome studies are performed. The aprotinin controversy underscores the potential risks of antifibrinolytics drugs in cardiac/CPB surgery. The small size of this study limits the ability to detect drug-related differences in these relatively rare adverse events. Several recent retrospective, observational studies, and meta-analyses suggest that the safety of EACA compares favorably with aprotinin.6,8,10,30,40 A recently published randomized controlled trial comparing lysine analogs and aprotinin also suggests that death and renal dysfunction may be more common with aprotinin.7 Nevertheless, the introduction of EACA into cardiac surgery has been flawed on many levels. Dose-ranging studies for EACA are clearly needed. Once an optimal dosing regimen has been established, larger clinical outcome trials using a saline-control group will need to be performed and should be powered to detect significant safety concerns (personal communication, Dwaine Rieves, MD, US FDA).41

Limitations These results may not be applicable to patients undergoing cardiac surgical procedures with greater risk for blood loss and transfusion, and it is unclear whether the benefits can be extrapolated to lower doses of EACA or the apparent safety can be extrapolated to larger doses of EACA. Our ability to reject the null hypothesis (EACA is inferior to aprotinin) was largely determined by the definition of the noninferiority margin. Our predefined tolerable threshold (24 h CTD difference of 350 mL) was based on a consensus of established clinicians and clinician-scientists. We note that drug licensing authorities have regarded larger noninferiority margins as adequate to establish drug equivalence in this context. Stafford-Smith et al.42 used a 12 h CTD difference of 400 mL in their Phase III cardiac/CPB trial conducted for FDA approval of heparinase as a replacement for protamine. The fact that this trial was conducted more than 7 years ago is a potential limitation. Although we cannot speak for other practices, we have not made any significant changes in our surgical or anesthetic technique that should have altered our findings had it ANESTHESIA & ANALGESIA

been repeated today. The primary change in most practices (related to this study) has been the increased use of a cell-saver device. This development would most likely have only further reduced any differences in blood transfusion among these low risk groups. In conclusion, EACA significantly reduces fibrinolysis and blood loss after primary, isolated CABG/ CPB surgery. High dose EACA appears as effective as full dose aprotinin when EACA plasma levels are maintained above the therapeutic target in a low (bleeding)-risk population. Further studies are needed to define the optimal EACA dosing regimen and to compare the risk:benefit ratios for EACA versus tranexamic acid. ACKNOWLEDGMENTS The authors are indebted to all the members of the cardiothoracic anesthesiology, surgical, and research teams that made this study possible. We would also like to acknowledge the guidance and critical review of the manuscript by Jay Horrow, MD, MS, and William Johnston, MD. REFERENCES 1. Ferraris VA, Ferraris SP, Saha SP, Hessel EA II, Haan CK, Royston BD, Bridges CR, Higgins RS, Despotis G, Brown JR, Spiess BD, Shore-Lesserson L, Stafford-Smith M, Mazer CD, Bennett-Guerrero E, Hill SE, Body S. Perioperative blood transfusion and blood conservation in cardiac surgery: the society of thoracic surgeons and the society of cardiovascular anesthesiologists clinical practice guideline. Ann Thorac Surg 2007;83: S27–S86 2. Davis R, Whittington R. Aprotinin. A review of its pharmacology and therapeutic efficacy in reducing blood loss associated with cardiac surgery. Drugs 1995;49:954 – 83 3. Longstaff C. Studies on the mechanisms of action of aprotinin and tranexamic acid as plasmin inhibitors and antifibrinolytic agents. Blood Coagul Fibrinolysis 1994;5:537– 42 4. Verstraete M. Clinical application of inhibitors of fibrinolysis. Drugs 1985;29:236 – 61 5. Slaughter TF, Faghih F, Greenberg CS, Leslie JB, Sladen RN. The effects of epsilon-aminocaproic acid on fibrinolysis and thrombin generation during cardiac surgery. Anesth Analg 1997;85: 1221– 6 6. Schneeweiss S, Seeger JD, Landon J, Walker AM. Aprotinin during coronary-artery bypass grafting and risk of death. N Engl J Med 2008;358:771– 83 7. Fergusson DA, Hebert PC, Mazer CD, Fremes S, MacAdams C, Murkin JM, Teoh K, Duke PC, Arellano R, Blajchman MA, Bussieres JS, Cote D, Karski J, Martineau R, Robblee JA, Rodger M, Wells G, Clinch J, Pretorius R. A comparison of aprotinin and lysine analogues in high-risk cardiac surgery. N Engl J Med 2008;358:2319 –31 8. Mangano DT, Tudor IC, Dietzel C. The risk associated with aprotinin in cardiac surgery. N Engl J Med 2006;354:353– 65 9. Mangano DT, Miao Y, Vuylsteke A, Tudor IC, Juneja R, Filipescu D, Hoeft A, Fontes ML, Hillel Z, Ott E, Titov T, Dietzel C, Levin J. Mortality associated with aprotinin during 5 years following coronary artery bypass graft surgery. JAMA 2007;297:471–9 10. Shaw AD, Stafford-Smith M, White WD, Phillips-Bute B, Swaminathan M, Milano C, Welsby IJ, Aronson S, Mathew JP, Peterson ED, Newman MF. The effect of aprotinin on outcome after coronary-artery bypass grafting. N Engl J Med 2008;358: 784 –93 11. Eberle B, Mayer E, Hafner G, Heinermann J, Dahm M, Prellwitz W, Dick W, Oelert H. High-dose epsilon-aminocaproic acid versus aprotinin: antifibrinolytic efficacy in first-time coronary operations. Ann Thorac Surg 1998;65:667–73 12. Casati V, Guzzon D, Oppizzi M, Cossolini M, Torri G, Calori G, Alfieri O. Hemostatic effects of aprotinin, tranexamic acid and epsilon-aminocaproic acid in primary cardiac surgery. Ann Thorac Surg 1999;68:2252– 6; discussion 2256 –7 Vol. 109, No. 1, July 2009

13. Menichetti A, Tritapepe L, Ruvolo G, Speziale G, Cogliati A, Di Giovanni C, Pacilli M, Criniti A. Changes in coagulation patterns, blood loss and blood use after cardiopulmonary bypass: aprotinin vs tranexamic acid vs epsilon aminocaproic acid. J Cardiovasc Surg (Torino) 1996;37:401–7 14. Penta de Peppo A, Pierri MD, Scafuri A, De Paulis R, Colantuono G, Caprara E, Tomai F, Chiariello L. Intraoperative antifibrinolysis and blood-saving techniques in cardiac surgery. Prospective trial of 3 antifibrinolytic drugs. Tex Heart Inst J 1995;22:231–6 15. Alderman EL, Levy JH, Rich JB, Nili M, Vidne B, Schaff H, Uretzky G, Pettersson G, Thiis JJ, Hantler CB, Chaitman B, Nadel A. Analyses of coronary graft patency after aprotinin use: results from the International Multicenter Aprotinin Graft Patency Experience (IMAGE) trial. J Thorac Cardiovasc Surg 1998;116:716 –30 16. Dietrich W, Schopf K, Spannagl M, Jochum M, Braun SL, Meisner H. Influence of high- and low-dose aprotinin on activation of hemostasis in open heart operations. Ann Thorac Surg 1998;65:70 –7; discussion 77– 8 17. Dietrich W, Spannagl M, Jochum M, Wendt P, Schramm W, Barankay A, Sebening F, Richter JA. Influence of high-dose aprotinin treatment on blood loss and coagulation patterns in patients undergoing myocardial revascularization. Anesthesiology 1990;73:1119 –26 18. Kaul S, Diamond GA. Good enough: a primer on the analysis and interpretation of noninferiority trials. Ann Intern Med 2006;145:62–9 19. Food and Drug Administration, HHS. International Conference on Harmonisation; choice of control group and related issues in clincal trials; availability. Notice. Fed Regist 2001;66:24390 –1 20. Horrow JC, Van Riper DF, Strong MD, Grunewald KE, Parmet JL. The dose-response relationship of tranexamic acid. Anesthesiology 1995;82:383–92 21. Roques F, Nashef SA, Michel P, Gauducheau E, de Vincentiis C, Baudet E, Cortina J, David M, Faichney A, Gabrielle F, Gams E, Harjula A, Jones MT, Pintor PP, Salamon R, Thulin L. Risk factors and outcome in European cardiac surgery: analysis of the EuroSCORE multinational database of 19030 patients. Eur J Cardiothorac Surg 1999;15:816 –22; discussion 822–3 22. Greilich PE, Brouse CF, Rinder CS, Smith BR, Sandoval BA, Rinder HM, Eberhart RC, Jessen ME. Effects of epsilonaminocaproic acid and aprotinin on leukocyte-platelet adhesion in patients undergoing cardiac surgery. Anesthesiology 2004; 100:225–33 23. Greilich PE, Brouse CF, Whitten CW, Chi L, Dimaio JM, Jessen ME. Antifibrinolytic therapy during cardiopulmonary bypass reduces proinflammatory cytokine levels: a randomized, double-blind, placebo-controlled study of epsilonaminocaproic acid and aprotinin. J Thorac Cardiovasc Surg 2003;126:1498 –503 24. Greilich PE, Okada K, Latham P, Kumar RR, Jessen ME. Aprotinin but not epsilon-aminocaproic acid decreases interleukin-10 after cardiac surgery with extracorporeal circulation: randomized, double-blind, placebo-controlled study in patients receiving aprotinin and epsilon-aminocaproic acid. Circulation 2001;104:I265–I269 25. Dentz ME, Slaughter TF, Mark JB. Early thrombus formation on heparin-bonded pulmonary artery catheters in patients receiving epsilon aminocaproic acid. Anesthesiology 1995;82:583– 6 26. Butterworth J, James RL, Lin Y, Prielipp RC, Hudspeth AS. Pharmacokinetics of epsilon-aminocaproic acid in patients undergoing aortocoronary bypass surgery. Anesthesiology 1999; 90:1624 –35 27. Curtin N, Highe G, Harris M, Braunstein A, Demattia F, Coss L. Extensive evaluation of the instrumentation laboratory IL test D-Dimer immunoturbidimetric assay on the ACL 9000 determines the D-Dimer cutoff value for reliable exclusion of venous thromboembolism. Lab Hematol 2004;10:88 –94 28. Nuttall GA, Fass DN, Oyen LJ, Oliver WC Jr, Ereth MH. A study of a weight-adjusted aprotinin dosing schedule during cardiac surgery. Anesth Analg 2002;94:283–9, table of contents 29. Despotis GJ, Joist JH, Goodnough LT, Santoro SA, Spitznagel E. Whole blood heparin concentration measurements by automated protamine titration agree with plasma anti-Xa measurements. J Thorac Cardiovasc Surg 1997;113:611–13 © 2009 International Anesthesia Research Society

23

30. Brown JR, Birkmeyer NJ, O’Connor GT. Meta-analysis comparing the effectiveness and adverse outcomes of antifibrinolytic agents in cardiac surgery. Circulation 2007;115:2801–13 31. Lasocki S, Provenchere S, Benessiano J, Vicaut E, Lecharny JB, Desmonts JM, Dehoux M, Philip I. Cardiac troponin I is an independent predictor of in-hospital death after adult cardiac surgery. Anesthesiology 2002;97:405–11 32. Karski JM, Teasdale SJ, Norman P, Carroll J, VanKessel K, Wong P, Glynn MF. Prevention of bleeding after cardiopulmonary bypass with high-dose tranexamic acid. Double-blind, randomized clinical trial. J Thorac Cardiovasc Surg 1995;110:835– 42 33. Hollander M, Wolfe DA. Nonparametric statistical methods. New York: John Wiley and Sons, 1973:75– 82 34. Hoylaerts M, Lijnen HR, Collen D. Studies on the mechanism of the antifibrinolytic action of tranexamic acid. Biochim Biophys Acta 1981;673:75– 85 35. Vander Salm TJ, Kaur S, Lancey RA, Okike ON, Pezzella AT, Stahl RF, Leone L, Li JM, Valeri CR, Michelson AD. Reduction of bleeding after heart operations through the prophylactic use of epsilon-aminocaproic acid. J Thorac Cardiovasc Surg 1996; 112:1098 –107 36. Ferraris VA, Ferraris SP, Moliterno DJ, Camp P, Walenga JM, Messmore HL, Jeske WP, Edwards FH, Royston D, Shahian DM, Peterson E, Bridges CR, Despotis G. The Society of Thoracic Surgeons practice guideline series: aspirin and other antiplatelet agents during operative coronary revascularization (executive summary). Ann Thorac Surg 2005;79:1454 – 61

24


37. Carless PA, Moxey AJ, Stokes BJ, Henry DA. Are antifibrinolytic drugs equivalent in reducing blood loss and transfusion in cardiac surgery? A meta-analysis of randomized head-to-head trials. BMC Cardiovasc Disord 2005;5:19 38. Bennett-Guerrero E, Sorohan JG, Gurevich ML, Kazanjian PE, Levy RR, Barbera AV, White WD, Slaughter TF, Sladen RN, Smith PK, Newman MF. Cost-benefit and efficacy of aprotinin compared with epsilon-aminocaproic acid in patients having repeated cardiac operations: a randomized, blinded clinical trial. Anesthesiology 1997;87:1373– 80 39. Bennett-Guerrero E, Sorohan JG, Canada AT, Ayuso L, Newman MF, Reves JG, Mythen MG. epsilon-Aminocaproic acid plasma levels during cardiopulmonary bypass. Anesth Analg 1997;85: 248 –51 40. Bridges CR. Valid comparisons of antifibrinolytic agents used in cardiac surgery. Circulation 2007;115:2790 –2 41. Myles PS, Smith J, Knight J, Cooper DJ, Silbert B, McNeil J, Esmore DS, Buxton B, Krum H, Forbes A, Tonkin A. Aspirin and Tranexamic Acid for Coronary Artery Surgery (ATACAS) Trial: rationale and design. Am Heart J 2008;155:224 –30 42. Stafford-Smith M, Lefrak EA, Qazi AG, Welsby IJ, Barber L, Hoeft A, Dorenbaum A, Mathias J, Rochon JJ, Newman MF. Efficacy and safety of heparinase I versus protamine in patients undergoing coronary artery bypass grafting with and without cardiopulmonary bypass. Anesthesiology 2005;103:229 – 40


Whole Blood Multiple Electrode Aggregometry Is a Reliable Point-of-Care Test of Aspirin-Induced Platelet Dysfunction Csilla Ja´mbor, MD* Christian F. Weber, MD‡ Konstanze Gerhardt‡ Wulf Dietrich, PhD* Michael Spannagl, PhD*† Bernhard Heindl, PhD* Bernhard Zwissler, PhD*

BACKGROUND: Aspirin is one of the most commonly ingested over-the-counter drugs. In addition to its analgesic and antiinflammatory actions, it also potently inhibits platelet aggregation. Evaluation of aspirin-induced platelet dysfunction is relevant in various clinical situations, including during complex surgeries with high bleeding risk in individuals who have ingested aspirin. In this study, we examined the suitability of multiple electrode aggregometry (MEA) for time course assessment of the antiplatelet effects of a single oral dose of 500 mg aspirin. We also determined the applicability of this method in the point-of-care (POC) setting by comparing the results of the test after different time intervals after blood sampling. METHOD: Twenty-four adult volunteers were enrolled in the study. After blood drawing at baseline, 500 mg aspirin was administered to all volunteers. Blood samples were taken at 4, 24, 56, 80, and 124 h after aspirin ingestion. At each time point, measurements were performed immediately and 30 and 60 min after drawing blood. Whole blood MEA was performed after stimulation with thrombin receptor activating peptide (TRAPtest, 32 ␮M) and arachidonic acid (ASPItest, 0.5 mM). Repeated measurement analysis of variance with a Bonferroni correction for multiple comparisons was performed to detect differences between time points. Assay imprecision was determined by calculating the coefficient of variation. The level of statistical significance was set to P ⬍ 0.05. RESULTS: Platelet aggregation by ASPItest was markedly decreased 4 h after aspirin intake. From the second day after aspirin intake, ASPItest values recovered with high interindividual variability, and 5 days after aspirin intake, ASPItest values did not differ significantly from baseline. TRAP-induced platelet aggregation (TRAPtest) showed no systematic changes during the study period. The resting time of the sample did not affect TRAPtest or ASPItest values. The coefficients of variation were 10% for the ASPItest and 7% for the TRAPtest. CONCLUSIONS: MEA reliably detected the effects of aspirin. Notably, 500 mg aspirin caused complete inhibition of arachidonic acid-induced platelet aggregation for 2 days in all volunteers. Aggregation returned to baseline values with a wide interindividual variation in time course by day 5. No resting time for the blood sample was required for ASPItest or TRAPtest. These assays can be implemented as real POC tests. The reproducibility of the assays studied here is within the range of modern POC analyzers. (Anesth Analg 2009;109:25–31)

A

spirin is one of the most commonly used drugs and can be used for prevention or treatment of cardiovascular disease, or as an over-the-counter medication for the treatment of pain and inflammation. Because of this wide-scale use, an increasing number of patients

From the *Clinic for Anesthesiology, †Department of Transfusion Medicine and Hemostaseology, University of Munich, Germany; and ‡Department of Anesthesiology, Intensive Care and Pain Medicine, Goethe-University Frankfurt am Main, Germany. Accepted for publication January 7, 2009. Supported, in part, by Dynabyte, Munich, Germany. Csilla Ja´mbor and Michael Spannagl have received speaking honoraria and research support from Dynabyte, Munich, Germany. Dr. Csilla Ja´mbor was responsible for study initiation and design, and Prof. Dr. Bernhard Zwissler approved the study protocol. Dr. Csilla Ja´mbor, Dr. Christain Weber, and Konstanze Gerhardt collected the data, Dr. Csilla Jámbor and PD Dr. Michael Spannagl analyzed the data and performed statistical analysis, and Dr. Csilla Vol. 109, No. 1, July 2009

undergoing surgery ingest aspirin preoperatively. The American College of Chest Physicians recommends that patients scheduled for coronary artery bypass grafting continue aspirin up to and throughout the time of coronary artery bypass grafting.1 This recommendation is made despite published reports of an Ja´mbor and PD Dr. B. Heindl wrote the manuscript. Dr. Christain Weber, PD Dr. Michael Spannagl, Prof. Dr. Wulf Dietrich, and Prof. Dr. Bernhard Zwissler counterchecked the manuscript. All authors approved the final version of the manuscript. Dr. Csilla Ja´mbor, PD Dr. Michael Spannagl, Prof. Dr. Wulf Dietrich, and PD Dr. Berhard Heindl were involved in the revision of the manuscript. Address correspondence and reprint requests to Dr. Csilla Ja´mbor, Clinic for Anesthesiology, University of Munich, MaxLebsche-Platz 32, D-81377 Munich, Germany. Address e-mail to [email protected]. Copyright © 2009 International Anesthesia Research Society DOI: 10.1213/ane.0b013e3181a27d10

25

increased risk of perioperative bleeding.2 Preoperative aspirin administration increases blood loss during bleeding-sensitive operations.3–5 Thus, the American College of Chest Physicians suggests that patients having noncardiac surgery who are at low risk for cardiac disease stop aspirin intake 7–10 days before surgery.1 Quantitative measurement of the antiplatelet effect of aspirin has clinical relevance, as platelet inhibition differs among individuals.6 Although laboratory aspirin resistance, a condition of inadequate inhibition of platelet function by aspirin, depends considerably upon the assessment method,7 it is associated with an increased incidence of major thromboembolic events.6 The detection of residual aspirin-related platelet dysfunction could result in postponement of complex bleeding-sensitive surgical interventions. However, the exclusion of residual aspirin effects in individuals with suspected aspirin intake or in aspirin-treated patients with unknown compliance could prevent unnecessary postponement of bleeding-sensitive surgical interventions or transfusion of platelet concentrates. Therefore, convenient methods for measuring the degree of platelet inhibition by aspirin or other platelet inhibitors are highly desirable. Multiple electrode aggregometry (MEA) (Multiplate®, Dynabyte, Munich, Germany) is a newly developed technique to test platelet function in whole blood, based on classical whole blood impedance aggregometry.8 MEA has recently been used to study the effects of aspirin,9,10 non-opioid analgesics,11 clopidogrel,12,13 anticoagulants,14 antifibrinolytics,15 colloids,16 and temperature17 on platelet aggregation. Platelet aggregation as determined by MEA is calculated from the AUC and quantified by arbitrary aggregation units over time (AU*min). A cutoff value of 300 AU*min for a therapeutic antiplatelet effect of aspirin was determined, as arachidonic acid-induced platelet aggregation was below 300 AU*min in 96% of the patients receiving long-term aspirin treatment.9 Values from aspirin-treated patients were significantly lower than those from healthy blood donors. However, thus far, no information is available regarding MEA determination of the time course of platelet inhibition after ingestion of a single high dose of aspirin. MEA does not require a specialized coagulation laboratory and may be useful for point-of-care (POC) analysis.10,13 As for most platelet function tests, a resting time of 30 min after blood sampling before testing is recommended from the manufacturer for MEA analysis. This resting time may impede immediate detection of platelet dysfunction intraoperatively. The degree of aspirin-induced platelet inhibition is of particular interest in the perioperative setting. We hypothesized that MEA could reliably detect the effect of aspirin at the POC. Therefore, we selected MEA for our study i) to assess the time course of the antiplatelet 26

Assessment of Aspirin Effects

effect before and up to 5 days after the ingestion of 500 mg aspirin, and ii) to reappraise the need of a 30-min resting interval before measurement.

METHODS Study Population After IRB approval, 24 healthy adult volunteers gave written informed consent and were included in the study. All participants denied the intake of any medication during the previous 10 days and had no bleeding history.

Blood Sampling and Intervention Blood was drawn from the antecubital vein by puncture without stasis using a 21G butterfly needle. The first 2 mL of blood was discarded. Blood was then collected into 4.5 mL tubes containing 25 ␮g/mL hirudin (Dynabyte, Munich, Germany) as an anticoagulant, according to the recommendations of the manufacturer. After blood drawing at baseline, all study participants took one tablet of 500 mg acetyl salicylic acid (Aspirin-Hexal®, Hexal AG, Holzkirchen, Germany) in the presence of a study advisor. Blood samples were then taken 4, 24, 56, 80, and 124 h after aspirin intake.

Platelet Function Assays Platelet function analysis was performed using the Multiplate analyzer, a novel whole blood impedance aggregometer (Dynabyte, Munich, Germany). The device has five MEA test cells for parallel testing, and each test cell incorporates two independent sensor units. One unit consists of two silver-coated, highly conductive copper wires. Analysis is based on platelet adhesion upon activation, a property that results in aggregation onto the metal sensor wires in the test cell, thus increasing the electrical impedance between the wires.8 For measurement, 300 ␮L of preheated saline (37°C) and 300 ␮L hirudin-anticoagulated whole blood were placed into the test cell, and the sample was stirred using a teflon-coated electromagnetic stirrer (800 rpm) over a 3-min incubation period. Platelet aggregation was initiated using arachidonic acid (AA) (ASPItest, 0.5 mM) or thrombin receptoractivating peptide (TRAP-6, TRAPtest, 32 ␮M) using reagents supplied by the manufacturer (Dynabyte, Munich, Germany). Increased impedance due to attachment of platelets to the electrodes was continuously measured for each sensor unit over a period of 6 min. Data were transformed to arbitrary AU and plotted as two separate aggregation curves versus time. Aggregation measured by MEA was quantified as the area under the aggregation curve (AUC, [AU*min]). Alternatively, the software of the analyzer allows for the expression of AUC values in [U], where 10 [AU*min] correspond to 1 [U]. The duplicate sensors served as an internal control to reduce the occurrence of systematic errors. Pearson’s correlation coefficients of the individual data points of the curves ANESTHESIA & ANALGESIA

as assessed by the two electrode pairs, and the differences between the AUC values detected by each sensor unit and the mean AUC were calculated. When the values were outside the acceptable range (correlation coefficient ⬍0.98, difference from the mean curve ⬎20%), the results were flagged and the measurement was repeated. Measurements were performed at the POC immediately before and at 4, 24, 56, 80, and 124 h after aspirin ingestion. To study the effect of the resting, both assays were also performed immediately after blood drawing (T0) and after resting intervals of 30 (T30) or 60 (T60) min.

Statistical Analysis Statistical analysis was performed using Sigma Stat software (Version 3.1, Jandel, San Rafael, CA). For sample size analysis, we defined differences in platelet aggregation between analyses at T0, T30, and T60 exceeding 15% as unacceptably high. A difference larger than 15% would suggest an impact of the sample resting time on the measurement results which would discard a real POC suitability of these assays. Thus, the minimum detectable differences in the means were 165 and 135 AU*min for TRAPtest and ASPItest, respectively (15% of the expected baseline values). To detect significant differences in platelet aggregation at T0, T30, and T60, sample size analyses (expected standard deviation 170 and 150 AU*min in TRAPtest and ASPItest, respectively, desired power 0.8, P ⬍ 0.05) resulted in sample sizes of 22 and 25 for TRAPtest and ASPItest, respectively. As the expected differences in platelet aggregation in the ASPItest before and after aspirin treatment were much higher than 15%, no sample size analysis for this objective was performed. One-way repeated measures analysis of variance was used to detect differences between the time points. If the normality test (Kolmogorov-Smirnov) failed, Friedman repeated measures analysis of variance on ranks was used. In case of significant differences in the means (or medians), a Bonferroni (or Dunn) post hoc multiple comparison procedure versus control was applied according to the distribution of the data. Variability of measurements was quantified using the mean of the standard deviations of the three consecutive measurements (T0, T30, and T60) at each measuring point and in each subject, and are expressed as a percentage of the mean values (coefficient of variation, CV, %). Results are expressed as the mean ⫾ sd or as the median. The level of statistical significance was set to P ⬍ 0.05.

RESULTS Of the 24 volunteers studied, 11 (46%) were men. The mean age was 31.1 ⫾ 3.7 yr and the mean Body Mass Index was 23.2 ⫾ 2.8 kg/m2. Bleeding and drug history were negative for all study participants. Vol. 109, No. 1, July 2009

The baseline range of MEA values in the different assays was determined as the range between the minimum and maximum values obtained from the baseline measurement 30 min after blood drawing, including the data from all volunteers. Baseline values were in the range from 596 to 1197 AU*min for ASPItest and 897-1469 AU*min for TRAPtest (Fig. 1). Four hours after aspirin ingestion each volunteer had MEA values below the cutoff of 300 AU*min for therapeutic aspirin effect in ASPItest (Fig. 1). Nonresponders were not observed. Platelet aggregation after stimulation with AA and TRAP (mean and positive sd) is shown in Figure 2.

Figure 1. Distribution of baseline values and values 4 h after aspirin ingestion in ASPItest and TRAPtest, obtained 30 min after drawing blood. The straight horizontal line shows the upper cutoff value for the full aspirin effect as determined by von Pape et al.9

Figure 2. Platelet aggregation after activation with arachidonic acid (AA) in ASPItest or with TRAP-6 in TRAPtest at T0, T30, and T60. The bars show the means, and error bars indicate the positive sd. *P ⬍ 0.001 as compared with baseline at T0, #P ⬍ 0.001 compared with baseline at T30, and §P ⬍ 0.001 as compared with baseline at T60. © 2009 International Anesthesia Research Society

27

Table 1. Mean ⫾ SD of Arachidonic Acid- and TRAP-Induced Platelet Aggregation in ASPItest and TRAPtest at all Measuring Points (Baseline, 4, 24, 56, 80, and 124 h) and Resting Intervals (T0, T30, T60), n ⫽ 24 ASPItest (AUC, 关AU*min兴) Rest Measuring point Baseline 4h 24 h 56 h 80 h 124 h

TRAPtest (AUC, 关AU*min兴)

T0 Mean ⫾ sd

T30 Mean ⫾ sd

T60 Mean ⫾ sd

T0 Mean ⫾ sd

T30 Mean ⫾ sd

T60 Mean ⫾ sd

839 ⫾ 164 137 ⫾ 44 151 ⫾ 71 207 ⫾ 144 485 ⫾ 167 711 ⫾ 138

889 ⫾ 157 155 ⫾ 44 160 ⫾ 63 245 ⫾ 126 527 ⫾ 172 798 ⫾ 156

942 ⫾ 154 168 ⫾ 39 169 ⫾ 64 228 ⫾ 129 509 ⫾ 161 829 ⫾ 156

1173 ⫾ 170 1210 ⫾ 209 1163 ⫾ 148 1209 ⫾ 143 1192 ⫾ 132 1160 ⫾ 127

1155 ⫾ 147 1204 ⫾ 144 1168 ⫾ 168 1186 ⫾ 104 1233 ⫾ 148 1167 ⫾ 146

1149 ⫾ 133 1185 ⫾ 165 1196 ⫾ 178 1171 ⫾ 99 1168 ⫾ 127 1149 ⫾ 148

T0 ⫽ immediate measurement after blood drawing; T30 ⫽ 30; T60 ⫽ 60 min after blood drawing; AUC ⫽ platelet aggregation 关AU*min兴;

The mean AA-induced platelet aggregation in ASPItest was significantly decreased 4 h after aspirin intake, with stepwise normalization of aggregation by 124 h after drug ingestion (Table 1). Aspirin-induced platelet inhibition was homogenous for up to 24 h after aspirin intake, as all study participants showed a therapeutic aspirin effect (AUC ⬍ 300 AU*min) at 4 and 24 h (Fig. 3). On day 3, at 56 h, platelet aggregation as measured by ASPItest was ⬍300 AU*min in 63% (15 of 24) of volunteers and on day 4 at 80 h in 8% (2 of 24) of the volunteers. None of the subjects showed a therapeutic antiplatelet effect of aspirin on day 6 (124 h) after drug administration. The individual course of the antiplatelet effect of aspirin was determined by calculating the individual aggregation responses at each measuring point relative to the baseline value (AUC4h/AUCbaseline*100, AUC24h/AUCbaseline*100, etc.), using the values obtained for ASPItest at T30. Aggregation response at 4 h was in the range of 7%–24% (mean 18%). At 24 h, uniform suppression of AA-induced platelet aggregation was still found, with an aggregation response in the range of 7%–35% (mean 18%). In contrast, marked interindividual variability was observed at 56 and 80 h after aspirin intake (Fig. 3). At 56 h, the individual aggregation response ranged from 9% to 61% (mean 28%), and at 80 h it ranged from 10% to 91% (mean 61%). On day 6 (124 h) after aspirin intake, the aggregation response reached in mean 89% of the baseline value (range, 47%–117%). Platelet aggregation returned to the baseline range (596 –1197 AU*min) in none of the volunteers by 56 h, in 33% by 80 h and in 88% by 124 h. Platelet aggregation in TRAPtest showed no significant changes during the 5-day study period after aspirin intake (Table 1, Fig. 2). There were no significant differences in AUC values when samples were measured immediately after drawing blood or after a 30 or 60 min rest (Fig. 2). Therefore, we treated the measurements at the three resting time points as triplicates to calculate the imprecision of the assays. We quantified this by determining the CV at the measuring points with mean values within the baseline ranges (baseline and 124 h 28


SD

⫽ standard deviation.

Figure 3. Individual platelet aggregation after stimulation with arachidonic acid (AA) (ASPItest). Each plot depicts the time course of the aspirin effect as determined by multiple electrode aggregometry (MEA) in one volunteer after a 30 min rest of the sample.

measuring points). CV was 10% for ASPItest and 7% for TRAPtest.

DISCUSSION The main findings of this study were that i) MEA reliably detected the antiplatelet effect of aspirin and AA-induced platelet aggregation returned to baseline values with a wide interindividual variation in time course by day 5, and ii) no resting time for the blood samples was required before determination of ASPItest and TRAPtest. Notably, 500 mg aspirin caused complete inhibition of AA-induced platelet aggregation in all volunteers (Fig. 1). A dose of 500 mg was selected for this study as it corresponds to the typical dose of aspirin ingested as an analgesic in Europe and is between 1 and 2 tablets of the over-the-counter United States dose (325 mg). Nonresponsiveness to aspirin was not observed in our study, probably due to the high dosage administered.18 The available prevalence data for aspirin resistance are inconsistent and vary from 5% to 80%, mainly depending on the diagnostic method used.6 In two recent studies, several assays were compared for ANESTHESIA & ANALGESIA

assessment of the effects of aspirin on platelet function,7,19 but only a moderate correlation was found between the different methods. Thus, definition of laboratory aspirin resistance demands clarification. Although 5%– 80% is a very wide range, laboratory aspirin resistance is nonetheless associated with impaired outcome inpatients with cardiovascular, cerebrovascular, or peripheral arterial diseases.6 Therefore, the clinical impact of aspirin resistance is significant, and good standardized and validated bedside tests are needed to reliably quantify the antiplatelet effect of aspirin. Some studies have linked MEA data to bleeding10,20 or thrombotic outcomes.21 A very recent investigation of 100 patients undergoing cardiac surgery suggested that preoperative ASPItest in MEA may be a more sensitive predictor of platelet transfusion than patient self-reporting on aspirin intake.10 The cutoff for an abnormal aggregation response of AUC ⬍510 AU*min determined in that study, however, is not applicable to our data because the authors used heparinized blood.10 For MEA analysis hirudin is the best investigated and standardized anticoagulant; however, some centers use heparinized blood and report reproducible results.10,22 Additionally, the manufacturer of the analyzer recommends the use of hirudin-anticoagulated blood collection tubes. Citrated samples are not recommended for MEA analysis, as this calciumchelating anticoagulant does not preserve the physiological concentrations of ionized calcium and magnesium, which are essential for platelet aggregation.8 In most studies using MEA, hirudin blood has been used9,11,12,14 –16,21,23; therefore, we also selected hirudin as the anticoagulant for this study. For the first 2 days after aspirin intake, we observed a more than 80% suppression of AA-induced platelet aggregation in all volunteers. This suppression was comparable with results reported for MEA during long-term aspirin treatment.9 We could also confirm the cutoff value for a therapeutic aspirin effect used in our study, as 300 AU*min clearly separated the baseline values from those obtained 4 h after aspirin ingestion (Fig. 1). Platelet function normalized gradually between the third and fifth days after drug ingestion. The time course of the antiplatelet effect of aspirin as assessed by MEA in our study was in accordance with results obtained via other monitoring techniques such as thromboxane B2 production24 or PFA-100,25 despite the fact that the administered doses in these studies were lower. Although various studies have confirmed a correlation of the results obtained by MEA and PFA-100 under aspirin treatment,9,13,26 assessment of platelet function by MEA might be superior in the perioperative setting. PFA-100 is not specific to the aspirinsensitive cyclooxygenase-1 (COX-1) pathway and has been reported to have limited predictivity for perioperative bleeding and transfusion requirements.27,28 Furthermore, PFA-100 closure-time has been shown Vol. 109, No. 1, July 2009

not to be reliable in cases of low platelet count (PC) or hematocrit,29 thus limiting its intraoperative usefulness inpatients with blood loss. In the present study, we observed a wide interindividual variability in the attenuation of the aspirin effect beginning at 56 h after administration (Fig. 3). Here, the possible effect of PC has to be discussed. Within normal ranges (164 –395/nL), no significant impact of PC on MEA results has been shown.9 As marked variability was observed only in ASPItest (intersubject variabilities of 58% and 33% at 56 and 80 h, respectively) but not in TRAPtest (intersubject variabilities of 10% and 11% at 56 and 80 h, respectively), differences in PC may not explain the observation of this phenomenon only in ASPItest. There are a number of further possible reasons for the variability of the aspirin effect, including those arising from patient compliance, individual platelet turnover after aspirin ingestion, the individual importance of the thromboxane A2-pathway, platelet hyperreactivity, molecular genetic variants of COX-1, and drug interactions with other COX-1 inhibitors.6,30 Considering that all study participants were young, healthy volunteers without any medications or comorbidity taking a witnessed dose of aspirin, it is likely that aspirintreated patients would show even greater variability in the time course of aspirin activity.31,32 The observed interindividual variability at 56 and 80 h after aspirin intake may have clinical implications for the perioperative management of aspirin-treated patients. At 56 h, only 63% of volunteers exhibited therapeutic suppression of platelet activity (AUC ⬍ 300 AU*min) after taking 500 mg aspirin. In these volunteers, a moderate, nontherapeutic aspirininduced platelet inhibition still has to be assumed, as platelet aggregation reached the normal range (596 –1197 AU*min) in none of these subjects at 56 h. Although the clinical relevance of this moderate aspirin effect is not known, discontinuing aspirin for more than 2 days before surgery might result in nearnormal platelet aggregation for a longer period than desirable. This could lead to unnecessary postponement of intervention or may increase the prothrombotic risk inpatients with cardiovascular diseases. Recommendations for an optimal time for cessation of aspirin treatment before surgery remain contradictory.1,25,33 The observed interindividual variability of the aspirin effect in the present study and the consequent need for individual risk assessment supports the importance of preoperative assessment of platelet function when patients have reported aspirin intake before invasive procedures. As expected, aspirin intake did not affect platelet aggregation in TRAPtest (Fig. 2). TRAP acts via the thrombin receptor, and the resulting massive thrombin stimulation of platelets bypasses the inhibition of COX-1 by aspirin. Thus, impairment of TRAPtest might be an indicator of global platelet function impairment. AA-induced platelet aggregation in © 2009 International Anesthesia Research Society

29

ASPItest can more specifically detect platelet inhibition by aspirin or other COX-1 inhibitors. Another important finding of this study was that platelet aggregation in response to AA and TRAP-6 was not affected by the resting time of the sample before testing, i.e., when samples were examined immediately or 30 or 60 min after collection. For nearly all platelet function tests, including MEA, it is recommended to wait 30 min after blood sampling before performing measurement.19 For appraisal of POC suitability, however, comparison of the results obtained after immediate measurement and different resting periods is essential. Our findings indicate that the recommended resting time of 30 min is not necessary for monitoring platelet function using ASPItest or TRAPtest. These assays can be implemented as real POC tests. The results obtained with MEA were consistent and reproducible, with assay imprecision (CV) values of 10% and 7% for ASPItest and TRAPtest, respectively; this is within the range of modern laboratory POC testing. There were some limitations to our study. First, our study was not designed to evaluate the clinical predictivity of MEA results. We could not determine whether patients with low aggregation results in ASPItest actually showed increased bleeding, or whether patients with high aggregation results in ASPItest actually showed more prothromboembolic events. Second, as mentioned we did not measure the PC, as young and healthy volunteers not taking any medications are very unlikely to have abnormal PC.34 Third, we did not compare our results to an established platelet function test other than MEA. However, this study was not intended to provide a comparison among different techniques. Although MEA is a relatively new method for whole blood impedance aggregometry, various studies have confirmed the correlation between the results obtained by MEA and light transmission aggregometry,12,13 and MEA and flow cytometry.23,35 Despite its limitations, our study has important clinical implications. We confirmed the duration of aspirin action as measured by MEA. MEA accurately detected the time-dependent antiplatelet effect of aspirin and provided reproducible platelet aggregation results. The time course of the attenuation of aspirin effects in this cohort of healthy volunteers was much more variable than the time of onset. Likewise, it is not necessary to allow blood samples to rest before analysis. Both ASPItest and TRAPtest can be determined as real POC assays immediately after blood drawing. ACKNOWLEDGMENTS The authors thank the study participants for their contribution to this research. We also thank Dr. A. Calatzis, University of Munich, for his valued commentary on this manuscript. 30


REFERENCES 1. Douketis JD, Berger PB, Dunn AS, Jaffer AK, Spyropoulos AC, Becker RC, Ansell J. The perioperative management of antithrombotic therapy: American College of Chest Physicians Evidence-Based Clinical Practice Guidelines (8th Edition). Chest 2008;133:299S–339S 2. Ferraris VA, Ferraris SP, Saha SP, Hessel EA II, Haan CK, Royston BD, Bridges CR, Higgins RS, Despotis G, Brown JR, Spiess BD, Shore-Lesserson L, Stafford-Smith M, Mazer CD, Bennett-Guerrero E, Hill SE, Body S. Perioperative blood transfusion and blood conservation in cardiac surgery: the Society of Thoracic Surgeons and The Society of Cardiovascular Anesthesiologists clinical practice guideline. Ann Thorac Surg 2007;83:S27–S86 3. Lawrence C, Sakuntabhai A, Tiling-Grosse S. Effect of aspirin and nonsteroidal antiinflammatory drug therapy on bleeding complications in dermatologic surgical patients. J Am Acad Dermatol 1994;31:988 –92 4. Nielsen JD, Holm-Nielsen A, Jespersen J, Vinther CC, Settgast IW, Gram J. The effect of low-dose acetylsalicylic acid on bleeding after transurethral prostatectomy–a prospective, randomized, double-blind, placebo-controlled study. Scand J Urol Nephrol 2000;34:194 – 8 5. Palmer JD, Sparrow OC, Iannotti F. Postoperative hematoma: a 5-year survey and identification of avoidable risk factors. Neurosurgery 1994;35:1061– 4 6. Zimmermann N, Hohlfeld T. Clinical implications of aspirin resistance. Thromb Haemost 2008;100:379 –90 7. Lordkipanidze` M, Pharand C, Schampaert E, Turgeon J, Palisaitis DA, Diodati JG. A comparison of six major platelet function tests to determine the prevalence of aspirin resistance in patients with stable coronary artery disease. Eur Heart J 2007;28:1702– 8 8. Toth O, Calatzis A, Penz S, Losonczy H, Siess W. Multiple electrode aggregometry: a new device to measure platelet aggregation in whole blood. Thromb Haemost 2006;96:781– 8 9. von Pape KW, Dzijan-Horn M, Bohner J, Spannagl M, Weisser H, Calatzis A. [Control of aspirin effect in chronic cardiovascular patients using two whole blood platelet function assays. PFA-100 and multiplate]. Hamostaseologie 2007;27:155– 60 10. Rahe-Meyer N, Winterhalter M, Hartmann J, Pattison A, Hecker H, Calatzis A, Solomon C. An evaluation of cyclooxygenase-1 inhibition before coronary artery surgery: aggregometry versus patient self-reporting. Anesth Analg 2008;107:1791–7 11. Ja´mbor C, Weber C, Lau A, Spannagl M, Zwissler B. Multiple electrode aggregometry for ex vivo detection of the anti-platelet effect of non-opioid analgesic drugs. Thromb Hemost 2009;101:207–9 12. Sibbing D, Braun S, Jawansky S, Vogt W, Mehilli J, Schomig A, Kastrati A, von Beckerath N. Assessment of ADP-induced platelet aggregation with light transmission aggregometry and multiple electrode platelet aggregometry before and after clopidogrel treatment. Thromb Haemost 2008;99:121– 6 13. Velik-Salchner C, Maier S, Innerhofer P, Streif W, Klingler A, Kolbitsch C, Fries D. Point-of-care whole blood impedance aggregometry versus classical light transmission aggregometry for detecting aspirin and clopidogrel: the results of a pilot study. Anesth Analg 2008;107:1798 – 806 14. Sibbing D, Busch G, Braun S, Jawansky S, Schomig A, Kastrati A, Ott I, von Beckerath N. Impact of bivalirudin or unfractionated heparin on platelet aggregation in patients pretreated with 600 mg clopidogrel undergoing elective percutaneous coronary intervention. Eur Heart J 2008;29:1504 –9 15. Mengistu AM, Rohm KD, Boldt J, Mayer J, Suttner SW, Piper SN. The influence of aprotinin and tranexamic acid on platelet function and postoperative blood loss in cardiac surgery. Anesth Analg 2008;107:391–7 16. Boldt J, Wolf M, Mengistu A. A new plasma-adapted hydroxyethylstarch preparation: in vitro coagulation studies using thrombelastography and whole blood aggregometry. Anesth Analg 2007;104:425–30 17. Scharbert G, Kalb M, Marschalek C, Kozek-Langenecker SA. The effects of test temperature and storage temperature on platelet aggregation: a whole blood in vitro study. Anesth Analg 2006;102:1280 – 4 18. Buerke M, Pittroff W, Meyer J, Darius H. Aspirin therapy: optimized platelet inhibition with different loading and maintenance doses. Am Heart J 1995;130:465–72


19. Paniccia R, Antonucci E, Gori AM, Marcucci R, Poli S, Romano E, Valente S, Giglioli C, Fedi S, Gensini GF, Abbate R, Prisco D. Comparison of different methods to evaluate the effect of aspirin on platelet function in high-risk patients with ischemic heart disease receiving dual antiplatelet treatment. Am J Clin Pathol 2007;128:143–9 20. Mengistu AM, Wolf MW, Boldt J, Rohm KD, Lang J, Piper SN. Evaluation of a new platelet function analyzer in cardiac surgery: a comparison of modified thromboelastography and whole-blood aggregometry. J Cardiothorac Vasc Anesth 2008;22:40 – 6 21. Muller-Schunk S, Linn J, Peters N, Spannagl M, Deisenberg M, Bru¨ckmann H, Mayer TE. Monitoring of clopidogrel-related platelet inhibition: correlation of nonresponse with clinical outcome in supra-aortic stenting. AJNR Am J Neuroradiol 2008;29:786 –91 22. Breugelmans J, Vertessen F, Mertens G, Gadisseur A, Van der Planken M. Multiplate whole blood impedance aggregometry: a recent experience. Thromb Haemost 2008;100:725– 6 23. Siller-Matula JM, Lang I, Christ G, Jilma B. Calcium-channel blockers reduce the antiplatelet effect of clopidogrel. J Am Coll Cardiol 2008;52:1557– 63 24. Patrono C, Ciabattoni G, Pinca E, Pugliese F, Castrucci G, De Salvo A, Satta MA, Peskar BA. Low dose aspirin and inhibition of thromboxane B2 production in healthy subjects. Thromb Res 1980;17:317–27 25. Cahill RA, McGreal GT, Crowe BH, Ryan DA, Manning BJ, Cahill MR, Redmond HP. Duration of increased bleeding tendency after cessation of aspirin therapy. J Am Coll Surg 2005;200:564 –73 26. Mueller T, Dieplinger B, Poelz W, Haltmayer M. Utility of the PFA-100 instrument and the novel multiplate analyzer for the assessment of aspirin and clopidogrel effects on platelet function in patients with cardiovascular disease. Clin Appl Thromb Hemost 2008;2008; [Epub ahead print]

Vol. 109, No. 1, July 2009

27. Cammerer U, Dietrich W, Rampf T, Braun SL, Richter JA. The predictive value of modified computerized thromboelastography and platelet function analysis for postoperative blood loss in routine cardiac surgery. Anesth Analg 2003;96:51–7 28. Fattorutto M, Pradier O, Schmartz D, Ickx B, Barvais L. Does the platelet function analyser (PFA-100) predict blood loss after cardiopulmonary bypass? Br J Anaesth 2003;90:692–3 29. Harrison P, Robinson MS, Mackie IJ, Joseph J, McDonald SJ, Liesner R, Savidge GF, Pasi J, Machin SJ. Performance of the platelet function analyser PFA-100 in testing abnormalities of primary haemostasis. Blood Coagul Fibrinolysis 1999;10:25–31 30. Cattaneo M. Aspirin and clopidogrel: efficacy, safety, and the issue of drug resistance. Arterioscler Thromb Vasc Biol 2004;24:1980 –7 31. Buchanan MR, Brister SJ. Individual variation in the effects of ASA on platelet function: implications for the use of ASA clinically. Can J Cardiol 1995;11:221–7 32. Schwartz KA, Schwartz DE, Ghosheh K, Reeves MJ, Barber K, DeFranco A. Compliance as a critical consideration in patients who appear to be resistant to aspirin after healing of myocardial infarction. Am J Cardiol 2005;95:973–5 33. Komatsu T, Tamai Y, Takami H, Yamagata K, Fukuda S, Munakata A. Study for determination of the optimal cessation period of therapy with anti-platelet agents prior to invasive endoscopic procedures. J Gastroenterol 2005;40:698 –707 34. Koscielny J, Ziemer S, Radtke H, Schmutzler M, Pruss A, Sinha P, Salama A, Kiesewetter H, Latza R. A practical concept for preoperative identification of patients with impaired primary hemostasis. Clin Appl Thromb Hemost 2004;10:195–204 35. Mueller T, Dieplinger B, Poelz W, Calatzis A, Haltmayer M. Utility of whole blood impedance aggregometry for the assessment of clopidogrel action using the novel Multiplate analyzer– comparison with two flow cytometric methods. Thromb Res 2007;121:249 –58


31

Echo Rounds

Aortic Sinotubular Ridge Calcification: A Common Transesophageal Echocardiography Finding with Uncertain Implications Martin M. Stechert, MD, PhD Martin J. London, MD

A

n 82-yr-old man with arterial hypertension, coronary artery disease, and critical aortic stenosis was scheduled for aortic valve replacement. The intraoperative transesophageal echocardiography (TEE) examination was notable for an estimated aortic valve peak pressure gradient of 116 mm Hg, calculated aortic valve area of 0.3 cm2 by the continuity equation, concentric left ventricular hypertrophy (16 mm left ventricular diastolic, septal and posterior wall thickness) and an impaired relaxation pattern based on transmitral Doppler examination. From the midesophageal TEE aortic long-axis view, a 1 ⫻ 0.5 cm, dense, echogenic, apparently calcified structure was seen at the sinotubular junction (Fig. 1, panel A, Video clip; please see video clips available at www.anesthesiaanalgesia.org). The same structure could be found using the midesophageal aortic short-axis view (Fig. 1, panel B, Video clip 2; please see video clips available at www.anesthesia-analgesia.org) and deep transgastric long-axis view (Fig. 1, panel C, Video clip 3; please see video clips available at www.anesthesia-analgesia. org). In the midesophageal short-axis view, a calcified “hump” was noted, adjacent to the anterior aortic wall between the left and right coronary cusp. Finally, on the deep transgastric long-axis view, the object was imaged at the anterior-lateral portion of the ascending This article has supplementary material on the Web site: www.anesthesia-analgesia.org.

From the Department of Anesthesiology and Perioperative Medicine, University of California, San Francisco, California. Accepted for publication January 7, 2009. Martin J. London is Section Editor of Cardiovascular and Thoracic Education for the Journal. This manuscript was handled by Charles W. Hogue, Jr., Associate Editor-in-Chief for Cardiovascular Anesthesiology, and Dr. London was not involved in any way with the editorial process or decision. Address correspondence and reprint requests to Martin M. Stechert, MD, PhD, Department of Anesthesiology (129), University of California, San Francisco, VA Medical Center, 4150 Clement St., San Francisco, CA 94121. Address e-mail to stechertm@anesthesia. ucsf.edu. Copyright © 2009 International Anesthesia Research Society DOI: 10.1213/ane.0b013e3181a27f8a

32

aorta at the sinotubular junction. The typical location and isolation of the structure from the aortic valve apparatus allowed for the identification of this finding as aortic sinotubular ridge calcification (ASTRIC). Calcification of portions of the transitional ridge that separates the sinus of Valsalva and the tubular segments of the ascending aorta has been described on autopsy. Tveter and Edwards1 demonstrated that these lesions were distinct from aortic atheromatous disease and histologically similar to aortic valve calcifications. Other investigators were able to demonstrate that ASTRIC nevertheless seems to be a marker for severe aortic atheromatous plaque formation.2 In a study of 101 patients with risk factors for embolic disease undergoing a TEE examination, calcification of the sinotubular junction was highly associated with total atherosclerosis (78% vs 44%, P ⫽ 0.019), more complex atherosclerosis (67% vs 34%, P ⫽ 0.021), and more ulcerated plaques (39% vs 10%, P ⫽ 0.0014) of the thoracic aorta than in patients without calcification. This relationship was found even more significantly different when only complex lesions of the aortic arch were investigated (61% vs 15%, P ⬍ 0.001). Microscopic examination revealed evidence of coronary flow obstruction by ASTRIC and coronary embolic material resembling the histology of ASTRIC. A predilection of the right sinus of Valsalva was noted, although, less frequently, multiple lesions involving two or all three sinus of Valsalva were seen.1 The preference for the right sinus of Valsalva was also noted by D’Cruz et al.3 in a study of 33 patients undergoing routine transthoracic echocardiography. Using the parasternal long-axis view, anterior sinotubular ridge calcification, corresponding to the right sinus of Valsalva, was noted in 19 of the 33 patients, with anterior and posterior lesions seen in 10 patients. Additional calcification of the aortic valve cusps was seen in 23 patients and mitral annulus calcification noted in 11 patients. Using TEE, ASTRIC is frequently found in patients with both aortic valve disease and coronary artery disease. From the midesophageal aortic long-axis view, the lesion is mostly found on the distal aortic Vol. 109, No. 1, July 2009

Figure 2. Midesophageal aortic long-axis view (ME-LAX at 127°). In a different patient, two separate lesions (white arrows) on opposing aortic walls are noted. This corresponds to the anterior and posterior aortic sinotubular ridge calcification (ASTRIC) location on transthoracic echocardiography.

Figure 3. Sagittal plane of an adult heart corresponding to

Figure 1. Upper Panel: Midesophageal aortic long-axis view (ME-LAX at 124°). An echo-dense structure (white arrow) is noted at the sinotubular junction of the ascending aorta (AA) in the far-field of the transesophageal echocardiography (TEE) scan, corresponding to an anterior location on transthoracic echocardiography (TTE). (LA ⫽ left atrium). Middle Panel: Midesophageal aortic short-axis view (ME-SAX). The same structure (white arrow) is found midway between the left and right sinus of Valsalva at approximately 5 o’clock adjacent to the ascending aortic wall and in proximity to the right ventricular outflow tract (RV). Lower Panel: Deep transgastric view at 0°. Once again, aortic sinotubular ridge calcification (ASTRIC) (white arrow) is recognized at the sinotubular junction of the ascending aorta.

sinotubular junction, corresponding to the right sinus of Valsalva (Fig. 1, Panel A and Fig. 2).4 The threedimensional localization can be further specified using the midesophageal aortic short-axis views (Fig. 1, Panel B). Deep transgastric views can be used to image ASTRIC, although the lesions are likely to be in the far-field with this technique (Fig. 1, Panel C). Anterior and posterior ASTRIC can be seen with TEE at both the proximal and distal sinotubular junctions (Figs. 2 and 3, Video clip 4; please see video clips available at Vol. 109, No. 1, July 2009

parasternal long-axis view on transthoracic echocardiography (TTE) (and thereby an axial mirror image on long-axis aortic view on transesophageal echocardiography [TEE]). Aortic valve with right sinus of Vasalva with right coronary ostium (R); posterior (P) “noncoronary” sinus of Vasalva; ascending aorta (Ao). The sinotubular ridge is apparent as a fine line above the coronary ostium and not calcified in the specimen shown. In the magnified sector image (left upper corner), the ridge has been digitally enhanced for better illustration. RVO ⫽ right ventricular outflow; LV ⫽ left ventricle; PM ⫽ posterior-medial papillary muscle; CS ⫽ coronary sinus; SVC ⫽ superior vena cava; whit arrows ⫽ pulmonary vein orifices; *Oblique sinus. Modified, with permission of the McGraw-Hill Companies, from Fuster V, Alexander RW, O’Rourke RA. Hurt’s the heart. 11th ed. McGraw-Hill, 2004:60.

www.anesthesia-analgesia.org). Color-flow or continuousflow Doppler evaluation should be performed to evaluate for obstructed coronary artery or ascending aortic flow. If a pressure gradient in the ascending aorta is present, which is not related to a stenotic aortic valve, the alternate diagnosis of supravalvular aortic stenosis should be considered. However, patients are usually younger and lesions more diffusely distributed.5 When cardiopulmonary bypass with aortic cannulation is planned, the finding of ASTRIC should lower the threshold for additional investigation of the aorta © 2009 International Anesthesia Research Society

33

by epiaortic ultrasound due to the frequent association with aortic arch atheromatous disease. Additional concern should be raised for the potential of coronary flow obstruction and embolization, in particular, when intravascular aortic instrumentation (catheterization, endo-aortic clamping) is planned. REFERENCES 1. Tveter KJ, Edwards JE. Calcified aortic sino-tubular ridge: a source of coronary ostial stenosis or embolism. J Am Coll Cardiol 1988;12:1510 –14

34

Echo Rounds

2. Finkelhor RS, Youssefi ME, Mohan SK, Bahler RC. Aortic sinotubular calcium as a marker of severe aortic atherosclerosis. Am J Cardiol 1998;82:1549 –52 3. D’Cruz IA, Calderon E, Clark R. Transthoracic echocardiographic visualization of calcification of the sinotubular ridge of the ascending aorta. Echocardiography 1998;15:425– 8 4. Khouzam RN, Minderman DP, D’Cruz IA. Aortic sinotubular ridge calcification: a recently recognized site of cardiac calcification. Echocardiography 2006;23:260 –2 5. Burch TM, McGowan FX, Kussman BD, Powell AJ, DiNardo JA. Congenital supravalvular aortic stenosis and sudden death associated with anesthesia. Anesth Analg 2008;107:1848 –54


Echo Rounds

Failure to Wean from Cardiopulmonary Bypass Due to Left Atrial Compression by Periaortic Hematoma Jayant Nick Pratap, MA, MRCPCH, FRCA Angus I. McEwan, FRCA

A

6-mo-old male presented for reoperative surgery after a prior incomplete arterial trunk repair. Full repair had not been possible in the neonatal period owing to intrauterine growth restriction (2.15 kg at 38 wk gestation), low cardiac output state, ventilator dependence, and unfavorable anatomy, including interrupted aortic arch and origin of the common trunk from the right ventricle (RV) (Fig. 1). The neonate had undergone reconstruction of the aortic arch and formation of a RV to pulmonary artery conduit (Fig. 2). Subsequently, at 2-mo-of-age, percutaneous stent angioplasty of the descending aorta was performed for obstruction at the site of interrupted arch repair. The truncal valve was dysplastic and bicuspid with mild regurgitation but no stenosis. For biventricular repair at 6 mo (weight 6.8 kg), surgery was performed with the aid of cardiopulmonary bypass (CPB) for 187 min and deep hypothermic circulatory arrest (DHCA) at 18°C for 41 min. The procedure included ventricular septal defect repair with bovine pericardium, further reconstruction of the aortic arch necessitating partial removal of the previously implanted stent and replacement of the RVpulmonary artery conduit with a valved homograft. After rewarming and correction of electrolyte abnormalities, bypass was weaned while infusing milrinone at 0.5 ␮g 䡠 kg⫺1 䡠 min⫺1 and epinephrine at up to 0.1 ␮g 䡠 kg⫺1 䡠 min⫺1. Despite administration of fluid boluses, the heart both filled and ejected poorly. On the third attempt bypass was discontinued but the patient remained hypotensive. Shortly afterwards, while modified ultrafiltration was in progress, the venous reservoir This article has supplementary material on the Web site: www.anesthesia-analgesia.org. From the Department of Anesthesia, Great Ormond St. Hospital, London, UK. Accepted for publication December 17, 2008. Reprints will not be available from the author. Address correspondence to Angus I. McEwan, FRCA, Department of Anesthesia, Great Ormond St. Hospital, Great Ormond St., London WC1N 3JH, UK. Address e-mail to [email protected]. Copyright © 2009 International Anesthesia Research Society DOI: 10.1213/ane.0b013e3181a210fc

Vol. 109, No. 1, July 2009

Figure 1. Schematic representation of Collett and Edwards Type II common arterial trunk with subtruncal ventricular septal defect (VSD) and Type B interruption of a left aortic arch. LA ⫽ left atrium; RA ⫽ right atrium; LV ⫽ left ventricle; RV ⫽ right ventricle; Asc Ao ⫽ ascending aorta; Desc Ao ⫽ descending aorta; LPA ⫽ left pulmonary artery; RPA ⫽ right pulmonary artery; IA ⫽ innominate artery; LCCA ⫽ left common carotid artery; LSA ⫽ left subclavian artery; Duct ⫽ arterial duct.

Figure 2. Schematic of neonatal palliative surgery including reconstruction of the aortic arch using a left subclavian arterial flap and pulmonary homograft, and connection of the right ventricle to the reconstructed pulmonary arteries using a 5 mm valveless Gore-tex conduit (RV-PA). LA ⫽ left atrium; RA ⫽ right atrium; LV ⫽ left ventricle; RV ⫽ right ventricle; Ao ⫽ aorta; LPA ⫽ left pulmonary artery; RPA ⫽ right pulmonary artery; IA ⫽ innominate artery; LCCA ⫽ left common carotid artery.

volume decreased much faster than expected, but no hemorrhage was apparent at the operative site. Transesophageal echocardiography (TEE) had been performed before the bypass run, but the probe was 35

Figure 3. Midesophageal four-chamber view (left) and with color flow Doppler of mitral valve (right) showing periaortic hematoma (white arrowheads) compressing the left atrium after cessation of cardiopulmonary bypass. Left upper pulmonary vein flow (red) marked with white arrow.

removed before DHCA and we were unable to reinsert it easily during rewarming. A further, now successful attempt, made to investigate continuing hypotension, revealed an area of low but homogenous echogenicity around the esophagus, located outside the heart but compressing the left atrium and impeding filling of left atrium and left ventricle (Fig. 3; Video Clip 1; please see video clip available at www.anesthesia-analgesia.org). Initially, the echocardiographic appearance was mysterious, as surgery had not been performed within this part of the thorax. We feared that the TEE probe had traumatized the esophagus with subsequent bleeding. However, CPB was reestablished, and reexploration showed perforation of the descending aorta at the site of removal of the previous aortic stent. Extravasating blood had formed a periaortic hematoma tracking through the posterior mediastinum. Using CPB for a further 188 min, including a 58-min interval of DHCA, the descending aorta was repaired, the hematoma was evacuated and the patient weaned from bypass. The sternum was closed 2 days after the surgery. A few weeks in the intensive care unit were required for management of initially low cardiac output and sequela, then left diaphragmatic paresis requiring plication, pleural effusion, and a narrow complex tachycardia, before the child was discharged home. Compression of isolated cardiac chambers by extrinsic hematoma is much less common than cardiac tamponade, and is a rare cause of failure to wean from CPB. Bleeding may be within the myocardium1 or outside,2 including into the pericardial sac where localized hematoma may form in the presence of adhesions.3 The left3 or right4 atrium may be subject to compression, as each is distended by a lower pressure than the ventricles and the arteries they perfuse. In congenital heart disease surgery, an extracardiac Fontan conduit is another possible site of compression.2 Echocardiography may mislead with images wrongly 36

Echo Rounds

interpreted as showing intraluminal thrombus.1,3,4 In the postoperative period such appearances have prompted anticoagulation, which may cause enlargement of what is, in fact, an extraluminal hematoma. In the case described we postulate that, despite an open mediastinum, blood at systemic arterial pressure was confined behind the left atrium by adhesions resulting from previous cardiotomy, distorting the low pressure left atrium and pulmonary veins such as to reduce venous return from the lungs. The clinical- and cost-effectiveness of intraoperative TEE in congenital heart disease surgery have been confirmed previously,5 but its use is less widespread than in adult practice, and varies among institutions. Several studies have found a significant number of cases in which repeat bypass runs are required for unsuspected abnormalities, in addition to a group in which changes in medical management are indicated. Routine use has therefore been advocated but may be limited by lack of equipment or operators. Although early studies were led primarily by cardiologists, appropriately experienced anesthesiologists may also provide a competent service with cardiology backup used in a small minority of cases. Some centers prefer to have one anesthesiologist dedicated to TEE alone for the duration of the case.5 TEE may itself lead to complications, particularly in small infants where the bulk of even the smallest commercially available probe is substantial relative to the size of the structures traversed. Indeed, it is our institutional practice both to avoid routine intraoperative TEE in children weighing ⬍5 kg and to remove the probe during DHCA in small children. Esophageal hematoma resulting in significant cardiac compression has been reported after thrombolysis6 but, to our knowledge, has not been reported in association with TEE. This case exemplifies the diagnosis in pediatric practice by TEE of a remediable, structural cause of ANESTHESIA & ANALGESIA

failure to wean from CPB which might otherwise have been overlooked, as it was both distant from the site of surgery and in a location difficult to inspect off bypass. REFERENCES 1. Momenah TS, McElhinney DB, Brook MM, Teitel DF, Hanley FL, Silverman NH. Intramyocardial hematoma causing cardiac tamponade after repair of Ebstein malformation: erroneous echocardiographic diagnosis as intracavitary thrombus. J Am Soc Echocardiogr 1998;11:1087–9 2. Lyons JM, Duffy JY, Manning PB, Pearl JM. Compression of an extracardiac fontan following classic fontan revision. J Card Surg 2004;19:254 –7

Vol. 109, No. 1, July 2009

3. Gologorsky E, Gologorsky A, Galbut DL, Wolfenson A. Left atrial compression by a pericardial hematoma presenting as an obstructing intracavitary mass: a difficult differential diagnosis. Anesth Analg 2002;95:567–9 4. Peterson MJ, Havemann LM, McKenzie ED, Miller-Hance WC. Unusual presentation of postcardiotomy hemorrhage in an infant with congenital heart disease. Anesth Analg 2005;100:1267– 8 5. Bettex DA, Preˆtre R, Jenni R, Schmid ER. Cost-effectiveness of routine intraoperative transesophageal echocardiography in pediatric cardiac surgery: a 10-year experience. Anesth Analg 2005;100:1271–5 6. Nault I, Bertrand OF. Severe haemodynamic compromise due to left atrial compression by oesophageal haematoma. Heart 2007;93:1190


37

Case Report

Intraoperative Autologous Transfusion of Hemolyzed Blood Tod B. Sloan, MD, MBA, PhD* Greg Myers, MD* Daniel J. Janik, MD* Evalina M. Burger, MD†

During two cases of lumbar spine surgery with instrumentation, we used intraoperative autologous transfusion (IAT), resulting in hemolysis during collection and hemoglobinuria and coagulation abnormalities after transfusion. Hemolysis during IAT collection can lead to hemoglobinuria and binding of nitric oxide, leading to vasoconstriction. The literature suggests that stroma from damaged cells and contact of the blood with the IAT device can lead to coagulation abnormalities and other morbidities, including adult respiratory distress syndrome. (Anesth Analg 2009;109:38 –42)

Vikas V. Patel, MD† Leslie C. Jameson, MD*

I

ntraoperative autologous transfusion (IAT) of washed blood collected from the operative field has been applied in several types of surgery, including major spine procedures. The major advantages of IAT when compared with allogeneic blood include rapid availability, reduction of exposure to infectious agents transmitted by homologous blood transfusion, and a decrease in immune modulation. Furthermore, IAT has been shown to be less costly than allogeneic transfusion in surgeries in which the blood loss exceeds 1000 mL or the infused IAT blood exceeds 750 mL.1,2 Although a large number of studies demonstrate the value of this technique in cardiac and vascular surgery, the process is not without risk and may not be advantageous for all orthopedic procedures. We recently had two cases that highlight the potential complications of infusing IAT blood when hemolysis has occurred during IAT collection. Despite a properly functioning IAT device and experienced machine operator, IAT led to hemoglobinuria and coagulopathy with the associated medical complications which raise questions about the overall value of IAT in these spine procedures.

CASE DESCRIPTIONS Case 1 The IRB determined permission is not required for this report. Permission to present cases was obtained from both patients. A 55-yr-old woman with a history of scoliosis and From the Departments of *Anesthesiology, and †Orthopedics, University of Colorado Denver, Aurora, Colorado. Accepted for publication February 25, 2009. Address correspondence and reprint requests to Tod B. Sloan, MD, MBA, PhD, Department of Anesthesiology, University of Colorado at Denver and Health Sciences Center, Anschutz Office West (AO1), PO Box 6511, 12631 E 17th Ave., Aurora, CO 80045. Address e-mail to [email protected]. Copyright © 2009 International Anesthesia Research Society DOI: 10.1213/ane.0b013e3181a335e4

38

hardware placement from T1 to S1 presented for L4 –5 hardware removal and reinstrumentation. She had no coexisting medical problems, was not taking any prescription or herbal medications known to interact with transfused products or hemostasis. General anesthesia with intraoperative neuromonitoring was conducted. After midazolam sedation, anesthesia and intubation of the trachea was accomplished with propofol, sufentanil, and vecuronium. Anesthesia was maintained with small-dose desflurane and infusions of dexmeditomidine and propofol to provide stable neuromonitoring responses. Approximately 3.5 h after induction the first IAT unit (200 mL) was transfused. Within 5–10 min, the patient had a brief episode of hypertension (systolic blood pressure exceeding 200 and diastolic blood pressure exceeding 120 mm Hg) with bradycardia, which recurred in 10 –15 min. Thirty-four minutes after the IAT transfusion, hemoglobinuria was noticed and free hemoglobin (Hb) was measured in the plasma at 96 mg/dL (normal ⬍5 mg/dL). The patient was treated with sodium bicarbonate, furosemide, and fluids to induce a diuresis. Blood drawn during the hypertension failed to form a hard clot after 1 h. A thromboelastogram (TEG威) done at that time was abnormal (␣ angle ⫽ 72.4 [normal 53– 67], coagulation index ⫺3.9 [normal ⫺3 to 3], coagulation time 6.7 [normal 3– 6], maximum amplitude 36.3 [normal 59 – 68], reaction time 14.8 [normal 10 –14]) and bleeding was noted from previously hemostatic wound edges and needle puncture sites. The platelet count was recorded at 135,000/␮L [normal 150,000 – 400,000] and serum calcium was 9.2 mg/dL [normal 8.5–10.3]. Visual inspection of a second IAT unit and residual from the first unit revealed gross hemolysis, indicating that hemolysis had occurred during the IAT process. The second IAT unit was not administered and the patient subsequently received 200 mg of calcium chloride, 2 U of packed red blood cells (PRBC), and 2 U of fresh frozen plasma (FFP) to treat coagulopathy and 1500 mL blood loss. The patient made an uneventful recovery. The cell saver device (Brat II, Cobe Cardiovascular, Quedgeley, UK) was inspected and found to be within company-specified quality assurance limits (same for Case 2).

Case 2 A 46-yr-old man had posterior T10 to S1 spinal osteotomy with instrumentation and fusion after removal of old hardware from T8 to S1. He had no co-existing medical problems, Vol. 109, No. 1, July 2009

DISCUSSION

Figure 1. Urine collected during Case 2. Shown is the urine collected from the start of the case (7:30) until 10:00 (left), the urine collected between 10:00 and noon, the urine collected between noon and 14:00, and the urine collected between 14:00 and 16:00. Shown is the progressive darkening of the urine associated with hemoglobinuria and the initial resolution after 14:00 with diuresis. was not taking any prescription or herbal medications known to interact with transfused products or the coagulation system or known to cause discoloration of the urine. General anesthesia with intraoperative neuromonitoring was conducted. After midazolam sedation, induction and intubation of the trachea were performed at 0730 using propofol, fentanyl, and vecuronium. Anesthesia was maintained with infusions of propofol, ketamine, and sufentanil to provide stable neuromonitoring responses. The initial TEG威 obtained before IAT was normal (␣ angle ⫽ 53.1, coagulation index ⫺4, coagulation time 5.3, maximum amplitude 55.7, reaction time 18.3). During the first 2.5 h, his urine output was clear (Fig. 1, 10:00), serum calcium was 8.6 mg/dL and platelet count of 139,000/␮L. The blood loss was approximately 400 mL/h; however, after the transfusion of a 300 mL IAT blood, there was a marked increase in the blood loss to approximately 600 mL/h reaching a maximum of 1600 mL/h at the end of the procedure. IAT blood continued to be transfused during surgery. By 4.5 h hemaglobinuria was documented and apparent on visual inspection (Fig. 1, noon). There was a progressive darkening of the urine (Fig. 1, 1400) and increasing coagulopathy. A TEG威 done at 1335 was abnormal (␣ angle ⫽ 38.5, coagulation index ⫺4.7, coagulation time 9.7, maximum amplitude 48.6, reaction time 12.2) and a coagulation profile shortly, thereafter, was also abnormal (prothrombin time (PT) ⫽ 16.9 s [normal ⫽ 12–14.6], partial thromboplastin time (PTT) ⫽ 33.7 s [normal ⫽ 23.8 –33.5], international normalized ratio (INR) ⫽ 1.4 [normal ⫽ 0.9 –1.1]). One gram of calcium chloride in divided doses, crystalloids, furosemide, mannitol, and additional blood products were administered. At the conclusion of surgery the coagulation profile demonstrated a severe coagulopathy (TEG威-␣ angle ⫽ 43.2, coagulation time ⫽ 9.3, maximum amplitude ⫽ 40.7, reaction time ⫽ 80; PT ⫽ 20 s, PTT ⫽ 48.8 s; INR of 1.2; fibrinogen ⫽ 37 mg/dL [normal 150 – 400]). He received 4500 mL IAT blood, 683 mL FFP, 2 U of PRBC, 251 mL platelets (8 U) for an estimated 6900 mL blood loss. His Hb ended at 10.1 gm/dL from an initial 15.2. Postoperatively in the intensive care unit, his laboratory values were consistent with disseminated intravascular coagulation (DIC) (PT ⫽ 27.3, PTT ⫽ 56.4, INR ⫽ 2.3, fibrinogen ⫽ 57 and DDimer ⫽ 9640 mg/dL [normal ⬍500]). He made a full recovery. Vol. 109, No. 1, July 2009

Both of these patients developed hemoglobinuria and abnormal clotting shortly after the first unit of IAT blood. Evaluation of the IAT processed blood revealed gross hemolysis before transfusion. Both patients had unanticipated intensive care unit stays. In Case 1, coagulation abnormalities and hemoglobinuria were noted after the first 200 mL of IAT blood, suggesting the hemolyzed blood contributed to both of these problems. In Case 2, the patient developed hemoglobinuria and DIC after transfusion of IAT blood. In this case the hemolyzed blood likely contributed to the early onset of hemoglobinuria, but due to marked blood loss it is unclear to what extent the transfusion of the hemolyzed blood contributed to the ultimate development of DIC. Unfavorable effects of IAT have been reported.3–5 In one study of 56 patients, the use of IAT was associated with a higher incidence of respiratory complications (40% vs 12.5%), greater transfusion of blood (5.8 vs 3.6 U) and FFP (4.4 vs 1.5 U), and a longer hospital stay (9.3 vs 5.6 days) than in patients who did not receive IAT.4 Of note, the amount of banked blood transfused was disproportionately increased in patients receiving more than 3000 mL of IAT. The authors recommended limiting IAT transfusions to 3000 mL. In another study of IAT in lumbar spine fusion surgery, the number of postoperative transfusions was not reduced such that the authors concluded IAT was neither necessary nor cost-effective.5 However, it is worthy to note that larger volumes of infused blood may also be associated with other factors leading to increased postoperative morbidity (e.g., longer surgery, more complex operations, more comorbidities). The IAT process may lead to these complications. Shed blood is suctioned from the operative field into heparinized or citrated isotonic crystalloid where cell trauma can occur. After filtration to remove large debris, including aggregates of white blood cells, platelets, and tissue fragments, the solution undergoes a centrifugal washing which causes red blood cells (RBCs) and materials of similar density to collect on the outer wall of the bowl, whereas less dense materials are washed away by a saline solution. Therefore, the blood may receive trauma during collection, is exposed to plastic surfaces, and unwanted products may be inadequately washed in the centrifugal washing device. Cell trauma and hemolysis are inevitable with collection. In a study of five IAT devices when used in cardiac surgery, free Hb in the collection reservoir ranged between 49.9 and 4689.9 mg/L (mean of 651 mg/L), demonstrating hemolysis with the collection process.6 The free Hb in the washed processed cells was between198.2 and 2645 mg/L (mean of 705 mg/L), demonstrating that the centrifugal processing and washing does not remove all stroma-free hemoglobin. © 2009 International Anesthesia Research Society

39

Hemolysis during collection may be caused by several mechanisms, and these may be more of a problem during spine surgery accounting for a difference in outcome between these applications. Since larger diameter suction devices and suctioning of large pools of blood are used in cardiac surgery and major joint replacement surgery, the effects of collection may be less traumatic on the cellular components than with spine surgery in which a smaller diameter device is used to “skim” blood from a large surface in the operative field in spine surgery. This technique causes more mechanical trauma and aspirates air, increasing the amount of blood-air interface which promotes cell rupture.7 Some studies have shown that dilution of the blood before suctioning results in less cell damage; however, the advantages of this technique have not been fully explored.7 Hemolysis can also occur secondary to the use of hypotonic irrigation fluids, excessive suction pressure (⬎100 mm Hg), and aspiration of clotted blood. Cell rupture can also be caused by the collection of the blood with povidineiodine, hydrogen peroxide, alcohol or bone cement.8 None of these factors was present in these two cases. The levels of free Hb in IAT may exceed 2000 mg/L, which are higher than in allogeneic PRBC (typically 2–20 mg/L).9 In addition, there are many more noncellular components in IAT than PRBC, making the total dose of free Hb and RBC stroma higher. Unlike potassium, the density of free Hb and ruptured RBC stroma causes some of these materials to be retained in the centrifugal washing device with the RBCs and included in the product to be transfused.9 The observation of gross hemolysis in the IAT units of Case 1 and the elevated serum-free Hb confirmed that hemolysis occurred. When infused, free Hb is rapidly bound to haptoglobin forming a complex which is removed by monocytes and macrophages.10 However, larger doses exhaust the haptoglobin levels in patients and free Hb rapidly increases in the plasma and then in the urine.11 The circulating free Hb binds nitric oxide (NO) producing nitrate and methemoglobin. The Hb-NO complex is thought to be the most important pathway for eliminating NO bioactivity.10,12 Since NO regulates vascular tone, smooth muscle tone and platelet activation, the reduction in NO can lead to hypertension, smooth muscle dystonias, gastrointestinal contractions, erectile dysfunction, and increased clot formation.10,12 The binding of NO by free Hb is 600 –1000 fold more than Hb confined within RBCs, and free Hb can diffuse into the subendothelial space, placing it closer to the source of NO thereby having a more profound effect on NO-dependent functions.10,13,14 The loss of NO results in vasoconstriction and a dose-dependent increase in systolic and diastolic blood pressure in the systemic and pulmonary circulation.10 These hypertensive effects (such as those seen in Case 1) have been described as the primary cause of morbidity and mortality with stroma-free Hb as a 40

Case Report

blood substitute such that the newest generation of Hb substitutes contain mutations designed to minimize the interaction with NO.12 These effects are also thought to play a role in the clinical problems and organ dysfunction seen in patients with paroxysmal nocturnal hemoglobinuria, sickle cell disease, thalassemias, malaria, and cardiopulmonary bypass.10,12–14 The reduction of NO increases the potential for thrombus formation.10 In addition, free Hb and its breakdown product (heme) promote procoagulant effects, including platelet activation and release of inflammatory substances, that can contribute to vascular obstruction.10 Intravascular hemolysis also occurs after transfusion of blood. This is most frequently due to attack by antibodies and complement from transfusion of ABOincompatible blood, although antibody-mediated destruction can occur extravascularly in the reticuloendothelial system. Similar to transfusion of already hemolyzed blood, renal injury and coagulation abnormalities are the major consequences. The activation of the intrinsic clotting system is thought to occur through erythrosin released with RBC stroma.15 Hypotension is common and thought to be due to the vasodilatation effect of bradykinin produced by activation of the kallikrein system.15 Because hemoglobinuria occurred in our patients before transfusion of non-IAT blood, the initial hemoglobinuria is consistent with the hemolysis seen in the IAT blood prepared for transfusion. However, since our patients had received transfusions in a prior surgery, it is possible that intravascular hemolysis from this effect might have played a role during the transfusion of banked blood in the later portions of these cases, especially Case 2. Of greater concern in these two cases was the development of a coagulopathy when no other cause was apparent. No preexisting bleeding tendency was known to be present and the degree of bony trauma was thought to be insufficient to release tissue plasminogen activator or urokinase, which activate the fibrinolytic system.16 Since the IAT devices were working properly, heparin would normally be adequately washed and unlikely a cause of coagulopathy.6 Reinfusion of substantial IAT volume (⬎3.5 L of IAT blood6 or where the transfused IAT blood exceeded 70 mL/kg17) produces a dilutional coagulopathy due to the removal of platelets and coagulation factors. This is unlikely in these patients because the onset of coagulopathy occurred well below these volumes. Coagulopathy and DIC have been documented with IAT. For example, since IAT blood is deficient in coagulation factors and platelets, a coagulopathy may develop if these other components are not transfused.18 One study in vascular and trauma cases noted this effect with large volumes of IAT.19 Specific coagulopathy attributed to IAT blood has been observed when combined with aprotinin in abdominal aortic ANESTHESIA & ANALGESIA

aneurysm surgery20 and in patients receiving transfusions of postoperative drainage blood or unwashed mediastinal blood collected after cardiopulmonary bypass surgery.21–24 Contaminated IAT blood has also been associated with coagulation abnormalities as seen in trauma patients receiving more than 15 U of IAT25 and in neuromuscular scoliosis patients in whom it was thought that osteopenic bone fragments were suctioned with the shed blood.17 Two of these reports noted hemoglobinuria, but the coagulation abnormalities were not attributed to transfusion of hemolyzed IAT blood.17,20 In addition, if chelating agents (e.g., EDTA) are used in the collection process or in transfused banked blood, a coagulopathy may develop due to deficient calcium levels. In our two cases, these factors do not seem to have been a factor in the early stages of the cases but may have played a role in the later aspects of the cases, particularly the second case. There, large volumes of IAT blood were used and the correction of hypocalcemia and the amounts of FFP and platelets would be important in correcting the contribution to the coagulopathy resulting from transfusing only RBCs (regardless of whether these were from IAT or banked blood). The clotting system can be activated in IAT by several means. Leukocytes and platelets damaged during collection or activated when they come in contact with the plastic IAT centrifuge bowl surface can release clotting activators.26 –29 This contact is postulated to increase when the surface of the bowl is not sufficiently covered by RBCs due to dilute collection fluid.26 –30 The released procoagulant, leukotactic, and inflammatory mediators are thought to contribute to the development of DIC and acute respiratory distress syndrome. Fortunately, activated polymorphonuclear leukocytes seem to be washed from the blood before transfusion.6,26,31,32 The coagulation process can also be initiated by cytokine and complement released from damaged RBCs and leukocytes.9,33,34 Consistent with this, complement, lipid mediators, pro and antiinflammatory cytokines, and leukocyte adhesion molecules have been detected in IAT blood.35 DIC can also be induced by the reinfusion of thrombin or Gelfoam collected from the wound by IAT.8,36 Coagulopathy with IAT blood in spine cases has been observed by Bull and Bull27 who coined the term “salvaged blood syndrome.” McKie and Herzenberg8 reported a coagulopathy which developed in a patient having scoliosis correction with IAT. DIC had its onset 30 min after infusion of the first IAT unit similar to our cases. This patient also developed cardiovascular collapse and respiratory distress syndrome similar to transfusion-related acute lung injury. Similarly, Gause et al.37 observed a significant increase in blood loss and the number of blood transfusions after IAT use. In our cases, the coagulopathy was noted by changes in Vol. 109, No. 1, July 2009

TEG威, PT, PTT, INR and clinical observation; however, in our second case, DIC may have also developed secondary to bony trauma or as a result of the multiple transfusions of IAT or banked blood products. Studies by Horlocker et al.38 suggest that the PT, PTT and INR are the best tests to detect a developing coagulopathy in instrumented spinal fusion. The substantial amount of free Hb in our cases is illustrated by the appearance of Hb in the urine of both patients who had no other cause of hemolysis when it started and hemolysis observed in the IAT blood (Fig. 1). There is a correlation between the total free Hb load and the subsequent renal dysfunction due to the precipitation of free Hb in the renal tubules.39 One prospective study in thoracic-aorta surgery concluded that more than 5 U of IAT blood contributed to renal failure.40 This hemoglobinuria suggests RBC damage with release of Hb and cell stroma into the IAT system. Cell damage during the initial collection and release of substances seemed to promote coagulopathy in our patients. The large volume of IAT transfused in Case 2 suggests an ongoing process that may explain the disproportionate morbidity seen in patients receiving large IAT volumes.4 Hemolysis potentially serves as a marker for the cellular trauma and release of RBC and leukocyte cell stroma that contributed to the coagulopathy. The American Association of Blood Banks has indicated that assessing the quality of IAT blood for residual albumin, free Hb, leukocytes, bacteria, and complement fractions may be of value; however, the time and resources involved in this would reduce the readily availability of IAT during surgery when active bleeding is occurring.41 Potassium could be used as a marker of hemolysis in the collection fluid, but not in the final product as it is effectively removed during IAT washing.9 These cases suggest IAT should be evaluated for its risk-benefit in individual cases. Concern should be raised if hemolysis is noted in the IAT blood before transfusion. Studies are necessary to determine ways to reduce hemolysis and cellular damage during the collection process or IAT processing to reduce these potential complications in instrumented spine surgery. REFERENCES 1. Kelley-Patteson C, Ammar AD, Kelley H. Should the Cell Saver Autotransfusion Device be used routinely in all infrarenal abdominal aortic bypass operations? J Vasc Surg 1993;18:261–5 2. Tawes RL Jr. Blood replacement and autotransfusion in major vascular surgery. In: Rutherford RB, ed. Vascular surgery. Philadelphia: WB Saunders, 1995 3. Freischlag JA. Intraoperative blood slavage in vascular surgeryworth the effort? Critical Care 2004;8:S53– 6 4. Posacioglu H, Apaydin AZ, Islamoglu F, Calkavur T, Yagdi T, Atay Y, Buket S. Adverse effects of cell saver in patients undergoing ruptured abdominal aortic aneurysm repair. Ann Vasc Surg 2002;16:450 –5 5. Reitman CA, Watters WC III, Sassard WR. The Cell Saver in adult lumbar fusion surgery: a cost-benefit outcomes study.[see comment]. Spine 2004;29:1580 –3 © 2009 International Anesthesia Research Society

41

6. Burman JF, Westlake AS, Davidson SJ, Rutherford LC, Rayner AS, Wright AM, Morgan CJ, Pepper JR. Study of five cell salvage machines in coronary artery surgery. Transfus Med 2002;12:173–9 7. Waters JH, Williams B, Yazer MH, Kameneva MV. Modification of suction-induced hemolysis during cell salvage. Anesth Analg 2007;104:684 –7 8. McKie JS, Herzenberg JE. Coagulopathy complicating intraoperative blood salvage in a patient who had idiopathic scoliosis. A case report. J Bone Joint Surg Am 1997;79:1391– 4 9. Szpisjak DF, Edgell DS, Bissonnette B. Potassium as a surrogate marker of debris in cell-salvaged blood. Anesth Analg 2000;91:40 –3 10. Rother RP, Bell L, Hillmen P, Gladwin MT. The clinical sequelae of intravascular hemolysis and extracellular plasma hemoglobin: a novel mechanism of human disease. JAMA 2005;293:1653– 62 11. Henn A, Hoffmann R, Muller HA. [Haptoglobin determination in the serum of patients following intraoperative autotransfusion using the Haemonetics Cell Saver III. Studies on the loading of patients with free hemoglobin in retransfused erythrocyte concentrate]. Anaesthesist 1988;37:741–5 12. Rochon G, Caron A, Toussaint-Hacquard M, Alayash AI, Gentils M, Labrude P, Stoltz JF, Menu P. Hemodilution with stroma-free [correction of stoma-free] hemoglobin at physiologically maintained viscosity delays the onset of vasoconstriction. Hypertension 2004;43:1110 –5 13. Kim-Shapiro DB, Schechter AN, Gladwin MT. Unraveling the reactions of nitric oxide, nitrite, and hemoglobin in physiology and therapeutics. Arterioscler Thromb Vasc Biol 2006;26:697–705 14. Schechter AN, Gladwin MT. Hemoglobin and the paracrine and endocrine functions of nitric oxide.[see comment]. N Engl J Med 2003;348:1483–5 15. Lopas H, Birndorf NI, Bell CE Jr, Robboy SJ, Colman RW. Immune hemolytic transfusion reactions in monkeys: activation of the kallikrein system. Am J Physiol 1973;225:372–9 16. Kahmann RD, Donohue JM, Bradford DS, White JG, Rao GH. Platelet function in adolescent idiopathic scoliosis. Spine 1992;17:145– 8 17. Murray DJ, Gress K, Weinstein SL. Coagulopathy after reinfusion of autologous scavenged red blood cells. Anesth Analg 1992;75:125–9 18. Page P. Perioperative autotransfusion and its correlation to hemostasis and coagulopathies. J Extra Corporeal Technol 1991;23:14 –21 19. Homann B, Klaue P, Hauptvogel S. [Intraoperative autotransfusion and its influence on the blood-clotting-system]. Anaesthesist 1977;26:606 20. Milne AA, Drummond GB, Paterson DA, Murphy WG, Ruckley CV. Disseminated intravascular coagulation after aortic aneurysm repair, intraoperative salvage autotransfusion, and aprotinin. Lancet 1994;344:470 –1 21. Dietrich W, Barankay A, Heinemann G, Blumel G, Wendt P, Richter JA. [Changes in blood clotting due to the autotransfusion of postoperative drainage blood]. Beitr Infusionther Klin Ernahr 1987;18:90 –3 22. Sieunarine K, Lawrence-Brown MM. Haematological effects of reinfused mediastinal blood after cardiac surgery. Med J Aust 1991;155:347 23. Griffith LD, Billman GF, Daily PO, Lane TA. Apparent coagulopathy caused by infusion of shed mediastinal blood and its prevention by washing of the infusate.[see comment]. Ann Thorac Surg 1989;47:400 – 6

42

Case Report

24. Schaff HV, Hauer JM, Bell WR, Gardner TJ, Donahoo JS, Gott VL, Brawley RK. Autotransfusion of shed mediastinal blood after cardiac surgery: a prospective study. J Thorac Cardiovasc Surg 1978;75:632– 41 25. Horst HM, Dlugos S, Fath JJ, Sorensen VJ, Obeid FN, Bivins BA. Coagulopathy and intraoperative blood salvage (IBS). J Trauma 1992;32:646 –52 26. Bull BS, Bull MH. Enhancing the safety of intraoperative RBC salvage. J Trauma 1989;29:320 –5 27. Bull BS, Bull MH. The salvaged blood syndrome: a sequel to mechanochemical activation of platelets and leukocytes? Blood Cells 1990;16:5–20 28. Bull BS, Bull MH. Hypothesis: disseminated intravascular inflammation as the inflammatory counterpart to disseminated intravascular coagulation. Proc Natl Acad Sci U S A 1994;91: 8190 – 4 29. Bull MH, Bull BS, Van Arsdell GS, Smith LL. Clinical implications of procoagulant and leukoattractant formation during intraoperative blood salvage. Arch Surg 1988;123:1073– 8 30. Ridler BMF, Thompson JF. The qualities of blood reinfused during cell salvage. Transfus Altern Trasfus Med 2003;5:466 –71 31. Innerhofer P, Wiedermann FJ, Tiefenthaler W, Schobersberger W, Klingler A, Velik-Salchner C, Oswald E, Salner E, Irschick E, Kuhbacher G. Are leukocytes in salvaged washed autologous blood harmful for the recipient? The results of a pilot study. Anesth Analg 2001;93:566 –72 32. Sandoval S, Alrawi S, Samee M, Satheesan R, Raju R, Cunningham JN, Acinapura AJ. A cytokine analysis of the effect of cell saver on blood in coronary bypass surgery. Heart Surg Forum 2001;4:113–7 33. Sieunarine K, Lawrence-Brown MM, Goodman MA, Prendergast FJ, Rocchetta S. Plasma levels of the lipid mediators, leukotriene B4 and lyso platelet-activating factor, in intraoperative salvaged blood. Vox Sang 1992;63:168 –71 34. Quinley ED. Immunohematology: principles and practice. New York: Lippincott Williams and Wilkins, 1998 35. Tawes RL Jr, Sydorak GR, Duvall TB, Scribner RG, Rosenman JE, Beare JP, Harris EJ. Avoiding coagulopathy in vascular surgery. Am J Surg 1990;160:212– 6 36. Ansell JE, Widrich WC, Johnson WC, Maizels M, Robbins AH, Nabseth DC, Deykin D. Gelfoam and autologous clot embolization: effect on coagulation. Invest Radiol 1978;13:115–20 37. Gause PR, Siska PA, Westrick ER, Zavatsky J, Irrgang JJ, Kang JD. Efficacy of intraoperative cell saver in decreasing postoperative blood transfusions in instrumented posterior lumbar fusion patients. Spine 2008;33:571–5 38. Horlocker TT, Nuttall GA, Dekutoski MB, Bryant SC. The accuracy of coagulation tests during spinal fusion and instrumentation. Anesth Analg 2001;93:33– 8 39. Klodell CT, Richardson JD, Bergamini TM, Spain DA. Does cell-saver blood administration and free hemoglobin load cause renal dysfunction? Am Surg 2001;67:44 –7 40. Godet G, Fleron MH, Vicaut E, Zubicki A, Bertrand M, Riou B, Kieffer E, Coriat P. Risk factors for acute postoperative renal failure in thoracic or thoracoabdominal aortic surgery: a prospective study. Anesth Analg 1997;85:1227–32 41. Guidelines for blood salvage and reinfusion in surgery and trauma. Bethesda, MD: American Association of Blood Banks, 1993


Pediatric Anesthesiology Section Editor: Peter J. Davis

The Association of Renal Dysfunction and the Use of Aprotinin in Patients Undergoing Congenital Cardiac Surgery Requiring Cardiopulmonary Bypass Ana Manrique, MD* Edmund H. Jooste, MB, ChB† Bradley A. Kuch, BS, RRT-NPS‡ Steven E. Lichtenstein, MD† Victor Morell, MD* Ricardo Munoz, MD§ Demetrius Ellis, MD储 Peter J. Davis, MD†

BACKGROUND: The use of large-dose aprotinin during cardiopulmonary bypass (CPB) in adult patients has been linked to postoperative renal dysfunction, but its effect on the pediatric population undergoing complex congenital cardiac operations is not well defined. METHODS: We used a retrospective cohort analysis to evaluate children undergoing cardiac surgery requiring CPB between July 2004 and July 2006. Demographic data and surgical risk quantified by the Aristotle surgical complexity level were analyzed as covariates. Renal dysfunction was defined according to the RIFLE criteria, an international consensus classification which defines three grades of increasing severity of acute kidney injury: risk (Class R), injury (Class I), and failure (Class F) based on serum creatinine values. A univariate and multivariate logistic regression analysis and a propensity score were used to analyze the data. The propensity score was developed using pretreatment covariates associated with the administration of aprotinin. A multivariate logistic regression was then used with the propensity score and intraoperative measures as covariates. A P value ⬍0.05 was considered statistically significant. RESULTS: Among 395 patients who underwent cardiac surgery, 55% received aprotinin and 45% did not. Thirty-one percent of the cohort had previous cardiac surgery; 17% were neonates. According to the RIFLE criteria, 80 of the patients (20.3%) had acute kidney injury in the postoperative period; 53 (13.4%) had risk of renal dysfunction with 23 (5.8%) having injury and four patients (0.7%) having failure. Those receiving aprotinin had a higher incidence of previous cardiac surgery (54.1% vs 5%), sepsis (6.9% vs.0.0%), heart failure (24.8% vs 12.4%), mechanical ventilation (25.2% vs 2.8%), or mechanical circulatory support (6.0% vs.0.6%). More patients had an Aristotle level of 4 (26.6% vs 2.8%) and were treated with diuretics (63.8% vs 26.6%), angiotensin converting enzyme inhibitors (21.1% vs 7.9%), milrinone (25.7% vs 4.5%), and inotropic support (16.1% vs 2.3%). Although there was a significant difference in the unadjusted risk of renal dysfunction, adjustment with the preoperative propensity score revealed that there was no association between aprotinin and renal dysfunction (OR 1.32; 95% CI 0.55–3.19). The duration of CPB was the only independent variable associated with the development of renal dysfunction (OR 1.0; 95% CI 1.009 –1.014). CONCLUSIONS: Patients who receive aprotinin are more likely to present with preoperative risk factors for the development of postoperative renal dysfunction. However, when associated risk factors are properly considered, the use of aprotinin does not seem to be associated with a higher risk of developing renal dysfunction in the immediate postoperative period in children. (Anesth Analg 2009;109:45–52)

R

eparative surgery and the perioperative management of pediatric patients with underlying congenital heart defects have undergone significant advances over the past few decades. Nonetheless, even with these

From the Departments of *Cardiovascular Surgery, †Anesthesiology, ‡Critical Care Medicine, §Pediatric Cardiac Critical Care, and 储Nephrology, Children’s Hospital of Pittsburgh, University of Pittsburgh School of Medicine, Pittsburgh, Pennsylvania. Accepted for publication January 29, 2009. Peter J. Davis is the Section Editor of Pediatric Anesthesiology for the Journal. This manuscript was handled by Steve L. Shafer, Editor-in-Chief, and Dr. Davis was not involved in any way with the editorial process or decision. Reprints will not be available from the author. Vol. 109, No. 1, July 2009

advances, significant risks of renal dysfunction,1– 4 bleeding diathesis, and neurological/cognitive impairments after cardiopulmonary bypass (CPB) still remain. To address the frequent coagulopathies and the morbidity associated with perioperative blood transfusions, anesthetic management strategies that minimize the need to administer blood products have

Address correspondence to Edmund Jooste, MB, ChB, Department of Anesthesia, Children’s Hospital, 7469 DeSoto Wing, 3705 5th Ave., Pittsburgh, PA 15213. Address e-mail to [email protected]. Copyright © 2009 International Anesthesia Research Society DOI: 10.1213/ane.0b013e3181a7f00a

45

evolved and frequently include the use of antifibrinolytic drugs. Aprotinin, a serine protease inhibitor with antifibrinolytic and antiinflammatory properties, has been used in an extensive, off-labeled manner in pediatric patients undergoing CPB. Studies of aprotinin in pediatric patients have resulted in conflicting information with respect to decreasing blood loss and administering blood products.5–9 Furthermore, there is scant published outcome data in the pediatric literature regarding aprotinin’s use and postoperative acute kidney injury. Two recent retrospective studies concluded that, after adjustment for preoperative and intraoperative variables, aprotinin was not an independent factor for postoperative renal dysfunction or the requirement of dialysis in the pediatric population.1,10 Studies in adult cardiac surgical patients have raised concerns regarding the drug’s safety. Recently, the BART study (Blood conservation using Antifibrinolytics: a randomized double-blind trial in 2331 adult high risk cardiac surgical patients) showed an increase in 30-day mortality in patients receiving aprotinin (relative risk, 1.55; 95% CI 0.99 –2.42) and an increase in the incidence of hemorrhage-related deaths.11 In studies by Mangano et al.12 and Karkouti et al.,13 aprotinin’s use was associated with a twofold increase in the risk of renal failure and the need for dialysis. However, in a meta-analysis of adult cardiac patients, Brown et al.14 described a dose-related effect of aprotinin on postoperative creatinine levels but did not note a significant increased risk of dialysisdependent renal failure. Fergusson et al.11 in a randomized controlled trial study also concluded that the use of aprotinin did not increase the need for postoperative renal dialysis. However, there was an increase in the proportion of patients exposed to aprotinin that had a doubling of postoperative serum creatinine levels. Furnay et al.,15 in a retrospective analysis of 23,105 patients, also noted that aprotinin was not a factor associated with postoperative renal failure. In view of the medical and legal implications of the November 2007 decision of Bayer Pharmaceuticals in conjunction with the Food and Drug Administration to temporarily stop the worldwide marketing of aprotinin,* coupled with the scant information available on aprotinin use in children, we reviewed and analyzed our institution’s congenital cardiac surgical database from July 2004 to July 2006 to determine the effect of aprotinin on renal outcome.

METHODS Design and Population After institutional approval, we analyzed a cohort of all patients undergoing congenital cardiac surgery at our institution from July 2004 to July 2006. Data *Bayer Healthcare Pharmaceuticals communication (Leverkussen GaWH, CT). http://www.trasylol.com, www.pharma.bayer. com.

46

Congenital Cardiac Surgery

collected included demographics, medical history including history of prematurity, medication use, prior surgical procedures, cardiac catheterizations, a history of renal dysfunction, preoperative status (elective, urgent, emergent), and current medications at the time of the surgical procedure. All patients undergoing cardiac surgery requiring CPB were included in the study. Serum creatinine values for the first 5 days after surgery and the need for dialysis before hospital discharge were recorded. The data were collected retrospectively using the existing cardiac intensive care unit database, patient’s electronic record, and the intraoperative anesthesia and perfusion records. The data were entered into the database by a single investigator.

Anesthetic Management Three anesthesiologists and two pediatric cardiac surgeons provided care during this period. The anesthetic management followed preestablished guidelines. In most patients, anesthesia was induced with inhaled sevoflurane and then maintained with isoflurane. In patients with preoperative hemodynamic instability, IV fentanyl was administered for both induction and maintenance. Rocuronium was administered to facilitate tracheal intubation and muscle relaxation. Except for the neonates and patients with contraindication to regional anesthesia, most patients received a single-shot caudal or epidural block administered with a combination of bupivacaine 0.25% with epinephrine (1 mL/kg up to 15 mL), clonidine (2 ␮g/kg), and preservative-free morphine (40 ␮g/kg). Intraoperative opioids were minimized in an effort to tracheally extubate all patients before leaving the operating room. In extubated patients, dexmedetomidine, and fentanyl were titrated to patient comfort. In neonates and in patients who would remain intubated, anesthesia was further supplemented with 10 –30 ␮g/kg fentanyl.

CPB Management Anticoagulation was initiated with heparin 350 – 400 U/kg with the target kaolin-ACT (HEMOCHRON威 Jr. Signature Microcoagulation System, International Technidyne Corporation, ITC) value more than 450 s before initiation of CPB. During surgery, blood was transfused if the hematocrit decreased below 25%. Nonpulsatile CPB was performed with roller pump and a microporous hollow fiber membrane oxygenator. Pump flow rates ranged from 3.0 to 3.2 L 䡠 min⫺1 䡠 m⫺2 for temperatures 37°C–30°C and from 2.0 to 2.5 L 䡠 min⫺1 䡠 m⫺2 for temperatures 29.9°C–18°C. Continuous mixed venous oxygen saturation was maintained in excess of 65%. When deep hypothermic cardiac arrest (DHCA) was used, the patients were cooled for a minimum of 20 min, ensuring that both bladder and esophageal temperatures were 18°C or less. A pH-stat acid-base management strategy for patients undergoing DHCA was used without the ANESTHESIA & ANALGESIA

addition of any continuous regional low-flow perfusion. Depending on the patient’s age, mean arterial blood pressure during bypass was maintained between 30 and 50 mm Hg. Dopamine 36 ␮g 䡠 kg⫺1 䡠 min⫺1 and milrinone 0.8 –1 ␮g 䡠 kg⫺1 䡠 min⫺1 were used on all patients to facilitate weaning from bypass. Nitroprusside 0.5–3 ␮g 䡠 kg⫺1 䡠 min⫺1 and epinephrine 0.05– 0.2 ␮g 䡠 kg⫺1 䡠 min⫺1 were added as clinically needed. Modified ultrafiltration (UF) was performed after termination of CPB in all children less than 15 kg. Furosemide 1 mg/kg up to 20 mg maximum was administered to all patients after termination of CPB. Platelets and packed red blood cells were given according to each anesthesiologist’s discretion. Protamine 4 mg/kg was given for reversal of heparin after termination of bypass and the MUF.

Aprotinin Protocol Aprotinin was used at the discretion of the anesthesiologist and in discussion with the surgeon. No other antifibrinolytic was used. In general, patients having simple lesions corrected, such as atrial septal defects, ventricular septal defects, endomyocardial cushion defects, and Tetralogy of Fallot repairs, did not receive aprotinin, whereas all reoperations, patients for complex repairs, such as transposition of the great vessels, Norwood procedures, interrupted aortic arch repairs, and patients undergoing planned DHCA received aprotinin. In all patients, except neonates, diphenhydramine (1 mg/kg up to 50 mg) was given just before aprotinin administration. Aprotinin (Trasylol威; Bayer Healthcare Pharmaceuticals, Pittsburgh, USA) containing 10,000 KIU/mL (Kallikrein Inhibitor Units) or 1.4 mg/mL was administered in the following two ways (based on the anesthesiologist’s preference): 1) in 86% of the cases, a 1 mL (1.4 mg) test dose was followed by a 170 mL/m2 or roughly 80,000 KIU/kg (maximum 200 mL) initial dose. The equivalent load was added to the CPB circuit prime, and a 30 mL 䡠 m⫺2 䡠 h⫺1 or roughly 15,000 KIU 䡠 kg⫺1 䡠 h⫺1 (maximum 25 mL/h) infusion was continued throughout the surgical period. In neonates, a standard 20 mL loading dose, a 20 mL initial dose to the pump prime and then a 6 mL/h infusion was administered; 2) in the remaining 14% of patients who received aprotinin, a 1 mL test dose was followed by a 2.7 mL/kg loading dose, an equivalent dose to the CPB circuit prime and then a 0.7 mL 䡠 kg⫺1 䡠 h⫺1 infusion for the duration of the surgery. Because of the small number of patients in the low-dose aprotinin group, we did not stratify patients according to the dose for the final analysis.

Definitions Normal creatinine and estimated creatinine clearance values were based on Ellis16 To define and classify acute kidney injury, the RIFLE criteria was used (Fig. 1).17,18 RIFLE defines three grades of increasing severity of acute kidney injury: risk (Class R), Vol. 109, No. 1, July 2009

Figure 1. RIFLE classification to determine the extent of acute kidney injury. injury (Class I), and failure (Class F), and two outcome classes (Loss (L) and End-stage (E) kidney disease). Each grade categorizes by the initial letter describing the injury or outcome: RIFLE. Grade 1, risk, requires a 1.5-fold increase in the creatinine level or a 25% decrease in glomerular filtration rate (GFR). Grade 2, injury, requires a twofold increase in serum creatinine or a 50% or more decrease in GFR. Grade 3, failure, requires a threefold increase in serum creatinine or a decrease in GFR more than 75%. The last two letters represent renal outcome; loss occurs if there is persistent acute renal failure for more than 4 wk, and end stage renal disease is defined as loss continued for more than 3 mo. We therefore used the first three RIFLE criteria in our analysis of acute kidney injury. To ensure we did not miss subtle renal changes, we defined acute kidney injury from the very first grade (risk) of the RIFLE criteria. There are more than 30 acute kidney injury definitions in the published literature, and there seems to be no consistent definition of renal dysfunction. The RIFLE criteria was proposed by the Acute Dialysis Quality Initiative Group in 2001, and its validity in children has been demonstrated.19,20 Furthermore, the RIFLE criteria have been shown to independently predict length of stay, costs, morbidity, and mortality in children.20 The Aristotle complexity score, designed to evaluate the quality of care in congenital heart procedures, was used to assess the complexity of each surgery. This complexity score is based on the type of surgical procedure as defined by the Society of Thoracic Surgeons and European Association for Cardiothoracic Surgery. The first step establishes the Basic Score, which assesses only the complexity of the procedures and is based on three factors: the potential for mortality, the potential for morbidity, and the anticipated technical difficulty. The basic score is derived from a sliding scale of increasing complexity with 3 points for a simple atrial septal defect, 7 points for a simple Tetralogy of Fallot repair, and 15 points for a Norwood procedure. The second step creates a Comprehensive Score, which further adjusts the complexity © 2009 International Anesthesia Research Society

47

according to the specific patient characteristics and can add 10 more points to the score. The maximum complexity score is 25 points. The Aristotle Basic Complexity Score is further divided into four levels, with Level 1 being the least complex and Level 4 being the most.21 The Aristotle score allows scoring of 145 congenital heart surgery procedures.22

Statistical Analysis Differences in baseline and clinical characteristics between groups were evaluated using nonparametric analysis. Categorical data that include binary baseline characteristics were compared using ␹2 and the Fisher’s exact test. Continuous data were compared using Student’s t-test and a Mann-Whitney U-test for normally distributed and nonnormally distributed data, respectively. Differences between preoperative and the highest postoperative creatinine value in the first 5 days were compared using a paired t-test. A P value of ⱕ0.05 was considered statistically significant. All analyses were performed using SPSS 12.0. Data were analyzed with univariate and multivariate logistic regression. Both models were fitted using preoperative and intraoperative covariates of acute kidney injury (P ⬍ 0.25). The multivariate analysis was performed using a forward stepwise logistic regression model. Upon completion of the final step, aprotinin and the Aristotle complexity level, which had been eliminated from the model during the stepwise logistic regression, were then “forced” back into the model reevaluating if they were associated with acute kidney injury after controlling for all other covariates. In addition to this analysis, we used a propensity score model to estimate treatment effect in patients receiving aprotinin. The score was developed using pretreatment/preoperative covariates of aprotinin administration. All variables found to have a P value ⬍0.5 during univariate analysis were included in the initial propensity score model as recommended by Luellen et al.23 A step-wise logistic regression (backward) was then performed using these covariates with the output being the probability of receiving aprotinin or propensity score. The initial univariate analysis for the development of the propensity score included the 27 preoperative variables of which 19 covariates remained in the final propensity score model. These variables were: age, height, body surface area, gender, race, history of renal dysfunction, previous cardiac surgery, sepsis, heart failure, cardiogenic shock, extracorporeal circulatory support, cardiopulmonary resuscitation before surgery, furosemide, antibiotics, digoxin, milrinone, inotropic support, vasodilators, and preoperative catheterization during the current admission. The model’s reliability and predictive ability were measured by the HosmerLemeshow test and C-index. A multivariate logistical regression was then fitted to evaluate if the association between aprotinin and acute kidney injury remained 48


significant, adjusting for propensity score and intraoperative covariates of acute kidney injury identified by univariate analysis (P ⬍ 0.25).

RESULTS Patient Characteristics and Preoperative and Intraoperative Variables Three hundred ninety-five patients having cardiothoracic surgeries between July 2004 and July 2006 were included. Sixty-seven patients (17%) were neonates. One hundred forty-five patients (36%) had other noncardiac congenital abnormalities. Intraoperative aprotinin was administered to 205 patients in 218 (55%) surgeries, whereas 176 patients undergoing 177 (45%) surgeries were not exposed to aprotinin. Those 13 patients who received more than one dose of aprotinin received the second dose on a different hospital admission, therefore their baseline characteristics for the propensity score were different and were considered new patients. Thirty-one percent of the cohort had previous cardiac surgery. Three patients of the aprotinin group had a history of milk allergy. No patients who received aprotinin had any evidence of an allergic reaction. Most of the baseline variables tested were different between patients who received aprotinin from those who did not (Table 4). Patients who received aprotinin were overall sicker or had undergone previous sternotomies. Significant differences between aprotinin and nonaprotinin groups were found in duration of CPB (103 ⫾ 66 min vs 74 ⫾ 37 min; P ⬍ 0.001), cross-clamp time (42 ⫾ 63 min vs 36 ⫾ 22 min; P ⬍ 0.002), circulatory arrest time (7.5 ⫾ 15 min vs 0; P ⬍ 0.001), and lowest temperature (28 ⫾ 7°C vs 31 ⫾ 2°C, P ⬍ 0.002). There was no difference in MUF volume between the groups.

Renal Dysfunction According to the RIFLE criteria, 80 of the patients (20.3%) had some grade of acute kidney injury in the postoperative period; 53 (13.4%) had risk of renal dysfunction, 23 (5.8%) had injury, and four patients (0.7%) had failure. Of the 80 patients meeting one of the RIFLE criteria, three patients needed renal dialysis within the first five postoperative days and two patients required renal dialysis within the following 2 wk. The apparent discrepancy between the four patients who met the failure criteria but five patients who required dialysis is because of the fact that we used the creatinine values in the first five postoperative days to categorize patients for the RIFLE criteria but report all patients requiring dialysis at any point in their admission. All five patients who required dialysis received aprotinin. The mortality within the group that received dialysis was 60%. Five percent of the patients in each group had a history of renal anatomic abnormality as reported by preoperative ANESTHESIA & ANALGESIA

Table 1. Univariate Regression Analysis of Preoperative and Intraoperative Variables and Acute Kidney Injury Variable preoperative/intraoperative

Acute kidney injury (n ⴝ 82)

No acute kidney injury (n ⴝ 313)

P

CPB time (min) Aprotinin (n) Gender (F/M)% Comorbidiy (n) Furosemide (n) Antibiotics (n) Sepsis (n) Heart failure (n) Cardiogenic shock (n) Extracorporeal cardiac life support (n) Cardiopulmonary resuscitation previous to surgery (n) Previous heart surgery (n) Mechanical ventilation (n)

115 ⫾ 90 54 12/21 42 50 32 6 25 15 5 7 35 19

83 ⫾ 42 26 86/74 103 136 57 9 51 26 9 6 89 41

0.018 0.365 0.011 0.058 0.490 0.811 0.782 0.765 0.873 0.242 0.061 0.081 0.681

All values represent total number of patients in each group with a specific preoperative or intraoperative variable except cardiopulmonary bypass (CPB) time which is in minutes and gender which is a represented by a percentage.

Table 2. Multivariate Logistic Regression Analysis of Preoperative and Intraoperative Predictors of Acute Kidney Injury (Hosmer-Lemeshow Test: 8.41; P ⫽ 0.39) 95% CI Variable

P

Exp(B)

Lower

Upper

Constant Gender (M) Comorbidity Preoperative cardiopulmonary resuscitation Cardiopulmonary bypass time

0.000 0.003 0.016 0.020

0.047 2.293 1.916 4.068

1.315 1.131 1.250

4.000 3.245 13.236

0.001

1.009

1.004

1.014

Table 3. Results with Aprotinin and Aristotle Complexity Score Placed Back in the Model show that Neither are Associated With the Risk of Developing Acute Kidney Injury (Hosmer-Lemeshow Test: 13.56; P ⫽ 0.09) 95% CI Variable

P

Exp (B)

Lower

Upper

Aprotinin Gender (M) Comorbidity Preoperative cardiopulmonary resuscitation Aristotle Cardiopulmonary bypass time Constant

0.347 0.004 0.013 0.027

1.326 2.263 1.959 3.848

0.736 1.294 1.152 1.163

2.389 3.957 3.330 12.732

0.282 0.015

1.205 1.007

0.858 1.001

1.692 1.012

0.000

0.030

were included. These results are shown in Tables 2 and 3 with longer duration of CPB time, male gender, presence of comorbidity, and preoperative cardiopulmonary resuscitation being significant predictors of acute kidney injury. Importantly, the use of aprotinin was not a predictor of acute kidney injury in the simple or the multivariate logistic regression analysis (Tables 2 and 3). Table 4 illustrates the preoperative variables used in the development of our propensity score analysis. The second column of P values are those calculated after propensity score adjustment (HosmerLemeshow test: 4.43; P ⫽ 0.816). The propensity score discriminated well between children who received aprotinin versus those who did not and had an area under the receiver operating characteristic curve, or C-index, of 0.95. The mean propensity score was 0.85 (95% CI 0.82– 0.89) in children receiving aprotinin and 0.36 (95% CI 0.15– 0.21) in children not receiving aprotinin. Our propensity score-adjusted model demonstrated no significant association between the administration of aprotinin and the development of acute kidney injury (OR 1.32; 95% CI 0.736 –2.38; P ⫽ 0.347). Furthermore, the multivariate analysis, including intraoperative covariates of acute kidney injury, found that only duration of CPB was associated with the postoperative development of acute kidney injury adjusting for the propensity score (OR 1.0 95% CI 1.009 –1.014; P ⬍ 0.001) (Table 5).

DISCUSSION renal ultrasounds or a history of acute kidney injury defined by the risk level of the RIFLE criteria.

Risk Modeling for Renal Insufficiency Univariate regression analysis of preoperative and intraoperative variables associated with acute kidney injury are shown in Table 1. Total CPB time and gender were the only two significant variables associated with acute kidney injury (P ⬍ 0.05). For the multivariate analysis, all variables with P values ⬍0.25 Vol. 109, No. 1, July 2009

In this retrospective, cohort study of pediatric patients requiring CPB for congenital heart surgery, univariate and multivariate logistic regression analysis and a propensity score model demonstrated that aprotinin was not related to the development of acute kidney injury. Although there are significant differences in the baseline characteristics of those patients who received aprotinin, the propensity score model adjusted well for those covariates. Thus, the development of acute kidney injury is more likely to be related © 2009 International Anesthesia Research Society

49

Table 4. Univariate Analysis of Patient Demographics and Preoperative Variables Between Patients who Received Aprotinin Versus Those Who Did Not Variable Age at surgery (yr) median (interquartile) Weight (kg) Height (cm) Body surface area (m2) Gender (F)%/(M)% Race (B)%/(W)%/(O)% History of renal dysfunction Other comorbidity Previous cardiac surgery Sepsis Heart failure Cardiogenic shock Extracorporeal circulatory support Cardiopulmonary resuscitation previous surgery Furosemide Additional diuretics Angiotensin converting enzyme inhibitor Antibiotics Digoxin Milrinone Dopamine and other inotropic support Vasodilators Mechanical ventilation Cardiac catheterization current admission Liver dysfunction Hospital admission before surgery


No aprotinin (n ⫽ 177)

P

Adjusted P

2.5 (0.01–7.24) 12 (3.8–27) 86 (53–125) 0.5 (0.24–0.95) 54/56 5/50/1 11/19 (5%) 84/145 (38.5%) 118/124 (54.1%) 15/15 (6.9%) 54/76 (24.8%) 37/41 (17%) 13/14 (6.0%) 11/13 (4.9%)

0.97 (0.33–5.09) 8 (12–16) 82 (53–125) 0.4 (0.29–0.73) 46/44 5/37/3 8/19 (4.5%) 61/145 (34.5%) 6/124 (5%) 0/15 22/76 (12.4%) 4/41 (2.3%) 1/14 (0.6%) 2/13 (1.1%)

0.64 0.56 0.81 0.61 0.57 0.25 0.8 0.4 ⬍0.001 ⬍0.001 0.002 ⬍0.001 0.004 0.03

0.27 0.85 0.53 0.57 0.26 0.32 0.35 0.76 0.14 0.99 0.23 0.26 0.99 0.99

139/186 (63.8%) 26/32 (11.9%) 46/60 (21.1%) 74/84 (33.9%) 23/47 (10.6%) 56/64 (25.7%) 35/39 (16.1%) 11 (5.0%) 55/60 (25.2%) 45/46 (20.6%)

47/186 (26.6%) 6/32 (3.4%) 14/60 (7.9%) 10/84 (5.6%) 24/47 (13.6%) 8/64 (4.5%) 4/39 (2.3%) 3 (1.7%) 5 (2.8%) 1/46 (0.6%)

⬍0.001 0.002 ⬍0.001 ⬍0.001 0.35 ⬍0.001 ⬍0.001 0.73 ⬍0.001 ⬍0.001

0.15 0.95 0.74 0.23 0.13 0.29 0.63 0.52 0.73 0.99

6/6 (2.8%) 95/109 (43.6%)

— 14/109 (7.9%)

0.035 ⬍0.001

0.99 0.77

These preoperative variables were used in the development of our propensity score analysis. Final column represents P value adjusted according to propensity score (to equalize groups).

Table 5. Results of a Multivariate Logistic Regression Analysis of Intraoperative Variables Causing Acute Kidney Injury Propensity score adjusted multivariate analysis Variable

Odds ratio (95% CI)

P

Cardiopulmonary bypass time Circulatory arrest Aprotinin Propensity score

1.0 (1.009–1.014)

⬍0.001*

1.0 (0.98–1.02) 0.96 (0.39–2.35) 2.07 (0.71–6.00)

0.90 0.93 0.18

Variables in the table represent those variables introduced in the multivariate model and were adjusted using the groups designed from the propensity score. * P ⬍ 0.05.

to the severity of illness and the duration of the CPB rather than to the administration of aprotinin. Acute renal failure is a well-documented complication after pediatric open heart surgery with an incidence of 10%–20%1–3,24 –27 and a mortality of 30%– 60% in patients requiring dialysis.24,26,28 This is consistent with our population who had a 20% incidence of acute kidney injury and a 60% mortality in the patients who required postoperative dialysis. There are two recently published reports on aprotinin use and renal dysfunction in children.1,10 Similar to our report, these studies were retrospective and used propensity score adjustments and multivariate regression in their analysis. Szekely et al.1 used similar criteria to ours for acute kidney injury and 50


noted a comparable incidence of renal dysfunction. In both studies, there was no association of aprotinin with an increased risk of renal dysfunction or the need for dialysis. It is interesting that in both these studies1,10 and this study there was an association with the duration of CPB and renal dysfunction. In a comparison of aprotinin to lysine analogs administered to 2331 high-risk adult cardiac surgical patients entered into a multicenter, blinded study, (BART study), Fergusson et al.11 noted that aprotinin was not associated with an increased risk of renal failure or the need for dialysis. They did, however, observe that aprotinin was associated with an increased proportion of patients who had a doubling in their serum creatinine values. This transient effect of aprotinin on kidney function may be related to its deposition and accumulation in the kidney’s proximal tubular cells where it interferes with normal tubule protease secretion29,30 and renal artery vasodilators.31 Aprotinin can also affect the bradykinin system and have renal effects similar to those of the angiotensin inhibitors. In this instance, the increase in serum creatinine resolved when the aprotinin was cleared from the kidneys.32,33 The main limitations of our observations are the single-center, retrospective, nonblinded, and nonrandomized study design. Furthermore, the small study sample (n ⫽ 395), although relatively large for ANESTHESIA & ANALGESIA

a pediatric cardiac cohort, and the relatively small number of patients who developed acute renal injury make the multivariate analysis less powerful. Because aprotinin was administered to sicker, more complex patients at the discretion of the treating physicians, there was a patient selection bias. Furthermore, all data were collected by a single data entry investigator who could have introduced coding errors. Because of major differences between the two groups’ preoperative status and surgical complexity, propensity score modeling was used to match patients who had similar likelihood of receiving either treatment according to their baseline characteristics. This was achieved by including covariates related to aprotinin administration in a logistic regression model, based on pretreatment variables. After matching patients according to their propensity score, a second regression adjustment was performed to evaluate the association between the treatment (exposure to aprotinin) and the outcome (renal dysfunction) by introducing the propensity score as the only confounding variable. After adjusting for risk factors related to aprotinin administration, no association was found between treatment with aprotinin and the development of acute renal injury. Furthermore, the only independent variable associated with postoperative renal injury was the CPB time with an OR of 1.0. Even though this is quite marginal, the confidence interval was positive for this event and this finding has been observed in previous studies.1,10 In conclusion, pediatric patients undergoing CPB for repair of their complex congenital cardiac lesions who received aprotinin had significantly more preoperative and intraoperative risk factors for the development of postoperative renal dysfunction. However, when the influence of confounding factors was adjusted using the propensity score, aprotinin was not associated with an increased risk of developing postoperative acute kidney injury. REFERENCES 1. Szekely A, Sapi E, Breuer T, Kertai MD, Bodor G, Vargha P, Szatmari A. Aprotinin and renal dysfunction after pediatric cardiac surgery. Paediatr Anaesth 2008;18:151–9 2. Sorof JM, Stromberg D, Brewer ED, Feltes TF, Fraser CD Jr. Early initiation of peritoneal dialysis after surgical repair of congenital heart disease. Pediatr Nephrol 1999;13:641–5 3. Dittrich S, Dahnert I, Vogel M, Stiller B, Haas NA, AlexiMeskishvili V, Lange PE. Peritoneal dialysis after infant open heart surgery: observations in 27 patients. Ann Thorac Surg 1999;68:160 –3 4. Alkan T, Akcevin A, Turkoglu H, Paker T, Sasmazel A, Bayer V, Ersoy C, Askn D, Aytac A. Postoperative prophylactic peritoneal dialysis in neonates and infants after complex congenital cardiac surgery. ASAIO J 2006;52:693–7 5. Chauhan S, Kumar BA, Rao BH, Rao MS, Dubey B, Saxena N, Venugopal P. Efficacy of aprotinin, epsilon aminocaproic acid, or combination in cyanotic heart disease. Ann Thorac Surg 2000;70:1308 –12 6. Carrel TP, Schwanda M, Vogt PR, Turina MI. Aprotinin in pediatric cardiac operations: a benefit in complex malformations and with high-dose regimen only. Ann Thorac Surg 1998;66:153– 8 Vol. 109, No. 1, July 2009

7. Bulutcu FS, Ozbek U, Polat B, Yalcin Y, Karaci AR, Bayindir O. Which may be effective to reduce blood loss after cardiac operations in cyanotic children: tranexamic acid, aprotinin or a combination? Paediatr Anaesth 2005;15:41– 6 8. Arnold DM, Fergusson DA, Chan AK, Cook RJ, Fraser GA, Lim W, Blajchman MA, Cook DJ. Avoiding transfusions in children undergoing cardiac surgery: a meta-analysis of randomized trials of aprotinin. Anesth Analg 2006;102:731–7 9. Eaton MP. Antifibrinolytic therapy in surgery for congenital heart disease. Anesth Analg 2008;106:1087–100 10. Backer CL, Kelle AM, Stewart RD, Suresh SC, Ali FN, Cohn RA, Seshadri R, Mavroudis C. Aprotinin is safe in pediatric patients undergoing cardiac surgery. J Thorac Cardiovasc Surg 2007;134:1421– 6; discussion 1426 – 8 11. Fergusson DA, Hebert PC, Mazer CD, Fremes S, MacAdams C, Murkin JM, Teoh K, Duke PC, Arellano R, Blajchman MA, Bussieres JS, Cote D, Karski J, Martineau R, Robblee JA, Rodger M, Wells G, Clinch J, Pretorius R; BART Investigators. A comparison of aprotinin and lysine analogues in high-risk cardiac surgery. N Engl J Med 2008;358:2319 –31 12. Mangano DT, Tudor IC, Dietzel C. The risk associated with aprotinin in cardiac surgery. N Engl J Med 2006;354:353– 65 13. Karkouti K, Beattie WS, Dattilo KM, McCluskey SA, Ghannam M, Hamdy A, Wijeysundera DN, Fedorko L, Yau TM. A propensity score case-control comparison of aprotinin and tranexamic acid in high-transfusion-risk cardiac surgery. Transfusion 2006;46:327–38 14. Brown JR, Birkmeyer NJ, O’Connor GT. Meta-analysis comparing the effectiveness and adverse outcomes of antifibrinolytic agents in cardiac surgery. Circulation 2007;115:2801–13 15. Furnary AP, Wu Y, Hiratzka LF, Grunkemeier GL, Page US III. Aprotinin does not increase the risk of renal failure in cardiac surgery patients. Circulation 2007;116:I127–33 16. Ellis D, Avner ED. Fluid and electrolyte disorders in pediatric patients. In: Puschett JB, ed. Disorders of Fluid and Electrolyte Balance. New York: Churchill Livingstone; 1985:217–30 17. Venkataraman R, Kellum JA. Defining acute renal failure: the RIFLE criteria. J Intensive Care Med 2007;22:187–93 18. Ricci Z, Cruz D, Ronco C. The RIFLE criteria and mortality in acute kidney injury: a systematic review. Kidney Int 2008;73: 538 – 46 19. Askenazi DJ, Bunchman TE. Pediatric acute kidney injury: the use of the RIFLE criteria. Kidney Int 2007;71:963– 4 20. Akcan-Arikan A, Zappitelli M, Loftis LL, Washburn KK, Jefferson LS, Goldstein SL. Modified RIFLE criteria in critically ill children with acute kidney injury. Kidney Int 2007;71:1028 –35 21. Jacobs JP, Lacour-Gayet FG, Jacobs ML, Clarke DR, Tchervenkov CI, Gaynor JW, Spray TL, Maruszewski B, Stellin G, Gould J, Dokholyan RS, Peterson ED, Elliott MJ, Mavroudis C. Initial application in the STS congenital database of complexity adjustment to evaluate surgical case mix and results. Ann Thorac Surg 2005;79:1635– 49; discussion 1635– 49 22. Lacour-Gayet F, Clarke D, Jacobs J, Gaynor W, Hamilton L, Jacobs M, Maruszewski B, Pozzi M, Spray T, Tchervenkov C, Mavroudis C. The Aristotle score for congenital heart surgery. Semin Thorac Cardiovasc Surg Pediatr Card Surg Annu 2004;7:185–91 23. Luellen JK, Shadish WR, Clark MH. Propensity scores: an introduction and experimental test. Eval Rev 2005;29: 530 –58 24. Werner HA, Wensley DF, Lirenman DS, LeBlanc JG. Peritoneal dialysis in children after cardiopulmonary bypass. J Thorac Cardiovasc Surg 1997;113:64 – 8; discussion 68 –70 25. Rigden SP, Barratt TM, Dillon MJ, De Leval M, Stark J. Acute renal failure complicating cardiopulmonary bypass surgery. Arch Dis Child 1982;57:425–30 26. Gomez-Campdera FJ, Maroto-Alvaro E, Galinanes M, Garcia E, Duarte J, Rengel-Aranda M. Acute renal failure associated with cardiac surgery. Child Nephrol Urol 1988;9:138 – 43 27. Baxter P, Rigby ML, Jones OD, Lincoln C, Shinebourne EA. Acute renal failure following cardiopulmonary bypass in children: results of treatment. Int J Cardiol 1985;7:235– 43 28. Book K, Ohqvist G, Bjork VO, Lundberg S, Settergren G. Peritoneal dialysis in infants and children after open heart surgery. Scand J Thorac Cardiovasc Surg 1982;16:229 –33 © 2009 International Anesthesia Research Society

51

29. Vio CP, Oestreicher E, Olavarria V, Velarde V, Mayfield RK, Jaffa AA. Cellular distribution of exogenous aprotinin in the rat kidney. Biol Chem 1998;379:1271–7 30. Rustom R, Grime S, Maltby P, Stockdale HR, Critchley M, Bone JM. A new method to measure renal tubular degradation of small filtered proteins in man using radiolabelled aprotinin (Trasylol). Clin Sci (Lond) 1992;83:289 –94 31. Maier M, Starlinger M, Zhegu Z, Rana H, Binder BR. Effect of the protease inhibitor aprotinin on renal hemodynamics in the pig. Hypertension 1985;7:32– 8

52


32. Fauli A, Gomar C, Campistol JM, Alvarez L, Manig AM, Matute P. Kidney-specific proteins in patients receiving aprotinin at high- and low-dose regimens during coronary artery bypass graft with cardiopulmonary bypass. Eur J Anaesthesiol 2005;22:666 –71 33. Coleman CI, Rigali VT, Hammond J, Kluger J, Jeleniowski KW, White CM. Evaluating the safety implications of aprotinin use: the retrospective evaluation of aprotinin in cardio thoracic surgery (REACTS). J Thorac Cardiovasc Surg 2007;133:1547–52


The Narcotrend Index Indicates Age-Related Changes During Propofol Induction in Children Sinikka Mu¨nte, MD, PhD* Jaakko Klockars, MD* Mark van Gils, PhD† Arja Hiller, MD, PhD* Michael Winterhalter, MD, PhD‡ Christina Quandt, MD‡ Matthias Gross, MD‡ Tomi Taivainen, MD, PhD*

BACKROUND: The Narcotrend威 electr oencephalogram monitor is designed to measure hypnotic state during anesthesia. We performed this study to evaluate the effectiveness and reliability of the Narcotrend monitor in assessing hypnotic state and loss of consciousness (LOC) during propofol anesthesia induction in children. METHODS: Sixty-two children, aged 1–5 (n ⫽ 17), 6 –12 (n ⫽ 23), and 13–16 (n ⫽ 21) yr, scheduled for elective surgery were studied. The patients were premedicated with oral midazolam 0.5 mg/kg. After IV access, propofol target controlled infusion (TCI) was started with 0.5 ␮g/mL and increased by 0.5 ␮g/mL increments every 2 min until the child did not respond to any verbal command or physical stimuli. A manual scheme was used for children weighing ⬍15 kg. Hypnotic state was measured every minute from the start of the propofol infusion using the University of Michigan Sedation Scale (UMSS). LOC was defined as a transition of UMSS scale value 2 to 3. The Narcotrend index (NI) was recorded before the start of induction and during the whole study period. NI values were noted simultaneously, yet independently of the sedation measurements. Prediction probability (PK) was used to assess the correspondence between NI and UMSS. Sensitivity and specificity of NI for differentiating between consciousness and unconsciousness were calculated. NI values at specific UMSS levels were compared between the different age groups and the relationships between TCI propofol concentrations and sedation levels were assessed using correlation analysis. RESULTS: A PK-value of 0.84 (95% CI [0.80 – 0.88]) of NI was calculated from the data for the detection of LOC. Similarly, a PK value of 0.82 (95% CI [0.78 – 0.86]) indicated agreement between NI and UMSS values. The average NI values differed between successive UMSS sedation levels 0 and 1 and levels 1 and 2 (P ⬍ 0.01). In the youngest age group, the NI discriminated between UMSS levels 2 and 3, in the second age group between levels 1 and 2 and 2 and 3, and in the oldest age group between 0 and 1. Furthermore, the NI values differed significantly between age groups at UMSS levels 1– 4 (P ⬍ 0.005), with the NI values being higher in younger compared with older children. The average NI value at LOC was 68. For the detection of consciousness, a sensitivity of 0.67 and specificity of 0.79 were achieved. Spearman correlation coefficients indicated higher association between TCI propofol concentrations and UMSS (0.96) than between NI and UMSS (⫺0.68). CONCLUSIONS: During propofol induction in children, the Narcotrend electroencephalogram monitor was capable of following changes in the sedation level of children to some extent, but also had a relatively high probability (0.18) of incorrectly predicting changes in conscious state. Therefore, the monitor should not solely be used to guide sedation and anesthesia. NI was age-dependent and younger children had higher NI-values than older children at the same level of sedation. (Anesth Analg 2009;109:53–9)

E

lectroencephalogram (EEG)-based devices to monitor loss of consciousness (LOC) and hypnotic state during anesthesia have been developed and

From the *Department of Anesthesiology and Intensive Care Medicine, Children’s Hospital, Helsinki University Clinics; †VTT Technical Research Centre of Finland, Finland; and ‡Department of Anesthesiology, Hanover Medical School, Hannover, Germany. Accepted for publication January 27, 2009. Supported by the Finnish National Founding (EVO), The Narcotrend威 was kindly lent by the developers Dr. Barbara Schulze and Dr. Arthur Schulze, Hannover, Germany. Address correspondence and reprint requests to Sinikka Mu¨nte, MD, PhD, Department of Anesthesiology and Intensive Care Medicine, Children’s Hospital, Helsinki University Clinics, PL 281, 00029 HUS, Finland. Address e-mail to [email protected]. Copyright © 2009 International Anesthesia Research Society DOI: 10.1213/ane.0b013e3181a49c98

Vol. 109, No. 1, July 2009

validated mainly in adults. However, children might benefit from depth of anesthesia monitoring for several reasons. The incidence of awareness during anesthesia is four- to eightfold in children over that of adults.1 Furthermore, the doses of anesthetics vary remarkably between children and adults, because of the higher clearance and volume of distribution in children for example.2 EEG monitoring might help the anesthesiologist better adjust doses of anesthetics to individual needs.3 The more precise dosing of anesthetics leading to improved recovery and home readiness might also yield cost savings.4 However, before evaluating the clinical usefulness of these devices for children, the essential question to answer is, how reliable are they in assessing depth of hypnosis and LOC during sedation and anesthesia?5 53

Table 1. Narcotrend Index Narcotrend index

Clinical description

100–95 90–80 79–65 64–37 36–13 12–0

Awake Sedation Light anesthesia General anesthesia General anesthesia with deep hypnosis Very deep hypnosis

Table 2. The University of Michigan Sedation Scale (UMSS)16 Value

Patient state

0 1

Awake and alert Minimally sedated: tired/sleepy, appropriate response to verbal conversation, and/or sound Moderately sedated: somnolent/sleeping, easily aroused with light tactile stimulation, or a simple verbal command Deeply sedated: deep sleep, aroused only with significant physical stimulation Unarousable

2 3 4

The Narcotrend EEG monitor (MT MonitorTechnik GmbH, Bad Bramstedt, Germany) evaluated in this study provides a computerized analysis of the EEG. The processed EEG results are calculated and given as a number, a Narcotrend index (NI) ranging from 0 (very deep hypnosis) to 100 (awake) (Table 1).6 Unlike other monitor devices, the age-related EEG changes are incorporated into the Narcotrend algorithm. Numerous evaluative studies of the Narcotrend have been published in adults, yielding mixed results,7–10 but there are only a few reports on the use of the Narcotrend in children who have been anesthetized with inhaled anesthetics.11–14 These studies suggest that, during sevoflurane or desflurane-based anesthesia, the Narcotrend EEG monitor can differentiate between children who are awake and those who are deeply anesthetized. Given that various anesthetic drugs may have different effects on EEG parameters15 and that previous studies do not answer the question as to whether the Narcotrend discriminates between increasing sedation levels, we felt another study using propofol was warranted. The main goal of this study was to evaluate the effectiveness of the NI in determining the transition between consciousness and unconsciousness during propofol anesthesia induction in children. Second, we investigated the ability of the NI to differentiate among increasing sedation levels. Various pediatric age groups were analyzed post hoc. The level of sedation and LOC (defined as a transition of University of Michigan Sedation Scale [UMSS] scale value from 2 to 3) were assessed using the UMSS.16 The UMSS differentiates among five clinical sedation levels from awake (0) to unarousable (4) (Table 2). Third, we determined the correlation between propofol target-controlled infusion (TCI) concentrations and sedation levels. We did not expect a strong correlation, 54

Narcotrend威 Validation Study in Children

because only plasma concentration (and not effect-site brain concentration) is calculated by the TCI software used in this study.

METHODS Patients and Study Design After approval from the institutional ethics committee (Hospital for Children and Adolescents, Helsinki University, Helsinki, Finland) informed consent of parents, and also, when appropriate, the respective child’s assent was obtained. Sixty-one pediatric patients aged between 1 and 16 yr, with ASA physical status of I-III and scheduled for elective surgery were studied. Patients were excluded from the study if they had a disease or used medication that influenced the central nervous, cardiovascular, or respiratory systems, impaired hearing, or if they were not able to respond adequately to verbal or tactile stimuli as required by the UMSS. The study phase was limited to the time immediately before and during induction of anesthesia until the patient could not be aroused (level 4 at UMSS ⫽ end of the study phase).

Anesthetic Regimen EMLA cream was topically applied to provide painless cannulation of a forearm vein and premedication with 0.5 mg/kg (maximum dose 15 mg) oral midazolam was administered 45 min before the estimated induction of anesthesia. After obtaining IV access, propofol TCI was started at 0.5 ␮g/mL and increased by 0.5 ␮g/mL steps every 2 min until the end of the study phase. We used the Alaris Asena TCI pump programmed according to Kataria et al.’s pharmacokinetic model (Alaris威 Medical Systems, Alaris Medical UK, Basingstoke, UK).17 Since this TCI device does not have the software to calculate plasma concentrations for children weighing ⬍15 kg, anesthesia was induced in these children with a continuous 20 mg 䡠 kg⫺1 䡠 h⫺1 propofol infusion. As the child became unconscious, supplemental oxygen was administered via face mask as needed. After the end of the study phase, the anesthesiologist was free to use any action or medication needed for other procedural purposes, such as tracheal intubation and maintenance of anesthesia. Monitoring during the study phase included recording of heart rate and peripheral oxygen saturation.

Narcotrend EEG-Monitoring Before starting the anesthesia induction, the Narcotrend EEG monitor (MT MonitorTechnik GmbH, Bad Bramstedt, software version 4.3) was connected to the patient’s forehead according to manufacturer’s instructions using “blue sensor” electrodes (Medicotest A/S, Ølstykke, Denmark). The skin was prepared at electrode sites with abrasive skin cleaning paste (Tyco/Healthcare, Mallinckrodt, Mirandola, Italy) to maintain impedances of ⬍6 k⍀. An EEG was recorded before the start of induction to obtain a base value of NI and during the whole study period. The Narcotrend monitor was placed out of view of the care ANESTHESIA & ANALGESIA

providers and observers to facilitate blinding. The Narcotrend values were noted before the induction and at every minute interval from the start of the propofol infusion until the end of the study phase. A second observer assigned UMSS scores simultaneously yet independently.

Sedation Measurement The anesthesiologist who evaluated the level of sedation according to UMSS was not aware of Narcotrend values or propofol TCI concentrations. Sedation levels were determined at 1-min intervals from the start of the propofol infusion according to the UMSS. The validity of the UMSS is well established for children and differentiates among five clinical sedation levels from awake (0) to unarousable (4).16,18 Measurements were discontinued after the patient had scored 4 on the UMSS on two successive occasions. NI values shown by the monitor’s display at the beginning of the sedation assessment were entered into the data analysis.

Statistical Analysis The sample size was based on our previous study with children.19 A power calculation based on the main goal of the study (to detect whether the mean NI at UMSS 2 is higher than the mean NI at UMSS 3) used an estimated mean NI at UMSS 2 of 70, and at UMSS 3 of 57, with estimated standard deviation of 20. With an ␣ error of 0.05 and ␤ error of 0.9, this gives a required sample size of 42. In this study, we used 61 cases which, post hoc, gives a power of 0.97. The ability and accuracy of the NI to distinguish among different hypnotic states were evaluated using the prediction probability (PK) value as a standard.20 PK values were calculated with an Excel (Microsoft威, Redmond, WA) Macro (PK Macro) per patient. Mean and standard deviations of these PK values were then used for further analysis. The PK value for separation between consciousness and unconsciousness was performed by transforming the UMSS values into a value of either 1 (when 0 ⬍ UMSS ⬍ ⫽ 2, before LOC) or 0 (after LOC). Thus, a dichotomous depth of consciousness scale was created. Using the scale and the NI values, the PK value for separating between consciousness and unconsciousness was obtained, again using PK values for individual patients. Since age-related effects on NI may be important, the calculations were done separately for the following age groups: Group I: 1–5 yr (n ⫽ 17), Group II: 6 –12 yr (n ⫽ 23), and Group III: 13–16 yr (n ⫽ 21). The comparisons of PK values among age groups were performed using the Mann–Whitney test. The selection of the age groups was done post hoc with the consideration of EEG changes caused by brain maturation and after consultation of the developers of the Narcotrend EEG monitor. Values are means (sd) unless stated otherwise. A PK value of 1.0 indicates that a parameter, such as NI, perfectly follows changes in the sedation level, Vol. 109, No. 1, July 2009

whereas a PK value of 0.5 indicates that the prediction is no better than chance alone. Spearman rank order correlation analysis was used to evaluate the relationship between propofol TCI concentrations and NI, propofol TCI concentrations and UMSS values and between NI and UMSS. Additionally, comparisons were made between NI values of successive UMSS levels using the Mann–Whitney U-test. The Kruskal–Wallis test was used to compare the distributions of NI values among age groups at each UMSS level. For these tests, for each patient one NI value per UMSS level was used, calculated as the average NI value observed at that UMSS level. The effectiveness of the NI was analyzed by comparing the last NI value before LOC with the first after LOC for each patient (i.e., in each patient the last NI at UMSS 2 was compared with the first NI at UMSS 3). These data were compared by using the paired Wilcoxon’s test. In all tests P ⬍ 0.05 was considered to be significant. We used a post hoc Bonferroni correction to adjust significance levels for P values using the number of tests performed. The sensitivity and specificity for detecting UMSS ⱖ3 for different threshold values of NI were calculated and a receiver operator characteristic (ROC) curve was generated to determine the best “cutoff” values for detecting awareness or detecting excessively deep anesthesia with high sensitivity (Fig. 2). Furthermore, to assess the relative effect of NI and age on the probability of UMSS ⱖ3 occurring, we developed a logistic regression model with “unconscious (UMSS ⱖ3)” as the outcome variable and NI and age as independents. Analyses were performed using SPSS V14.0 (SPSS, Chicago, IL) and Matlab V7.1 (The Mathworks, Natick MA).

RESULTS The demographic data of the study are shown in Table 3. The study had to be discontinued for three children in the youngest age group (ages 1–5 yr) because of agitation, two children because of propofol injection pain, and one child because the impedance values of ⬍6 k⍀ could not be attained. The results are based on data collected from 61 patients. There were 744 UMSS assessments in total, and of these, 663 had corresponding and comparable NI measurements that gave the actual PK value. There were 81 (11%) missing NI values because of procedural/study artifacts. The basal NI values could not be recorded for eight patients (13.1%). In this study, the PK value for the Narcotrend to distinguish among different hypnotic states was 0.82 (95% CI [0.78 – 0.86]). Children (n ⫽ 52) sedated using the TCI pump had a very high correlation (0.96 Spearman correlation) between TCI concentration and UMSS sedation levels. The Spearman correlation between NI and UMSS was ⫺0.68. There was also a © 2009 International Anesthesia Research Society

55

Table 3. Demographic Data (means,

Mean ages (yr) Mean weight (kg) Gender (M:F) ASA (1:2:3)

SD)

for all Subjects and for Different Age Groups

Group I (1–5 yr) n ⫽ 17

Group II (6–12 yr) n ⫽ 23

Group III (12 ⬍16 yr) n ⫽ 21

All n ⫽ 61

2.8 (1.4) 14.0 (3.4) 13:4 11:4:2

9.5 (1.9) 33.0 (10.8) 11:12 15:7:1

13.6 (1.3) 51.0 (7.1) 11:10 16:5:0

9.0 (4.6) 33.9 (16.7) 35:26 42:16:3

Table 4. Relationship Between UMSS, NI, and TCI1 Age

1—Pk (NI versus UMSS)

Rho spearman (NI versus TCI)

Rho spearman (TCI versus UMSS)

Rho spearman (NI versus UMSS)

1–5 yr 6–12 yr 13–16 yr Overall

0.86 (0.75–0.96) 0.82 (0.74–0.91) 0.80 (0.73–0.87) 0.82 (0.78–0.87)

⫺0.85 (⫺0.96 to ⫺0.73) ⫺0.73 (⫺0.85 to ⫺0.60) ⫺0.68 (⫺0.81 to ⫺0.55) ⫺0.74 (⫺0.81 to ⫺0.67)

0.96 (0.95–0.98) 0.95 (0.94–0.97) 0.94 (0.92–0.96) 0.95 (0.94–0.96)

⫺0.84 (⫺0.98 to ⫺0.71) ⫺0.69 (⫺0.84 to ⫺0.53) ⫺0.63 (⫺0.77 to ⫺0.49) ⫺0.68 (⫺0.77 to ⫺0.59)

The table shows prediction probability (PK)-values of the relationship between UMSS (University of Michigan Sedation Scale16) and NI (Narcotrend Index), Spearman correlation coefficient between NI and TCI (target controlled infusion), TCI and UMSS, and NI and UMSS in various age groups. Values are means of individual correspondence measures and in brackets 95% CI of mean.

Figure 1. The relationship between Narcotrend Index and University of Michigan Sedation Score. Each box shows the median (horizontal line inside the box) and quartiles (the bottom of the box represents first quartile, the top the third quartile). The whiskers (the lines that extend out the top and bottom of the box) represent the highest and lowest values that are not outliers or extreme cases (outliers are values between 1.5 and 3 time interquartile range, extreme values are more than three times the interquartile range from the top or bottom of the box).

significant negative correlation (⫺0.73 Spearman correlation) between NI and propofol concentrations. The PK values (relationship between NI and UMSS score) and relationships between TCI and UMSS and NI, respectively, are shown in Table 4. Figure 1 shows box plots depicting the distribution of NI values separately at each UMSS sedation level. The Mann–Whitney test showed significant differences in average NI values between successive UMSS sedation levels 0 and 1 and levels 1 and 2 (P ⬍ 0.01). Separate calculations for each age group revealed that average NI significantly discriminated between UMSS levels 2 and 3 in the youngest age group, between 56


Figure 2. Receiver operating characteristic curve for Narcotrend Index for separating between University of Michigan Sedation Scale (UMSS) ⱕ2 vs UMSS ⱖ3, area under curve is 0.84 (95% confidence interval 0.81– 0.87). Several example points on the curve are: sensitivity ⫽ 0.26, specificity ⫽ 0.94 for NI ⫽ 40; sensitivity ⫽ 0.61, specificity ⫽ 0.83 for n ⫽ 60; sensitivity ⫽ 0.67, specificity ⫽ 0.79 for NI ⫽ 68; sensitivity ⫽ 0.74, specificity ⫽ 0.74 for NI ⫽ 76; and sensitivity ⫽ 0.81, specificity ⫽ 0.71 for NI ⫽ 80.

levels 1 and 2 and 2 and 3 in the second age group, and between levels 0 and 1 in the oldest age group. Furthermore, the Kruskall–Wallis test showed that the NI values differed significantly among age groups at UMSS levels 1– 4 (P ⬍ 0.005). A PK value of 0.84 (95% CI [0.80 – 0.86]) for the detection of LOC (transition of UMSS values from 2 to 3) was calculated from these data. Furthermore, we tested the ability of the NI to discriminate between moderate sedation and deep sedation. There was a significant difference between NI for UMSS 2 vs 3 (P ⫽ 0.004). ANESTHESIA & ANALGESIA

Table 5. NI Values (mean,

SD)

and TCI Concentrations (Mean,

NI values at LOC† TCI concentration (␮g/mL)‡

SD)

at LOC (transition of UMSS 2–3)

Group I (1–5 yr) n ⫽ 17*

Group II (6–12 yr) n ⫽ 23

Group III (12 ⬍ 16 yr) n ⫽ 21

All n ⫽ 61

77 (19) 2.25 (0.48)

71 (18) 1.78 (0.41)

59 (16) 1.37 (0.44)

68 (19) 1.71 (0.53)

NI ⫽ Narcotrend index; TCI ⫽ target controlled infusion; LOC ⫽ loss of consciousness; UMSS ⫽ University of Michigan Sedation Scale.16 * In this age group eight children (⬎15 kg weight) were sedated using a TCI pump. † The NI values differed significantly, P ⫽ 0.0001 (but Group I is not significantly different from II). ‡ The TCI concentrations differed significantly among age groups, P ⬍ 0.01 (although Group II is not significantly different from III).

To help define the NI values for separating between LOC and transition from moderate to deep sedation, a ROC curve was constructed using a UMSS score threshold ⬎3 (Fig. 2). The cutoff value to use for the NI depends on a trade-off between sensitivity and specificity depending on which situation is regarded as more important to detect an awake situation when it is there or to detect “not awake.” If the importance of both are weighted equally, a suggested cutoff value for the NI is 76, where both sensitivity and specificity are 0.74. The mean NI values (Table 5) indicating LOC (⫽ transition from UMSS 2 to 3) were calculated for all cases and separately for each age group. The NI values differed significantly (P ⬍ 0.0001) among groups except between Groups I and II. The average NI value at LOC is 68, associated with a sensitivity of 0.67 and specificity of 0.79 as can be seen on the ROC curve. The 52 children sedated using the TCI pump had a very high correlation between propofol concentrations and sedation levels (Spearman correlation coefficient of 0.96). Table 5 shows the mean propofol TCI concentrations at LOC (transition from UMSS 2 to UMSS 3) for different age groups. In 5 of 61 patients (one with manual infusion scheme), the UMSS indicated lightening of the sedation level despite constant or increasing concentration of propofol. Children between 3 and 5 yr had higher propofol concentrations (P ⬍ 0.01) at LOC than older children. The logistic regression model was highly significant (␹2 233, P ⫽ 1 ⫻ 10⫺51), with coefficients NI, age and constant, all highly significantly differing from 0 (all P ⬍ 1 ⫻ 10⫺5). The logistic regression equation for predicted unconsciousness is log(p/(1⫺p)) ⫽ 5.361– 0.0659 NI– 0.0090 age (with age in months). The associated odds ratios are 0.936 (95% CI [0.926 – 0.946]) for NI and 0.991 (95% CI [0.987– 0.995]) for age.

DISCUSSION First, the current study shows that the Narcotrend monitor is capable of determining the depth of hypnosis and LOC during propofol sedation in children. However, there is a large overlap of NI values at each sedation level, and a relatively high probability (0.18) of incorrectly predicting changes in conscious states as determined by UMSS. Second, the NI is age-dependent: higher mean NI values are associated with younger children compared with Vol. 109, No. 1, July 2009

older children at the same sedation level and also at LOC. In addition, better agreement was observed between TCI concentrations and sedation (UMSS) than between NI and UMSS. Since hemodynamic variables, such as arterial blood pressure or heart rate, are of limited value as indicators of anesthetic depth both in adults21 and in children22; therefore, these were not included in our study. This study is the first to show that the Narcotrend is capable of discriminating LOC and between moderate and deep sedation levels during increasing propofol concentrations in children. To be able to do this clearly is essential in clinical work, if this device is used to guide the level of hypnosis adequate for diagnostic and therapeutic procedures. In adults, mixed results have been reported about the capability of Narcotrend to distinguish among hypnotic states. For example, the first studies using implicit memory task7 and observers assessment of anesthesia and sedation scale indicated that Narcotrend is able to distinguish among hypnotic states (PK ⫽ 0.92),23 but subsequent studies using the Isolated Forearm Technique demonstrated that the Narcotrend monitor is not able to discriminate between conscious and unconscious states during anesthesia.8,10 In children, only a few evaluative studies of the Narcotrend have been published, and these used sevoflurane and desflurane-based anesthesia techniques.11–14 Generally, these studies suggest good sensitivity of the NI to distinguish between awake and anesthetized patients and significant negative correlations with NI and end-tidal gas concentrations at nonsteady-state. Different methods used to measure sedation and consciousness may contribute to the better performance of the Narcotrend in these earlier studies compared with ours. The earlier studies suggested a perfect ability of NI to predict consciousness and unconsciousness (PK ⫽ 1.0), but the comparisons were made between deeply anesthetized and fully awake children and it is not clear how LOC was measured.11–13 A recent study reports a PK-value of 0.73 for the NI (and a PK of 0.93 when a 0 – 6 mo age group was excluded) to differentiate between consciousness and unconsciousness in children awakening from sevoflurane based anesthesia.14 The awakening was determined by response to verbal command, phonation, or continuous purposeful movement. We used the UMSS because it is © 2009 International Anesthesia Research Society

57

currently the best sedation score validated for children.16,18 In contrast to previous studies, comparisons were not made between completely awake and deeply anesthetized children. Instead, LOC was determined as the transition of the UMSS scale value from 2 to 3, as this corresponds to a change of hypnotic state in which patients could respond (UMSS 2) to a state (UMSS 3) in which they could no longer respond to verbal command or to light tactile stimulation. In our study, a PK value of 0.84 (0.80 – 0.88) for detection of LOC was determined. In contrast to other depth of anesthesia monitors, the age-related EEG changes are incorporated into the Narcotrend algorithm, but the impact of age on the calculation of the NI is proprietary. The setup menu requires the patient’s date of birth before further proceeding. The recommended NI values to provide adequate anesthesia for children are the same as for adults. However, in the current study, the NI values varied markedly across age groups at each sedation level, with younger children having higher NI values compared with those of older children. Also, the scatter of NI values was high and the mean NI values at LOC were higher in younger than in older (⬎12 yr) children. In the youngest age group, the boys were over-represented and we cannot exclude that this may have affected our results, but the impact of gender on NI is unknown. Furthermore, the UMSS always requires a subjective interpretation of patient responses to stimuli, which may vary among observers and may have, at least in part, influenced our results.18 Two earlier studies examined the effect of age on NI values in children.11–14 In agreement with our results, higher bispectral index (BIS) and NI values during anesthesia in children aged between 7 and 18 mo were obtained in one study,14 whereas the other reported no between-group differences of NI values at LOC in preschool children.11 Age-related changes have also been reported for BIS and entropy values during sevoflurane-based anesthesia. These studies reported low preawakening values in infants19,24,25 and higher values in small children (6 mo– 6 yr) compared with older children.14,24,26,27 Moreover, the scatter of the individual values for BIS, entropy, and NI have been shown to be age-dependent. One study reported that at awakening the spread of BIS and NI decreases as the age of the children increases.14 It was not the primary interest of this study to investigate relationships between TCI plasma propofol concentrations and NI values or UMSS scores. However, we found a better Spearman rank correlation between UMSS and TCI concentrations (0.96) than between UMSS and NI (⫺0.68), which suggests that the TCI is a valuable tool in guiding the level of hypnosis. The relationship between propofol TCI concentrations and the level of hypnosis is poorly investigated in children. Similar effect-site concentrations of propofol aimed at obtaining a BIS value of 50 were reported for children (3.65 ␮g/mL) and adults (3.75 58


␮g/mL) and were associated with a deep hypnosis (UMSS value 4) in both populations.28 One other study reports a moderate correlation between BIS and predicted plasma concentration of propofol (1.5 ␮g/mL) during emergence in children aged between 2 and 7 yr.29 In this study, children aged between 3 and 6 yr seemed to need higher TCI concentrations of propofol (2.25 ␮g/mL) to become unconscious than older children (1.78 Group II vs 1.37 ␮g/mL, Group III). However, comparability between these studies is difficult, since different pharmacokinetic models for TCI were used. Furthermore, the predictability of the plasma concentrations from the pharmacokinetic models is limited, especially in children. Kataria et al.’s model17 used in this study has been reported to under-estimate the measured plasma concentrations in older children (10 yr)30; therefore, the differences we observed among age groups at LOC might be over-estimated. The problem of artifacts and the high number of missing values has been reported in previous studies.8,11 We could not record the basal NI value for eight patients (13.1%), which is a smaller proportion than those reported in other studies (⬎50%).11,13,14 According to the manufacturer’s information, the Narcotrend does not show any values in the first few minutes after the start of measurement. Therefore, we worked diligently to get initial NI values by waiting and changing electrodes. In total, 11% of the measurements showed “⬎30 Hz” or “EMG” because of artifacts, and the EEG stage could not be obtained. In one study comparing BIS and Narcotrend monitors in children, the data acquisition and missing values were more a problem with Narcotrend than with the BIS monitor.14 In conclusion, the NI indicates transition between consciousness and unconsciousness and the hypnotic effect of propofol in children, but the accuracy is limited. Furthermore, the impact of age on the NI is unclear. Presently, the manufacturers of various EEG devices recommend the same values of “adequate anesthesia” for children and adults but, as shown by our study, values indicating LOC differ in children of various ages. In the future, children might benefit from EEG monitoring providing that it includes algorithms developed specifically for children. These should be able to distinguish between hypnotic states and not just between deeply anesthetized and awake children, which can be done without EEG monitoring.31 Titrating anesthetic dose against monitor indices, such as the Narcotrend, might be harmful for patients, since in attempting to minimize costs anesthesia providers may inadvertently provide insufficient levels of anesthesia in the mistaken belief that monitor readings are 100% reliable and valid. Until more studies confirm the validity of the Narcotrend and other devices, the clinical usefulness of depth of anesthesia monitoring in children is uncertain and in this population it should be used with caution. ANESTHESIA & ANALGESIA

REFERENCES 1. Davidson AJ, Huang GH, Czarnecki C, Gibson MA, Stewart SA, Jamsen K, Stargatt R. Awareness during anesthesia in children: a prospective cohort study. Anesth Analg 2005;100:653– 61 2. Mazoit JX. Pharmacokinetic/pharmacodynamic modeling of anesthetics in children: therapeutic implications. Paediatr Drugs 2006;8:139 –50 3. Powers KS, Nazarian EB, Tapyrik SA, Kohli SM, Yin H, van der Jagt EW, Sullivan JS, Rubenstein JS. Bispectral index as a guide for titration of propofol during procedural sedation among children. Pediatrics 2005;115:1666 –74 4. Davidson AJ. Monitoring the anaesthetic depth in children—an update. Curr Opin Anaesthesiol 2007;20:236 – 43 5. Watcha MF. investigations of the bispectral index monitor in pediatric anesthesia: first things first (editorial). Anesth Analg 2001;92:805–7 6. Schultz B, Grouven U, Schultz A. Automatic classification algorithms of the EEG monitor Narcotrend for routinely recorded EEG data from general anesthesia: a validation study. Biomed Tech (Berl) 2002;47:9 –13 7. Mu¨nte S, Mu¨nte TF, Grotkamp J, Haeseler G, Raymondos K, Piepenbrock S, Kraus G. Implicit memory varies as a function of hypnotic electroencephalogram stage in surgical patients. Anesth Analg 2003;97:132– 8 8. Schneider G, Kochs EF, Horn B, Kreuzer M, Ningler M. Narcotrend威 does not adequately detect the transition between awareness and unconsciousness in surgical patients. Anesthesiology 2004;101:1105–11 9. Kreuer S, Bruhn J, Larsen R, Bialas P, Wilhelm W. Comparability of Narcotrend™ index and bispectral index during propofol anaesthesia. Br J Anaesth 2004;93:235– 40 10. Russell IF. The Narcotrend “depth of anaesthesia” monitor cannot reliably detect consciousness during general anaesthesia: an investigation using the isolated forearm technique. Br J Anaesth 2006;96:346 –52 11. Weber F, Hollnberger H, Gruber M, Frank B, Taeger K. Narcotrend威 depth of anesthesia monitoring in infants and children. Can J Anaesth 2004;51:855– 6 12. Weber F, Gruber M, Taeger K. The correlation of the Narcotrend威 Index and classical electroencephalographic parameters with endtidal desflurane concentrations and hemodynamic parameters in different age groups. Paediatr Anaesth 2005;15:378 – 84 13. Weber F, Hollnberger H, Gruber M, Frank B, Taeger K. The correlation of the Narcotrend Index with endtidal sevoflurane concentrations and hemodynamic parameters in children. Pediatr Anesth 2005;15:727–32 14. Wallenborn J, Kluba K, Olthoff D. Comparative evaluation of Bispectral Index and Narcotrend Index in children below 5 years of age. Paediatr Anaesth 2007;17:140 –7 15. Ganesh A, Watcha MF. Bispectral index monitoring in pediatric anesthesia. Curr Opin Anaesthesiol 2004;17:229 –34 16. Malviya S, Voepel-Lewis T, Tait AR, Merkel S, Tremper K, Naughton N. Depth of sedation in children undergoing computed tomography: validity and reliability of the University of Michigan Sedation Scale (UMSS). Br J Anaesth 2002;88:241–5

Vol. 109, No. 1, July 2009

17. Kataria BK, Ved SA, Nicodemus HF, Hoy GR, Lea D, Dubois MY, Mandema JW, Shafer SL. The pharmacokinetics of propofol in children using three different data analysis approaches. Anesthesiology 1994;80:104 –22 18. Malviya S, Voepel-Lewis T, Tait AR. A Comparison of observational and objective measures to differentiate depth of sedation in children from birth to 18 years of age. Anesth Analg 2006; 102:389 –94 19. Klockars JG, Hiller A, Ranta S, Talja P, van Gils MJ, Taivainen T. Spectral entropy as a measure of hypnosis in children. Anesthesiology 2006;104:708 –17 20. Smith WD, Dutton RC, Smith N. Measuring the performance of anesthetic depth indicators. Anesthesiology 1996;84:38 –51 21. Schmidt GN, Bischoff P, Standl T, Issleib M, Voigt M, Schulte Am Esch J. ARX-derived auditory evoked potential index and bispectral index during the induction of anesthesia with propofol and remifentanil. Anesth Analg 2003;97:139 – 44 22. McCann ME, Bacsik J, Davidson A, Auble S, Sullivan L, Laussen P. The correlation of bispectral index with end-tidal sevoflurane concentration and haemodynamic parameters in preschoolers. Paediatr Anaesth 2002;12:519 –25 23. Bauerle K, Greim C, Schroth M, Geisselbrecht M, Ko¨bler A, Roewer N. Prediction of depth of sedation and anaesthesia by the Narcotrend™ EEG monitor. Br J Aaesth 2004;92:842–5 24. Davidson AJ. Performance of entropy and Bispectral Index as measures of anaesthesia effect in children of different ages. Br J Anaesth 2005;95:674 –9 25. Malviya S, Voepel-Lewis T, Tait AR, Watcha MF. Effect of age and sedative agent on the accuracy of bispectral index in detecting depth of sedation in children. Pediatrics 2007;120:461–70 26. Kim HS, Oh AY, Kim CS, Kim SD, Seo KS, Kim JH. Correlation of bispectral index with end-tidal sevoflurane concentration and age in infants and children. Br J Anaesth 2005;95:362– 6 27. Wodey E, Tirel O, Bansard JY, Terrier A, Chanavaz C, Harris R, Ecoffey C, Senhdji L. Impact of age on both BIS values and EEG bispectrum during anaesthesia with sevoflurane in children. Br J Anaesth 2005;94:810 –20 28. Munõz HR, Cortínez LI, Ibacache ME, Leoń PJ. Effect site concentrations of propofol producing hypnosis in children and adults: comparison using the bispectral index. Acta Anaesthesiol Scand 2006;50:882–7 29. Park H, Kim YL, Kim CS, Kim SD, Kim HS. Changes of bispectral index during recovery from general anesthesia with 2% propofol and remifentanil in children. Paediatr Anaesth 2007;17:353–7 30. Rigouzzo A, Girault L, Louvet N, Servin F, De-Smet T, Piat V, Seeman R, Murat I, Constant I. The relationship between Bispectral Index and propofol during target-controlled infusion anesthesia: a comparative study between children and young adults. Anesth Analg 2008;106:1109 –16 31. Schneider G, Kochs EF. The search for structures and mechanisms controlling anesthesia-induced unconsciousness. Anesthesiology 2007;107:195– 8


59

Review Article

Perioperative Management of Children with Obstructive Sleep Apnea Deborah A. Schwengel, MD*† Laura M. Sterni, MD† David E. Tunkel, MD*†‡ Eugenie S. Heitmiller, MD*†

Obstructive sleep apnea syndrome (OSA) affects 1%–3% of children. Children with OSA can present for all types of surgical and diagnostic procedures requiring anesthesia, with adenotonsillectomy being the most common surgical treatment for OSA in the pediatric age group. Thus, it is imperative that the anesthesiologist be familiar with the potential anesthetic complications and immediate postoperative problems associated with OSA. The significant implications that the presence of OSA imposes on perioperative care have been recognized by national medical professional societies. The American Academy of Pediatrics published a clinical practice guideline for pediatric OSA in 2002, and cited an increased risk of anesthetic complications, though specific anesthetic issues were not addressed. In 2006, the American Society of Anesthesiologists published a practice guideline for perioperative management of patients with OSA that noted the pediatric-related risk factor of obesity, and the increased perioperative risk associated with adenotonsillectomy in children younger than 3 yr. However, management of OSA in children younger than 1 yr-of-age was excluded from the guideline, as were other issues related specifically to the pediatric patient. Hence, many questions remain regarding the perioperative care of the child with OSA. In this review, we examine the literature on pediatric OSA, discuss its pathophysiology, current treatment options, and recognized approaches to perioperative management of these young and potentially high-risk patients. (Anesth Analg 2009;109:60 –75)

M

anagement of the child known to have or is suspected of having obstructive sleep apnea syndrome (OSA) is a challenge for the anesthesiologist. Children with OSA have recurrent episodes of partial or complete airway obstruction during sleep, resulting in hypoxemia, hypercapnia, and sleep disruption, and approximately 1%–3% of all children are thought to have OSA.1–5 Children who carry the diagnosis of OSA usually have been evaluated by an otolaryngologist, pulmonologist, or sleep specialist, often in preparation for adenotonsillectomy. However, children scheduled for other types of surgical procedures may not be as thoroughly evaluated; OSA in these patients may be missed by surgeons or primary care providers.6 – 8 Anesthesiologists are aware of the concerns in caring for children diagnosed with OSA, and 82% of surveyed anesthesiologists reported that guidelines would assist them in caring for them.9 The American From the Departments of *Anesthesiology/Critical Care Medicine, †Pediatrics, and ‡Otolaryngology Head and Neck Surgery, Johns Hopkins University School of Medicine, Baltimore, Maryland. Accepted for publication December 19, 2008. Reprints will not be available from the author. Address correspondence to Deborah Schwengel, MD, Johns Hopkins Medical Institutions, Department of Anesthesiology/Critical Care Medicine, 600 North Wolfe St., Blalock 1412, Baltimore, MD 212878711. Address e-mail to [email protected]. Copyright © 2009 International Anesthesia Research Society DOI: 10.1213/ane.0b013e3181a19e21

60

Society of Anesthesiology (ASA) Task Force on Perioperative Management of Patients with OSA suggests a scoring system to estimate perioperative risk based on severity of sleep apnea, invasiveness of procedure, and postoperative opioid requirement. This task force also supports postoperative admission for children ⬍3 yr-of-age who undergo adenotonsillectomy. However, medical evidence is lacking as to whether a monitored bed is required for some patients with OSA. Because specific guidelines are lacking regarding such patients, local institutional policies often lead to controversies among anesthesiologists, surgeons, and third-party payers regarding the need for, and level of, postoperative monitoring.10 This review will summarize what is currently known about childhood OSA, including diagnosis, treatment, and strategies for perioperative care of these patients.

DIAGNOSIS OSA presents differently in young children than it does in teenagers and adults (Table 1). Adults and teenagers with OSA are often obese and have daytime somnolence; younger children may have normal weight or failure to thrive and behavior disorders such as hyperactivity, attention problems, and enuresis.11,12 Both adult and pediatric patients have an increased risk for complications during or after surgery.10 Vol. 109, No. 1, July 2009

Table 1. Childhood Versus Adult Obstructive Sleep Apnea Syndrome Features Children Presentation Age Gender Obesity Tonsils and adenoids Daytime sleepiness Sleep Obstruction Sleep architecture Arousals with obstruction Treatment Surgical Medical (positive airway pressure)

Adults

2–6-yr peak Male ⫽ female Few Often enlarged Less common than in adults but can be seen

Increased elderly Males ⬎ females Most Rarely enlarged Common

Obstructive apnea or hypoventilation Usually normal May not be seen

Obstructive apnea Decreased delta and REM At end of each apnea

Definitive therapy in most patients

Minority of cases with inconsistent results Most common therapy

Selected patients

Adapted from Sterni and Tunkel, Pediatr Clin North Am 2003;50:427– 43. REM ⫽ rapid eye movement.

A history of snoring is a sensitive, though not specific, symptom that supports the diagnosis of OSA.13 Primary snorers are those who snore but have no obstructive apnea, gas exchange abnormalities, or multiple arousals on polysomnography (sleep study). Approximately 10% of children have primary snoring.14 –17 Primary snoring does not seem to progress to OSA and may resolve over time,13,18,19 and although generally considered benign, it may be associated with neurobehavioral changes, such as attention disorders, mild cognitive problems, and anxious or depressive symptoms.20 OSA in children may manifest with obstructive apneas or obstructive hypoventilation (OH).21 Children with OH exhibit snoring with continuous partial upper airway obstruction during sleep, leading to paradoxical respiratory effort, hypercarbia, and hypoxemia in some.22 As many as 40% of snoring children who are referred to a sleep clinic or otolaryngologist will have OSA,1–5 and children with OSA almost always snore and have increased respiratory effort during sleep.3,23,24 Screening children for OSA is recommended by the American Academy of Pediatrics as part of routine health maintenance. Evaluation should be pursued for children with a history of nightly snoring. Although several pediatric studies were unable to show that questionnaires differentiated those with OSA from those with primary snoring,3,5,25–29 two independent studies of Chinese children reported validated questionnaires for OSA.30,31 Direct observation of a child having apnea and labored respirations during sleep may support the diagnosis of OSA. However, parents may not be able to distinguish primary snoring from obstructive snoring; as well, obstructive episodes that occur during rapid eye movement (REM) sleep may go unwitnessed.32 Videotapes, audiotapes with pulse oximetry, questionnaires, and daytime nap studies may be used but do not exclude OSA when negative.3,5,31,33– 40 Unattended home polysomnography Vol. 109, No. 1, July 2009

Table 2. Components of Polysomnography Recommended by The American Thoracic Society Respiratory effort—assessed by abdominal and chest wall movement Airflow at nose, mouth, or both Arterial oxygen saturation End-tidal CO2 or transcutaneous CO2 (recommended specifically for pediatric polysomnography to detect hypoventilation) Electrocardiograph Electromyography (tibial) to monitor arousals Electroencephalography, electrooculography, and electromyography for sleep staging From Standards and indications for cardiopulmonary sleep studies in children. American Thoracic Society. Am J Respir Crit Care Med 1996;153:866 –78, with permission.

might prove to be a useful testing modality,41,42 but validation of these studies is needed. The “gold standard” for making the diagnosis of OSA in children is use of overnight polysomnography. However, there are significant differences in the criteria for the performance, scoring, and interpretation of pediatric versus adult polysomnograms; for complete efficacy, it is essential that polysomnography laboratories have experience in the performance and interpretation of these studies in children.43 Use of sedatives and sleep deprivation is not recommended.44,45 If sufficient sleep and REM time are captured, a single overnight study is usually adequate. Obstructive events in children with OSA occur primarily during REM sleep,46 although adult patients exhibit non-REM preponderance or equal REM and non-REM obstruction.47 Obstruction is thought to worsen over the course of a night, however, evaluation of upper airway muscles in children has not shown clinically significant muscle fatigue.48 The effect of the perioperative period on this process has not been adequately studied. Postoperative REM rebound may worsen OSA in some patients.47,49 The components of polysomnography recommended by the American Thoracic Society are listed in Table 2. Respiratory events that may be seen during © 2009 International Anesthesia Research Society

61

Table 3. Respiratory Events that can be Seen During Polysomnography Event Central apnea Obstructive apnea Obstructive hypopnea Mixed apneas Obstructive hypoventilation

Definition Pause in airflow with absent respiratory effort, scored when ⬎20 s or two missed breaths and a ⬎3% drop in oxygen saturation ⬎90% reduction of airflow despite continuing respiratory effort, scored when event lasts at least two missed breaths in children ⬎50% reduction of airflow with associated with respiratory effort, scored when at least two missed breaths and ⬎3% drop in oxygen saturation or arousal ⱖ90% reduction in airflow, lasting at least two missed breaths, and containing absent respiratory effort initially (a central apneic pause), followed by resumption of respiratory effort without a resumption of airflow (an obstructive apnea) End-tidal CO2 ⬎50 mm Hg for ⬎25% of the total sleep time with paradoxical respirations, snoring, and no baseline lung disease

polysomnography are listed and defined in Table 3 and depicted in Figure 1. Often, obstructive apneas and hypopneas identified during a sleep study are combined to provide the apnea– hypopnea index (AHI), defined as the number of discrete obstructive events per hour. Many sleep laboratories will report an AHI or a respiratory disturbance index (RDI) that will include the number of all scored respiratory events (including central apneas) per hour. Indices that combine central and obstructive events cannot be used to diagnose OSA in children, as normal children have more frequent central apneas when compared with adults. Only obstructive event indices should be used to identify pediatric OSA.50,51 Data correlating polysomnography parameters with clinical outcomes in children are lacking, and there are no standard guidelines for classifying the severity of OSA in children. Polysomnography data from nonsnoring children have defined OSA as more than one obstructive event per hour.37,52,53 It is important to note that the scoring does not consider the length of time of the obstructive event. The AHI may be misleadingly low in children who have OH rather than discrete apnea. OH is not scored as an event but diagnosed as shown in Table 3. Thus, children with OH may have significant disease with a low AHI if the periods of OH are few but long. For this reason, other factors must be considered. In our Pediatric Sleep Laboratory, we classify the severity of OSA based on total clinical picture, number of obstructive events per hour, duration of elevated end-tidal CO2, and frequency and severity of oxygen desaturation. We classify OSA as severe if the patient has an AHI of ⱖ10/h because of the increased risk of respiratory compromise after adenotonsillectomy.54 Oxygen saturation nadir ⬍80% is also suggestive of severe disease and postadenotonsillectomy respiratory morbidity55–58 Suggested guidelines for assessing severity of OSA based on polysomnography are listed in Table 4.

PATHOPHYSIOLOGY The essential feature of OSA in children is increased upper airway resistance during sleep.59 Adenotonsillar hypertrophy, allergic rhinitis, turbinate hypertrophy, deviated septum, and maxillary constriction 62

Obstructive Sleep Apnea in Children

cause airway narrowing in children.60 Enlarged tonsils can lead to collapse of the hypopharynx at the level of the soft palate because of posterior displacement of the tongue and descent of the tonsils.61 Although enlarged tonsils and adenoids are clearly an important risk factor in children,62 there is no absolute correlation between the size of the tonsils and adenoids and the presence of OSA.63– 66 Enlarged soft tissues from obesity or lymphoid tissues contribute to OSA in many children. Other factors include abnormal central arousal threshold, abnormal bony anatomy, disordered neural control of airway caliber or sensation, and decreased pharyngeal tone that may be seen in certain types of cerebral palsy.67 Dynamic airway narrowing or collapse occurs at multiple sites in children with OSA.59 Thus, OSA is often a multifactorial disorder with overlapping influences that together predispose the patient to obstructed breathing (Fig. 2).22 Genetics affect the risk of developing OSA. The incidence of OSA is higher in first-degree relatives of index patients with OSA.68 Infants of families with multiple members affected by sudden infant death syndrome, apparent life-threatening event, and OSA are more likely to have OSA than those in families with only one case of sudden infant death syndrome or apparent life-threatening event; these infants were found to have OSA in their first year of life.69 Craniofacial abnormalities with altered airway anatomy are associated with abnormal breathing function.70 The combination of enlarged tonsils and craniofacial abnormalities (Table 5) can predict development of OSA.71,72 Obesity is an important and increasingly common risk factor for OSA in children. Deposition of adipose tissue around the upper airway and external compression from the excess soft tissue around the neck and jaw lead to upper airway narrowing. Decreased chest wall compliance and upward displacement of the diaphragm by the obese abdomen when the individual is supine lead to smaller lung volumes during sleep, decreased oxygen stores, and increased risk of desaturation with obstructive events.73 Obese children continue to have an increased risk of continuing OSA after adenotonsillectomy.74 ANESTHESIA & ANALGESIA

Figure 1. Typical polysomnographic recordings demonstrating: (A) Obstructive apnea. Note how the nasal thermistor and expired CO2 tracings flatten while paradoxical respiratory efforts occur. The event is accompanied by a decrease in arterial oxygen saturations. (B) Obstructive hypopnea. The hypopnea is characterized by a decrease in nasal pressure signal associated with continuing paradoxical respiratory effort and desaturation. (C) Central apnea is distinguished from obstructive apnea by the absence of respiratory effort, which leads to the loss of airflow/pressure signals. This tracing shows one central apnea episode.

Congenital syndromes carry an increased risk of OSA (Table 6). Patients with Trisomy 21 may be predisposed to OSA,75–78 particularly those with midface and mandibular hypoplasia, a small upper Vol. 109, No. 1, July 2009

airway combined with relatively large and medially positioned tonsils, macroglossia, glossoptosis, increased secretions, obesity, and generalized hypotonia.79 © 2009 International Anesthesia Research Society

63

Figure 1. Continued. Chronic OSA can have both reversible and irreversible consequences for the airway. Children with OSA have increased pharyngeal collapsibility and lose normal neuromotor responses to hypercarbia and negative pressure.22,80,81 This increased collapsibility occurs with even mild OSA and may predispose these patients to more severe pharyngeal collapse later in life.81 Reversible and irreversible effects on the cardiovascular system are also possible. Chronic obstructed breathing leads to chronic hypoxemia, hypercarbia or both and, if left untreated, can eventually lead to pulmonary hypertension.23,24,70 Pulmonary hypertension develops from vasoconstriction of the pulmonary arterial vessels in response to the chronic nocturnal hypoxemia, hypercarbia, and acidosis that accompany severe untreated OSA.82– 86 In a series of 92 children with adenotonsillar hypertrophy, 3.3% developed pulmonary hypertension that was reversed with adenotonsillectomy.87 Cardiovascular morbidity is associated with endothelial dysfunction. Gozal et al.88 showed that soluble CD40 ligand levels (sCD40L triggers inflammatory and procoagulant states) were elevated in children with polysomnography-proven OSA; tonsillectomy produced both significant improvements in AHI and sCD40L in all children except those with significant family histories of cardiovascular diseases. Cardiac dysfunction associated with OSA may be manifested by structural and functional changes of both ventricles. Right ventricular dysfunction develops from chronically elevated pulmonary pressure and negative intrathoracic pressure created by breathing against a partially closed upper airway over 64


time.70 If pulmonary hypertension is left untreated, cor pulmonale is eventual and is diagnosed by echocardiographic findings of right ventricular hypertrophy, ventricular enlargement, pulmonary and tricuspid valve insufficiency, decreased ejection fraction, and dilation of the pulmonary artery.70,89,90 In time, right atrial pressure increases, and decreased venous return to the heart results in peripheral edema, hepatic congestion, and ascites. Fortunately, right ventricular dysfunction and cor pulmonale may be reversible with surgical treatment of OSA.83,86,89 Although right ventricular dysfunction is classic, biventricular hypertrophy can develop. It is more likely to be seen in patients with severe OSA, but has been reported in patients with only mild OSA.91 Children with OSA show signs of enhanced sympathetic activity,92 autonomic dysfunction93, and endothelial dysfunction.88 Systemic and diastolic hypertension or a trend toward higher arterial blood pressures has been documented in children with OSA.24,94 –96 Higher blood pressures, especially at night, have been associated with increased severity of OSA that is particularly related to desaturation events,97,98 even though the blood pressures reported might not have been high enough to require treatment with antihypertensives. Other studies show evidence of endothelial dysfunction with OSA but no hypertension.88 In adults, it has been shown that intermittent hypoxia, not hypercapnia, is the critical stimulus for OSAassociated sympathetic activation, endothelial dysfunction, oxidative stress and inflammation, which produce cardiovascular dysfunction.99 Right ventricular dysfunction and overload leads to bowing of the ventricular septum and subsequent ANESTHESIA & ANALGESIA

Figure 2. Pathophysiology of pediatric obstructive sleep apnea

increased left end-diastolic pressure that can result in pulmonary edema and pulmonary parenchymal damage.70,100,101 Chronic hypoxia is an independent risk factor for the development of left ventricular hypertrophy,91 which is a known risk factor for future cardiovascular disease.102,103 Severe OSA doubles the risk of congestive heart failure in adult patients.104 Patients with OSA can also experience postobstructive pulmonary edema due to either acute airway obstruction and generation of marked negative inspiratory pressures or in the relief of significant chronic airway obstruction. In both cases, physical damage to pulmonary capillaries, release of vasoactive mediators, and hydrostatic forces that result in fluid transudation to the pulmonary parenchyma may occur.105 Impaired growth has been seen in some children with OSA23–25 and has been thought to be related to increased work of breathing during sleep.106 These children have been shown to have impaired secretion of nocturnal growth hormone.107 Improved growth has been reported after treatment with adenotonsillectomy.106 –109

syndrome. The different influences on the airway are additive. Adapted from Marcus, Respir Physiol 2000;119:143–54.

TREATMENT

Table 4. A Severity Ranking System Based on Polysomnography Apnea-hypopnea index Normal Mild OSA Moderate OSA Severe OSA

0–1 2–4 5–9 ⬎10

Oxygen saturation nadir ⬎92 ⬍80

Peak ETCO2 values and percent of time spent with ETCO2 ⬎50 mm Hg should also be considered when assessing severity. OSA ⫽ obstructive sleep apnea syndrome.

Table 5. Facial and Airway Features Suggestive of Obstructive Sleep Apnea Small triangular chins Retro-position of the mandible Steep mandibular plane High palate Long, oval-shaped face Long soft palate Large tonsils in association with the above facial features Adapted from Guilleminault et al., Pediatrics 1996;98:871– 82; Guilleminault et al., Otolaryngol Head Neck Surg 2007;136:169 –75.

Table 6. Some Congenital and Medical Conditions Associated with Obstructive Sleep Apnea Syndrome Achondroplasia Apert syndrome Beckwith–Wiedemann syndrome Cerebral palsy Choanal stenosis Cleft palate patients after repair Crouzon syndrome Cystic hygroma Down syndrome Hallermann–Streiff syndrome Hypothyroidism Klippel–Feil syndrome Mucopolysaccharidosis Obesity Osteopetrosis Papillomatosis (oropharyngeal) Pierre Robin syndrome Pfeiffer syndrome Pharyngeal flap surgery Prader–Willi syndrome Sickle cell disease Treacher–Collins syndrome From Sterni and Tunkel, Pediatr Clin North Am 2003;50:427– 43, with permission.

Vol. 109, No. 1, July 2009

Adenotonsillectomy is the treatment of choice for most children with OSA. Improvement in sleeprelated airway obstruction and quality-of-life measures is estimated to occur in more than 75% of children after adenotonsillectomy.110,111 However, persistent respiratory abnormalities may be seen in obese children and children with the most severe cases of OSA.74,112 Suen et al.26 suggested that a RDI ⬎19.1 may be predictive of persistent OSA after adenotonsillectomy. The range of surgical techniques for adenotonsillectomy reflects attempts to reduce the considerable postoperative discomfort and minimize the risk of hemorrhage after surgery. The use of radiofrequency volume reduction or powered intracapsular tonsillectomy can reduce perioperative pain by avoiding trauma to surrounding peritonsillar tissues. These procedures provide relief of disordered breathing in children with OSA, and recovery is more rapid than with total tonsillectomy techniques113–115; however, there is a risk of regrowth of tonsillar tissue.116 Children with abnormal craniofacial anatomy or abnormalities of neuromotor tone may require additional treatment of persistent OSA, including pharyngeal surgery, craniofacial surgery, and even tracheostomy.117 Uvulopalatopharyngoplasty has been used for treatment of persistent OSA in children with neuromotor disease (such as cerebral palsy) or with craniofacial anomalies (such as those seen with Trisomy 21). Uvulopalatopharyngoplasty includes resection of the uvula, part of the soft palate and tonsillar pillars, with the goal of reducing upper airway obstruction at the level of the palate and oropharyngeal and nasopharyngeal levels. Tongue reduction procedures have been used in syndromic children with © 2009 International Anesthesia Research Society

65

obstructive macroglossia, and flap takedown can be performed for children with sleep-related airway obstruction after pharyngeal flap surgery for velopharyngeal insufficiency.118 Craniofacial procedures, such as mandibular distraction/advancement, genioglossus advancement, and midfacial advancement, have been used to treat OSA that results from craniofacial structural abnormalities.119 Noninvasive nasal positive-pressure ventilation is a common medical treatment for OSA in children. Continuous positive airway pressure (CPAP) delivers a constant pressure to the airway; bi-level positive airway pressure (BiPAP) applies pressure that decreases during exhalation. The positive pressure mechanically stents the airway open and leads to improved functional residual capacity.120 Both CPAP and BiPAP are used safely and successfully for children who have contraindications to adenotonsillectomy, persistent OSA after adenotonsillectomy, minimal adenotonsillar tissue, or prefer nonsurgical interventions.120 –127 The level of positive pressure required to eliminate obstructive apneas and normalize ventilation and night-time oxygen saturation must be determined in the sleep laboratory. Serial evaluation and adjustment of CPAP is required for growing children, as their pressure requirements change with time.121,126 Complications of CPAP and BiPAP are usually minor and include local discomfort or irritation from poor mask fit, eye irritation, conjunctivitis, congestion, and skin ulceration. Children using noninvasive ventilation should have regular assessment of facial development. Midfacial hypoplasia has been reported with longterm use.128 Regular evaluation of mask fit can help to avoid these difficulties. Infrequently, hypoventilation can be seen; BiPAP with a back-up rate can be used.122 Pneumothorax and clinically significant reductions in cardiac output have not been reported in children treated for OSA.121 The greatest limitation to the use of noninvasive positive-pressure ventilation in children is poor compliance.121,125 Nocturnal oxygen supplementation has been used as a temporary treatment for patients with significant hypoxia associated with OSA until definitive therapy can be provided.121,129 In most cases, sleep-disordered breathing is not worsened. However, in some patients, supplemental oxygen may suppress the hypoxic ventilatory drive and worsen hypercapnia.121 Nocturnal oxygen therapy for children with OSA should be initiated only under monitored conditions, including assessment of CO2 exchange. The use of orthodontic devices in the treatment of OSA in children seems promising. Orthodontic maxillary expansion can improve sleep-related airway obstruction in children with narrow palates,72,130,131 but further studies are necessary to define the indications, proper candidates, and effectiveness of orthodontic treatment in the care of children with OSA. 66


Table 7. Key Questions to Ask Parents Does your child have difficulty breathing during sleep? Have you observed symptoms of apnea? Have you observed sweating while your child sleeps? Does your child have restless sleep? Does your child breathe through his/her mouth when awake? Are you worried about your child’s breathing at night? Do you have any family history of obstructive sleep apnea, sudden infant death syndrome, or apparent life-threatening events? Does your child have behavioral problems? Adapted from Li et al., Pediatr Pulmonol 2006;41:1153– 60; Brouilette et al., J Pediatr 1984;105:10 –14; Messner, Otolaryngol Clin North Am 2003;36:519 –30; McNamara and Sullivan, J Pediatr 2000;136:318 –23; Whiteford et al., Arch Dis Child 2004;89:851–5.

ANESTHETIC MANAGEMENT Children with OSA present for all types of surgical and diagnostic procedures. Although the most studied procedure for this patient population is adenotonsillectomy, there is evidence that perioperative complications are increased in OSA patients after all types of surgery.132

Preoperative Assessment Primary care providers do not routinely screen patients for OSA, nor do they demonstrate adequate knowledge about pediatric sleep disorders.6 – 8 The prudent anesthesiologist should screen patients beginning with the question: Does your child snore? A history of nightly snoring is a sensitive (91%) but not completely specific (75%) marker of OSA.32 If a patient regularly snores, additional focused questions may help to clinically identify those with OSA (Table 7), especially in patients with known OSA risk (Table 6). Children ⬍1-yr-of-age comprise a special group of patients that has not been adequately studied, but there is OSA in this age group. Presenting signs are snoring, apnea, failure to thrive, developmental delay, and recurrent respiratory infections.133 It is not clear if infants are in a subset of OSA with more severe disease, more comorbidities, or genetic predisposition. Some infants studied at 2-mo-of-age showed resolution of OSA between 6 and 12-mo-old.69 The physical examination should include an airway assessment: nasal anatomy, ability to breathe through the nose, presence of elongated facies, oral aperture size, mandibular size, intermaxillary distance, thyromental distance, tonsillar size, tongue volume, body habitus, and Mallampati score.134 Guilleminault et al.71 showed increased risk of sleep-disordered breathing in children with the physical features listed in Table 5. The examination of the patient should also include an assessment of muscle tone, handling of oral secretions, and observation of facial malformations. When history, physical examination, and sleep laboratory data are combined, risk assessment is more accurate.56 Most children presenting for adenotonsillectomy do not need cardiac evaluation. However, adult patients with multiple episodes of severe hypoxemia, ANESTHESIA & ANALGESIA

Table 8. Clinical Features that Predict Respiratory Compromise After Adenotonsillectomy and, in Some Cases, Persistent Obstructive Sleep Apnea Severe obstructive sleep apnea on polysomnography History of prematurity, especially with respiratory disease Age ⬍3 yr Morbid obesity Nasal problems (deviated septum, enlarged turbinates) Mallampati score 3 or 4 Neuromuscular disorders/disordered pharyngeal tone Genetic or chromosomal disorders Craniofacial disorders Enlarged lingual tonsils Upper respiratory infection within 4 wk of surgery Cor pulmonale Systemic hypertension Marked obstruction on inhalational induction Disordered breathing in the postanesthesia care unit Difficulty breathing during sleep Growth impairment due to chronic obstructed breathing Adapted from Blum and McGowan, Paediatr Anaesth 2004;14:75– 83; Guilleminault et al., Otolaryngol Head Neck Surg 2007;136:169 –75; Gerber et al., Arch Otolaryngol Head Neck Surg 1996;122:811–14; McGowan et al., Pediatr Pulmonol 1992;13:222– 6; Fricke et al., Pediatr Radiol 2006;36:518 –23.

defined as oxygen saturation ⬍70%, are at risk of left ventricular dysfunction135 and childhood OSA is associated with hypertension and arterial blood pressure dysregulation.136,137 Patients with cardiac involvement are at increased risk of perioperative cardiopulmonary complications.70,82,83 Although the available data are somewhat limited in the pediatric population, we recommend cardiac evaluation for any child with signs of right ventricular dysfunction, systemic hypertension, or multiple episodes of desaturation below 70%. Electrocardiogram and chest radiograph are not sensitive evaluation tools; echocardiography is recommended.102,138 Routine blood gas analysis is not recommended, but a basic metabolic panel can identify a patient with compensatory metabolic alkalosis in response to chronic hypercarbia, and a hemoglobin level may identify the patient with severe chronic hypoxemia.139 Preoperative CPAP has been used to reduce the postoperative complication rate and increase airway patency in adult OSA patients,140,141 and may be beneficial for certain pediatric patients.121 In our practice, children with very severe OSA who are at risk for persistent OSA (Table 8) and those with cardiovascular complications from OSA are considered for preoperative CPAP/BiPAP therapy. Effective CPAP/BiPAP therapy may improve pulmonary hypertension and reduce the patient’s surgical risks.142 This therapy is usually initiated by a pediatric pulmonologist, and the pressures required to treat the patient’s OSA are determined in the sleep laboratory. The child’s preoperative CPAP/BiPAP regimen can also be used in postoperative care, and the patient’s response to adenotonsillectomy and need for long-term CPAP can be determined several weeks postoperatively. Vol. 109, No. 1, July 2009

Intraoperative Management Studies that compare one anesthetic to another for adenotonsillectomy are few, and no technique is preferred. Sedative and anesthetic medications alter the CO2 response curve, theoretically placing OSA patients at higher risk of sedation and anesthesiainduced respiratory complications.143 Patients with OSA rescue themselves during obstructive episodes by arousal from sleep, but sedatives or residual anesthetics may make it impossible for patients to arouse themselves during obstructive episodes. Consequently, short-acting anesthetics should be chosen. The use of sedatives in pediatric OSA patients has not been well studied, and the few studies that exist do not include control groups. Preoperative administration of midazolam 0.5 mg/kg to 70 children undergoing adenotonsillectomy for OSA (diagnosed as severe in 40% of subjects by polysomnography) resulted in two children having respiratory events; one had a self-limited desaturation event before surgery, and one had postoperative obstruction with desaturation, requiring a nasal airway.144 In another small series, patients with Trisomy 21 were successfully sedated for magnetic resonance imaging studies with dexmeditomidine and ketamine without airway instrumentation.79 Sedatives used before the induction of general anesthesia may delay emergence in patients, especially for short cases.145,146 Without more evidence, we conclude that patients with OSA can receive sedatives but require monitoring until recovery can be demonstrated. There is no consensus regarding a best anesthetic induction strategy in OSA, but common sense dictates that some patients with OSA require approaches different from those used with normal patients. Children with altered bony anatomy or syndromes are at higher risk for having a difficult airway. Induction of anesthesia with volatile anesthetics results in airway collapse from relaxation of the genioglossus muscle, thus placing the OSA patient at high risk for airway obstruction.147–149 Positioning in an upright or lateral position, use of a jaw thrust maneuver, delivery of positive pressure by face mask and placement of an oral airway may aid in relieving the obstruction.150,151 In cases when the patient is only partially anesthetized and suffering from airway collapse, an airway device may not be tolerated. If the patient remains in a state of obstruction, desaturation ensues. Severe airway obstruction in a spontaneously breathing patient may result in a very high negative inspiratory force generated although the patient is inhaling against the collapsed pharynx or closed glottis; the increased pulmonary blood flow and pulmonary microvascular pressure that ensues can result in postobstructive pulmonary edema.70,117 IV induction can be used to rapidly induce a deep plane of anesthesia ready for airway instrumentation. This technique may be preferable for patients with very severe OSA. Children with simple adenotonsillar hypertrophy with normal © 2009 International Anesthesia Research Society

67

body habitus and without maxillofacial malformations are often easily mask-ventilated once an oral airway is placed, and endotracheal intubation is likewise usually straightforward. Children with OSA in addition to craniofacial abnormalities or other significant airway disorders must be evaluated for potential difficult intubation. A recent study of 40,000 adults showed OSA to be an independent predictor of impossible mask ventilation (Kheterpal S et al. Incidence, predictors, and management of impossible mask ventilation; A review of 40,000 anesthetics, 2008 ASA annual meeting, abstract A1243). Whether this association applies to the pediatric population with OSA is not known. Laryngeal mask airways are used by some anesthesiologists for adenotonsillectomy,152–154 but studies are lacking regarding their use in patients with OSA for this procedure. They can also be used in OSA patients with bony abnormalities for the purpose of managing a difficult airway.155,156 Upon completion of the procedural anesthetic, patients should be awake and have adequate strength to maintain the upper airway before tracheal extubation. We do not recommend deep extubation; patients with severe OSA and those with comorbidities are at risk of persistent OSA after surgery (Table 8). Extubation should take place in a controlled environment with appropriate personnel and equipment available; once stable, the extubated patient can be transported to the appropriate care unit. Before extubation, we sometimes place nasal airways in patients with severe OSA. The child who continues to have significant obstructive episodes after extubation can be positioned in the lateral decubitus or prone position to help relieve the obstruction. CPAP or BiPAP can be used to assist ventilation and relieve airway collapse. The placement of a nasal airway before extubation might be considered in more severe cases. Reintubation may be required in an occasional patient. Efforts are made to reduce the risk of postoperative vomiting and pain after adenotonsillectomy. A study that compared anesthetic techniques using sevoflurane and propofol with muscle relaxant and fentanyl showed no statistical difference in postoperative vomiting.157 Steroids have been shown to improve postoperative oral intake and reduce pain and vomiting.158–161 A dosing study of IV dexamethasone for adenotonsillectomy showed that low dose (0.0625 mg/kg) is just as effective as high dose (1 mg/kg) in reducing postoperative pain and vomiting.162 In another study, adult tonsillectomy patients who were treated with steroids had less pain, less nausea, and vomiting and improved healing as compared with those who were not treated.163 Although nasal steroids have been used to treat pediatric OSA, with improvement in sleepdisordered breathing,164 no studies have examined the use of nasal steroids to reduce perioperative swelling. A review and meta-analysis of antiemetics showed that antiserotonergic drugs and dexamethasone were 68


most effective for reducing postoperative vomiting after tonsillectomy. A randomized, double-blind study comparing the antiemetic effects of metoclopramide (0.5 mg/kg) versus ondansetron (0.1 mg/kg) in children after tonsillectomy showed ondansetron to be superior.165,166 Antibiotics are often given to children perioperatively to reduce postoperative complications of adenotonsillectomy. A meta-analysis examining the use of antibiotics for these patients showed an associated 1-day reduction in time required for return to normal oral intake.167

Postoperative Management and Analgesia Children with OSA usually need pain medication after surgery, yet chronic hypoxemia renders them more susceptible to the respiratory depressant effects of opioids.168,169 OSA patients have been shown to have a higher incidence of apnea after administration of 0.5 ␮g/kg of fentanyl, and diminished minute ventilation during spontaneous ventilation under general anesthesia with inhaled anesthetics, when compared with non-OSA patients.170 OSA patients may also be more sensitive to the analgesic effects of opioids. Children with oxygen saturation nadir of ⬍85% on polysomnography required half the morphine dose as those with less desaturation to achieve the same level of analgesia.171 If high doses of opioids are required for an OSA patient after a surgical procedure, intensive postoperative cardiopulmonary monitoring must be considered. When possible, regional anesthesia and/or analgesia should be used. For short procedures, one approach is to minimize opioids intraoperatively and then titrate them to effect when the child is awake, tracheally extubated, and in a monitored setting.70 Less painful procedures may only require non-opioid analgesics, such as acetaminophen or nonsteroidal antiinflammatory drugs (NSAIDS). Methods for managing posttonsillectomy pain without opioids have been studied. Surgical options to spare the tonsillar bed have been shown to reduce the need for pain medications.172,173 Mixed results have been reported with the use of local anesthetic infiltration to the tonsillar bed, and this strategy carries the risk of intravascular injection.174,175 One group reported administering ketamine to the tonsillar region, finding it to be an effective analgesic with no sedating side effects in patients without OSA.176,177 Tramadol, a synthetic selective mu1 agonist, is an alternate medication for moderate pain management with potentially less respiratory depression, but it is not currently approved by the Food and Drug Administration for pediatric patients in the United States. In an Australian study, tonsillectomy patients with moderate OSA who received morphine had a significantly higher risk of desaturation than those given Tramadol, although they were more comfortable.178 The use of NSAIDS in posttonsillectomy patients has been avoided because of reports of associated postoperative bleeding139,179; ANESTHESIA & ANALGESIA

Figure 3. Algorithm for risk assessment and disposition planning.

however, a systematic review did not find an increased risk of reoperation for bleeding and, additionally, found less vomiting when NSAIDS were part of the analgesic regimen.180 A different meta-analysis showed that a significant risk of perioperative bleeding is associated with the use of aspirin for analgesia but no increased risk with other NSAIDS.181 Furthermore, a large retrospective review on the subject concluded that ibuprofen should be used for the relief of pain after adenotonsillectomy.182 Others have written that the use of NSAIDS after attainment of hemostasis is reasonable, given the evidence and pharmacokinetics of the nonaspirin NSAIDS.183 Consequently, the bulk of evidence supports the use of nonaspirin NSAIDS for postoperative analgesia. Numerous authors have demonstrated that pediatric patients with OSA are at increased risk for postoperative respiratory complications.17,56,184 –186 Rates of complications range from 6.4% to 27% but depend on age, severity of OSA, uniformity of diagnosis, and comorbidities.17,57,186,187 Children ⬍3-yr-of-age have twice the risk of children who are 3– 6 yr of age.186 Children with OSA had a 23% rate of respiratory difficulties after adenotonsillectomy, with the greatest risk seen in children ⬍3 yr and those with preoperative RDI ⬎10.54 Nursing intervention was required to treat complications in 60% of children with severe OSA; complications included oxygen desaturation ⬍90%, increased work of breathing and changes on a chest radiograph (edema, atelectasis, infiltrate, pneumothorax, pneumomediastinum, or pleural effusion).55,57,187 Other complications associated with severe OSA include laryngospasm, apnea, pulmonary edema, pulmonary hypertensive crisis, pneumonia, Vol. 109, No. 1, July 2009

and perioperative death.54,56,188,189 Although adenotonsillectomy improves most patients, children with disorders of pharyngeal tone or craniofacial anatomy may have residual airway obstruction during the postoperative period and require close observation to assess the need for intervention.134 The surgeon and anesthesiologist must agree on the discharge plan (Fig. 3). We recommend that polysomnography be performed to help determine the postoperative disposition of patients with OSA. Without objective evidence of the severity of OSA, patients cannot be discharged with confidence. Polysomnography is essential for patients with comorbidities and high-risk features (Table 8). Children with OSA who are identified as high risk for respiratory compromise require overnight inpatient monitoring after surgery in a setting where signs of respiratory depression and airway obstruction can be recognized and prompt intervention can occur.54,120,190 Two study groups have reported the onset of respiratory compromise during sleep ⱖ5 h postoperatively in children with OSA, and those with severe OSA had significantly more overnight obstructive episodes on the first postoperative night when compared with children who had mild OSA.187,191 Because REM rebound is a possibility beyond the first postoperative night,192 careful thought must also be given to whether it is safe to discharge patients with severe OSA on day 2, especially if opioids are needed to control pain.171 Postoperative intensive care unit admission is reserved for very severe OSA, very young children and those with comorbidities that cannot be managed on the floor (Fig. 3). © 2009 International Anesthesia Research Society

69

CONCLUSIONS

Table 9. American Society of Anesthesiologists Risk Assessment, Scoring System Points Severity of sleep apnea None Mild Moderate Severe Invasiveness of surgery and anesthesia Superficial surgery/local anesthesia and no sedation Superficial surgery/moderate sedation or general anesthesia Peripheral surgery with regional anesthesia and moderate sedation Peripheral surgery with general anesthesia Airway surgery with moderate sedation Major surgery/general anesthesia Airway surgery/general anesthesia Requirement for postoperative opioids None Low-dose oral opioids High-dose oral, parenteral, or neuraxial opioids Total score

0 1 2 3 0 1 1 2 2 3 3 0 1 3

This table is a possible format developed by the ASA Task Force. It has not been subjected to prospective study, especially in the pediatric population. Risk stratification was determined by a panel of experts; a score of 4 suggests possible increased risk, and scores of ⱖ5 suggest significantly increased risk. From Gross et al., Anesthesiology 2006;104:1081–93, with permission.

Patients with mild-to-moderate obstructive disease (defined as AHI ⬍10) and no comorbidities can usually be discharged home the same day if they are ⬎3-yr-of-age. Polysomnography performed on otherwise healthy children with mild OSA on the first night after adenotonsillectomy showed that the number of apnea events decreased, and oxygen saturation during sleep improved immediately after surgery.139,193 If a child with OSA is to be discharged on the day of surgery, an early morning operative time has been recommended by some,191 and the ASA practice guidelines suggest monitoring patients longer than those without OSA.10 We recommend a 2-h minimum postanesthesia care unit stay. Research is greatly needed in this area. Determining which patients are at high risk for postoperative respiratory complications is a challenge because polysomnography is not a standard preoperative test for patients with suspected OSA, even though it is recommended by the American Thoracic Society and the American Academy of Pediatrics.120,194 One study showed that ⬍12% of school-aged children who underwent adenotonsillectomy for OSA had prior polysomnography.195 Furthermore, it is clear that although sleep studies may be improved by tonsillectomy, they do not return to normal immediately after surgery, and certain patients continue to be at high risk of perioperative complications.187 The ASA OSA Task Force attempted to use a scoring system identify patients at highest risk (Table 9); however, this system uses scoring of the severity of sleep apnea, which may not be available because of lack of polysomnography testing.10 70


Patients with significant OSA are clearly at higher anesthetic risk in the perioperative period than are patients with normal upper airways. Anesthesiologists should routinely screen patients for snoring, airway dysfunction, airway anatomic disorders and other coexisting diseases that can increase risk from OSA in the postoperative period. Rational, safe, costeffective decision making for individual patients hinges on accurate risk assessment in the preoperative period. Preoperative diagnostic techniques that are affordable and readily available are needed. Anesthesiologists, along with pulmonologists and otolaryngology surgeons, should strive to develop and evaluate ways of identifying children at high risk to determine safe disposition in the postoperative period. ACKNOWLEDGMENTS The authors would like to thank Dr. Myron Yaster for his thoughtful review of this manuscript and Tzipora Sofare, MA, for her expert editorial assistance. REFERENCES 1. Anuntaseree W, Rookkapan K, Kuasirikul S, Thongsuksai P. Snoring and obstructive sleep apnea in Thai school-age children: prevalence and predisposing factors. Pediatr Pulmonol 2001;32:222–7 2. Brunetti L, Rana S, Lospalluti ML, Pietrafesa A, Francavilla R, Fanelli M, Armenio L. Prevalence of obstructive sleep apnea syndrome in a cohort of 1,207 children of southern Italy. Chest 2001;120:1930 –5 3. Carroll JL, McColley SA, Marcus CL, Curtis S, Loughlin GM. Inability of clinical history to distinguish primary snoring from obstructive sleep apnea syndrome in children. Chest 1995;108: 610 –18 4. Gislason T, Benediktsdo´ttir B. Snoring, apneic episodes, and nocturnal hypoxemia among children 6 months to 6 years old. An epidemiologic study of lower limit of prevalence. Chest 1995;107:963– 6 5. Goldstein NA, Sculerati N, Walsleben JA, Bhatia N, Friedman DM, Rapoport DM. Clinical diagnosis of pediatric obstructive sleep apnea validated by polysomnography. Otolaryngol Head Neck Surg 1994;111:611–17 6. Tamay Z, Akcay A, Kilic G, Suleyman A, Ones U, Guler N. Are physicians aware of obstructive sleep apnea in children? Sleep Med 2006;7:580 – 4 7. Uong EC, Jeffe DB, Gozal D, Arens R, Holbrook CR, Palmer J, Cleveland C, Schotland HM. Development of a measure of knowledge and attitudes about obstructive sleep apnea in children (OSAKA-KIDS). Arch Pediatr Adolesc Med 2005;159: 181– 6 8. Owens JA. The practice of pediatric sleep medicine: results of a community survey. Pediatrics 2001;108:E51 9. Turner K, VanDenkerkhof E, Lam M, Mackillop W. Perioperative care of patients with obstructive sleep apnea - a survey of Can anesthesiologists. Can J Anaesth 2006;53:299 –304 10. Gross JB, Bachenberg KL, Benumof JL, Caplan RA, Connis RT, Cote´ CJ, Nickinovich DG, Prachand V, Ward DS, Weaver EM, Ydens L, Yu S; American Society of Anesthesiologists Task Force on Perioperative Management. Practice guidelines for the perioperative management of patients with obstructive sleep apnea: a report by the American Society of Anesthesiologists Task Force on Perioperative Management of patients with obstructive sleep apnea. Anesthesiology 2006;104:1081–93; quiz 1117–18 11. Rosen CL. Obstructive sleep apnea syndrome (OSAS) in children: diagnostic challenges. Sleep 1996;19(suppl 10):S274 –S277 12. Marcus CL. Clinical and pathophysiological aspects of obstructive sleep apnea in children. Pediatr Pulmonol Suppl 1997;16: 123– 4


13. Marcus CL, Hamer A, Loughlin GM. Natural history of primary snoring in children. Pediatr Pulmonol 1998;26:6 –11 14. Ali NJ, Pitson DJ, Stradling JR. Snoring, sleep disturbance, and behaviour in 4 –5 year olds. Arch Dis Child 1993;68:360 – 6 15. Corbo GM, Fuciarelli F, Foresi A, De Benedetto F. Snoring in children: association with respiratory symptoms and passive smoking. BMJ 1989;299:1491– 4 16. Owen GO, Canter RJ, Robinson A. Snoring, apnoea and ENT symptoms in the paediatric community. Clin Otolaryngol Allied Sci 1996;21:130 – 4 17. Schechter MS;. Section on Pediatric Pulmonology, Subcommittee on Obstructive Sleep Apnea Syndrome. Technical report: diagnosis and management of childhood obstructive sleep apnea syndrome. Pediatrics 2002;109:e69 18. Ali NJ, Pitson D, Stradling JR. Natural history of snoring and related behaviour problems between the ages of 4 and 7 years. Arch Dis Child 1994;71:74 – 6 19. Topol HI, Brooks LJ. Follow-up of primary snoring in children. J Pediatr 2001;138:291–3 20. O’Brien LM, Mervis CB, Holbrook CR, Bruner JL, Klaus CJ, Rutherford J, Raffield TJ, Gozal D. Neurobehavioral implications of habitual snoring in children. Pediatrics 2004;114:44 –9 21. Rosen CL, D’Andrea L, Haddad GG. Adult criteria for obstructive sleep apnea do not identify children with serious obstruction. Am Rev Respir Dis 1992;146:1231– 4 22. Marcus CL. Pathophysiology of childhood obstructive sleep apnea: current concepts. Respir Physiol 2000;119:143–54 23. Brouillette RT, Fernbach SK, Hunt CE. Obstructive sleep apnea in infants and children. J Pediatr 1982;100:31– 40 24. Guilleminault C, Korobkin R, Winkle R. A review of 50 children with obstructive sleep apnea syndrome. Lung 1981;159:275– 87 25. Leach J, Olson J, Hermann J, Manning S. Polysomnographic and clinical findings in children with obstructive sleep apnea. Arch Otolaryngol Head Neck Surg 1992;118:741– 4 26. Suen JS, Arnold JE, Brooks LJ. Adenotonsillectomy for treatment of obstructive sleep apnea in children. Arch Otolaryngol Head Neck Surg 1995;121:525–30 27. Wang RC, Elkins TP, Keech D, Wauquier A, Hubbard D. Accuracy of clinical evaluation in pediatric obstructive sleep apnea. Otolaryngol Head Neck Surg 1998;118:69 –73 28. Dreher A, de la Chaux R, Klemens C, Werner R, Baker F, Barthlen G, Rasp G. Correlation between otorhinolaryngologic evaluation and severity of obstructive sleep apnea syndrome in snorers. Arch Otolaryngol Head Neck Surg 2005;131:95– 8 29. Brietzke SE, Katz ES, Roberson DW. Can history and physical examination reliably diagnose pediatric obstructive sleep apnea/hypopnea syndrome? A systematic review of the literature. Otolaryngol Head Neck Surg 2004;131:827–32 30. Chau KW, Ng KK, Kwok KL, Cheung MY. Survey of children with obstructive sleep apnea syndrome in Hong Kong of China. Chin Med J (Engl) 2004;117:657– 60 31. Li AM, Cheung A, Chan D, Wong E, Ho C, Lau J, Wing YK. Validation of a questionnaire instrument for prediction of obstructive sleep apnea in Hong Kong Chinese children. Pediatr Pulmonol 2006;41:1153– 60 32. Chau KW, Ng DK, Kwok CK, Chow PY, Ho JC. Clinical risk factors for obstructive sleep apnoea in children. Singapore Med J 2003;44:570 –3 33. Chervin RD, Weatherly RA, Garetz SL, Ruzicka DL, Giordani BJ, Hodges EK, Dillon JE, Guire KE. Pediatric sleep questionnaire: prediction of sleep apnea and outcomes. Arch Otolaryngol Head Neck Surg 2007;133:216 –22 34. Brouilette R, Hanson D, David R, Klemka L, Szatkowski A, Fernbach S, Hunt C. A diagnostic approach to suspected obstructive sleep apnea in children. J Pediatr 1984;105:10 –14 35. Lamm C, Mandeli J, Kattan M. Evaluation of home audiotapes as an abbreviated test for obstructive sleep apnea syndrome (OSAS) in children. Pediatr Pulmonol 1999;27:267–72 36. Brouillette RT, Morielli A, Leimanis A, Waters KA, Luciano R, Ducharme FM. Nocturnal pulse oximetry as an abbreviated testing modality for pediatric obstructive sleep apnea. Pediatrics 2000;105:405–12 37. Marcus CL, Omlin KJ, Basinki DJ, Bailey SL, Rachal AB, Von Pechmann WS, Keens TG, Ward SL. Normal polysomnographic values for children and adolescents. Am Rev Respir Dis 1992;146:1235–9 Vol. 109, No. 1, July 2009

38. Saeed MM, Keens TG, Stabile MW, Bolokowicz J, Davidson Ward SL. Should children with suspected obstructive sleep apnea syndrome and normal nap sleep studies have overnight sleep studies? Chest 2000;118:360 –5 39. Sivan Y, Kornecki A, Schonfeld T. Screening obstructive sleep apnoea syndrome by home videotape recording in children. Eur Respir J 1996;9:2127–31 40. Brietzke SE, Mair EA. Acoustical analysis of pediatric snoring: what can we learn? Otolaryngol Head Neck Surg 2007;136: 644 – 8 41. Jacob SV, Morielli A, Mograss MA, Ducharme FM, Schloss MD, Brouillette RT. Home testing for pediatric obstructive sleep apnea syndrome secondary to adenotonsillar hypertrophy. Pediatr Pulmonol 1995;20:241–52 42. Morielli A, Ladan S, Ducharme FM, Brouillette RT. Can sleep and wakefulness be distinguished in children by cardiorespiratory and videotape recordings? Chest 1996;109:680 –7 43. Redline S, Budhiraja R, Kapur V, Marcus CL, Mateika JH, Mehra R, Parthasarthy S, Somers VK, Strohl KP, Sulit LG, Gozal D, Wise MS, Quan SF. The scoring of respiratory events in sleep: reliability and validity. J Clin Sleep Med 2007;3: 169 –200 44. Canet E, Gaultier C, D’Allest AM, Dehan M. Effects of sleep deprivation on respiratory events during sleep in healthy infants. J Appl Physiol 1989;66:1158 – 63 45. Hershenson M, Brouillette RT, Olsen E, Hunt CE. The effect of chloral hydrate on genioglossus and diaphragmatic activity. Pediatr Res 1984;18:516 –19 46. Goh DY, Galster P, Marcus CL. Sleep architecture and respiratory disturbances in children with obstructive sleep apnea. Am J Respir Crit Care Med 2000;162:682– 6 47. Loadsman JA, Wilcox I. Is obstructive sleep apnoea a rapid eye movement-predominant phenomenon? Br J Anaesth 2000;85: 354 – 8 48. Karamessinis L, Galster P, Schultz B, Elliott J, Mason TA 2nd, Brooks LJ, Gallagher PR, Marcus CL. Relationship between REM density, duty cycle, and obstructive sleep apnea in children. Sleep 2007;30:837– 43 49. Kaw R, Michota F, Jaffer A, Ghamande S, Auckley D, Golish J. Unrecognized sleep apnea in the surgical patient: implications for the perioperative setting. Chest 2006;129:198 –205 50. Tang JP, Rosen CL, Larkin EK, DiFiore JM, Arnold JL, Surovec SA, Youngblut JM, Redline S. Identification of sleepdisordered breathing in children: variation with event definition. Sleep 2002;25:72–9 51. Rosen CL, Palermo TM, Larkin EK, Redline S. Health-related quality of life and sleep-disordered breathing in children. Sleep 2002;25:657– 66 52. Traeger N, Schultz B, Pollock AN, Mason T, Marcus CL, Arens R. Polysomnographic values in children 2–9 years old: additional data and review of the literature. Pediatr Pulmonol 2005;40:22–30 53. Uliel S, Tauman R, Greenfeld M, Sivan Y. Normal polysomnographic respiratory values in children and adolescents. Chest 2004;125:872– 8 54. McColley SA, April MM, Carroll JL, Naclerio RM, Loughlin GM. Respiratory compromise after adenotonsillectomy in children with obstructive sleep apnea. Arch Otolaryngol Head Neck Surg 1992;118:940 –3 55. Brown KA, Morin I, Hickey C, Manoukian JJ, Nixon GM, Brouillette RT. Urgent adenotonsillectomy: an analysis of risk factors associated with postoperative respiratory morbidity. Anesthesiology 2003;99:586 –95 56. Rosen GM, Muckle RP, Mahowald MW, Goding GS, Ullevig C. Postoperative respiratory compromise in children with obstructive sleep apnea syndrome: can it be anticipated? Pediatrics 1994;93:784 – 8 57. Wilson K, Lakheeram I, Morielli A, Brouillette R, Brown K. Can assessment for obstructive sleep apnea help predict postadenotonsillectomy respiratory complications? Anesthesiology 2002;96:313–22 58. Nixon GM, Kermack AS, Davis GM, Manoukian JJ, Brown KA, Brouillette RT. Planning adenotonsillectomy in children with obstructive sleep apnea: the role of overnight oximetry. Pediatrics 2004;113:e19 – e25 59. Katz ES, D’Ambrosio CM. Pathophysiology of pediatric obstructive sleep apnea. Proc Am Thorac Soc 2008;5:253– 62 © 2009 International Anesthesia Research Society

71

60. Guilleminault C, Li K, Quo S, Inouye RN. A prospective study on the surgical outcomes of children with sleep-disordered breathing. Sleep 2004;27:95–100 61. Richardson MA, Seid AB, Cotton RT, Benton C, Kramer M. Evaluation of tonsils and adenoids in Sleep Apnea syndrome. Laryngoscope 1980;90:1106 –10 62. Lam YY, Chan EY, Ng DK, Chan CH, Cheung JM, Leung SY, Chow PY, Kwok KL. The correlation among obesity, apneahypopnea index, and tonsil size in children. Chest 2006;130: 1751– 6 63. Fernbach SK, Brouillette RT, Riggs TW, Hunt CE. Radiologic evaluation of adenoids and tonsils in children with obstructive sleep apnea: plain films and fluoroscopy. Pediatr Radiol 1983;13:258 – 65 64. Laurikainen E, Erkinjuntti M, Alihanka J, Rikalainen H, Suonpaä¨. Radiological parameters of the bony nasopharynx and the adenotonsillar size compared with sleep apnea episodes in children. Int J Pediatr Otorhinolaryngol 1987;12:303–10 65. Li AM, Hui S, Wong E, Cheung A, Fok TF. Obstructive sleep apnoea in children with adenotonsillar hypertrophy: prospective study. Hong Kong Med J 2001;7:236 – 40 66. Mahboubi S, Marsh RR, Potsic WP, Pasquariello PS. The lateral neck radiograph in adenotonsillar hyperplasia. Int J Pediatr Otorhinolaryngol 1985;10:67–73 67. Kosko JR, Derkay CS. Uvulopalatopharyngoplasty: treatment of obstructive sleep apnea in neurologically impaired pediatric patients. Int J Pediatr Otorhinolaryngol 1995;32:241– 6 68. Messner AH. Treating pediatric patients with obstructive sleep disorders: an update. Otolaryngol Clin North Am 2003;36:519–30 69. McNamara F, Sullivan CE. Obstructive sleep apnea in infants: relation to family history of sudden infant death syndrome, apparent life-threatening events, and obstructive sleep apnea. J Pediatr 2000;136:318 –23 70. Blum RH, McGowan FX Jr. Chronic upper airway obstruction and cardiac dysfunction: anatomy, pathophysiology and anesthetic implications. Paediatr Anaesth 2004;14:75– 83 71. Guilleminault C, Pelayo R, Leger D, Clerk A, Bocian RC. Recognition of sleep-disordered breathing in children. Pediatrics 1996;98:871– 82 72. Villa MP, Malagola C, Pagani J, Montesano M, Rizzoli A, Guilleminault C, Ronchetti R. Rapid maxillary expansion in children with obstructive sleep apnea syndrome: 12-month follow-up. Sleep Med 2007;8:128 –34 73. Koenig SM. Pulmonary complications of obesity. Am J Med Sci 2001;321:249 –79 74. Mitchell RB, Kelly J. Outcome of adenotonsillectomy for obstructive sleep apnea in obese and normal-weight children. Otolaryngol Head Neck Surg 2007;137:43– 8 75. Dyken ME, Lin-Dyken DC, Poulton S, Zimmerman MB, Sedars E. Prospective polysomnographic analysis of obstructive sleep apnea in down syndrome. Arch Pediatr Adolesc Med 2003;157:655–60 76. Levanon A, Tarasiuk A, Tal A. Sleep characteristics in children with Down syndrome. J Pediatr 1999;134:755– 60 77. Marcus CL, Keens TG, Bautista DB, von Pechmann WS, Ward SL. Obstructive sleep apnea in children with Down syndrome. Pediatrics 1991;88:132–9 78. Shott SR, Amin R, Chini B, Heubi C, Hotze S, Akers R. Obstructive sleep apnea: should all children with Down syndrome be tested? Arch Otolaryngol Head Neck Surg 2006;132:432– 6 79. Luscri N, Tobias JD. Monitored anesthesia care with a combination of ketamine and dexmedetomidine during magnetic resonance imaging in three children with trisomy 21 and obstructive sleep apnea. Paediatr Anaesth 2006;16:782– 6 80. Marcus CL, Katz ES, Lutz J, Black CA, Galster P, Carson KA. Upper airway dynamic responses in children with the obstructive sleep apnea syndrome. Pediatr Res 2005;57:99 –107 81. Fregosi RF, Quan SF, Morgan WL, Goodwin JL, Cabrera R, Shareif I, Fridel KW, Bootzin RR. Pharyngeal critical pressure in children with mild sleep-disordered breathing. J Appl Physiol 2006;101:734 –9 82. Brown OE, Manning SC, Ridenour B. Cor pulmonale secondary to tonsillar and adenoidal hypertrophy: management considerations. Int J Pediatr Otorhinolaryngol 1988;16:131–9 83. Gorur K, Doven O, Unal M, Akkus N, Ozcan C. Preoperative and postoperative cardiac and clinical findings of patients with adenotonsillar hypertrophy. Int J Pediatr Otorhinolaryngol 2001;59:41– 6

72


84. Shintani T, Asakura K, Kataura A. The effect of adenotonsillectomy in children with OSA. Int J Pediatr Otorhinolaryngol 1998;44:51– 8 85. Meyrick BO, Perkett EA. The sequence of cellular and hemodynamic changes of chronic pulmonary hypertension induced by hypoxia and other stimuli. Am Rev Respir Dis 1989;140: 1486 –9 86. Duman D, Naiboglu B, Esen HS, Toros SZ, Demirtunc R. Impaired right ventricular function in adenotonsillar hypertrophy. Int J Cardiovasc Imaging 2008;24:261–7 87. Wilkinson AR, McCormick MS, Freeland AP, Pickering D. Electrocardiographic signs of pulmonary hypertension in children who snore. Br Med J (Clin Res Ed) 1981;282:1579 – 81 88. Gozal D, Kheirandish-Gozal L, Serpero LD, Sans Capdevila O, Dayyat E. Obstructive sleep apnea and endothelial function in school-aged nonobese children: effect of adenotonsillectomy. Circulation 2007;116:2307–14 89. Miman MC, Kirazli T, Ozyurek R. Doppler echocardiography in adenotonsillar hypertrophy. Int J Pediatr Otorhinolaryngol 2000;54:21– 6 90. Sofer S, Weinhouse E, Tal A, Wanderman KL, Margulis G, Leiberman A, Gueron M. Cor pulmonale due to adenoidal or tonsillar hypertrophy or both in children. Noninvasive diagnosis and follow-up. Chest 1988;93:119 –22 91. Amin RS, Kimball TR, Bean JA, Jeffries JL, Willging JP, Cotton RT, Witt SA, Glascock BJ, Daniels SR. Left ventricular hypertrophy and abnormal ventricular geometry in children and adolescents with obstructive sleep apnea. Am J Respir Crit Care Med 2002;165:1395–9 92. Baharav A, Kotagal S, Rubin BK, Pratt J, Akselrod S. Autonomic cardiovascular control in children with obstructive sleep apnea. Clin Auton Res 1999;9:345–51 93. O’Brien LM, Gozal D. Autonomic dysfunction in children with sleep-disordered breathing. Sleep 2005;28:747–52 94. Marcus CL, Greene MG, Carroll JL. Blood pressure in children with obstructive sleep apnea. Am J Respir Crit Care Med 1998;157:1098 –103 95. Serratto M, Harris VJ, Carr I. Upper airways obstruction. Presentation with systemic hypertension. Arch Dis Child 1981;56:153–5 96. Enright PL, Goodwin JL, Sherrill DL, Quan JR, Quan SF; Tucson Children’s Assessment of Sleep Apnea study. Blood pressure elevation associated with sleep-related breathing disorder in a community sample of white and Hispanic children: the Tucson Children’s Assessment of Sleep Apnea study. Arch Pediatr Adolesc Med 2003;157:901– 4 97. Li AM, Au CT, Sung RY, Ho C, Ng PC, Fok TF, Wing YK. Ambulatory blood pressure in children with obstructive sleep apnoea: a community based study. Thorax 2008;63:803–9 98. Leung LC, Ng DK, Lau MW, Chan CH, Kwok KL, Chow PY, Cheung JM. Twenty-four-hour ambulatory BP in snoring children with obstructive sleep apnea syndrome. Chest 2006;130: 1009 –17 99. Khayat R, Patt B, Hayes D Jr. Obstructive sleep apnea: the new cardiovascular disease. Part I: obstructive sleep apnea and the pathogenesis of vascular disease. Heart Fail Rev 2008 Sep 20 [epub ahead of print] 100. Badke FR. Left ventricular dimensions and function during right ventricular pressure overload. Am J Physiol 1982;242: H611–H618 101. Laver MB, Strauss HW, Pohost GM. Herbert Shubin Memorial Lecture. Right and left ventricular geometry: adjustments during acute respiratory failure. Crit Care Med 1979;7:509 –19 102. Prineas RJ, Rautaharju PM, Grandits G, Crow R; MRFIT Research Group. Independent risk for cardiovascular disease predicted by modified continuous score electrocardiographic criteria for 6-year incidence and regression of left ventricular hypertrophy among clinically disease free men: 16-year follow-up for the multiple risk factor intervention trial. J Electrocardiol 2001;34:91–101 103. Vakili BA, Okin PM, Devereux RB. Prognostic implications of left ventricular hypertrophy. Am Heart J 2001;141:334 – 41 104. Naughton MT. Common sleep problems in ICU: heart failure and sleep-disordered breathing syndromes. Crit Care Clin 2008;24:565– 87, vii–viii 105. Guffin TN, Har-el G, Sanders A, Lucente FE, Nash M. Acute postobstructive pulmonary edema. Otolaryngol Head Neck Surg 1995;112:235–7


106. Marcus CL, Carroll JL, Koerner CB, Hamer A, Lutz J, Loughlin GM. Determinants of growth in children with the obstructive sleep apnea syndrome. J Pediatr 1994;125:556 – 62 107. Nieminen P, Lopponen T, Tolonen U, Lanning P, Knip M, Lopponen H. Growth and biochemical markers of growth in children with snoring and obstructive sleep apnea. Pediatrics 2002;109:e55 108. Stradling JR, Thomas G, Warley AR, Williams P, Freeland A. Effect of adenotonsillectomy on nocturnal hypoxaemia, sleep disturbance, and symptoms in snoring children. Lancet 1990;335:249 –53 109. Williams EF III, Woo P, Miller R, Kellman RM. The effects of adenotonsillectomy on growth in young children. Otolaryngol Head Neck Surg 1991;104:509 –16 110. Gozal D, Kheirandish-Gozal L. Sleep apnea in children– treatment considerations. Paediatr Respir Rev 2006;7(suppl 1):S58 –S61 111. De Serres LM, Derkay C, Sie K, Biavati M, Jones J, Tunkel D, Manning S, Inglis AF, Haddad J Jr, Tampakopoulou D, Weinberg AD. Impact of adenotonsillectomy on quality of life in children with obstructive sleep disorders. Arch Otolaryngol Head Neck Surg 2002;128:489 –96 112. Tauman R, Gulliver TE, Krishna J, Montgomery-Downs HE, O’Brien LM, Ivanenko A, Gozal D. Persistence of obstructive sleep apnea syndrome in children after adenotonsillectomy. J Pediatr 2006;149:803– 8 113. Coticchia JM, Yun RD, Nelson L, Koempel J. Temperaturecontrolled radiofrequency treatment of tonsillar hypertrophy for reduction of upper airway obstruction in pediatric patients. Arch Otolaryngol Head Neck Surg 2006;132:425–30 114. Koltai PJ, Solares CA, Koempel JA, Hirose K, Abelson TI, Krakovitz PR, Chan J, Xu M, Mascha EJ. Intracapsular tonsillar reduction (partial tonsillectomy): reviving a historical procedure for obstructive sleep disordered breathing in children. Otolaryngol Head Neck Surg 2003;129:532– 8 115. Tunkel DE, Hotchkiss KS, Carson KA, Sterni LM. Efficacy of powered intracapsular tonsillectomy and adenoidectomy. Laryngoscope 2008;118:1295–302 116. Celenk F, Bayazit YA, Yilmaz M, Kemaloglu YK, Uygur K, Ceylan A, Korkuyu E. Tonsillar regrowth following partial tonsillectomy with radiofrequency. Int J Pediatr Otorhinolaryngol 2008;72:19 –22 117. Sterni LM, Tunkel DE. Obstructive sleep apnea in children: an update. Pediatr Clin North Am 2003;50:427– 43 118. Maturo SC, Mair EA. Submucosal minimally invasive lingual excision: an effective, novel surgery for pediatric tongue base reduction. Ann Otol Rhinol Laryngol 2006;115:624 –30 119. Sundaram S, Bridgman SA, Lim J, Lasserson TJ. Surgery for obstructive sleep apnoea. Cochrane Database Syst Rev 2005:CD001004 120. Section on Pediatric Pulmonology, Subcommittee on Obstructive Sleep Apnea Syndrome. American Academy of Pediatrics. Clinical practice guideline: diagnosis and management of childhood obstructive sleep apnea syndrome. Pediatrics 2002;109:704 –12 121. Marcus CL, Ward SL, Mallory GB, Rosen CL, Beckerman RC, Weese-Mayer DE, Brouillette RT, Trang HT, Brooks LJ. Use of nasal continuous positive airway pressure as treatment of childhood obstructive sleep apnea. J Pediatr 1995;127:88 –94 122. Waters KA, Everett FM, Bruderer JW, Sullivan CE. Obstructive sleep apnea: the use of nasal CPAP in 80 children. Am J Respir Crit Care Med 1995;152:780 –5 123. Guilleminault C, Pelayo R, Clerk A, Leger D, Bocian RC. Home nasal continuous positive airway pressure in infants with sleep-disordered breathing. J Pediatr 1995;127:905–12 124. Padman R, Hyde C, Foster P, Borkowski W Jr. The pediatric use of bilevel positive airway pressure therapy for obstructive sleep apnea syndrome: a retrospective review with analysis of respiratory parameters. Clin Pediatr (Phila) 2002;41:163–9 125. Marcus CL, Rosen G, Ward SL, Halbower AC, Sterni L, Lutz J, Stading PJ, Bolduc D, Gordon N. Adherence to and effectiveness of positive airway pressure therapy in children with obstructive sleep apnea. Pediatrics 2006;117:e442– e451 126. McNamara F, Sullivan CE. Obstructive sleep apnea in infants and its management with nasal continuous positive airway pressure. Chest 1999;116:10 –16 Vol. 109, No. 1, July 2009

127. Downey R III, Perkin RM, MacQuarrie J. Nasal continuous positive airway pressure use in children with obstructive sleep apnea younger than 2 years of age. Chest 2000;117:1608 –12 128. Li KK, Riley RW, Guilleminault C. An unreported risk in the use of home nasal continuous positive airway pressure and home nasal ventilation in children: mid-face hypoplasia. Chest 2000;117:916 –18 129. Aljadeff G, Gozal D, Bailey-Wahl SL, Burrell B, Keens TG, Ward SL. Effects of overnight supplemental oxygen in obstructive sleep apnea in children. Am J Respir Crit Care Med 1996;153:51–5 130. Pirelli P, Saponara M, Guilleminault C. Rapid maxillary expansion in children with obstructive sleep apnea syndrome. Sleep 2004;27:761– 6 131. Cozza P, Gatto R, Ballanti F, Prete L. Management of obstructive sleep apnoea in children with modified monobloc appliances. Eur J Paediatr Dent 2004;5:24 –9 132. Hwang D, Shakir N, Limann B, Sison C, Kalra S, Shulman L, Souza Ade C, Greenberg H. Association of sleep-disordered breathing with postoperative complications. Chest 2008;133: 1128 –34 133. Leiberman A, Tal A, Brama I, Sofer S. Obstructive sleep apnea in young infants. Int J Pediatr Otorhinolaryngol 1988;16:39 – 44 134. Guilleminault C, Huang YS, Glamann C, Li K, Chan A. Adenotonsillectomy and obstructive sleep apnea in children: a prospective survey. Otolaryngol Head Neck Surg 2007;136:169–75 135. Fung JW, Li TS, Choy DK, Yip GW, Ko FW, Sanderson JE, Hui DS. Severe obstructive sleep apnea is associated with left ventricular diastolic dysfunction. Chest 2002;121:422–9 136. Dincer HE, O’Neill W. Deleterious effects of sleep-disordered breathing on the heart and vascular system. Respiration 2006;73:124 –30 137. Kwok KL, Ng DK, Chan CH. Cardiovascular changes in children with snoring and obstructive sleep apnoea. Ann Acad Med Singapore 2008;37:715–21 138. Yilmaz MD, Onrat E, Altuntas A, Kaya D, Kahveci OK, Ozel O, Derekoy S, Celik A. The effects of tonsillectomy and adenoidectomy on pulmonary arterial pressure in children. Am J Otolaryngol 2005;26:18 –21 139. Bandla P, Brooks LJ, Trimarchi T, Helfaer M. Obstructive sleep apnea syndrome in children. Anesthesiol Clin North America 2005;23:535– 49, viii 140. Rennotte MT, Baele P, Aubert G, Rodenstein DO. Nasal continuous positive airway pressure in the perioperative management of patients with obstructive sleep apnea submitted to surgery. Chest 1995;107:367–74 141. Ryan CF, Lowe AA, Li D, Fleetham JA. Magnetic resonance imaging of the upper airway in obstructive sleep apnea before and after chronic nasal continuous positive airway pressure therapy. Am Rev Respir Dis 1991;144:939 – 44 142. Arias MA, Garcia-Rio F, Alonso-Fernandez A, Martinez I, Villamor J. Pulmonary hypertension in obstructive sleep apnoea: effects of continuous positive airway pressure: a randomized, controlled cross-over study. Eur Heart J 2006; 27:1106 –13 143. Strauss SG, Lynn AM, Bratton SL, Nespeca MK. Ventilatory response to CO2 in children with obstructive sleep apnea from adenotonsillar hypertrophy. Anesth Analg 1999;89:328 –32 144. Francis A, Eltaki K, Bash T, Cortes S, Mojdehi K, Goldstein NA. The safety of preoperative sedation in children with sleep-disordered breathing. Int J Pediatr Otorhinolaryngol 2006;70:1517–21 145. Viitanen H, Annila P, Viitanen M, Tarkkila P. Premedication with midazolam delays recovery after ambulatory sevoflurane anesthesia in children. Anesth Analg 1999;89:75–9 146. Viitanen H, Annila P, Viitanen M, Yli-Hankala A. Midazolam premedication delays recovery from propofol-induced sevoflurane anesthesia in children 1–3 yr. Can J Anaesth 1999;46:766 –71 147. Helfaer MA, Wilson MD. Obstructive sleep apnea, control of ventilation, and anesthesia in children. Pediatr Clin North Am 1994;41:131–51 148. Ochiai R, Guthrie RD, Motoyama EK. Differential sensitivity to halothane anesthesia of the genioglossus, intercostals, and diaphragm in kittens. Anesth Analg 1992;74:338 – 44 149. Litman RS, Kottra JA, Gallagher PR, Ward DS. Diagnosis of anesthetic-induced upper airway obstruction in children using respiratory inductance plethysmography. J Clin Monit Comput 2002;17:279 – 85 © 2009 International Anesthesia Research Society

73

150. Arai YC, Nakayama M, Kato N, Wakao Y, Ito H, Komatsu T. The effects of jaw thrust and the lateral position on heart rate variability in anesthetized children with obstructive sleep apnea syndrome. Anesth Analg 2007;104:1352–5, table of contents 151. Arai YC, Fukunaga K, Hirota S, Fujimoto S. The effects of chin lift and jaw thrust while in the lateral position on stridor score in anesthetized children with adenotonsillar hypertrophy. Anesth Analg 2004;99:1638 – 41, table of contents 152. Clarke MB, Forster P, Cook TM. Airway management for tonsillectomy: a national survey of UK practice. Br J Anaesth 2007;99:425– 8 153. Anderson BJ, Pearce S, McGann JE, Newson AJ, Holford NH. Investigations using logistic regression models on the effect of the LMA on morphine induced vomiting after tonsillectomy. Paediatr Anaesth 2000;10:633– 8 154. Hatcher IS, Stack CG. Postal survey of the anaesthetic techniques used for paediatric tonsillectomy surgery. Paediatr Anaesth 1999;9:311–15 155. Infosino A. Pediatric upper airway and congenital anomalies. Anesthesiol Clin North America 2002;20:747– 66 156. Bucx MJ, Grolman W, Kruisinga FH, Lindeboom JA, Van Kempen AA. The prolonged use of the laryngeal mask airway in a neonate with airway obstruction and Treacher Collins syndrome. Paediatr Anaesth 2003;13:530 –3 157. Simurina T, Mikulandra S, Mraovic B, Sonicki Z, Kovacic M, Dzelalija B, Rudic M. The effect of propofol and fentanyl as compared with sevoflurane on postoperative vomiting in children after adenotonsillectomy. Coll Antropol 2006;30:343–7 158. Pappas AL, Sukhani R, Hotaling AJ, Mikat-Stevens M, Javorski JJ, Donzelli J, Shenoy K. The effect of preoperative dexamethasone on the immediate and delayed postoperative morbidity in children undergoing adenotonsillectomy. Anesth Analg 1998;87:57– 61 159. Kaan MN, Odabasi O, Gezer E, Daldal A. The effect of preoperative dexamethasone on early oral intake, vomiting and pain after tonsillectomy. Int J Pediatr Otorhinolaryngol 2006;70:73–9 160. Hirunwiwatkul P. Pain-relieving effect of local steroid injection in uvulopalatopharyngoplasty. J Med Assoc Thai 2001;84 (suppl 1):S384 –S390 161. April MM, Callan ND, Nowak DM, Hausdorff MA. The effect of intravenous dexamethasone in pediatric adenotonsillectomy. Arch Otolaryngol Head Neck Surg 1996;122:117–20 162. Kim MS, Cote CJ, Cristoloveanu C, Roth AG, Vornov P, Jennings MA, Maddalozzo JP, Sullivan C. There is no doseescalation response to dexamethasone (0.0625–1.0 mg/kg) in pediatric tonsillectomy or adenotonsillectomy patients for preventing vomiting, reducing pain, shortening time to first liquid intake, or the incidence of voice change. Anesth Analg 2007;104:1052– 8, tables of contents 163. Al-Shehri AM. Steroid therapy for post-tonsillectomy symptoms in adults: a randomized, placebo-controlled study. Ann Saudi Med 2004;24:365–7 164. Brouillette RT, Manoukian JJ, Ducharme FM, Oudjhane K, Earle LG, Ladan S, Morielli A. Efficacy of fluticasone nasal spray for pediatric obstructive sleep apnea. J Pediatr 2001;138:838 – 44 165. Bolton CM, Myles PS, Carlin JB, Nolan T. Randomized, double-blind study comparing the efficacy of moderate-dose metoclopramide and ondansetron for the prophylactic control of postoperative vomiting in children after tonsillectomy. Br J Anaesth 2007;99:699 –703 166. Bolton CM, Myles PS, Nolan T, Sterne JA. Prophylaxis of postoperative vomiting in children undergoing tonsillectomy: a systematic review and meta-analysis. Br J Anaesth 2006;97:593– 604 167. Iyer S, DeFoor W, Grocela J, Kamholz K, Varughese A, Kenna M. The use of perioperative antibiotics in tonsillectomy: does it decrease morbidity? Int J Pediatr Otorhinolaryngol 2006;70: 853– 61 168. Moss IR, Belisle M, Laferriere A. Long-term recurrent hypoxia in developing rat attenuates respiratory responses to subsequent acute hypoxia. Pediatr Res 2006;59:525–30 169. Moss IR, Brown KA, Laferriere A. Recurrent hypoxia in rats during development increases subsequent respiratory sensitivity to fentanyl. Anesthesiology 2006;105:715–18

74


170. Waters KA, McBrien F, Stewart P, Hinder M, Wharton S. Effects of OSA, inhalational anesthesia, and fentanyl on the airway and ventilation of children. J Appl Physiol 2002;92: 1987–94 171. Brown KA, Laferriere A, Lakheeram I, Moss IR. Recurrent hypoxemia in children is associated with increased analgesic sensitivity to opiates. Anesthesiology 2006;105:665–9 172. Ericsson E, Wadsby M, Hultcrantz E. Pre-surgical child behavior ratings and pain management after two different techniques of tonsil surgery. Int J Pediatr Otorhinolaryngol 2006;70:1749 –58 173. Hultcrantz E, Ericsson E. Pediatric tonsillotomy with the radiofrequency technique: less morbidity and pain. Laryngoscope 2004;114:871–7 174. Stuart JC, MacGregor FB, Cairns CS, Chandrachud HR. Peritonsillar infiltration with bupivacaine for paediatric tonsillectomy. Anaesth Intensive Care 1994;22:679 – 82 175. Jebeles JA, Reilly JS, Gutierrez JF, Bradley EL Jr, Kissin I. The effect of pre-incisional infiltration of tonsils with bupivacaine on the pain following tonsillectomy under general anesthesia. Pain 1991;47:305– 8 176. Erhan OL, Goksu H, Alpay C, Bestas A. Ketamine in posttonsillectomy pain. Int J Pediatr Otorhinolaryngol 2007;71: 735–9 177. Canbay O, Celebi N, Uzun S, Sahin A, Celiker V, Aypar U. Topical ketamine and morphine for post-tonsillectomy pain. Eur J Anaesthesiol 2008;25:287–92 178. Hullett BJ, Chambers NA, Pascoe EM, Johnson C. Tramadol vs morphine during adenotonsillectomy for obstructive sleep apnea in children. Paediatr Anaesth 2006;16:648 –53 179. Marret E, Flahault A, Samama CM, Bonnet F. Effects of postoperative, nonsteroidal, antiinflammatory drugs on bleeding risk after tonsillectomy: meta-analysis of randomized, controlled trials. Anesthesiology 2003;98:1497–502 180. Cardwell M, Siviter G, Smith A. Non-steroidal antiinflammatory drugs and perioperative bleeding in paediatric tonsillectomy. Cochrane Database Syst Rev 2005:CD003591 181. Krishna S, Hughes LF, Lin SY. Postoperative hemorrhage with nonsteroidal anti-inflammatory drug use after tonsillectomy: a meta-analysis. Arch Otolaryngol Head Neck Surg 2003;129: 1086 –9 182. Jeyakumar A, Brickman TM, Williamson ME, Hirose K, Krakovitz P, Whittemore K, Discolo C. Nonsteroidal antiinflammatory drugs and postoperative bleeding following adenotonsillectomy in pediatric patients. Arch Otolaryngol Head Neck Surg 2008;134:24 –7 183. Dsida R, Cote´ CJ. Nonsteroidal antiinflammatory drugs and hemorrhage following tonsillectomy: do we have the data? Anesthesiology 2004;100:749 –51; author reply 751–2 184. Brown KA, Laferriere A, Moss IR. Recurrent hypoxemia in young children with obstructive sleep apnea is associated with reduced opioid requirement for analgesia. Anesthesiology 2004;100:806 –10; discussion 5A 185. Sanders JC, King MA, Mitchell RB, Kelly JP. Perioperative complications of adenotonsillectomy in children with obstructive sleep apnea syndrome. Anesth Analg 2006;103:1115–21 186. Statham MM, Elluru RG, Buncher R, Kalra M. Adenotonsillectomy for obstructive sleep apnea syndrome in young children: prevalence of pulmonary complications. Arch Otolaryngol Head Neck Surg 2006;132:476 – 80 187. Nixon GM, Kermack AS, McGregor CD, Davis GM, Manoukian JJ, Brown KA, Brouillette RT. Sleep and breathing on the first night after adenotonsillectomy for obstructive sleep apnea. Pediatr Pulmonol 2005;39:332– 8 188. Biavati MJ, Manning SC, Phillips DL. Predictive factors for respiratory complications after tonsillectomy and adenoidectomy in children. Arch Otolaryngol Head Neck Surg 1997;123:517–21 189. Bower CM, Gungor A. Pediatric obstructive sleep apnea syndrome. Otolaryngol Clin North Am 2000;33:49 –75 190. Gerber ME, O’Connor DM, Adler E, Myer CM III. Selected risk factors in pediatric adenotonsillectomy. Arch Otolaryngol Head Neck Surg 1996;122:811–14 191. Koomson A, Morin I, Brouillette R, Brown KA. Children with severe OSAS who have adenotonsillectomy in the morning are less likely to have postoperative desaturation than those operated in the afternoon. Can J Anaesth 2004;51:62–7


192. Rosenberg J, Wildschiodtz G, Pedersen MH, von Jessen F, Kehlet H. Late postoperative nocturnal episodic hypoxaemia and associated sleep pattern. Br J Anaesth 1994;72:145–50 193. Helfaer MA, McColley SA, Pyzik PL, Tunkel DE, Nichols DG, Baroody FM, April MM, Maxwell LG, Loughlin GM. Polysomnography after adenotonsillectomy in mild pediatric obstructive sleep apnea. Crit Care Med 1996;24:1323–7 194. Cardiorespiratory sleep studies in children. Establishment of normative data and polysomnographic predictors of morbidity. American Thoracic Society. Am J Respir Crit Care Med 1999;160:1381–7

Vol. 109, No. 1, July 2009

195. Weatherly RA, Mai EF, Ruzicka DL, Chervin RD. Identification and evaluation of obstructive sleep apnea prior to adenotonsillectomy in children: a survey of practice patterns. Sleep Med 2003;4:297–307


75

Ambulatory Anesthesiology Section Editor: Peter S. A. Glass

An Evaluation of the Efficacy of Licorice Gargle for Attenuating Postoperative Sore Throat: A Prospective, Randomized, Single-Blind Study Anil Agarwal, MD* Devendra Gupta, MD* Ghanshyam Yadav, MD* Puneet Goyal, MD, DM* Prabhat K. Singh, MD* Uttam Singh, PhD†

BACKGROUND: Postoperative sore throat (POST) contributes to postoperative morbidity. Licorice has been used as an expectorant in cough and cold preparations. In this study, we evaluated the efficacy of licorice gargle for attenuating POST. METHODS: Forty adults (18 – 60 yr), ASA physical status I and II of either sex, undergoing elective lumber laminectomy were randomized into two groups of 20 each. Group C: received water; Group L: received 0.5 g licorice in water. Both groups received a 30 mL mixture for 30 s, 5 min before anesthesia which was standardized. The incidence and severity of POST at rest and on swallowing and side effects were assessed at 0, 2, 4, and 24 h, postoperatively. Severity of POST was assessed by visual analog scale (between 0 and 100 mm; where 0 means no sore throat and 100 means worst imaginable sore throat). Postextubation cough was assessed immediately after tracheal extubation. Data were analyzed by Z test and Fisher’s exact test. P ⬍ 0.05 was considered as significant. RESULTS: POST (incidence and severity) was reduced in the Group L compared with Group C at rest and on swallowing for all time points (P ⬍ 0.05), except that the severity of POST at rest, at 24 h, was similar in both groups (P ⬎ 0.05). Postextubation cough was reduced in Group L compared with Group C (P ⬍ 0.05). There was no difference in side effects between groups (P ⬎ 0.05). CONCLUSION: Licorice gargle performed 5 min before anesthesia is effective in attenuating the incidence and severity of POST. (Anesth Analg 2009;109:77–81)

P

ostoperative sore throat (POST) is a wellrecognized complication after anesthesia contributing to postoperative morbidity and patient dissatisfaction.1,2 In patients who have had an endotracheal tube inserted, the incidence of POST varies from 40% to 100%.3–5 POST was rated by patients as the 8th most undesirable outcome in the postoperative period.6 Similarly, postextubation coughing (PEC) causes considerable patient discomfort and can result in a number of undesirable side effects, including hypertension, tachycardia, tachyarrhythmia, increased intracranial pressure, and increased intraocular pressure.7 Nonpharmacological measures for minimizing POST are smaller-sized endotracheal tubes, lubricating the endotracheal tube with water-soluble jelly, careful airway instrumentation, intubation after full relaxation, gentle oropharyngeal suctioning, minimizing intracuff pressure, and extubation in deeper plane From the *Department of Anesthesiology and †Biostatics, Sanjay Gandhi Post Graduate Institute of Medical Sciences, Lucknow, India. Accepted for publication January 26, 2009. Address correspondence and reprint requests to Anil Agarwal, MD, Type V B/11, Sanjay Gandhi Post Graduate Institute of Medical Sciences, Lucknow 226 014, India. Address e-mail to [email protected]. Copyright © 2009 International Anesthesia Research Society DOI: 10.1213/ane.0b013e3181a6ad47

Vol. 109, No. 1, July 2009

of anesthesia and with a fully deflated tracheal tube cuff.8 Pharmacological measures for attenuating POST are beclomethasone inhalation and gargling with azulene sulfonate, aspirin, and benzydamine hydrochloride, local spray with lidocaine and intracuff administration of alkalized lignocaine.4,9 –12 However, all these techniques have their own limitations and variable success rate. Licorice, derived from the root of Glycyrrhiza glabra, has been used for many millennia as an alternative medicine for treatment of inflammation, allergies, and gastric and duodenal ulcers.13 We therefore hypothesized that licorice could decrease POST because of its topical antiirritant, antiinflammatory, and a peripheral and central antitussive effects.14 This study was therefore planned to evaluate the efficacy of licorice gargle for prevention of POST after orotracheal intubation.

METHODS This prospective, randomized, single-blind, placebocontrolled study was conducted after approval from the institute’s ethics committee and written informed consent from the patients. We considered patients for inclusion if they were aged between 16 and 60 yr, ASA physical status I and II, of either sex or undergoing elective lumbar laminectomy in the prone position 77

under general anesthesia. During the preoperative visit by an anesthesiologist consultant (PG), patients with a history of preoperative sore throat, upper respiratory tract infection, common cold, gastroesophageal reflux, regurgitation, known allergy to licorice, patients with recent nonsteroidal antiinflammatory drug medication, pregnancy, anticipated difficult intubation, and Mallampati grade ⬎2 were excluded from study. After exclusion, patients were randomized into two groups of 20 each with the help of a computergenerated table of random numbers. The control group gargled with 30 mL of water. The licorice group gargled with 0.5 g licorice (Yastimadhu 60 g manufactured by Tansukh Herbal, Lucknow, India) made in 30 mL of water. According to their randomization, patients were asked to gargle the study medication for 30 s, 5 min before induction of anesthesia. The dose of licorice chosen for the study was based on the dosage recommended in other studies.15–16 The gargle solution was prepared by the process of decoction, which involves boiling the licorice powder (5 g) in 300 mL of water and filtering the decoction. This decoction was used within 24 h of its preparation and was used for gargling at room temperature. The solution was not compounded with any other additives like sugar or alcohol. It has a distinct sweet taste and patients did not find it unpleasant. Patients were clearly instructed to gargle with this solution and none of the patients swallowed it. The solution was odorless and investigators were therefore blinded to the preparation used for gargle. Medications for gargle were placed in an opaque container by a staff nurse who was not involved in subsequent management of these patients. Patients requiring ⬎1 attempt for intubation that could not be extubated at the end of the surgery because of any reason or were reintubated within the study period were considered drop-outs. Patients were premeditated with oral lorazepam 0.04 mg/kg the night and 2 h before the induction of anesthesia with sips of water. Anesthesia was induced with fentanyl 3 ␮g/kg and propofol 2 mg/kg. Tracheal intubation was facilitated with vecuronium bromide 0.1 mg/kg. Direct laryngoscopy was done with the use of a Macintosh laryngoscope blade by applying minimal pressure and a soft seal cuffed sterile polyvinyl chloride endotracheal tube (Portex, CT 21 6JL,UK); 7.5 mm inner diameter for females and 8.5 mm inner diameter for males was used. Laryngoscopy and endotracheal intubation were performed by an anesthesia senior resident who was blinded to group allocation. The endotracheal tube was lubricated with sterile water-soluble jelly (Lubic (K); Neon Laboratories, Mumbai, India). The cuff was inflated with air maintaining cuff pressure of 18 –22 cm of water. Cuff pressure was measured with the help of a pressuremonitoring transducer (Edward Life Sciences, LIC Irvine, CA) by connecting it to a pilot balloon of the endotracheal tube, which then provided a continuous 78

Licorice Gargle Attenuates POST

digital display of the intracuff pressure on the monitor. All patients’ lungs were ventilated with O2 in air (fraction of inspired oxygen [Fio2] ⫺0.4), keeping the end-tidal CO2 between 32 and 35 mm Hg. A heat and moist exchanger was used to provide humidification of the anesthetic gases. Anesthesia was maintained with propofol infusion 100 –150 ␮g 䡠 kg⫺1 䡠 min⫺1 and intermittent fentanyl and vecuronium as required. At the end of surgery, neuromuscular blockade was antagonized by a combination of neostigmine 0.05 mg/kg and glycopyrolate 0.01 mg/kg. After gentle suctioning of oral secretions by a 12 F suction catheter, patients were tracheally extubated and transferred to the postanaesthesia care unit. Patients received fentanyl via IV patient-controlled analgesia for their postoperative pain. The primary end point was POST (incidence and severity at rest and on swallowing); secondary end points were occurrence of PEC and side effects, if any. All the end points were assessed by an independent observer (DG) who was blinded to group allocation. PEC was assessed immediately after tracheal extubation in the operation room. The incidence of POST was obtained by inquiring from the patients the presence or absence of soreness in the throat at rest and on swallowing, with the help of a question framed in local language. Severity of POST was assessed using a visual analog scale (between 0 and 100; where 0 means no sore throat and 100 means worst imaginable sore throat). POST and side effects were assessed at 0, 2, 4, and 24 h postoperatively. Calculation of sample size was based on the presumption that the incidence of POST after licorice gargle would decline to 35% compared with 70% in the control group. For the results to be of clinical significance with ␣ ⫽ 0.05 and ␤ ⫽ 0.80, one needed to recruit 16 patients in each group. The method of analysis was decided prospectively and incorporated the intention-to-treat principle. Demographic data were analyzed with one way analysis of variance for continuous variables and ␹2 test for categorical variables. The incidence of POST was analyzed by Z test, whereas severity of POST was analyzed by Mann– Whitney test. The incidence of side effects was analyzed with Fisher’s exact test. SPSS 14.0 (SPSS, Chicago, IL) was used for statistical analysis. P ⬍ 0.05 was considered as significant.

RESULTS Forty-nine consecutive patients were evaluated between September 2007 to April 2008, of which nine patients were excluded from the study because of a history of preoperative upper respiratory tract infection, use of recent nonsteroidal antiinflammatory drug medication, Mallampati grade ⬎2, anticipated difficult intubation, or gastroesophageal reflux. Forty patients were included in this study and received preoperative gargles after randomization. Thirtyseven patients completed the study (18 in the control ANESTHESIA & ANALGESIA

Figure 1. Study design.

Table 1. Demographic Data Presented Either as Mean ⫾ SD or Numbers. No Significant Differences Between the Groups by One Way ANOVA for Continuous Variables and ␹2 Test for Categorical Variables (P ⬎ 0.05) Variables⶿groups

Control (n ⴝ 20)

Licorice (n ⴝ 20)

Age (yr) Sex (M/F) Weight (kg) Height (cm) Duration of anesthesia (min)

42.7 ⫾ 15.5 13/5 56.4 ⫾ 8.9 168.7 ⫾ 7.1 135 ⫾ 13.5

43.4 ⫾ 15.1 15/4 57.4 ⫾ 6.8 166.1 ⫾ 5.2 133 ⫾ 13.6

ANOVA ⫽ analysis of variance.

group and 19 in the licorice group) because three patients were dropped from the study (Fig. 1). There was no difference between groups regarding to age, sex, weight, height, and duration of anesthesia (P ⬎ 0.05) (Table 1). The incidence of POST was reduced in the licorice group compared with the control group both at rest and on swallowing at all time points (P ⬍ 0.05) (Fig. 2). The severity of POST was reduced in the licorice group compared with the control group both at rest (0, 2, and 4 h) and on swallowing at (0, 2, 4, and 24 h) postoperatively (P ⬍ 0.05) (Fig. 3). The severity of POST at rest, at 24 h, was similar in both groups (P ⬎ 0.05) (Fig. 3). The number of patients having PEC was significantly reduced in the licorice group, i.e., two, when compared with six in the control group (P ⬍ 0.05). There was no significant difference regarding the side effects in any group (P ⬎ 0.05). Vol. 109, No. 1, July 2009

Absolute risk reduction and the number-needed-totreat in the licorice group were 57% and 2%, respectively, in relation to POST during swallowing at 0 h.

DISCUSSION We observed a significant reduction in the incidence and severity of POST and PEC after a licorice gargle with similar side effect profile compared with a control group. The etiology of POST is multifactorial and this may be related to patient, technique of anesthesia, and type of surgery.17 POST might be a consequence of localized trauma, leading to inflammation of pharyngeal mucosa secondary to either laryngoscopy or an endotracheal intubation or both.4 Licorice, also known as sweet-wood or Glycyrrhiza glabra, belongs to the Leguminosae family.18 A number of active ingredients have been isolated from Licorice, such as glycyrrhizin, glycyrrhizic acid, liquilitin, liquiritigenin glabridin, and hispaglabridins.14 Licorice has been reported to have antiinflammatory and antiallergic properties due to glycyrrhizin.19 Glycyrrhizic acid has been demonstrated to retard the inflammatory process by inhibiting cyclooxygenase activity, prostaglandin formation, and inhibition of platelet aggregation.20 Liquilitin and liquiritigenin have been reported to have peripheral and central antitussive properties.21 Glabridin has significant antioxidant and ulcer-healing properties,14 which might be helpful in minimizing the extent of ischemic injury © 2009 International Anesthesia Research Society

79

Figure 2. Incidence of postoperative sore throat. Data are presented as numbers; analyzed by Z test; *Denotes P ⬍ 0.05 versus control.

Figure 3. Severity of postoperative sore throat. Data presented as median visual analog scale scores (interquartile range); analyzed by Mann–Whitney test; *Denotes P ⬍ 0.05 versus control. to the pharyngeal and tracheal mucosa and expedite their healing. Injury to the pharyngeal and tracheal mucosa might be caused secondary to laryngoscopy, intubation, and endotracheal tube cuff inflation. The observed reduction in incidence and severity of POST and PEC after licorice gargle might be because of any of these above-mentioned mechanisms in isolation or their additive or synergistic effect. Ogata et al.10 observed that preoperative gargling with azunol reduced the incidence of POST from 65% to 25% when patients performed gargle with 4 mg azunol diluted with 100 mL tap water. Similarly, Canbay et al.22 reported a reduction in POST from 78% to 40% after ketamine gargle performed 5 min before induction of anesthesia with 40 mg in saline 30 mL mixture for 30 s. Agarwal et al.4 in their previous study stated that aspirin and benzydamine hydrochloride gargle made in 30 mL of water performed after patients’ arrival in the operation room for 30 s were 80

Licorice Gargle Attenuates POST

equally effective in reducing POST from 80% to 26%. The methodology of the this study is similar to the previously conducted trials.4,10,22 Results of the this study are also similar to earlier studies as we observed a reduction in POST (rest) from 78% to 21%. We also observed a reduction in POST (swallowing) from 83% to 21% after licorice gargle. The study conducted by Ogata et al. was double-blind, whereas the rest of the trials were performed in a single-blind fashion.4,10,22 Licorice extract, due to its ingredient glycyrrhizin, is 50 times sweeter than sugar and is therefore used as a sugar substitute in food.19 In modern medicine, licorice extracts are often used as a flavoring agent to mask the bitter taste of medicines and as an expectorant in cough and cold preparations.14 The sweet taste of licorice increases its acceptability as a gargle and can be used without any added flavors. Licorice is cost-effective as the cost of a 60 g pack is around 1 USD (INR 50) in India, and one pack could be used in ⬎100 patients for prevention of POST and PEC. The ANESTHESIA & ANALGESIA

presence of ulcers in the mouth is not a contraindication for the use of licorice gargle; on the contrary, these patients are reported to benefit from licorice.23 The limitations of the this study are that it could not be performed in a double-blind manner because of the distinct sweet taste of licorice. At times patients might find it inconvenient to perform gargles in the operation room, especially if they have been well premedicated. Excessive sedation during premedication could increase the chance of aspiration, thus limiting the use of licorice gargle. Licorice gargle is also difficult to perform in uncooperative patients and children. To conclude, licorice gargle performed 5 min before induction of anesthesia is an effective method for attenuating both the incidence and severity of POST. REFERENCES 1. McHardy FE, Chung F. Postoperative sore throat: cause, prevention and treatment. Anaesthesia 1999;54:444 –53 2. Joshi GP, Inagaki Y, White PF, Taylor-Kennedy L, Wat LI, Gevirtz C, McCraney JM, McCulloch DA. Use of the laryngeal mask airway as an alternative to the tracheal tube during ambulatory anaesthesia. Analg Anesth 1997;85:573–7 3. Mencke T, Knoll H, Schreiber JU, Echternach M, Klein S, Noeldge-Schomburg G, Silomon M. Rocuronium is not associated with more vocal cord injuries than succinylcholine after rapid-sequence induction: a randomized, prospective, controlled trial. Anesth Analg 2006;102:943–9 4. Agarwal A, Nath SS, Goswami D, Gupta D, Dhiraaj S, Singh PK. An evaluation of the efficacy of aspirin and benzydamine hydrochloride gargle for attenuating postoperative sore throat: a prospective randomized, single– blind study. Anesth Analg 2006;103:1–3 5. Chen KT, Tzeng JI, Lu CL, Liu KS, Chen YW, Hsu CS, Wang JJ. Risk factors associated with postoperative sore throat after tracheal intubation: an evaluation in the postanesthetic recovery room. Acta Anaesthesiol Taiwan 2004;42:3– 8 6. Monroe MC, Gravenstein N, Saga-Rumley S. Postoperative sore throat: effect of oropharyngeal airway in orotracheally intubated patients. Anesth Analg 1990;70:512–16 7. Minogue SC, Ralph J, Lampa MJ. Laryngotracheal topicalization with lidocaine before intubation decreases the incidence of coughing on emergence from general anesthesia. Anesth Analg 2004;99:1253–7

Vol. 109, No. 1, July 2009

8. Al-Qahtani AS, Messahel FM. Quality improvement in anesthetic practice—incidence of sore throat after using small tracheal tube. Middle East J Anesthesiol 2005;18:179 – 83 9. el Hakim M. Beclomethasone prevents postoperative sore throat. Acta Anaesthesiol Scand 1993;37:250 –2 10. Ogata J, Minami K, Horishita T, Shiraishi M, Okamoto T, Terada T, Sata T. Gargling with sodium azulene sulfonate reduces the postoperative sore throat after intubation of the trachea. Anesth Analg 2005;101:290 –3 11. Estebe JP, Dollo G, Le Corre P, Le Naoures A, Chevanne F, Le Verge R, Ecoffey C. Alkanization of intracuff lidocaine improves endotracheal tube-induced emergence phenomena. Anesth Analg 2002;94:227–30 12. Le´vy B, Mouillac F, Quilichini D, Schmitz J, Guadart J, Gouin F. [Topical methylprednisolone vs lidocaine for the prevention of postoperative sore throat]. Ann Fr Anesth Reanim 2003;22:595–9 13. Shin YW, Bae EA, Lee B, Lee SH, Kim JA, Kim YS, Kim DH. In vitro and in vivo antiallergic effects of Glycyrrhiza glabra and its components. Planta Med 2007;73:257– 61 14. Glycyrrhiza glabra. Monograph. Altern Med Rev 2005;10:230 –7 15. Yamamura Y, Kawakami J, Santa T, Kotaki H, Uchino K, Sawada Y, Tanaka N, Iga T. Pharmacokinetic profile of glycyrrhizin in healthy volunteers by a new high performance liquid chromatographic method. J Pharm Sci 1992;81:1042– 6 16. Blumenthal M, Gruenwald J, Hall T, Riggins C, Richter R. The complete German commission E monographs: therapeutic guide to herbal meds. Austin, TX: American Botanical Society, 1998 17. Higgins PP, Chung F, Mezei G. Postoperative sore throat after ambulatory surgery. Br J Anaesth 2002;88:582– 4 18. Ody P. The complete medicinal herbal. New York, NY: Dorling Kindersley, 1998 19. Aly AM, Al-Alousi L, Salem HA. Licorice: a possible antiinflammatory and anti-ulcer drug. AAPS PharmSciTech 2005;6:E74 –E82 20. Okimasu E, Moromizato Y, Watanabe S, Sasaki J, Shiraishi N, Morimoto YM, Miyahara M, Utsumi K. Inhibition of phospholipase A2 and platelet aggregation by glycyrrhizin, an antiinflammation drug. Acta Med Okayama 1983;37:385–91 21. Kamei J, Saitoh A, Asano T, Nakamura R, Ichiki H, Iiduka A, Kubo M. Pharmacokinetic and pharmacodynamic profiles of the antitussive principles of Glycyrrhizae radix (licorice), a main component of the Kampo preparation Bakumondo-to (Maimen-dong-tang). Eur J Pharmacol 2005;507:163– 8 22. Canbay O, Celebi N, Sahin A, Celiker V, Ozgen S, Aypar U. Ketamine gargle for attenuating postoperative sore throat. Br J Anaesth 2008;100:490 –3 23. Furuhashi I, Iwata S, Shibata S, Sato T, Inoue H. Inhibition by licochalcone A, a novel flavonoid isolated from liquorice root, of IL-1beta-induced PGE2 production in human skin fibroblasts. J Pharm Pharmacol 2005;57:1661– 6


81

Anesthetic Pharmacology Preclilnical Pharmacology

Clinical Pharmacology

section Editor: Marcel E. Durieux

Section Editor: Tony Gin

A Comparison of the Effects on Respiratory Carbon Dioxide Response, Arterial Blood Pressure, and Heart Rate of Dexmedetomidine, Propofol, and Midazolam in Sevoflurane-Anesthetized Rabbits Cheng Chang, MD Akinori Uchiyama, MD, PhD Ling Ma, MD Takashi Mashimo, MD, PhD Yuji Fujino, MD, PhD

BACKGROUND: Dexmedetomidine, propofol, and midazolam are commonly used sedative-hypnotic drugs. Using a steady-state method, we examined the CO2 ventilatory response, mean arterial blood pressure (MAP) and heart rate (HR) effects of these three drugs in sevoflurane-anesthetized rabbits. METHODS: New Zealand white rabbits weighing 2.9 ⫾ 0.2 kg (mean ⫾ sd) were used. After anesthetic induction and tracheostomy, the animals inhaled 2% sevoflurane to ensure a stable level of sedation throughout the experiment. After preparation, the rabbits were randomly assigned to four groups (n ⫽ 10 ⫻ 4) and received the following drugs: Group C, control; Group D, dexmedetomidine infused at 2 ␮g 䡠 kg⫺1 䡠 h⫺1; Group P, propofol with the plasma concentration maintained at 15 ␮g/mL; Group M, midazolam initial IV 0.3 mg/kg bolus dose, followed by infusion at 1.86 mg 䡠 kg⫺1 䡠 h⫺1. At 15 minutes after the start of infusion, for 20 min periods, in random sequences, gas including 0%, 1%, 2%, 3%, 4%, or 5% of CO2 was delivered to each animal. Fraction of inspired oxygen was maintained at 0.9. We did intergroup comparisons of minute ventilation (MV), respiratory rate, MAP, and HR during the final minute of each inspiratory carbon dioxide concentration (FiCO2) period. RESULTS: For Groups P and M, the rightward shift of plots for MV against FiCO2 indicated significant respiratory depression compared with Group C. There was also significantly more depression than in Group D. We found no significant differences between Groups P and M or between Groups C and D in the plots of MV against FiCO2. No significant differences among the four groups were apparent for respiratory rate. Paco2-MV response plots were derived from linear regression analysis of data for mean MV and mean Paco2 at each FiCO2 to compute apneic CO2 thresholds and CO2 sensitivities. The apneic CO2 thresholds of Groups P and M were larger than those of Groups C and D. The CO2 sensitivities of Group D were slightly lower than in Group C. No similar significant difference between the CO2 sensitivities of other group pairs was apparent. MAP in Group D was lower than in Groups C and M. In Group D, HR was lower than in Groups C, P, and M. CONCLUSIONS: The major finding is that, during sevoflurane anesthesia in rabbits, dexmedetomidine slightly altered the ventilatory response to CO2. It decreased MAP more than propofol and midazolam, which both significantly depressed the ventilatory response to CO2. (Anesth Analg 2009;109:84 –9)

P

ropofol, midazolam, and dexmedetomidine are often used to achieve satisfactory sedation for critically ill intensive care unit patients.1 Although several human

From the Department of Anesthesiology and Intensive Care Medicine, Osaka University Graduate School of Medicine, Yamadaoka, Suita, Japan. Accepted for publication January 7, 2009. Address correspondence and reprint requests to Akinori Uchiyama, MD, PhD, Department of Anesthesiology and Intensive Care Medicine, Osaka University Graduate School of Medicine, 2-15 Yamadaoka, Suita, Osaka 565-0871, Japan. Address e-mail to [email protected]. Copyright © 2009 International Anesthesia Research Society DOI: 10.1213/ane.0b013e3181a2ad5f

84

and animal studies have presented evidence that each of the three commonly used sedatives cause respiratory depression,2– 8 no direct comparison of the respiratory depression associated with dexmedetomidine, propofol, and midazolam had previously been performed. Knowing that CO2 ventilatory response is affected by many factors, such as hypoxia, exercise, state of wakefulness, sleep, or anesthesia,9 in the groups under comparison, we aimed to maintain a uniform level of anesthesia, which has varied in previous reports for dexmedetomidine.7,8,10 To clarify the depressive effects on respiration, we designed a steady-state protocol using a sevoflurane-anesthetized rabbit model to examine and the different effects of dexmedetomidine, Vol. 109, No. 1, July 2009

propofol, and midazolam on mean arterial blood pressure (MAP) and heart rate (HR) and how each drug affects ventilatory responses to CO2.

METHODS The Laboratory Investigation Committee of Osaka University Medical School approved this study. Animals were cared for in accordance with the University’s standards for care and use of laboratory animals.

Animal Preparation Forty adult New Zealand white rabbits weighing 2.9 ⫾ 0.2 kg (mean ⫾ sd) were used. After a marginal ear vein was cannulated, anesthesia was induced by administering 3%–5% sevoflurane in 4 L/min of 100% oxygen through a facemask. When the corneal reflex was depressed, the animals were placed in a supine position on a heating pad. After infiltration of 1% lidocaine (1–2 mL), tracheostomy was performed. A 3.5 mm (body weight ⬍3.0 kg) or 4.0 mm (body weight ⬎3.0 kg) inner diameter endotracheal tube was inserted into the trachea. The tube was connected to a heat-moisture exchanger (Pneumoist PMH⫹N, TKB International, Costa Mesa, CA) and a pneumotachograph (Hans Rudolph, Kansas City, MO) which, to measure respiratory flow, was connected to a differential pressure transducer (TP-602T, ⫾5 cm H2O, Nihon Kohden, Tokyo, Japan). To measure inspiratory carbon dioxide concentration (FiCO2), fraction of inspired oxygen, and concentration of sevoflurane, inspiratory gas was sampled with an anesthesia gas monitor (Capnomac Ultima; GE Healthcare, Buckinghamshire, UK). The right internal carotid artery was cannulated with a 20-gauge catheter to measure MAP. Blood samples were later aspirated through this carotid arterial line and measured with a blood gas analyzer (ABL505, Radiometer, Copenhagen, Denmark). Respiratory-flow wave and arterial-pressure wave signals were amplified (AP-641G, AR-601G, Nihon Kohden, Tokyo, Japan) and recorded to a computer at a sampling rate of 100 Hz using an analog-digital converter (DI-220, Dataq Instruments, Akron, OH) and data acquisition software (WinDaq, Dataq Instruments, Akron, OH). Each animal received lactate Ringer’s solution at 8 mL 䡠 kg⫺1 䡠 h⫺1 through the venous line and a normal saline solution, including 4 U/mL heparin at 2 mL/h through the arterial line. To maintain the animal’s body temperature at 39°C, body temperature was monitored rectally (Mon-a-therm/Model 6510, Mallinckrodt Medical, St. Louis, MO) and through the heating pad. Measuring devices were calibrated before each experiment.

Experimental Protocol Throughout the experiment, to ensure a stable level of anesthesia, the animals inhaled 2% sevoflurane, which suppressed any voluntary movement of the extremities while maintaining stable conditions for Vol. 109, No. 1, July 2009

respiration and circulation. After inhalation for 1 h to stabilize anesthesia, each rabbit was randomly assigned, according to the drug administered, to one of four groups of 10 animals: control group Group C, distilled water; Group D, dexmedetomidine infused at 2 ␮g 䡠 kg⫺1 䡠 h⫺1; Group P, propofol emulsion, which was administered using target-controlled infusion simulation software (Rugloop simulation software, Version 3.28, University Hospital Ghent, Belgium) to control propofol-in-plasma concentration at 15 ␮g/mL (actual administration was bolus 3.4 mg/kg within 30 s followed by infusion at between 30 to 50 mg 䡠 kg⫺1 䡠 h⫺1); and Group M, midazolam administered as an initial bolus dose of 0.3 mg/kg IV, followed by infusion at 1.86 mg 䡠 kg⫺1 䡠 h⫺1. Throughout the experiment, all animals were allowed to breathe spontaneously. At 15 minutes after the start of the infusion, gas including 0%, 1%, 2%, 3%, 4%, or 5% of CO2 was delivered to each animal in a random sequence of 20-min periods. Fraction of inspired oxygen was maintained throughout at 0.9. The absence of any further increase of end-tidal CO2 and respiratory rate (RR) was taken as confirmation of the achievement of steady-state. Baseline data were recorded just before the beginning of the drug infusion and respiratory-flow waves and arterial-pressure waves were recorded, and arterial blood gases were measured during the final minute of each FiCO2 period. Immediately after sampling, each sample was stored at ⫺70°C for later analysis. For six Group P animals, the plasma concentration of propofol was measured using high performance liquid chromatography. For six Group M animals, the plasma concentration of midazolam was similarly measured using high performance liquid chromatography.

Data Analysis Wave form analysis software (WinDaq playback, Dataq Instruments, Akron, OH) was used for analyzing the stored data. Minute ventilation (MV) and RR were calculated from the recorded respiratory wave forms. MAP and HR were calculated from the recorded arterial-pressure wave forms. All results are expressed as mean ⫾ sd. Plotting the Paco2-MV response has proven useful for assessing respiratory depression caused by drugs.9,11 Linear regression analysis was performed from data for mean MV and mean Paco2 during each FiCO2 period. The relationship between MV and Paco2 is expressed by the following formula9:

MV ⴝ S ⴛ Paco2 ⴙ K Changes in the Paco2-MV response plot are characterized by changes in the X intercept (⫺K/S) and changes in slope (S). The X intercept is derived by extrapolating the Paco2-MV response plot line to zero ventilation and represents the apneic CO2 threshold of the respiratory center.9,12 When the X intercept value © 2009 International Anesthesia Research Society

85

increases, there is a means rightward displacement of the Paco2-MV response plot, which indicates respiratory depression. The Paco2-MV response slope indexes the CO2 sensitivity of the respiratory system: the lower the slope value, the less the CO2 sensitivity and the greater the respiratory depression. Statistical analyses were performed using statistical software (SPSS 13.0, SPSS, Chicago, IL). Repeated analysis of variance was applied for group comparisons, followed by a Tukey honestly significant difference test for post hoc analysis. Using linear regression analysis data, Paco2-ventilatory response was plotted from the data for mean MV and mean Paco2 during each FiCO2 period. Differences among the regression plots of each drug were checked by analysis of covariance. P ⬍ 0.05 was considered statistically significant.

RESULTS We found no statistically significant differences among the four groups in the baseline data of body weight, MV, RR, MAP, HR, and blood gas data (data not shown). Paco2 at the baseline of all the animals was 32.8 ⫾ 5.3 mm Hg. Neither were any statistically significant differences apparent in the measured plasma concentrations of propofol or midazolam at each FiCO2 setting; mean plasma concentrations in the samples from six animals of the relevant groups were

Figure 1. MV (mL/kg body weight) plotted against inspiratory carbon dioxide concentration (FiCO2); MV ⫽ minute ventilation.

Figure 3. Paco2-MV response curves. Linear regression analysis data were plotted for mean MV and mean Paco2 during each FiCO2 period. MV ⫽ minute ventilation. propofol 12.6 ⫾ 2.7 ␮g/mL and midazolam 717.4 ⫾ 115.6 ng/mL. Plots of MV (mL/kg body weight) against FiCO2 in Groups P and M show that MV was statistically significantly lower than in Groups C and D (Fig. 1). Analyzing MV plotted against FiCO2, we found no statistically significant differences between Groups P and M or between Groups C and D. Figure 2 shows the relationship of RR to FiCO2 for each of the groups. We found no statistically significant differences for RR among any of the four groups. Plots for Paco2-MV response were derived from linear regression analysis data for mean MV and mean Paco2 during each FiCO2 period (Fig. 3). Table 1 shows the slope (S) and X intercept (K/S) findings of the linear regression lines. The slope values (CO2 sensitivity) for Group D were statistically significantly lower than that for Group C (P ⬍ 0.05). No similar statistically significant differences between the slope values of the other group pairs was apparent. The X intercepts (apneic CO2 threshold) of Groups P and M were larger than those of Groups C and D (P ⬍ 0.001); our results revealed that greater respiratory depression is associated with propofol and midazolam. Figure 4 shows MAP plotted against FiCO2. In Group D, MAP was statistically significantly lower than in Groups C and M. As Figure 5 shows, across FiCO2 levels, HR for Group D was statistically significantly lower than for Groups C, P, and M.

DISCUSSION

Figure 2. RR (per min) plotted inspiratory carbon dioxide concentration (FiCO2); RR ⫽ respiratory rate.

86

Dexmedetomidine, Propofol, and Midazolam

Our major findings are that dexmedetomidine produces less ventilatory depression but more hypotension and reduction of HR than propofol or midazolam in a sevoflurane-anesthesia rabbit model. Although ventilatory depression was greater with propofol and midazolam than with sevoflurane anesthesia alone, there was much less ventilatory depression when dexmedetomidine was added to sevoflurane. Previous studies have shown that dexmedetomidine produces relatively mild respiratory depression and our results are in line with the findings of those studies.7,8,10 We found that dexmedetomidine, producing a Paco2-MV response plot with a slope (CO2 ANESTHESIA & ANALGESIA

Table 1. Slope (S) and X Intercepts (⫺K/S) of PaCO 2-MV Response Curves Control ⫺1

⫺1

Slope (S) (mL 䡠 kg BW 䡠 mm Hg X intercept (⫺K/S) (mm Hg)

)

24.8 9.21

Dexmedetomidine 17.1* 2.18

Propofol

Midazolam

19.5 21.52*†

20.3 22.07*†

The equation MV (mL/kg body weight) ⫽ S ⫻ PaCO2 ⫹ K is derived from linear-regression analysis of the results shown in Figure 3. The X intercepts (⫺K/S) are derived by extrapolating the PaCO2-MV plot line to zero ventilation and mark the apneic CO2 threshold of the respiratory center. The slopes of PaCO2-MV response curves (S) show the CO2 sensitivity of the respiratory system. MV ⫽ minute ventilation. *P ⬍ 0.05 compared with control. †P ⬍ 0.05 compared with dexmedetomidine group.

Figure 4. MAP plotted against inspiratory carbon dioxide concentration (FiCO2). MAP ⫽ mean arterial blood pressure.

Figure 5. HR plotted against inspiratory carbon dioxide concentration (FiCO2). HR ⫽ heart rate. sensitivity) only slightly different than the control, had little effect on the X intercept (apneic CO2 threshold). In a rabbit rebreathing model, Nishida et al.10 found that administration of dexmedetomidine did not change CO2 sensitivity. Our differing results reflect a different state of wakefulness and different method of measurement. Even in the absence of CO2 inhalation, the Nishida et al. study showed wide RR variation, ranging from 80 to 200 breaths/min. This variation might have obscured any changes in the character of the CO2 sensitivity. In addition, there are two methods for estimating the ventilatory response to CO2: the rebreathing method and the steady-state method. The CO2 sensitivities obtained with the rebreathing method are generally greater than those obtained with the steady-state method.13–15 In two studies, one of welltrained dogs7 and the other of humans,8 bolus dexmedetomidine had an effect on the CO2 sensitivity similar to that in the present study. In the human study, Vol. 109, No. 1, July 2009

dexmedetomidine was also associated with minor increases in the apneic CO2 threshold, which the authors speculated was affected by variations in sedation and sleep status.8 In the present study, because we maintained a uniform state of anesthesia through all the settings, the apneic CO2 threshold may have been little affected by dexmedetomidine. Because the apneic CO2 thresholds for Groups P and M were larger than those of Groups C and D, our results revealed that greater respiratory depression is associated with propofol and midazolam. Although previous studies have similarly shown that propofol and midazolam produce respiratory depression, reports of the effects of sedatives on the Paco2-MV response plot are still confusing. For example, administration of diazepam is reported to shift the apneic CO2 threshold in the same way as our results16; however, Power et al.4 reported that midazolam has little effect on either the CO2 sensitivity or the apneic CO2 threshold in humans. Forster et al.5 presented evidence that midazolam reduces CO2 sensitivities in humans. Moreover, although Goodman et al.17 reported that propofol reduces CO2 sensitivity but does not change the apneic CO2 threshold in humans, Nieuwenhuijs et al.18 found, with volunteers, that propofol reduces CO2 sensitivity and changes apneic CO2 threshold. The X intercept marks the apneic CO2 threshold of the respiratory center.9,12 Because it is difficult to suppress spontaneous breathing effort in the conscious state, it is difficult to actually measure the apneic CO2 threshold. Hickey et al.19 have plotted Paco2-MV response and actually measured apneic CO2 thresholds in humans inhaling volatile anesthetics. They found the measured apneic CO2 thresholds, which were higher than the X intercept of Paco2-MV response plots, were 5–9 mm Hg lower than Paco2 at rest.19 Although the X intercept is not exactly the same as Paco2 at rest, changes in the X intercept may indicate the effects of sedatives on Paco2 at rest. In a resting state, midazolam and propofol are more likely than dexmedetomidine to increase Paco2. The slope of the Paco2-MV response plot shows the CO2 sensitivity of peripheral chemoreceptors and the respiratory center. For example, inhaled anesthetics usually cause ventilatory depression, which is indicated by decreased slope values.11 Our results indicate that the effects of dexmedetomidine are not likely to have much effect on the Paco2-MV response plot. © 2009 International Anesthesia Research Society

87

The pathways through which sedatives affect respiratory depression have yet to be properly clarified. Propofol and midazolam are thought to depress respiration by stimulating central ␥-aminobutyric acid (GABA) A receptors.5,18 This could implicate the caudal ventrolateral medulla, which contains GABAergic neurons that modulate the activity of the rostal ventrolateral medulla which, in turn, is involved in respiratory regulatory within the brainstem.20 By contrast, several studies have reported that ␣2-adrenergic receptors in the brainstem are also involved in respiration. Errchidi et al.21 reported that the activity of the medullary respiratory rhythm generator is modulated by the neurons of the pontine A5 area via ␣2-adrenergic receptors located in the rostal ventrolateral medulla. Oyamada et al.22 reported that the respiratory-phased inhibition of locus coeruleus neurons depends on activation of an ␣2-adrenergic pathway. The present study does not provide any obvious clues why dexmedetomidine, propofol, and midazolam have different respiratory depressive effects. We can only conjecture that the working of GABA A receptors and ␣2-adrenergic receptors in the brainstem may explain our results. In this study, we assume that changes in MV were caused mainly by the change of tidal volume. The RR plot is not significantly different for any of the four groups. Moreover, our findings show that RR changed relatively little as FiCO2 increased. A previous rabbit study found that increasing FiCO2 led to decreasing RR in both nonsedated and dexmedetomidine groups.10 Why the effect on RR in rabbits studies and humans might be different is unknown. To compare CO2 response, it would be ideal to achieve a steady-state of three factors: level of Paco2, blood drug concentration, and level of sedation. In the classical steady-state study,23 the CO2 inhalation time was 20 min, so we decided to allow 20 min CO2 inhalation, which was actually longer than in previous animal studies.13,24 In addition, the absence of any further increase of end-tidal CO2 and RR was taken as confirmation of the achievement of a steady state. Blood concentration of the drugs in each FiCO2 should also be kept at the same level. Stable blood concentration of drugs is best achieved by continuous drug infusion, and a long period of infusion is required to sustain stable concentration. To avoid lengthy infusion, we sought to minimize any differences in drug concentration by randomizing the FiCO2 sequence. Rather than absolute control of drug concentration, given the practical circumstances, we think that this aspect of the protocol was more important for yielding valid data. We administered propofol using target-controlled infusion software control. Midazolam was administered by bolus plus continuous infusion. Bolus administration of dexmedetomidine causes transient circulatory change6,25 so, to avoid these unpredictable effects, we did not use 88

Dexmedetomidine, Propofol, and Midazolam

bolus injection of dexmedetomidine. Plasma concentrations of propofol and midazolam were confirmed to be almost the same at each FiCO2 phase. We did not measure dexmedetomidine plasma concentration, which was assumed to be acceptably similar at each sequencerandomized FiCO2. The results of this study are limited because they are derived from a model using sevoflurane-inhaling rabbits. When investigating the effects of sedatives on respiration, obviously it would be better if control animals did not receive any drug that affects respiration. In animals without sedatives, however, CO2 response would be difficult to measure because exercise, state of wakefulness, and incidental stress or pain are likely to stimulate respiration. Consequently, to stabilize basal conditions, we designed a protocol using inhaled sevoflurane to maintain minimal basal anesthesia in each of the groups. In rabbits, the minimum alveolar concentration (MAC) for sevoflurane is 3.7%, so we administered 2% sevoflurane, or about 0.54 MAC,26 which produces no significant depression of phrenic nerve activity27 and has little effect on hemodynamics.6,10,28 Nguyen et al.7 have reported that, compared with 1.5% isoflurane alone, 3 ␮g/kg of dexmedetomidine during 1.5% isoflurane (1.15 MAC in dogs) anesthesia caused a small increase in Paco2 at rest and decrease in the Paco2-sensitivity. Inhaled anesthetics are assumed to have synergic effects with propofol, midazolam, or dexmedetomidine. Because the present study compared the effects of sedative drugs used in conjunction with sevoflurane, there is always a possibility that sevoflurane might bias the results, instead of merely minimizing the effects of exercise, state of wakefulness, and any stress or pain. In addition, the results exhibited the relative pharmacological characteristics of the three drugs in rabbits. Our results cannot be directly applied to humans in the intensive care unit. Because assessment of the sedation level of animals presents practical difficulties, we also cannot be sure whether the doses of the three drugs were comparable in terms of their sedative effects. In the present study, we used three criteria to evaluate target sedation levels: the drugs did not induce any critical problems in respiration and circulation; the animals exhibited minimal voluntary movement; and the animals maintained response to noxious stimuli. Without sedation, because inhalation of CO2 in itself is stimulating in rabbits, it would be impossible to keep the animals awake and calm. Also, because tests to assess level of sedation would probably affect the response to CO2, it was not practical to evaluate sedation. Consequently, we can only assume that our drug-administration protocols sedated the animals at adequately similar sedation levels. Throughout each protocol, each animal met the target sedation level criteria. Compared with control, administration of dexmedetomidine resulted in lower MAP and decreased HR. Our findings in MAP and HR for dexmedetomidine ANESTHESIA & ANALGESIA

are consistent with other studies.6,10,25,29,30 We found no statistically significant differences for MAP and HR data among Groups C, P, and M. We found that dexmedetomidine induced hypotension and reduced HR more than propofol and midazolam. In conclusion, in a sevoflurane-anesthetized rabbit CO2 inhalation model, the results for midazolam and propofol were similar but, compared with propofol and midazolam, dexmedetomidine produced less respiratory depression, more hypotension and greater reduction of HR. REFERENCES 1. Jacobi J, Fraser GL, Coursin DB, Riker RR, Fontaine D, Wittbrodt ET, Chalfin DB, Masica MF, Bjerke HS, Coplin WM, Crippen DW, Fuchs BD, Kelleher RM, Marik PE, Nasraway SA Jr, Murray MJ, Peruzzi WT, Lumb PD. Clinical practice guidelines for the sustained use of sedatives and analgesics in the critically ill adult. Crit Care Med 2002;30:119 – 41 2. Bellman MH, Pleuvry BJ. Comparison of the respiratory effects of ICI 35868 and thiopentone in the rabbit. Br J Anaesth 1981;53:425–9 3. Goodman NW, Black AM, Carter JA. Some ventilatory effects of propofol as sole anaesthetic agent. Br J Anaesth 1987;59:1497–503 4. Power SJ, Morgan M, Chakrabarti K. Carbon dioxide response curves following midazolam and diazepam. Br J Anaesth 1983;55:837– 41 5. Forster A, Gardaz JP, Suter PM, Gemperie M. Respiratory depression by midazolam and diazepam. Anesthesiology 1980;53:494 –7 6. Zornow MH. Ventilatory, hemodynamic and sedative effects of the ␣2 adrenergic agonist, dexmedetomidine. Neuropharmacology 1991;30:1065–71 7. Nguyen D, Abdul-Rasool I, Ward D, Hsieh J, Kobayashi D, Hadlock S, Singer F, Bloor B. Ventilatory effects of dexmedetomidine, atipamezole, and isoflurane in dogs. Anesthesiology 1992;76:573–9 8. Belleville JP, Ward DS, Bloor BC, Maze M. Effects of intravenous dexmedetomidine in humans. I. Sedation, ventilation, and metabolic rate. Anesthesiology 1992;77:1125–33 9. Cunningham DLC, Robbins PA, Wolef CB. Integration of respiratory response to change in alveolar partial pressures of CO2 and O2 and in arterial pH. In: Goerke J, Clements JA, eds. Handbook of physiology. The respiratory system 2. Bethesda: American Physiological Society, 1986:475–507 10. Nishida T, Nishimura M, Kagawa K, Hayashi Y, Mashimo T. The effects of dexmedetomidine on the ventilatory response to hypercapnia in rabbits. Intensive Care Med 2002;28:969 –75 11. Pavlin ED, Hornbein T. Anesthesia and the control of ventilation. In: Goerke J, Clements JA, eds. Handbook of physiology. The respiratory system 2. Bethesda: American Physiological Society, 1986:793– 813 12. Belville JW, Seed JC. The effect of drugs on the respiratory response to carbon dioxide. Anesthesiology 1960;21:727– 41

Vol. 109, No. 1, July 2009

13. Steven C. Methadone-induced respiratory depression in the dog: comparison of steady-state and rebreathing techniques and correlation with serum drug concentration. J Pharmacol Exp Ther 1978;207:109 –22 14. Berkenbosch A, Bovill JG, Dahan A. The ventilatory CO2 sensitivities form Read’s rebreathing method and the steadystate method are not equal in man. J Physiol 1989;411:367–77 15. Bourke DL, Warley A. The steady-state and rebreathing methods compared during morphine administration in humans. J Physiol 1989;419:509 –17 16. Dalen JE, Evans GL, Banas JS Jr, Brooks HL, Paraskos JA, Dexter L. The hemodynamic and respiratory effects of diazepam (Valium). Anesthesiology 1969;30:259 – 63 17. Goodman NW, Black AM, Carter JA. Some ventilatory effects of propofol as sole anaesthetic agent. Br J Anaesth 1987;59:1497–503 18. Nieuwenhuijs D, Sarton E, Teppema LJ, Kruyt E, Olievier I, van Kleef J, Dahan A. Respiratory sites of action of propofol: absence of depression of peripheral chemoreflex loop by low-dose propofol. Anesthesiology 2001;95:889 –95 19. Hickey RF, Fourcade HE, Eger EI 2nd, Larson CP Jr, Bahlman SH, Stevens WC, Gregory GA, Smith NT. The effects of ether, halothane, and Forane on apneic thresholds in man. Anesthesiology 1971;35:32–7 20. Mandel DA, Schreihofer AM. Central respiratory modulation of barosensitive neurones in rat caudal ventrolateral medulla. J Physiol 2006;572:881–96 21. Errchidi S, Monteau R, Hilaire G. Noradrenergic modulation of the medullary respiratory rhythm generator in the newborn rat: an in vitro study. J Physiol 1991;443:477–98 22. Oyamada Y, Ballantyne D, Mu¨ckenhoff K, Scheid P. Respirationmodulated membrane potential and chemosensitivity of locus coeruleus neurones in the in vitro brainstem-spinal cord of the neonatal rat. J Physiol 1998;513:381–98 23. Jordan C. Assessment of the effects of drugs on respiration. Br J Anaesth 1982;54:763– 82 24. Honda Y. Ventilatory response to CO2 during hypoxia and hyperoxia in awake and anesthetized rabbits. Respir Physiol 1968;5:279 – 87 25. Bloor BC, Ward DS, Belleville JP, Maze M. Effects of intravenous dexmedetomidine in humans. II. Hemodynamic changes. Anesthesiology 1992;77:1134 – 42 26. Scheller MS, Saidman LJ, Partridge BL. MAC of sevoflurane in humans and the New Zealand white rabbit. Can J Anaesth 1988;35:153– 6 27. Ma D, Chakrabarti MK, Whitwam JG. The combined effects of sevoflurane and remifentanil on central respiratory activity and nociceptive cardiovascular responses in anesthetized rabbits. Anesth Analg 1999;89:453– 61 28. Flecknell PA, Roughan JV, Hedenqvist P. Induction of anaesthesia with sevoflurane and isoflurane in the rabbit. Lab Anim 1999;33:41– 6 29. Ickeringill M, Shehabi Y, Adamson H, Ruettimann U. Dexmedetomidine infusion without loading dose in surgical patients requiring mechanical ventilation: haemodynamic effects and efficacy. Anaesth Intensive Care 2004;32:741–5 30. Ebert TJ, Hall JE, Barney JA, Uhrich TD, Colinco MD. The effects of increasing plasma concentrations of dexmedetomidine in humans. Anesthesiology 2000;93:382–94


89

Predicted Propofol Effect-Site Concentration for Induction and Emergence of Anesthesia During Early Pregnancy Nicolas Mongardon, MD* Fre´de´rique Servin, MD, PhD* Mathilde Perrin, MD* Ennoufous Bedairia, MD* Sylvie Retout, PhD† Chadi Yazbeck, MD, PhD‡ Philippe Faucher, MD‡ Philippe Montravers, MD, PhD* Jean-Marie Desmonts, MD* Jean Guglielminotti, MD*

BACKGROUND: Pregnancy is associated with decreased hypnotic requirement, allegedly related to progesterone. However, the effects of pregnancy and progesterone on propofol requirement have not been thoroughly investigated. We conducted this study to determine whether propofol dose and predicted effect-site concentration for loss of consciousness (LOC) during induction of anesthesia, and eye opening during emergence from anesthesia, are decreased during early pregnancy. We also investigated whether blood progesterone was correlated with propofol dose and effect-site concentration for LOC. METHODS: We studied 57 ASA I-II women patients undergoing elective termination of pregnancy and 55 control patients undergoing transvaginal oocyte puncture for in vitro fertilization. Anesthesia was induced by administration of a 1% propofol infusion at 200 mL/min. Propofol dose and calculated effect-site concentration (Schnider model) were recorded at the time of LOC during induction. We also calculated effect-site concentration at the time of eye opening upon emergence from anesthesia. Blood progesterone was measured after surgery. RESULTS: Mean (⫾1 sd) propofol dose at LOC was significantly reduced in the pregnant patients compared with the nonpregnant control patients (108.57 ⫾ 20.04 vs 117.59 ⫾ 17.98 mg, respectively; P ⫽ 0.014). Similarly, the calculated propofol effect-site concentration at LOC was significantly lower in the pregnant patients than the nonpregnant control patients (4.59 ⫾ 0.72 vs 5.01 ⫾ 0.64 ␮g/mL, respectively; P ⫽ 0.0014). There was no difference in the calculated effect-site concentration on eye opening upon emergence. No significant relationship was observed between blood progesterone and propofol dose or calculated propofol effect-site concentration at LOC. CONCLUSION: Propofol dose and predicted propofol effect-site concentration at LOC are decreased during early pregnancy. Progesterone does not explain this result. (Anesth Analg 2009;109:90 –5)

T

he recovery profile of propofol makes it the anesthetic of choice for short-duration day-case surgical procedures. In this setting, recovery is optimized by use of a target-controlled infusion (TCI) that targets the propofol concentration at the site of drug effect.1– 4 The currently available propofol pharmacokinetic models running TCI infusion pumps considers patient characteristics: weight for the Marsh model or age, weight, sex, and height for the Schnider model.5,6 From the *De´partement d’Anesthe´sie et de Reánimation Chirurgicale; †De´partement d’Epide´miologie, Biostatistique et Recherche Clinique; and ‡Service de Gynećologie et d’Obste´trique, Assistance Publique-Hoˆpitaux de Paris, Hoˆpital Bichat-Claude Bernard, Paris, France. Accepted for publication January 5, 2009. Presented, in part, at the annual meeting of the American Society of Anesthesiologists, Chicago, October 14 –18, 2006. Address correspondence and reprint requests to Jean Guglielminotti, MD, De´partement d’Anesthe´sie et de Reánimation Chirurgicale, Assistance Publique-Hoˆpitaux de Paris, Hoˆpital Bichat-Claude Bernard, 46 rue Henri Huchard, 75018 Paris, France. Address e-mail to [email protected]. Copyright © 2009 International Anesthesia Research Society DOI: 10.1213/ane.0b013e3181a1a700

90

The models do not account for the possibility that the patient is pregnant. Early and late pregnancy reduces the minimum alveolar anesthetic concentration (MAC) of volatile anesthetics, and the dose of thiopental that abolishes response to verbal command in 95% of patients (the “ED95”).7–11 The cause for the change in anesthetic potency in pregnancy is controversial,9,11 but some investigators attribute it to increased progesterone.12,13 However, the effect of pregnancy on propofol dose or effect-site concentration (Ce) required for loss of consciousness (LOC) has not been thoroughly investigated. Higuchi et al.14 reported that the concentration of propofol associated with 50% probability of LOC, C50, is not decreased during early pregnancy. Unfortunately, calculation of C50 was based on venous propofol concentrations, which do not reflect Ce or the rapid increase in concentration during anesthetic induction.15 Real-time calculation of propofol Ce is now possible with new TCI devices. We conducted this study to determine whether propofol dose and predicted Ce at LOC during induction of anesthesia, and at eye opening upon emergence Vol. 109, No. 1, July 2009

from anesthesia, are decreased in early pregnancy. Our second goal was to determine whether the propofol dose and Ce associated with our study end points correlated with blood progesterone concentration.

METHODS Patients The study was approved by the Ethics Committee of Cochin Hospital, Assistance Publique-Hoˆpitaux de Paris. Written informed consent was obtained from each patient. Consecutive women patients scheduled to undergo either early elective termination of pregnancy (TOP group) or transvaginal oocyte retrieval for in vitro fertilization (IVF group) were approached for participating in the study. Patients were excluded from the study for any of the following conditions: age younger than 18 yr, ASA physical status III or IV, regular medication with a ␤ blocker, anxiolytic, antidepressant or opioid, alcohol or drug abuse, obesity (Body Mass Index ⬎30 kg/m2), and cardiac, neurological, or psychiatric disease. The diagnosis of pregnancy and estimation of fetal age was made with ultrasonic abdominal examination. The upper fetal age limit for legal elective TOP in France is 14 wk. For each patient, lean body mass was calculated with the following formula: (1.07 ⫻ weight ⫺ 148 ⫻ [weight/height]2).6 Study Design All IVF patients underwent the same drug protocol before oocyte retrieval. Follicular development was induced with follitropine beˆta (Puregon®, Organon, Puteaux, France) and induction of ovulation with choriogonadotrophin ␣ (Ovitrelle®, Merck Lipha Sante´, Lyon, France). None of the patients received exogenous progesterone. On the morning of surgery, no preanesthetic medication was given to IVF patients. TOP patients received an oral combination of cimetidine and citrate 2 h before surgery. On arrival in the operative room, a 20-gauge IV catheter was inserted on the left hand. A three-way stopcock was connected to the catheter and devoted to propofol infusion. Patients were monitored with noninvasive arterial blood pressure, five-lead electrocardiogram, pulse oximetry, and Bispectral Index (BIS XP, Aspect Medical Systems, Newton, MA). The smoothing interval for BIS analysis was 15 s. We did not consider the lag time between the clinical effect and the BIS reading. The BIS values presented are the values read at the time of LOC during induction and eye opening during emergence. Propofol was infused using the Base Primea pump (Fresenius-Vial Company, Brezins, France). The control screen of the infusion pump numerically displays real-time predicted propofol Ce according to Schnider’s Vol. 109, No. 1, July 2009

pharmacokinetic model,6 the total infused dose of propofol, and the time elapsed from the start of the infusion. Acetaminophen (1 g) was administered IV at the beginning of the case. After administration of oxygen, anesthesia was induced with a continuous 1% propofol infusion (Diprivan®, Astra-Zeneca, RueilMalmaison, France) at 200 mL/h until LOC. Prior IV lidocaine administration was forbidden. LOC was defined as loss of verbal contact, and was tested every 15 s from the start of propofol infusion by asking the patient with a normal voice, but without tactile stimulation, to say her name. After LOC, the continuous infusion was switched to an effect-site TCI. Immediately before the start of surgery, a single 10 ␮g/kg alfentanil bolus was administered. During surgery, a propofol effect-site target was selected to maintain the BIS value between 45 and 55. Until LOC, the airway was managed with a face mask without jaw advancement to avoid patient stimulation. After LOC, patients breathed spontaneously 100% oxygen with a face mask and, if necessary, with jaw advancement. No oral device was used. Duration of surgery was defined as the time elapsed between the introduction of the transvaginal echographic probe in the IVF group, or vaginal insertion of the speculum in the TOP group, and its removal. Propofol infusion was stopped after the removal of the speculum or of the echographic probe. After propofol infusion had been stopped, the patient was asked every 15 s with a normal voice and without tactile stimulation to open her eyes. Duration of anesthesia was defined as time elapsed between LOC and eyes opening after propofol discontinuation. Time for eyes opening was defined as the time elapsed between propofol discontinuation and eyes opening. At the time of LOC on induction of anesthesia and eye opening upon emergence from anesthesia, 1 of the 2 principal investigators (JG or NM) recorded from the TCI pump the following variables: predicted propofol Ce, total propofol dose, time elapsed from the start of infusion, and BIS value. As two observers could record Ce displayed on the Base Primea control screen, reproducibility between the two observers was assessed as follows. For 11 patients at the time of LOC, Ce was simultaneously and independently recorded by the two observers. Alfentanil concentration at the time of eye opening upon emergence from anesthesia was calculated with pkpdtools (C Minto and T Schnider, pkpdtools.com) using Maitre et al.16 pharmacokinetic model. Venous blood samples were obtained on arrival in the postanesthesia care unit. Blood progesterone concentration was measured with a competitive chemiluminescent enzyme immunoassay method (Progesterone II Elecsys test, Roche Diagnostics GmbH, Mannheim, Germany). © 2009 International Anesthesia Research Society

91

Table 1. Characteristics of the two Groups of Female Patients TOP group IVF group (n ⫽ 57) (n ⫽ 55) Age (yr) Height (cm) Weight (kg) Lean body mass (kg) Fetal age Blood progesterone concentration (nmol/L) Duration of anesthesia Duration of surgery (min) Time for eyes opening (min)

P

27 ⫾ 8 166 ⫾ 6 59 ⫾ 11 44 ⫾ 5 10.6 ⫾ 1.6 69 ⫾ 3

32 ⫾ 4 165 ⫾ 5 61 ⫾ 10 44 ⫾ 5 — 42 ⫾ 4

⬍0.0001 0.51 0.45 0.64 — ⬍0.0001

14 ⫾ 4

19 ⫾ 9

0.0005

7⫾3

11 ⫾ 5

⬍0.0001

7⫾3

8⫾6

0.47

Results are presented as mean ⫾ 1 SD. Lean body mass was calculated as 1.07 ⫻ weight ⫺ 148 ⫻ (weight/height)2.6 TOP ⫽ termination of pregnancy; IVF ⫽ in vitro fertilization.

Statistical Analysis Results are expressed as mean ⫾1 sd (95% confidence interval). Categorical data were compared with the ␹2 test and quantitative data with the Student’s t-test. Association between quantitative variables was tested with simple regression analysis. P values ⬍0.05 were considered statistically significant. Analysis was conducted with Statview software (SAS Institute, Cary, NC). Ce at the time of LOC measured in a pilot study of 21 TOP patients was 4.5 ⫾ 0.76 ␮g/mL (mean ⫾ 1 sd). To demonstrate a 10% difference between TOP and IVF patients with a 0.05 ␣ risk, a 0.20 ␤ risk and a 2-tailed test, 46 patients in each group were required. The investigator (JG) who performed data analysis was not blinded to the pregnancy state.

RESULTS From July to December 2005, 120 patients were enrolled. Eight patients were excluded: two had problems with the propofol infusion, and eight had unrecognized noninclusion criteria. Data of 112 patients were therefore analyzed, 57 in the TOP group and 55 in the IVF group. The characteristics of the two groups are presented in Table 1. The TOP group was significantly younger and had higher blood progesterone concentration than the IVF group. Elective TOP duration was shorter than oocyte retrieval.

The mean propofol dose at the time of LOC was significantly lower in the TOP group compared with the IVF group (108.57 ⫾ 20.04 vs 117.59 ⫾ 17.98 mg, respectively; P ⫽ 0.014) as was time elapsed from the start of infusion to LOC (193.8 ⫾ 31.5 vs 209.5 ⫾ 32.8 s, respectively; P ⫽ 0.026) (Table 2). Similarly, Ce for propofol at LOC was significantly lower in the TOP group (4.59 ⫾ 0.72 vs 5.01 ⫾ 0.64 ␮g/mL, respectively; P ⫽ 0.0014) (Table 2). BIS values at the time of LOC did not differ between TOP and IVF groups (80 ⫾ 8 vs 78 ⫾ 9, respectively; P ⫽ 0.20). Owing to shorter duration of surgery (Table 1), the mean propofol dose infused at the time of eye opening upon emergence was lower in the TOP group compared with the IVF group (260 ⫾ 68 vs 323 ⫾ 116 mg, respectively; P ⫽ 0.0017). At the time of eye opening, the mean Ce for propofol did not differ between TOP and IVF groups (1.68 ⫾ 0.46 vs 1.66 ⫾ 0.3, respectively; P ⫽ 0.82) as did BIS (71 ⫾ 9 vs 70 ⫾ 10, respectively; P ⫽ 0.57). However, the mean predicted alfentanil Ce was higher in the TOP group (Table 2). No association was observed at the time of LOC between blood progesterone concentration and propofol dose (r ⫽ 0.02 and P ⫽ 0.89 in the 57 TOP patients and r ⫽ 0.05 and P ⫽ 0.58 in the 112 TOP and IVF patients) and between blood progesterone concentration and Ce (r ⫽ 0.02 and P ⫽ 0.87 in the 57 TOP patients and r ⫽ 0.11 and P ⫽ 0.23 in the 112 TOP and IVF patients) (Fig. 1). The reproducibility between recordings by the two observers was estimated with the intraclass correlation coefficient. It was 0.992, indicating a very good reproducibility.

DISCUSSION The propofol dose required for LOC is reduced in TOP patients compared with nonpregnant patients undergoing oocyte retrieval. Similarly, predicted propofol Ce at LOC was decreased in TOP patients. There was no relationship between propofol dose or Ce for LOC and progesterone blood concentration. At the time of eye opening upon emergence from anesthesia, propofol Ce did not differ between groups. Previous studies have demonstrated decreased hypnotic requirements for both induction and maintenance of anesthesia during pregnancy. A 20% to 30% decrease in MAC for several inhaled anesthetics has

Table 2. Results TOP group (n ⫽ 57)

IVF group (n ⫽ 55)

Propofol dose at LOC (mg) 108.57 ⫾ 20.04 (103.25–113.89) 117.59 ⫾ 17.98 (112.73–122.45) Time to LOC (s) 193.8 ⫾ 31.5 (183.5–204.0) 209.5 ⫾ 32.8 (200.0–219.0) Propofol Ce at LOC (␮g/mL) 4.59 ⫾ 0.72 (4.40–4.78) 5.01 ⫾ 0.64 (4.84–5.18) Propofol dose at EO (mg) 260 ⫾ 68 (240–280) 323 ⫾ 116 (289–357) Propofol Ce at EO (␮g/mL) 1.68 ⫾ 0.46 (1.54–1.81) 1.66 ⫾ 0.36 (1.55–1.76) Alfentanil Ce at EO (ng/mL) 22.6 ⫾ 8.9 (19.9–25.3) 18.1 ⫾ 6.2 (16.3–20.0)

P

Mean difference

0.014 ⫺9.02 (⫺16.16 to ⫺1.89) 0.026 ⫺15.7 (⫺29.5 to ⫺1.9) 0.0014 ⫺0.42 (⫺0.67 to ⫺0.17) 0.0017 ⫺63 (⫺102 to ⫺24) 0.82 0.02 (⫺0.15 to 0.19) 0.0075 4.4 (1.2–7.7)

Data are presented as mean ⫾ 1 SD (95% confidence interval). TOP ⫽ termination of pregnancy; IVF ⫽ in vitro fertilization; LOC ⫽ loss of consciousness; EO ⫽ eye opening; Ce ⫽ predicted effect-site concentration.

92

Early Pregnancy and Propofol


Figure 1. Association between blood progesterone concentration and propofol dose (top) and blood progesterone concentration and predicted propofol effect-site concentration (Ce) (bottom) for loss of consciousness in the pooled 112 patients of the elective termination of pregnancy and in vitro fertilization groups.

been observed as early as the 8th wk of gestation and up to the first 48 h after delivery.7–9,11 Similarly, thiopenthal’s ED 95 for loss of responsiveness to verbal command is reduced by 20% at the 7th wk of gestation.10 In the current study, we observed a decrease in both mean propofol dose and propofol Ce for LOC in the elective TOP group when compared with the IVF group. It contrasts with Higuchi et al.’s14 results, who did not observe decreased venous propofol concentration for LOC during early pregnancy. However, a few differences between the two studies may account for these apparently contradictatory results. First, the fetal age was younger in Higuchi et al.’s study (median: 8 vs 11 wk). Second, Higuchi et al. infused lidocaine before starting the propofol infusion. Third, Higuchi et al. measured venous propofol concentration, which does not accurately reflect arterial concentration or Ce.15 Although statistically significant, the small difference seen in propofol dose and concentration at LOC with pregnancy (9 mg and 0.42 ␮g/mL, respectively) Vol. 109, No. 1, July 2009

is likely of little clinical significance. The wide intragroup and intergroup variability of both dose and Ce suggests that the propofol dose or targeted Ce for induction of anesthesia should be based on individual titration to a specific end point. Moreover, the increase in cardiac output observed as early as the 8th wk of gestation may have limited the expected decrease in propofol dose and concentration for LOC in TOP patients because of an inverse relationship between cardiac output and propofol dose and concentration during induction of anesthesia.17,18 The effect-site is a hypothetical compartment with a negligible mass at which the anesthetic exerts its pharmacological effect. Its concept was developed to explain the hysteresis observed between the time course of blood concentrations and clinical effect.19 –21 Whereas many studies have measured blood propofol concentration required for LOC, fewer have examined the Ce at LOC. Struys et al. reported a mean Ce value for LOC of 4.7 ␮g/mL with Schnider model.2 This is close to the 4.59 ␮g/mL observed in our IVF control group. Predicted Ce with pharmacokinetic software is an estimation whose accuracy may be questioned when a measured concentration is lacking. Owing to the discrepancy between venous and arterial concentration and Ce, measuring concentration would have implied arterial sampling which was unethical for young patients undergoing short day-case procedures.15 To minimize the bias resulting from the application of a pharmacokinetic model in a population in which the kinetic set was not derived, we included young, healthy, and nonobese patients. We cannot exclude the possibility that early pregnancy affects the accuracy of the Schnider model. However, the slow infusion rate we used for inducing anesthesia should decrease the variability of the pharmacokinetic model.22,23 Moreover, propofol dose or time required for LOC are measured and not calculated variables. The observed decrease of both dose and time to LOC in TOP patients suggests that the aforementioned limitations regarding Ce calculation had little effect on the results. During the first minute of propofol infusion, prediction of Ce with classic pharmacokinetic models may be inaccurate.24 However, time from start of infusion to LOC was around 3 min in our patients, which may have avoided this bias. The decreased Ce at LOC in the TOP patients may have resulted from pharmacokinetic changes related to hemodynamic changes of pregnancy, ke0 change or altered patient sensitivity (a pharmacodynamic change). The design of our study does not allow us to determine the involved mechanisms. Propofol is highly bound to albumin.25 Hypoalbuminemia increases unbound propofol fraction and may therefore influence Ce. However, the decrease in plasma albumin during the first trimester of pregnancy is small, around 13%.26,27 Moreover, one previous study suggests that even profound hypoalbuminemia has little or no effect on the accuracy of TCI models.28 © 2009 International Anesthesia Research Society

93

Predicted Ce for propofol at eye opening upon emergence from anesthesia did not differ between the elective TOP and IVF groups. There are several possible explanations. First, sample-size calculation was based on demonstrating decreased Ce in the IVF group for LOC but not for eye opening upon emergence. We cannot therefore exclude a Type 1 error. To demonstrate a 10% difference between TOP and IVF patients in Ce at eye opening upon emergence with a 0.05 ␣ risk, a 0.20 ␤ risk and a 2-tailed test, 125 patients in each group would have been required. Second, predicted Ce for alfentanil was higher in the TOP group owing to a shorter duration of surgery in the TOP group than in the IVF group. A higher opioid concentration may have masked a lower propofol concentration on eye opening, due to the synergy between opioids and hypnotics on consciousness.29,30 However, alfentanil concentration at the end of surgery was small (mean for the 112 patients: 25 ng/mL) as was the difference between groups (mean difference between groups: 4.2 ng/mL). Moreover, the synergistic hypnotics-opioid interaction is stronger for tracheal intubation or skin incision than for LOC or eye opening.29,30 It is therefore unlikely that alfentanil accounted for the lack of difference in Ce at the time of eye opening. The mechanisms of a pregnancy-induced decrease in anesthetic requirement remain hypothetical. Although animal experiments suggest progesterone involvement,12 human studies do not support this finding.7,9,11 Clinical trials have not demonstrated any relationship between maternal blood progesterone concentration and MAC for volatile anesthetics both during early pregnancy and after delivery. This is in close agreement with our results. The effect of progesterone on hypnotic requirements may be different in other situations, as suggested by Erden et al.13 They demonstrated that an increased progesterone level observed during the luteal phase of the menstrual cycle is associated with decreased sevoflurane requirements for maintaining anesthesia using a BIS-driven algorithm. We cannot exclude that estrogen played a role, because of a negative correlation between estrogen level and decreased sensitivity to barbiturates has been demonstrated.31 In conclusion, propofol dose and predicted Ce for LOC on induction of anesthesia, but not for eye opening upon emergence from anesthesia, are decreased during early pregnancy. Progesterone does not explain this result. REFERENCES 1. Passot S, Servin F, Allary R, Pascal J, Prades JM, Auboyer C, Molliex S. Target-controlled versus manually-controlled infusion of propofol for direct laryngoscopy and bronchoscopy. Anesth Analg 2002;94:1212–16 2. Struys MM, De Smet T, Depoorter B, Versichelen LF, Mortier EP, Dumortier FJ, Shafer SL, Rolly G. Comparison of plasma compartment versus two methods for effect compartmentcontrolled target-controlled infusion for propofol. Anesthesiology 2000;92:399 – 406

94

Early Pregnancy and Propofol

3. Wakeling HG, Zimmerman JB, Howell S, Glass PS. Targeting effect compartment or central compartment concentration of propofol: what predicts loss of consciousness? Anesthesiology 1999;90:92–7 4. Struys M, Versichelen L, Thas O, Herregods L, Rolly G. Comparison of computer-controlled administration of propofol with two manually controlled infusion techniques. Anaesthesia 1997;52:41–50 5. Marsh B, White M, Morton N, Kenny GN. Pharmacokinetic model driven infusion of propofol in children. Br J Anaesth 1991;67:41– 8 6. Schnider TW, Minto CF, Gambus PL, Andresen C, Goodale DB, Shafer SL, Youngs E. The influence of method of administration and covariates on the pharmacokinetics of propofol in adult volunteers. Anesthesiology 1998;88:1170 – 82 7. Chan MT, Gin T. Postpartum changes in the minimum alveolar concentration of isoflurane. Anesthesiology 1995;82:1360 –3 8. Chan MT, Mainland P, Gin T. Minimum alveolar concentration of halothane and enflurane are decreased in early pregnancy. Anesthesiology 1996;85:782– 6 9. Gin T, Chan MT. Decreased minimum alveolar concentration of isoflurane in pregnant humans. Anesthesiology 1994;81:829 –32 10. Gin T, Mainland P, Chan MT, Short TG. Decreased thiopental requirements in early pregnancy. Anesthesiology 1997;86:73– 8 11. Zhou HH, Norman P, DeLima LG, Mehta M, Bass D. The minimum alveolar concentration of isoflurane in patients undergoing bilateral tubal ligation in the postpartum period. Anesthesiology 1995;82:1364 – 8 12. Datta S, Migliozzi RP, Flanagan HL, Krieger NR. Chronically administered progesterone decreases halothane requirements in rabbits. Anesth Analg 1989;68:46 –50 13. Erden V, Yangin Z, Erkalp K, Delatiog˘lu H, Bahçeci F, Seyhan A. Increased progesterone production during the luteal phase of menstruation may decrease anesthetic requirement. Anesth Analg 2005;101:1007–11 14. Higuchi H, Adachi Y, Arimura S, Kanno M, Satoh T. Early pregnancy does not reduce the C50 of propofol for loss of consciousness. Anesth Analg 2001;93:1565–9 15. Coetzee JF, Glen JB, Wium CA, Boshoff L. Pharmacokinetic model selection for target controlled infusions of propofol. Assessment of three parameter sets. Anesthesiology 1995;82:1328 – 45 16. Maitre PO, Vozeh S, Heykants J, Thomson DA, Stanski DR. Population pharmacokinetics of alfentanil: the average doseplasma concentration relationship and interindividual varialibity in patients. Anesthesiology 1987;66:3–12 17. Capeless EL, Clapp JF. Cardiovascular changes in early phase of pregnancy. Am J Obstet Gynecol 1989;161:1449 –53 18. Upton RN, Ludbrook GL, Grant C, Martinez AM. Cardiac output is a determinant of the initial concentrations of propofol after short-infusion administration. Anesth Analg 1999;89:545–52 19. Sheiner LB, Beal S, Rosenberg B, Marathe VV. Forecasting individual pharmacokinetics. Clin Pharmacol Ther 1979;26:294 –305 20. Sneyd JR, Rigby-Jones AE. Effect site: who needs it? Br J Anaesth 2007;98:701– 4 21. Minto CF, Schnider TW, Gregg KM, Henthorn TK, Shafer SL. Using the time of maximum effect site concentration to combine pharmacokinetics and pharmacodynamics. Anesthesiology 2003;99:324 –33 22. Hu C, Horstman DJ, Shafer SL. Variability of target-controlled infusion is less than the variability after bolus injection. Anesthesiology 2005;102:639 – 45 23. Doufas AG, Bakhshandeh M, Bjorksten AR, Shafer SL, Sessler DI. Induction speed is not a determinant of propofol pharmacodynamics. Anesthesiology 2004;101:1112–21 24. Avram MJ, Krejcie TC. Using front-end kinetics to optimize target-controlled drug infusions. Anesthesiology. 2003;99:1078 – 86 25. Mazoit JX, Samii K. Binding of propofol to blood components: implications for pharmacokinetics and for pharmacodynamics. Br J Clin Pharmacol 1999;47:35– 42 26. Coryell MN, Beach EF, Robinson AR, Macy IG, Mack HC. Metabolism of women during the reproductive cycle. XVII. Changes in electrophoretic patterns of plasma proteins throughout the cycle and following delivery. J Clin Invest 1950;29:1559 – 67 27. Mendenhall HW. Serum protein concentrations in pregnancy. I. Concentrations in maternal serum. Am J Obstet Gynecol 1970;106:388 –99


28. Cavaliere F, Conti G, Moscato U, Meo F, Pennisi MA, Costa R, Proietti R. Hypoalbuminaemia does not impair Diprifusor performance during sedation with propofol. Br J Anaesth 2005;94:453–58 29. Vuyk J, Engbers FH, Burm AG, Vletter AA, Griever GE, Olofsen E, Bovill JG. Pharmacodynamic interaction between propofol and alfentanil when given for induction of anesthesia. Anesthesiology 1996;84:288 –99

Vol. 109, No. 1, July 2009

30. Vuyk J, Mertens MJ, Olofsen E, Burm AG, Bovill JG. Propofol anesthesia and rational opioid selection: determination of optimal EC50-EC95 propofol-opioid concentrations that assure adequate anesthesia and a rapid return of consciousness. Anesthesiology 1997;87:1549 – 62 31. Bergman A, Niv D, David M P, Yanai J. Barbiturate narcosis and estrogen levels in women. Gynecol Obstet Invest 1987;23:167–71


95

The Shivering Threshold in Rabbits with JM-1232(ⴚ), a New Benzodiazepine Receptor Agonist Taishi Masamune, MD* Hiroaki Sato, MD* Katsumi Okuyama, MD, PhD† Yusuke Imai, MD‡ Hironobu Iwashita, MD, PhD* Tadahiko Ishiyama, MD, PhD* Takeshi Oguchi, MD, PhD† Daniel I. Sessler, MD§ Takashi Matsukawa, MD, PhD†

BACKGROUND: JM-1232(⫺) is a novel isoindoline derivative which shows sedative and hypnotic activities through the benzodiazepine site of ␥-aminobutyric acid type A (GABAA) receptors. Typical doses of midazolam, another GABAA receptor agonist, slightly reduce the shivering threshold in humans. We thus determined the extent to which JM-1232(⫺) decreases the shivering threshold. METHODS: Eighteen rabbits, lightly anesthetized with isoflurane 0.2 minimum alveolar anesthetic concentration (MAC), were randomly assigned to infusions of 1) saline (control), 2) 0.01 mg 䡠 kg⫺1 䡠 min⫺1 JM-1232(⫺), or 3) 0.1 mg 䡠 kg⫺1 䡠 min⫺1 JM-1232(⫺). Body temperature was reduced at a rate of 2–3°C/h by perfusing water at 10°C though a U-shaped plastic tube positioned in the colon. Cooling continued until shivering was observed by an investigator blinded to treatment, or until core temperature reached 34°C. Core temperatures were recorded from the distal esophagus, and core temperature at the onset of shivering defined the threshold. Data were analyzed by one-way analysis of variance with Student-Newman-Keuls tests. Results are presented as means ⫾ sd; P ⬍ 0.05 was considered statistically significant. RESULTS: The rabbits given a saline infusion shivered at 36.5 ⫾ 0.3°C. Five of the six rabbits given JM-1232(⫺) at a rate of 0.01 mg 䡠 kg⫺1 䡠 min⫺1 shivered at 35.7 ⫾ 0.8°C, and one of these rabbits failed to shiver at 34.0°C. None of the rabbits given JM-1232(⫺) at a rate of 0.1 mg 䡠 kg⫺1 䡠 min⫺1 shivered before reaching the 34.0°C cutoff temperature. CONCLUSION: A low dose of JM-1232(⫺) reduced the shivering threshold in rabbits approximately 0.8°C which is similar to the effects in humans given premedication doses of midazolam. In contrast, a 10-fold larger dose reduced the threshold more than 2.5°C. This is a substantial decrement and might facilitate induction of therapeutic hypothermia. (Anesth Analg 2009;109:96 –100)

P

erioperative hypothermia is common and causes numerous serious complications.1–3 Hypothermia results from the combination of drug-induced impairment of thermoregulatory control4 and exposure to a cool environment. In practice, however, thermoregulatory impairment and subsequent redistribution of body heat from core to peripheral tissues is the primary cause.5 General anesthetics profoundly impair thermoregulatory defenses, but thermoregulatory defenses are also From the *Operating Theater, Yamanashi University Hospital, Yamanashi, Japan; †Department of Anesthesiology, University of Yamanashi, Japan; ‡Department of Anesthesia, Kanoiwa General Hospital, Yamanashi, Japan; and §Department of Outcomes Research, The Cleveland Clinic, Cleveland, Ohio. Accepted for publication December 17, 2008. Supported by the Joseph Drown Foundation (Los Angeles, CA). Maruishi Pharmaceutical CO, Ltd. (Osaka, Japan) donated the JM-1232(⫺). There are no potential conflicts of interest arising from associations with commercial or corporate interests in connection with the work submitted. None of the authors has a personal financial interest in this research. Address correspondence and reprint requests to Taishi Masamune, MD, Operating Theatre, Yamanashi University Hospital. 1110 Shimokato,Chuo,Yamanashi409-3898,Japan.Addresse-mailtoezn00202@ nifty.ne.jp or www.or.org. Copyright © 2009 International Anesthesia Research Society DOI: 10.1213/ane.0b013e3181a1a5ed

96

impaired by sedatives, including opioids,6 central ␣ agonists7,8, and nefopam.9 Typical doses of midazolam, a ␥-aminobutyric acid type A (GABAA) receptor agonist, slightly reduce the shivering threshold in humans. For example, midazolam at a steady-state plasma concentration of 0.3 ␮L/mL reduced the shivering threshold only 0.6°C in volunteers.10 This concentration corresponds to an extremely high dose, about 40 mg over the course of 4 h, or about 20 times a typical premedication dose. A subsequent study demonstrated that premedication doses of midazolam slightly reduce core temperature in volunteers.11 JM-1232(⫺) is a novel isoindoline derivative which shows sedative and hypnotic activities through the benzodiazepine site of GABAA receptors (GABAA-R), although it acts at a site distinct from midazolam. The drug is water soluble and seems to possess a wide therapeutic index.12 It also has a shorter elimination half-life than midazolam. We thus tested the hypothesis that JM-1232(⫺) has a dose-dependent effect on the shivering thresholds in rabbits.

METHODS With the approval of the Committee on Animal Research at the Faculty of Medicine, University of Vol. 109, No. 1, July 2009

Table 1. Principle Results

Shivered (#/n) MAP (mm Hg) Heart rate (beats/min) Respiratory rate (breaths/min) Arterial pH Paco2 (mm Hg) Pao2 (mm Hg) Shivering threshold (°C)

Control

JM-1232(⫺) 0.01 (mg 䡠 kg⫺1 䡠 mim⫺1)

JM-1232(⫺) 0.1 (mg 䡠 kg⫺1 䡠 mim⫺1)

6/6 97 ⫾ 14 253 ⫾ 13 39 ⫾ 6 7.48 ⫾ 0.07 23 ⫾ 2 537 ⫾ 77 36.5 ⫾ 0.3

5/6 97 ⫾ 15 216 ⫾ 36‡ 43 ⫾ 2 7.44 ⫾ 0.03 31 ⫾ 6‡ 503 ⫾ 53 35.7 ⫾ 0.8

0/6 69 ⫾ 21*† 190 ⫾ 22* 50 ⫾ 11 7.39 ⫾ 0.02 34 ⫾ 3* 548 ⫾ 83 —

Results for the control group and the rabbits given 0.01 mg . kg⫺1 . min⫺1 JM-1232(⫺) are at the shivering threshold; results for the rabbits given 0.1 mg . kg⫺1 . min⫺1 JM-1232(⫺) were recorded at a core temperature of 34°C. Data are reported as means ⫾ SD . MAP ⫽ mean arterial blood pressure. * P ⬍ 0.01 compared with control group. † P ⬍ 0.01 compared with 0.01 mg . kg⫺1 . min⫺1 JM-1232(⫺). ‡ P ⬍ 0.05 compared with control group.

Yamanashi, we studied 18 male Japanese white rabbits, weighing 2.9 –3.8 (mean, 3.4) kg. The daytime core temperature in these rabbits is usually approximately 39°C. Ambient temperature was maintained near 24°C throughout the study. To exclude the effects of circadian variation on body temperature, studies started about 10:00 and typically concluded near 15:00.

Protocol Animal preparation was performed as previously described.13–18 Briefly, the animals were anesthetized by inhalation of 3%–5% isoflurane and 67% nitrous oxide in oxygen. Each rabbit was intubated with a 3-mm endotracheal tube after local anesthesia by 8% lidocaine spray and subsequently allowed to breathe spontaneously. Nitrous oxide was discontinued and the end-tidal concentration of isoflurane adjusted to 0.2 minimum alveolar anesthetic concentration in 100% oxygen (NORMAC AA-102; GE Health care, Andover, MA). A catheter was inserted in the marginal ear vein and 3– 4 mL 䡠 kg⫺1 䡠 h⫺1 lactated Ringer’s solution was infused throughout the study. A catheter was inserted into a femoral artery. The animals were loosely restrained in an experimental chamber during the study. Between 15 and 30 min after the drug or saline infusion was started, core temperature was cooled at a rate of 2 to 3°C/h by perfusing water at 10°C through a U-shaped plastic tube (thermode) positioned in the colon. The animals were randomly assigned to 1 of 3 infusion regimens: 1) saline (control); 2) 0.01 mg 䡠 kg⫺1 䡠 min⫺1 JM-1232(⫺); or 3) 0.1 mg 䡠 kg⫺1 䡠 min⫺1 JM-1232(⫺). The study ended when shivering was detected or core temperature reached 34°C. At the end of the experiment, each animal was killed by KCl infusion. The concentration of JM-1232(⫺) was based on a preliminary study during which we observed that an infusion of 0.1 mg 䡠 kg⫺1 䡠 min⫺1 JM-1232(⫺) produced an approximately 20 point decrease in Bispectral Index (BIS) (from a lightly anesthetized control baseline value) and eliminated shivering even at a temperature of 34°C. We thus used 0.1 mg 䡠 kg⫺1 䡠 min⫺1 Vol. 109, No. 1, July 2009

JM-1232(⫺) as our higher dose, and one-tenth that amount as our lower dose.

Measurements Core temperatures were recorded from the distal esophagus (MGA 3–219; Nihon Kohden, Tokyo, Japan). Shivering was evaluated via inspection by an observer blinded to the treatment. Arterial blood was sampled for gas analysis (ABL 700TM; Radiometer, Tokyo, Japan) at the time of shivering or at the temperature of 34°C in each rabbit. Before induction of anesthesia, the animals’ heads were shaved. The purpose-built electroencephalogram (EEG) electrodes (BIS sensor XP small size; Aspect Medical Systems, Norwood, MA) were positioned according to the method of Martín-Cancho et al.19 The electrodes were connected to a BIS monitor (A-2000; Aspect Medical Systems, Norwood, MA).

Data Analysis A power analysis indicated that 6 rabbits per group were sufficient to detect a 0.8°C difference in esophageal temperature among groups with an ␣ ⫽ 0.05 and a power of 0.8 based on a sd of 0.4°C. We performed a ␹2 test to compare our data with theoretical values and found that our data were normally distributed. The core temperature triggering sustained vigorous shivering defined the shivering threshold. Results were analyzed with one-way analysis of variance and Student-Newman-Keuls tests. Data are expressed as means ⫾ sd; P ⬍ 0.05 was considered statistically significant.

RESULTS Responses in each group at the time of shivering or at the temperature of 34°C are shown in Table 1. Except for mean arterial blood pressure, heart rate and Paco2, hemodynamic and respiratory responses at the time of shivering or at a temperature of 34°C were not significantly different. All six rabbits given a saline infusion shivered at an average of 36.5 ⫾ 0.3°C. In contrast, 5 of the 6 rabbits © 2009 International Anesthesia Research Society

97

Figure 2. Steady-state Bispectral Index values before core Figure 1. Shivering thresholds in each group. Squares show shivering thresholds in individual rabbits; diamonds indicate that a minimum core temperature of 34°C was reached without triggering shivering. Circles with sd error bars indicate the mean shivering thresholds in each group. Shivering was induced easily in the control group and JM1232(⫺) 0.01 mg 䡠 kg⫺1 䡠 min⫺1 group, but was obliterated by JM-1232(⫺) 0.1 mg 䡠 kg⫺1 䡠 min⫺1 infusion. *P ⬍ 0.01 vs Control; †P ⬍ 0.05 vs 0.01 mg 䡠 kg⫺1 䡠 min⫺1 JM-1232(⫺) infusion group.

given 0.01 mg 䡠 kg⫺1 䡠 min⫺1 of JM-1232(⫺) shivered at 35.7 ⫾ 0.8°C; one rabbit in this group failed to shiver at a minimum core temperature of 34.0°C. In contrast, none of the rabbits shivered at a minimum core temperature of 34.0°C with the larger 0.1 mg 䡠 kg⫺1 䡠 min⫺1 dose of JM-1232(⫺). The shivering thresholds were significantly different among the groups (Fig. 1). Steady-state (lightly anesthetized) BIS values before core cooling are shown in Figure 2. BIS in the control groups was 75 ⫾ 6. Both doses of JM-1232(⫺) significantly reduced the BIS to approximately 50; however, there were no statistically significant nor clinically important differences in BIS between the two treatment groups, even though the doses differed by an order of magnitude.

DISCUSSION JM-1232(⫺) produces sedative and hypnotic activities through the benzodiazepine site of GABAA-R. It therefore seemed likely that JM-1232(⫺) would also somewhat decrease the shivering thresholds as midazolam does. Indeed, a 0.01 mg 䡠 kg⫺1 䡠 min⫺1 JM1232(⫺) infusion reduced the shivering threshold 0.8°C and a 0.1 mg 䡠 kg⫺1 䡠 min⫺1 JM-1232(⫺) infusion reduced the threshold ⬎2.5°C. Unlike midazolam, highdose JM-1232(⫺) thus markedly reduces the shivering threshold. 98

Shivering Threshold with JM-1232

cooling in each group at an end-tidal isoflurane concentration of 0.2 minimum alveolar anesthetic concentration (MAC). Circles show this in individual rabbits; diamonds with sd error bars indicate the means. This in the control groups was 75 ⫾ 5.9. In contrast, that in 0.01 mg 䡠 kg⫺1 䡠 min⫺1 (53 ⫾ 8.8) and 0.1 mg 䡠 kg⫺1 䡠 min⫺1 (48 ⫾ 9.5) JM-1232(⫺) infusion groups was significantly lower than that in control groups. *P ⬍ 0.01 vs Control.

We quantified the shivering threshold. However, shivering is only one important cold defense in humans; the other is arteriovenous shunt vasoconstriction. Shunt vasoconstriction is an effective defense that constrains metabolic heat to the thermal core, thus preserving core temperature. With the exception of meperidine20 and nefopam,9 the vasoconstriction and shivering thresholds are reduced in parallel by drugs that impair central thermoregulatory control. Typically, the vasoconstriction is about 1°C more than the shivering threshold,21 a relationship that is maintained at any given dose of various drugs.4,22 It is thus most likely that the reduction in vasoconstriction induced by JM-1232(⫺) parallels that observed for shivering, although this theory remains to be confirmed. Thermoregulatory defenses are generally salutary in that they help maintain core temperature homeostasis. However, mild hypothermia (i.e., 2–3°C reductions in tissue temperature) provides considerable protection against ischemia in numerous animal models.23,24 Hypothermia has also been proven beneficial during recovery from cardiac arrest25,26 and in asphyxiated neonates.27,28 It is easy to induce and maintain hypothermia in these populations, but it can be challenging in the context of acute myocardial infarction or stroke, which are potential (although yet to be proven) targets of therapeutic hypothermia. Some drugs or drug combinations have been shown to induce a degree of thermal tolerance,29,30 but typically at the risk of some complications. Investigators are ANESTHESIA & ANALGESIA

thus actively looking for other, less toxic ways to blunt thermoregulatory defenses. The higher dose of JM-1232(⫺) obliterated shivering, at least to our cutoff temperature of 34°C. It remains to be determined whether such high doses will be tolerated in humans without respiratory or other complications. But to the extent that high doses can be used safely, JM-1232(⫺) might facilitate induction of therapeutic hypothermia. Few studies have been reported evaluating the use of BIS assessments in veterinary practice, despite its extensive validation and use in humans.31 However, rabbits have been used to study the EEG effects of anesthetics, and EEG has been used to evaluate the depth of anesthesia when using injectable combinations in this species. Although the BIS algorithm of humans may not be applicable to that of rabbits, BIS values may at least roughly help determine anesthetic depth in rabbits.19,32 Furthermore, BIS values correlate with the Observers’ Assessment of Alertness/Sedation score in patients given midazolam for sedation.33 Nonetheless, hypnotic depth in our rabbits, as evaluated with BIS monitoring, should thus be considered only a rough estimate of anesthetic effect. Before core cooling, BIS values were approximately 75 in the rabbis given saline and 0.2 minimum alveolar concentration isoflurane. BIS significantly decreased approximately 25 points with each drug dose, but, despite 10 times the concentration of the drug, BIS levels between the 0.01 mg 䡠 kg⫺1 䡠 min⫺1 and 0.1 mg 䡠 kg⫺1 䡠 min⫺1 drug-infusion groups did not differ significantly. This apparent ceiling effect with respect to sedation differs markedly from the drug’s thermoregulatory effect, which showed a distinct dose response. Therefore, the reduction in the shivering threshold seems to be independent from the depth of sedation. JM-1232(⫺) remains in preclinical evaluation and is therefore unavailable for use in volunteers or patients. A consequent limitation of our study is that it was conducted in rabbits rather than humans. What the analogous doses in humans might be remain unknown. It is thus possible that the doses we tested, although spanning a 10-fold range, may not cover the range subsequently determined to be therapeutic in humans. In summary, a low dose of JM-1232(⫺) reduced the shivering threshold approximately 0.8°C, which is similar to the effects in humans given premedication doses of midazolam. In contrast, a dose 10-fold larger reduced the threshold more than 2.5°C. This is a substantial decrement and might facilitate induction of therapeutic hypothermia. ACKNOWLEDGMENTS We thank Professor and Chair Makoto Ozaki, MD, PhD, from the Department of Anesthesiology at Tokyo Women’s Medical University, Tokyo, Japan, for his valuable comments regarding our study. Vol. 109, No. 1, July 2009

REFERENCES 1. Kurz A, Sessler DI, Lenhardt RA, Study of wound infections and temperature group. Perioperative normothermia to reduce the incidence of surgical-wound infection and shorten hospitalization. N Engl J Med 1996;334:1209 –15 2. Frank SM, Fleisher LA, Breslow MJ, Higgins MS, Olson KF, Kelly S, Beattie C. Perioperative maintenance of normothermia reduces the incidence of morbid cardiac events: a randomized clinical trial. JAMA 1997;277:1127–34 3. Rajagopalan S, Mascha E, Na J, Sessler DI. The effects of mild perioperative hypothermia on blood loss and transfusion requirement: a meta-analysis. Anesthesiology 2008;108:71–7 4. Xiong J, Kurz A, Sessler DI Plattner O, Christensen R, Dechert M, Ikeda T. Isoflurane produces marked and non-linear decreases in the vasoconstriction and shivering thresholds. Anesthesiology 1996;85:240 –5 5. Matsukawa T, Sessler DI, Sessler AM, Schroeder M, Ozaki M, Kurz A, Cheng C. Heat flow and distribution during induction of general anesthesia. Anesthesiology 1995;82:662–73 6. Kurz A, Go JC, Sessler DI, Kaer K, Larson MD, Bjorksten AR. Alfentanil slightly increases the sweating threshold and markedly reduces the vasoconstriction and shivering thresholds. Anesthesiology 1995;83:293–9 7. Horn EP, Standl T, Sessler DI, von Knobelsdorff G, Bu¨chs C, Schulte am Esch J. Physostigmine prevents postanesthetic shivering as does meperidine or clonidine. Anesthesiology 1998;88:108 –13 8. Talke P, Tayefeh F, Sessler DI Jeffrey R, Noursalehi M, Richardson C. Dexmedetomidine does not alter the sweating threshold, but comparably and linearly reduces the vasoconstriction and shivering thresholds. Anesthesiology 1997;87:835– 41 9. Alfonsi P, Adam F, Passard A, Guignard B, Sessler DI, Chauvin M. Nefopam, a non-sedative benzoxazocine analgesic, selectively reduces the shivering threshold in unanesthetized subjects. Anesthesiology 2004;100:37– 43 10. Kurz A, Sessler DI, Annadata R Dechert M, Christensen R, Bjorksten R. Midazolam minimally impairs thermoregulatory control. Anesth Analg 1995;81:393– 8 11. Matsukawa T, Hanagata K, Ozaki M, Iwashita H, Koshimizu M, Kumazawa T. I.m. midazolam as premedication produces a concentration dependent decrease in core temperature in male volunteers. Br J Anaesth 1997;78:396 –9 12. Kanamitsu N, Osaki T, Itsuji Y Yoshimura M, Tsujimoto H, Soga M. Novel water-soluble sedative-hypnotic agents: isoindolin-1one derivatives. Chem Pharm Bull 2007;55:1682– 8 13. Matsukawa T, Kashimoto S, Kumazawa T Miyaji T, Hashimoto M, Iriki M. Effects of halothane and enflurane on the peripheral vasoconstriction and shivering induced by internal body cooling in rabbits. J Anesth 1994;8:311–15 14. Hanagata K, Matsukawa T, Sessler DI Miyaji T, Funayama T, Koshimizu M, Kashimoto S, Kumazawa T. Isoflurane and sevoflurane produce a dose-dependent reduction in the shivering threshold in rabbits. Anesth Analg 1995;81:581– 4 15. Iwashita H, Matsukawa T, Ozaki M Sessler DI, Imamura M, Kumazawa T. Hypoxemia decreases the shivering threshold in rabbits anesthetized with 0.2 MAC isoflurane. Anesth Analg 1998;87:1408 –11 16. Imamura M, Matsukawa T, Ozaki M Sessler DI, Nishiyama T, Okuyama K, Kumazawa T. Nitrous oxide decreases the shivering threshold less than isoflurane in rabbits. Br J Anaesth 2003;90:88 –90 17. Okuyama K, Matsukawa T, Ozaki M Sessler DI, Nishiyama T, Imamura M, Kumazawa T. Doxapram produces a dosedependent reduction in the shivering threshold in rabbits. Anesth Analg 2003;97:759 – 62 18. Imai Y, Yamakage M, Sato H Okuyama K, Ishiyama T, Matsukawa T. Isovolaemic haemodilution decreases the shivering threshold in rabbits. Eur J Anaesth 2008;25:450 –3 19. Martín-Cancho MF, Lima JR, Luis L Crisosto´mo V, CarrascoJimeńez MS, Usoń-Gargallo J. Relationship of bispectral index values, haemodynamic changes and recovery times during sevoflurane or propofol anaesthesia in rabbits. Lab Anim 2006;40:28 – 42 20. Kurz A, Ikeda T, Sessler DI, Larson MD, Bjorksten AR, Dechert M, Christensen R. Meperidine decreases the shivering threshold twice as much as the vasoconstriction threshold. Anesthesiology 1997;86:1046 –54 © 2009 International Anesthesia Research Society

99

21. Lopez M, Sessler DI, Walter K, Emerick T, Ozaki M. Rate and gender dependence of the sweating, vasoconstriction, and shivering thresholds in humans. Anesthesiology 1994;80:780 – 8 22. Matsukawa T, Kurz A, Sessler DI, Bjorksten AR, Merrifield B, Cheng C. Propofol linearly reduces the vasoconstriction and shivering thresholds. Anesthesiology 1995;82:1169 – 80 23. Logue ES, McMichael MJ, Callaway CW. Comparison of the effects of hypothermia at 33 degrees C or 35 degrees C after cardiac arrest in rats. Acad Emerg Med 2007;14:293–300 24. Ohta H, Terao Y, Shintani Y, Kiyota Y. Therapeutic time window of post-ischemic mild hypothermia and the gene expression associated with the neuroprotection in rat focal cerebral ischemia. Neurosci Res 2007;57:424 –33 25. Bernard SA, Gray TW, Buist MD, Jones BM, Silvester W, Gutteridge G, Smith K. Treatment of comatose survivors of out-of-hospital cardiac arrest with induced hypothermia. N Engl J Med 2002;346:557– 63 26. Hypothermia after cardiac arrest study group. Mild therapeutic hypothermia to improve the neurologic outcome after cardiac arrest. N Engl J Med 2002;346:549 –56 27. Shankaran S, Laptook AR, Ehrenkranz RA, Tyson JE, McDonald SA, Donovan EF, Fanaroff AA, Poole WK, Wright LL, Higgins RD, Finer NN, Carlo WA, Duara S, Oh W, Cotton CM, Stevenson DK, Stoll BJ, Lemons JA, Guillet R, Jobe AH, National institute of child health and human development neonatal research network. Wholebody hypothermia for neonates with hypoxic-ischemic encephalopathy. N Engl J Med 2005;353:1574–84

100

Shivering Threshold with JM-1232

28. Gluckman PD, Wyatt JS, Azzopardi D, Edwards AD, Ferriero DM, Polin RA, Robertson CM, Thoresen M, Whitelaw A, Gunn AJ. Selective head cooling with mild systemic hypothermia after neonatal encephalopathy: multicentre randomised trial. Lancet 2005;365:663–70 29. Mokhtarani M, Mahgob AN, Morioka N, Doufas AG, Dae M, Shaughnessy TE, Bjorksten AR, Sessler DI. Buspirone and meperidine synergistically reduce the shivering threshold. Anesth Analg 2001;93:1233–9 30. Doufas AG, Lin CM, Suleman MI, Liem EB, Lenhardt R, Morioka N, Akca O, Shah YM, Bjorksten AR, Sessler DI. Dexmedetomidine and meperidine additively reduce the shivering threshold in humans. Stroke 2003;34:1218 –23 31. Myles PS, Leslie K, McNeil J, Forbes A, Chan MT. Bispectral index monitoring to prevent awareness during anaesthesia: the B-Aware randomised controlled trial. Lancet 2004;363:1757– 63 32. Shibuya K, Ishiyama T, Ichikawa M, Sato H, Okuyama K, Sessler DI, Matsukawa T. The direct effects of propofol on pial microvessels in rabbits. J Neurosurg Anesthesiol 2009;21:40 – 6 33. Liu J, Singh H, White PF. Electroencephalogram bispectral analysis predicts the depth of midazolam-induced sedation. Anesthesiology 1996;84:64 –9


Helium Breathing Provides Modest Antiinflammatory, but No Endothelial Protection Against Ischemia-Reperfusion Injury in Humans In Vivo Eliana Lucchinetti, PhD* Johannes Wacker, MD† Christian Maurer, MD† Marius Keel, MD‡ Luc Ha¨rter, PhD‡ Kathrin Zaugg, MD, PhD§ Michael Zaugg, MD储

BACKGROUND: The noble gas helium is devoid of anesthetic effects, and it elicits cardiac preconditioning. We hypothesized that inhalation of helium provides protection against postocclusive endothelial dysfunction after ischemia-reperfusion of the forearm in humans. METHODS: Eight healthy male subjects were enrolled in this study with a crossover design. Each volunteer was randomly exposed to 15 min of forearm ischemia in the presence or absence of helium inhalation. Helium was inhaled at an end-tidal concentration of 50 vol% from 15 min before ischemia until 5 min after the onset of reperfusion (“helium conditioning”). Hyperemic reaction, a marker of nitric oxide bioavailability and endothelial function, was determined at 15 and 30 min of reperfusion on the forearm using venous occlusion plethysmography. Expression of the proinflammatory markers CD11b, ICAM-1, PSGL-1, and L-selectin (CD62L) on leukocytes and P-selectin (CD62P), PSGL-1, and CD42b on platelets were measured by flow cytometry during reperfusion. RESULTS: Ischemia-reperfusion consistently reduced the postocclusive endotheliumdependent hyperemic reaction at 15 and 30 min of reperfusion. Periischemic inhalation of helium at 50 vol% did not improve postocclusive hyperemic reaction. Helium decreased expression of the proinflammatory marker CD11b and ICAM-1 on leukocytes and attenuated the expression of the procoagulant markers CD42b and PSGL-1 on platelets. CONCLUSIONS: Although inhalation of helium diminished the postischemic inflammatory reaction, our data indicate that human endothelium, which is a component of all vital organs, is not amenable to protection by helium at 50 vol% in vivo. This is in contrast to sevoflurane, which protects human endothelium at low subanesthetic concentrations. (Anesth Analg 2009;109:101–8)

E

ndothelium-mediated vasodilator response is markedly diminished by mental and physical stress, pain, diabetes, hypercholesterolemia, and hypertension.1–3

From the *Department of Anesthesiology and Pain Medicine, University of Alberta, Edmonton, Canada; †Institute of Anesthesiology, ‡Department of Trauma Surgery, §Department of Radiation Oncology, University Hospital Zurich, Zurich, Switzerland; and 储Department of Anesthesiology and Pain Medicine, University of Alberta and Perioperative Translational Medicine, Mazankowski Alberta Heart Institute, Edmonton, Canada. Accepted for publication January 7, 2009. Supported by the Grant #3200B0-103980/1 of the Swiss National Science Foundation, Berne, Switzerland, the Swiss Society of Anesthesiology and Reanimation, Berne, Switzerland, the 5th Frontiers in Anesthesia Research Award from the International Anesthesia Research Society, Cleveland, OH, and a grant from the Mazankowski Alberta Heart Institute, Edmonton, Canada. EL, JW, KZ, and MZ contributed similarly to this work. Please see supplementary material available at www. anesthesia-analgesia.org. Reprints will not be available from the author. Address correspondence to Michael Zaugg, MD, DEAA, FRCPC, Department of Anesthesiology and Pain Medicine, University of Alberta, CSB Room 8-120, Edmonton AB T6G 2G3. Address e-mail to [email protected]. Copyright © 2009 International Anesthesia Research Society DOI: 10.1213/ane.0b013e3181a27e4b

Vol. 109, No. 1, July 2009

The same conditions are also associated with an increased incidence of perioperative cardiovascular complications.4,5 This is not surprising as endothelial dysfunction during ischemia-reperfusion exacerbates vasospasm and thus is a critical determinant of the extent of ischemic organ injury.6,7 Because the endothelium is a key component of all vital organs, it can be speculated that interventions aiming at improving endothelial function potentially protect the whole body. Using a human forearm model of simulated ischemiareperfusion injury, we recently demonstrated that the halogenated ether sevoflurane administered at low subanesthetic but sedative doses (ⱕ1 vol%) abolished postocclusive endothelial dysfunction and activation of inflammatory white blood cells.8 Small-dose sevoflurane inhalation was further capable of inhibiting agonist-induced formation of thrombogenic granulocyte-platelet aggregates9 and of decreasing the expression of the proinflammatory L-selectin 24 h after sevoflurane inhalation in volunteers,10 which is consistent with the occurrence of a “second window of protection” in humans. An increasing body of evidence now suggests that the noble gas helium, which 101

Figure 1. Study protocols. The forearm was rendered hypoxic for 15 min. Hyperemic blood flow response was measured using venous occlusion plethysmography after 15 and 30 min of reperfusion to assess endothelial function. Blood samples for flow cytometry were collected after 5, 10, and 30 min of reperfusion. CTL ⫽ control protocol.

is devoid of any sedative effects, may provide similar and efficient protection in the heart by activating signaling pathways previously described for halogenated ethers.11–13 Experimental data imply that helium concentrations well below 50 vol% are cardioprotective. However, it is unclear whether helium specifically protects the endothelium in humans similar to halogenated ethers8 or, in other words, whether endothelial protection is a component of heliuminduced organ protection. Therefore, using our previously established model,8 we tested in volunteers whether helium could prevent endothelial dysfunction in vivo and, if so, whether this protection would be related to a reduced expression of proinflammatory markers on leukocytes and platelets. Specifically, we hypothesized that periischemic helium inhalation, i.e., “helium conditioning,” would improve postocclusive endothelial dysfunction by inhibiting the inflammatory response on white blood cells.

METHODS Study Subjects This study was performed in accordance with the Declaration of Helsinki (2000) and was approved by the local ethics committee. Eight healthy male volunteers (35 ⫾ 7 (25– 45) yr) with a Body Mass Index of 21 ⫾ 2 kg/m2 and normal hematological variables (hemoglobin 14.5 ⫾ 0.5 g/dL, leukocytes 5.54 ⫾ 0.82 ⫻ 103 per ␮L, platelets 245 ⫾ 34 ⫻ 103 per ␮L) gave informed signed consent. All subjects were nonsmokers and refrained from caffeine and dark chocolate 24 h before study participation. 102

Helium and Human Endothelium

Study Protocol Ischemia-reperfusion of the forearm was used as a model of endothelial dysfunction.7,8,14 Figure 1 depicts the time course of the experiments. An IV cubital line was placed on the nondominant arm, and 100 mL Ringer’s solution was administered. Studies were performed in a temperature-controlled and quiet room (24°C) during the early morning hours. The nondominant forearm was rendered ischemic for 15 min by inflating a 12-cm-wide blood pressure cuff placed around the upper arm to 180 mm Hg to induce endothelial dysfunction (index ischemia). In all protocols, the spontaneously breathing volunteers used a tightly fitting cushioned face mask connected to the common gas outlet of an anesthesia machine (Siemens Servo 900D ventilator; Siemens Life Support Systems, Sona, Sweden). To overcome the internal resistance of the ventilator and connected tubing and to obtain the most convenient conditions for each volunteer, a pressure support of 2– 4 mbar was used and the inspiration trigger was set to 0 –3 mbar. At the beginning and at the end of the protocols, all subjects inhaled nitrogen in 50 vol% oxygen (Fig. 1). Helium (Messer Schweiz AG, Lenzburg, Switzerland) or nitrogen in 50 vol% oxygen was inhaled from 15 min before forearm ischemia until 5 min after reperfusion. Blood samples were taken from the cubital vein of the nondominant arm before test ischemia (baseline) and after 5, 10, and 30 min of reperfusion. Monitoring consisted of intermittent noninvasive arterial blood pressure measurements, 5-lead electrocardiogram, end-tidal CO2, ANESTHESIA & ANALGESIA

O2, and N2 concentrations (Infinity Delta XL, Draeger Medical Systems, Danvers, MA). Assessment of Resistance Vessel Endothelial Function Using Hyperemic Blood Flow Response in the Forearm We have previously described this procedure in detail.8 Mercury-in-silastic strain-gauge venous occlusion plethysmography (Vasoquant 4000; ELCAT GmbH, Wolfratshausen, Germany), which is considered to be the “gold standard” for the early detection of endothelial dysfunction, was used to measure the hyperemic blood flow response in the forearm.15–18 Briefly, volunteers were placed in a supine position with both arms extended and elevated. Venous congestion was achieved by inflating a cuff around the upper arm to 40 mm Hg. The recorded period of blood flow consisted of four cycles of venous occlusion followed by deflation (each 5 s) while the hands were excluded from the circulation. Reactive hyperemia was induced by 4 min of blood flow arrest using a blood pressure cuff inflated 20 mm Hg above the systolic blood pressure. Early hyperemic reaction (EHR, peak flow at the onset of reperfusion representing the first measurement after deflation of the cuff) and late hyperemic reaction (LHR, maintenance of hyperemia representing the mean of the three subsequent measurements) were recorded. Reactive hyperemia was determined at baseline before index ischemia on the dominant nonexperimental forearm to avoid ischemic preconditioning before the experiment, 15 and 30 min after test ischemia on both forearms. In the helium protocol, reactive hyperemia was also determined 5 min after initiation of helium inhalation to test whether helium itself without prior index ischemia would change the hyperemic blood flow response. Determination of Proinflammatory (CD11b, ICAM-1, CD62L, and PSGL-1) and Procoagulant (CD42b and CD62P) Markers Using Flow Cytometry Heparinized blood samples were immediately processed for flow cytometry. The expression of the proinflammatory molecules was determined at baseline, 5, 10, and 30 min during reperfusion in all eight volunteers.8,9 Three microliters of the primary fluorochrome-labeled antibody was added to 50 ␮L of blood in endotoxin-free tubes and incubated in the dark for 10 min at room temperature. Lysis buffer (450 ␮L) (Becton Dickinson, Basel Switzerland) was added and incubated for an additional 20 min at room temperature. The lysates were fixed for 30 min in 0.5 mL 0.2% paraformaldehyde solution at room temperature. The samples were centrifuged, and the cell pellets were suspended in 0.5 mL of TLR buffer and stored at 4°C in the dark. The FACSCalibur (Becton Dickinson, Basel Switzerland) flow cytometer was used to measure R-phycoerythrin (PE)-fluorescence at 580 nm and FITC-fluorescence at 515 nm. White Vol. 109, No. 1, July 2009

blood cells and platelets were distinguished from each other by typical physical characteristics, resulting in well-delineated cellular subpopulations that are easily identified on forward and side-scatter plots. Monoclonal antibodies for polymorphnuclear granulocytes (CD15, PE-labeled, clone 80H5, Immunotech, Marseille France), monocytes (CD14, FITC-labeled, clone 61D3, eBioscience, Wembley, UK), and platelets (CD62P, PE-labeled, clone AK-4, Santa Cruz Biotechnology, CA) further served for identification of cellular subgroups. The following additional surface markers were used: CD62L (FITC-labeled, clone AN51, DAKO, Glostrup Denmark), CD11b (PE-labeled, clone 2LPM19C, DAKO, Glostrup Denmark), ICAM-1 (FITC-labeled, clone sc-107, Santa Cruz Biotechnology, CA), PSGL-1 (FITC-labeled, clone sc-32302, Santa Cruz Biotechnology, CA), and CD42b (GPIIb, clone AN51, FITC-labeled, DAKO, Glostrup Denmark). Results were compared with isotype-matched antibodies staining as controls (PE-labeled IgG, eBioscience, Wembley, UK and FITC-labeled IgG, Becton Dickinson, Basel, Switzerland). A minimum of 20,000 events was counted on each sample. Data are shown as fold-change of mean fluorescence intensities (MFIs) from the respective baseline values.

Statistical Analysis Forearm blood flow was measured in milliliters per 100 g tissue per minute and expressed as percent change compared with baseline flow measurements. The sample size was calculated based on published data of endothelial protection by sevoflurane in humans, as measured by venous occlusion plethysmography.8 With an expected difference of 50% between group means (percent change from baseline), i.e., between the control protocol without helium inhalation and the protocol with helium inhalation (Heliox), and a standard deviation of 30% in each group, an ␣ ⫽ 0.05, and a ␤ ⫽ 0.80, a sample size of eight volunteers (control and treatment arm) was necessary in this study with a crossover design. For each participating subject, activation of leukocytes, and platelets on the ischemic limb was expressed as the ratio of MFIs between time-matched control and helium blood samples of the investigated surface markers, as follows: MFISi HELIOX(tj) , ratio(subject Si; timepoint tj) ⫽ MFISi CTL(tj) where i ⫽ 1, 2, . . . 8 and j ⫽ baseline, 5 min, 10 min, 30 min Data were subsequently normalized to baseline.8 Paired t-test and repeated-measures analysis of variance were used for comparison. P ⬍ 0.05 was considered significant. Data are given as mean ⫾ sd after testing for normality. Analyses were performed using SigmaStat Version 2 (SPSS Science, Chicago, IL). © 2009 International Anesthesia Research Society

103

Table 1. Recorded Variables During Inhalation of Gas Mixtures Subjects Before Helium SBP (mm Hg) HR (bpm) Spo2 (%) During Helium SBP (mm Hg) HR (bpm) Spo2 (%) ETo2 (vol%) ETco2 (vol%)

V_1

V_2

V_3

V_4

V_5

V_6

V_7

V_8

133 73 100

110 60 99

114 56 100

112 60 99

126 56 100

128 58 100

116 70 100

120 73 100

130 66 100 42 5

111 65 100 44 4.8

110 56 99 45 4.4

105 52 100 47 4.5

130 55 100 43 4.2

132 58 100 45 5.2

117 60 99 46 3.8

104 70 100 45 3.7

SBP ⫽ systolic blood pressure at baseline and 10 min after initiation of helium (50 vol%) inhalation; HR ⫽ mean heart rate; ETO2⫽ mean end-tidal O2 concentration; ETCO2 ⫽ mean end-tidal CO2; SpO2 ⫽ mean oxygen saturation (finger tip); bpm ⫽ beats per minute; V1– 8 ⫽ individual volunteers.

Figure 2. Late hyperemic blood flow response (LHR) in the ischemic arm. The percent change in blood flow is shown for each subject at 15 and 30 min after test ischemia. Mean values and sds are depicted next to the individual measurements. There was a significant reduction in LHR after ischemia-reperfusion injury versus baseline (*P ⬍ 0.05). However, helium breathing did not improve postocclusive hyperemic reaction. CTL ⫽ control protocol; V1– 8 ⫽ individual volunteers.

RESULTS All study subjects tolerated the procedures well without complications. The 35 min period of helium inhalation at 50 vol% had no effect on arterial blood pressure, heart rate, and end-tidal CO2 concentrations (Table 1). There was no sedation in any of the participants, nor was there any analgesic effect on the discomfort associated with the 15 min of ischemiareperfusion of the forearm when helium was inhaled compared with nitrogen in the control protocol. Despite equal ventilation settings in the control and helium protocols for each of the eight participants, seven of them reported a subjective reduction in the resistive work of breathing during helium inhalation.

Ischemia-Reperfusion in the Forearm Markedly Reduces Endothelium-Dependent Hyperemic Blood Flow Response in Humans Fifteen minutes of ischemia followed by reperfusion was used to induce endothelial dysfunction. Hyperemic blood flow response was measured by venous occlusion plethysmography after 15 and 30 min of reperfusion on both the ischemic and the nonischemic side. Although peak flow at reopening of the forearm vessels (EHR) was unaffected by test ischemia when compared with baseline (at 15 min of reperfusion: ⫺6% ⫾ 23%, at 30 min of reperfusion: ⫺12% ⫾ 15%), maintenance of hyperemic reaction 104


(LHR) was markedly (P ⬍ 0.05) reduced versus baseline (at 15 min of reperfusion: ⫺27% ⫾ 15%, at 30 min of reperfusion: ⫺29% ⫾ 12%) (Fig. 2). Hyperemic blood flow response was unaffected by index ischemia on the nonischemic control arm, i.e., there was no induction of “remote” effects on the neighboring endothelium after ischemia-reperfusion injury (data not shown). Original tissue flow measurements are available in Supplementary Table S1, available at www. anesthesia-analgesia.org.

Periischemic Breathing of Helium does not Restore Postocclusive Endothelial Dysfunction of the Forearm in Humans Each volunteer underwent the same procedure with and without helium inhalation in a random fashion. Helium was inhaled from 15 min before ischemia until 5 min after reperfusion, mimicking a combination of pre- and postconditioning (“conditioning”). There was no postocclusive improvement of LHR (Fig. 2). Interestingly, helium itself did not affect EHR versus baseline (3% ⫾ 6%) or LHR versus baseline (⫺1% ⫾ 11%), nor did it affect EHR on the nonischemic limb versus control after index ischemia (at 15 min of reperfusion with helium 1% ⫾ 8% vs 5% ⫾ 11% in control, at 30 min of reperfusion with helium ⫺4% ⫾ 12% vs 5% ⫾ 3% in control) or LHR on the nonischemic arm after index ischemia (at 15 min of reperfusion with helium 4% ⫾ 15% vs 2% ⫾ 17% in ANESTHESIA & ANALGESIA

CD62P (P-selectin) expression were observed, reductions in PSGL-1 and CD42b expressions on platelets were detected during reperfusion (Fig. 4). Because this study used the identical experimental protocols to assess endothelial function, as previously reported for sevoflurane by our group,8 we directly compared the endothelial protection by helium (50 vol%) with sevoflurane (ⱕ1 vol%). The comparison shows that helium can only provide modest antiinflammatory actions against ischemia-reperfusion injury. However, it is incapable of improving postocclusive endothelial dysfunction in humans (Fig. 5).

DISCUSSION

Figure 3. Proinflammatory markers on leukocytes. CD11b, CD62L (L-selectin), ICAM-1, and PSGL-1 expressions were used to assess leukocyte activation after ischemiareperfusion injury in the ischemic arm. There was only a modest decrease in CD11b and ICAM-1 expression with helium inhalation. Data are mean ⫾ sd. CTL ⫽ control protocol; PMNs ⫽ polymorphnuclear granulocytes; MFI ⫽ mean fluorescence intensity; HELIOX ⫽ 50 vol% helium/50 vol% oxygen. *P ⬍ 0.05 versus baseline. control, at 30 min of reperfusion with helium 0% ⫾ 23% vs ⫺1% ⫾ 21% in control). These data further exemplify the inability of helium to directly affect endothelial function in humans.

Helium Breathing only Modestly Reduces the Proinflammatory and Procoagulant Cell Surface Markers After Ischemia-Reperfusion Injury To test whether helium affects the expression of the proinflammatory markers on leukocytes and the procoagulant markers on platelets, blood was collected during reperfusion from the injured side and compared with baseline conditions. Baseline flow cytometry data are available in Supplementary Table S2, available at www.anesthesia-analgesia.org. The markers were expressed as the ratio of MFI between time-matched control and helium blood samples. Helium reduced the ratio of CD11b expression on monocytes and ICAM-1 on granulocytes and monocytes during reperfusion (Fig. 3). Although no changes in Vol. 109, No. 1, July 2009

The present study shows the following salient findings. First, in contrast to our hypothesis, helium inhalation in humans elicited no postocclusive improvement of endothelial dysfunction after ischemiareperfusion injury on the forearm. Moreover, and unlike previous observations with sevoflurane,8 helium did not affect vasomotion on the nonischemic limb, indicating no remote systemic “preconditioninglike” effects on the vasculature. Second, although helium modestly attenuated the postischemic expression of inflammatory cell surface markers on leukocytes and platelets, it was unable to restore endothelial function at the used concentration (50 vol%). Nonetheless, the observed antiinflammatory and anticoagulant actions elicited by helium may be, at least in part, responsible for its previously reported organ protective effects.12,13,19 Venous occlusion plethysmography is a reliable noninvasive tool to investigate vascular function and is regarded to be the gold standard for evaluating endothelial function.16,17 In the present study, we determined reactive hyperemia, which is mediated by the endothelial release of nitric oxide,20 using venous occlusion plethysmography after prolonged test ischemia of the forearm. With the aid of this well-established model,1,7,8 we previously showed that hyperemic blood flow remains unchanged after nitroglycerin application consistent with an endothelium-dependent, but smooth muscle-independent, vasomotion effect.8 These studies further demonstrated that late, as opposed to early hyperemic blood flow response, more reliably reflects endothelial function and vascular integrity. The findings of the present study show that helium, as opposed to sevoflurane, is unable to prevent postocclusive endothelial dysfunction. Helium is a biologically inert noble gas without anesthetic properties.21,22 Its application as an oxygen carrier in the clinical setting is largely based on its low density (⫺250% if nitrogen in air is replaced by helium) and rapid diffusion due to its low atomic weight, as compared with nitrogen, which markedly reduces the resistive work of breathing.23 Improved CO2 elimination and reduced oxygen requirements combined with antiinflammatory actions lead to an © 2009 International Anesthesia Research Society

105

Figure 4. Procoagulant markers on platelets. CD42b, CD62P (P-selectin), and PSGL-1 expressions were used to assess platelet activation after ischemia-reperfusion injury in the ischemic arm. There was a significant decrease in CD42b and PSGL-1 expression with helium inhalation. Data are mean ⫾ sd. CTL ⫽ control protocol; MFI ⫽ mean fluorescence intensity; HELIOX ⫽ 50 vol% helium/50 vol% oxygen. *P ⬍ 0.05 versus baseline.

Figure 5. Effects of helium and sevoflurane inhalation on endothelial protection. Using the identical model of ischemia-reperfusion injury on the forearm of volunteers in a crossover study, sevoflurane (⬍1 vol%) but not helium (50 vol%) prevented postocclusive endothelial dysfunction. Panel A: Late hyperemic reaction (LHR). Mean values and standard deviations are depicted next to the individual measurements. n ⫽ 5 for sevoflurane (SEVO), n ⫽ 8 for HELIOX (50 vol% helium/50 vol% oxygen). Sevoflurane data are reproduced from Ref. 8 Figure 2B. Panel B: CD11b expression on polymorphnuclear granulocytes (PMNs) and monocytes. *P ⬍ 0.05 versus baseline. #P ⬍ 0.05 versus time-matched HELIOX sample. Sevoflurane data are reproduced from Lucchinetti et al.8 with permission from Wolters Kluwer Health, Figure 5A and 5B. Data are mean ⫾ sd.

increased efficiency of ventilation and suggest a potential lung-protective strategy.19 Recently, the nonsedative helium also emerged as an attractive preconditioning-mimicking agent in other vital organs. Using helium inhalation in an in vivo rat model of focal brain ischemia, Pan et al.12 reported a reduction in infarct size, as assessed by triphenyltetrazolium staining, from 36% to 4% after a middle cerebral artery occlusion for 2 h and a reperfusion period of 1 h. Similarly, using an in vivo rabbit model of coronary artery ligation, Pagel et al.13 reported an infarct size reduction by 50% (from 45% to 23%) if three cycles of helium breathing were administered 106


before a lethal ischemia of 3 h. Blocker experiments from these studies suggest that helium activates prosurvival kinases and mitochondrial KATP channels, ultimately inhibiting mitochondrial permeability transition pore opening,24 as previously reported for halogenated ethers.25,26 In other experimental studies comparing helium versus carbon dioxide pneumoperitoneum, helium ameliorated pneumoperitoneum-associated renal dysfunction27 and reduced tumor recurrence and spreading.28,29 Interestingly, in our study helium breathing predominantly decreased ICAM-1 expression on leukocytes, a cellular surface marker that plays a pivotal ANESTHESIA & ANALGESIA

role in leukocyte-endothelium interaction.30 Also, the adhesion promoting PSGL-1, mainly expressed on leukocytes but also on activated platelets,31 and to a lesser degree the platelet receptor for the von Willebrand factor (CD42b), were diminished by periischemic helium inhalation. In contrast to these promising reports and despite the fact that vascular endothelium and leukocyte- and plateletendothelial interactions are critically involved in many steps of tissue damage originating from ischemia-reperfusion injury, our human in vivo study could only show modest helium-induced antiinflammatory effects, which were not capable of preventing ischemia-induced endothelial dysfunction. With respect to our previous study on endothelial protection by sevoflurane,8 it could be speculated that either sevoflurane-induced endothelial preconditioning and antiinflammatory actions are mediated by separate pathways (of which helium also uses the latter), or the mild antiinflammatory actions of helium are mediated by pathways distinct from those of sevoflurane and apparently are unrelated to endothelial protection by preconditioning. Nevertheless, these beneficial actions of helium could still contribute to the previously reported organ protection12,13,19,27 elicited by helium and potentially provide synergism in protection if helium was combined with halogenated ethers. In fact, helium as a carrier gas for halogenated ethers32,33 and oxygen may bring along several substantial advantages. First, the endothelial protective but yet sedative dose of approximately 1/2 minimum alveolar concentration8 required for sevofluraneinduced (endothelial) protection could be markedly reduced, although still maintaining the same degree of protection. Second, because of the optimized respiratory mechanics and the specific gas properties of helium, airway patency could be guaranteed more safely. Third, denitrogenation by itself might protect mitochondria from ischemia-reperfusion injury, because swelling of these organelles from nitrogen is thought to be a possible cause of reperfusion damage.34 Hence, these obvious advantages make the implementation of helium into a protective inhalation treatment strategy highly attractive in nonsedated at-risk patients undergoing coronary or vascular interventions. However, helium bears a certain risk of hypothermia (fivefold higher conductivity than air), diffusion hypoxia (33-fold lower solubility than nitrous oxide), tension pneumothorax and bubble formation, as observed in decompression sickness. Specific Remarks and Study Limitations: To overcome interindividual variability, we have chosen a crossover study design, and confounding variables were carefully excluded. As observed in our own data, some long-term (days to weeks) reproducibility studies evaluating venous occlusion plethysmography over time showed relatively high coefficients of variation (20%– 40%), mainly because of the different placement of the strain-gauge and venous occlusion Vol. 109, No. 1, July 2009

cuff on the forearm.17 However, short-term (hours) reproducibility of venous occlusion plethysmography, particularly without manipulating the strain-gauge and blood pressure cuff, is excellent (5%–10%).17 Also, marked changes in absolute values of basal forearm blood flow in an individual may occur as a result of changes in sympathetic tone.15 Although these changes may affect absolute responses to interventions, they do not alter the percent response to interventions calculated from the ratio of blood flow. Thus, the same intervention in a single subject recorded on different days gives similar results. Conversely, observed differences in the ratio of blood flow of the same individual in response to a specific intervention during two protocols recorded on different days are likely to be true treatment effects. In our study, the postischemic tissue perfusion measurements used for comparison with baseline perfusion were recorded within 1 h, and the percent changes were subsequently compared between the control and the helium protocol. Hence, it is unlikely that long-term variations in tissue perfusion may have biased our results. There was a small improvement in LHR on an average at 15 min of reperfusion after helium inhalation, and four of eight volunteers showed an increase in LHR after helium inhalation. If these small changes were consistently detectable and clinically relevant, our study would be under-powered. However, despite helium inhalation, tissue perfusion remained below baseline conditions after ischemia-reperfusion injury in three of four volunteers, who showed small improvements, indicating the failure of this therapy to restore endothelial function and to normalize postischemic tissue perfusion. In the present study, we used helium breathing at 50 vol%. Therefore, we cannot completely exclude that helium inhalation at 70 vol% or higher concentrations might provide endothelial protection. However, gas mixtures with low oxygen content are unlikely to be acceptable by clinicians, particularly in at-risk patients suffering from ischemia-reperfusion injury. Because our ischemia forearm model does not induce necrosis of endothelial cells, but rather a state of “endothelial stunning,” which resolves spontaneously within 60 min,15,18 it is unlikely that more substantial endothelial injury, which usually occurs in clinical situations, might be protected by helium. In conclusion, helium inhalation at 50 vol% is unable to specifically provide protection to the endothelium in humans. However, future studies inpatients should evaluate the potential synergism in vital organ protection between helium and the highly protective halogenated ethers. ACKNOWLEDGMENTS The authors thank the PACU nurses, colleagues, and volunteers, who participated in this study. © 2009 International Anesthesia Research Society

107

REFERENCES 1. Ghiadoni L, Donald AE, Cropley M, Mullen MJ, Oakley G, Taylor M, O’Connor G, Betteridge J, Klein N, Steptoe A, Deanfield JE. Mental stress induces transient endothelial dysfunction in humans. Circulation 2000;102:2473– 8 2. Jambrik Z, Santarcangelo EL, Rudisch T, Varga A, Forster T, Carli G. Modulation of pain-induced endothelial dysfunction by hypnotisability. Pain 2005;116:181– 6 3. Sanada H, Higashi Y, Goto C, Chayama K, Yoshizumi M, Sueda T. Vascular function in patients with lower extremity peripheral arterial disease: a comparison of functions in upper and lower extremities. Atherosclerosis 2005;178:179 – 85 4. Zaugg M, Schaub MC, Foex P. Myocardial injury and its prevention in the perioperative setting. Br J Anaesth 2004; 93:21–33 5. Zaugg M, Schulz C, Wacker J, Schaub MC. Sympathomodulatory therapies in perioperative medicine. Br J Anaesth 2004;93:53– 62 6. Hearse DJ. Myocardial protection during ischemia and reperfusion. Mol Cell Biochem 1998;186:177– 84 7. Kharbanda RK, Peters M, Walton B, Kattenhorn M, Mullen M, Klein N, Vallance P, Deanfield J, MacAllister R. Ischemic preconditioning prevents endothelial injury and systemic neutrophil activation during ischemia-reperfusion in humans in vivo. Circulation 2001;103:1624 –30 8. Lucchinetti E, Ambrosio S, Aguirre J, Herrmann P, Harter L, Keel M, Meier T, Zaugg M. Sevoflurane inhalation at sedative concentrations provides endothelial protection against ischemia-reperfusion injury in humans. Anesthesiology 2007;106:262– 8 9. Wacker J, Lucchinetti E, Jamnicki M, Aguirre J, Ha¨rter L, Keel M, Zaugg M. Delayed inhibition of agonist-induced granulocyteplatelet aggregation after low-dose sevoflurane inhalation in humans. Anesth Analg 2008;106:1749 –58 10. Lucchinetti E, Aguirre J, Feng J, Zhu M, Suter M, Spahn DR, Harter L, Zaugg M. Molecular evidence of late preconditioning after sevoflurane inhalation in healthy volunteers. Anesth Analg 2007;105:629 – 40 11. Zaugg M, Lucchinetti E, Uecker M, Pasch T, Schaub MC. Anaesthetics and cardiac preconditioning. Part I. Signalling and cytoprotective mechanisms. Br J Anaesth 2003;91:551– 65 12. Pan Y, Zhang H, VanDeripe DR, Cruz-Flores S, Panneton WM. Heliox and oxygen reduce infarct volume in a rat model of focal ischemia. Exp Neurol 2007;205:587–90 13. Pagel PS, Krolikowski JG, Shim YH, Venkatapuram S, Kersten JR, Weihrauch D, Warltier DC, Pratt PF Jr. Noble gases without anesthetic properties protect myocardium against infarction by activating prosurvival signaling kinases and inhibiting mitochondrial permeability transition in vivo. Anesth Analg 2007;105:562–9 14. Broadhead MW, Kharbanda RK, Peters MJ, MacAllister RJ. KATP channel activation induces ischemic preconditioning of the endothelium in humans in vivo. Circulation 2004;110:2077– 82 15. Benjamin N, Calver A, Collier J, Robinson B, Vallance P, Webb D. Measuring forearm blood flow and interpreting the responses to drugs and mediators. Hypertension 1995;25:918 –23 16. Higashi Y, Sasaki S, Nakagawa K, Matsuura H, Kajiyama G, Oshima T. A noninvasive measurement of reactive hyperemia that can be used to assess resistance artery endothelial function in humans. Am J Cardiol 2001;87:121–5, A9 17. Thijssen DH, Bleeker MW, Smits P, Hopman MT. Reproducibility of blood flow and post-occlusive reactive hyperaemia as measured by venous occlusion plethysmography. Clin Sci (Lond) 2005;108:151–7 18. Tousoulis D, Antoniades C, Stefanadis C. Evaluating endothelial function in humans: a guide to invasive and non-invasive techniques. Heart 2005;91:553– 8

108


19. Nawab US, Touch SM, Irwin-Sherman T, Blackson TJ, Greenspan JS, Zhu G, Shaffer TH, Wolfson MR. Heliox attenuates lung inflammation and structural alterations in acute lung injury. Pediatr Pulmonol 2005;40:524 –32 20. Englelke KA, Halliwill JR, Proctor DN, Dietz NM, Joyner MJ. Contribution of nitric oxide and prostaglandins to reactive hyperemia in human forearm. J Appl Physiol 1996;81:1807–14 21. Harris PD, Barnes R. The uses of helium and xenon in current clinical practice. Anaesthesia 2008;63:284 –93 22. Fink JB. Opportunities and risks of using heliox in your clinical practice. Respir Care 2006;51:651– 60 23. Eves ND, Petersen SR, Haykowsky MJ, Wong EY, Jones RL. Helium-hyperoxia, exercise, and respiratory mechanics in chronic obstructive pulmonary disease. Am J Respir Crit Care Med 2006;174:763–71 24. Pagel PS, Krolikowski JG, Pratt PF Jr., Shim YH, Amour J, Warltier DC, Weihrauch D. Reactive oxygen species and mitochondrial adenosine triphosphate-regulated potassium channels mediate helium-induced preconditioning against myocardial infarction in vivo. J Cardiothorac Vasc Anesth 2008;22:554 –9 25. Feng J, Lucchinetti E, Ahuja P, Pasch T, Perriard JC, Zaugg M. Isoflurane postconditioning prevents opening of the mitochondrial permeability transition pore through inhibition of glycogen synthase kinase 3beta. Anesthesiology 2005;103:987–95 26. Zaugg M, Lucchinetti E, Spahn DR, Pasch T, Schaub MC. Volatile anesthetics mimic cardiac preconditioning by priming the activation of mito KATP channels via multiple signaling pathways. Anesthesiology 2002;97:4 –14 27. Carmona M, Lopes RI, Borba M, Omokawa M, Naufal R, Miyaji K, Matsumura N, Vieira N, Pereira PR. Comparison of the effects of carbon dioxide and helium pneumoperitoneum on renal function. J Endourol 2008;22:1077– 82 28. Schmeding M, Schwalbach P, Reinshagen S, Autschbach F, Benner A, Kuntz C. Helium pneumoperitoneum reduces tumor recurrence after curative laparoscopic liver resection in rats in a tumor-bearing small animal model. Surg Endosc 2003;17:951–9 29. Brundell SM, Tsopelas C, Chatterton B, Touloumtzoglou J, Hewett PJ. Experimental study of peritoneal blood flow and insufflation pressure during laparoscopy. Br J Surg 2002; 89:617–22 30. Springer TA. Traffic signals on endothelium for lymphocyte recirculation and leukocyte emigration. Annu Rev Physiol 1995;57:827–72 31. Frenette PS, Denis CV, Weiss L, Jurk K, Subbarao S, Kehrel B, Hartwig JH, Vestweber D, Wagner DD. P-Selectin glycoprotein ligand 1 (PSGL-1) is expressed on platelets and can mediate platelet-endothelial interactions in vivo. J Exp Med 2000;191:1413–22 32. Lucchinetti E, Hofer C, Bestmann L, Hersberger M, Feng J, Zhu M, Furrer L, Schaub MC, Tavakoli R, Genoni M, Zollinger A, Zaugg M. Gene regulatory control of myocardial energy metabolism predicts postoperative cardiac function in patients undergoing off-pump coronary artery bypass graft surgery: inhalational versus intravenous anesthetics. Anesthesiology 2007;106:444 –57 33. Julier K, da Silva R, Garcia C, Bestmann L, Frascarolo P, Zollinger A, Chassot PG, Schmid ER, Turina MI, von Segesser LK, Pasch T, Spahn DR, Zaugg M. Preconditioning by sevoflurane decreases biochemical markers for myocardial and renal dysfunction in coronary artery bypass graft surgery: a doubleblinded, placebo-controlled, multicenter study. Anesthesiology 2003;98:1315–27 34. VanDeripe DR. The swelling of mitochondria from nitrogen gas; a possible cause of reperfusion damage. Med Hypotheses 2004;62:294 – 6


Thiopental Inhibits Lipopolysaccharide-Induced Tissue Factor Expression Matthias Hartmann, Priv Doz Dr med* ¨ zlu¨gedik, Dr med† Semih O Juergen Peters, Prof Dr med*

BACKGROUND: During Gram-negative sepsis, lipopolysaccharide (LPS) stimulates toll-like receptor 4, resulting in an activation of the immune system and the expression of tissue factor on monocytes. As a consequence, intravascular coagulation, ischemia, and multiorgan dysfunction may occur. Because thiopental has been described to modulate the immune system, we tested the hypothesis that thiopental alters the LPS-induced tissue factor expression. METHODS: (i) Citrated whole blood samples were incubated with thiopental (0, 0.25, 0.5, 1 mg/mL) and LPS (100 ␮g/mL) for 4 h. After recalcification, clotting time (CT) was determined by rotational thrombelastometry. (ii) The mechanism of the LPS-induced shortening of CT was investigated using the tissue factor blocker active-site inhibited factor VIIa and the protein synthesis inhibitor cycloheximide. (iii) A concentration response curve for the effect of tissue factor on CT was generated. RESULTS: LPS shortened CT from 618 ⫾ 122 s to 192 ⫾ 33 s (n ⫽ 6; P ⬍ 0.05). Shortening of CT was mediated by synthesis of tissue factor, because both inhibition of protein synthesis and blockade of tissue factor effects abolished this effect of LPS. Thiopental markedly inhibited the LPS-induced shortening of CT (372 ⫾ 86 s; n ⫽ 6; P ⬍ 0.001). Comparison of CT with a tissue factor standard curve demonstrated that thiopental reduced the LPS-induced tissue factor activity up to 86%. A direct effect of thiopental on coagulation was excluded, because tissue factor-induced CT was not affected by the barbiturate. CONCLUSIONS: Thiopental markedly inhibits the LPS-induced tissue factor expression in whole blood samples. (Anesth Analg 2009;109:109 –13)

G

ram-negative sepsis leads to an activation of the immune system and coagulation, resulting in disseminated intravascular coagulation (DIC).1,2 This disturbance of hemostasis is of outstanding pathophysiological importance as it can induce organ ischemia and multiorgan dysfunction. Accordingly, the diagnosis of DIC has been demonstrated to be an independent risk factor for death in septic patients.3 Further proof for the pathophysiological importance of DIC was derived from animal experiments in which coagulation was inhibited by pharmacological means. Inhibition of coagulation using the tissue factor pathway inhibitor, antithrombin, heparin, or activated protein C reduced the mortality in

From the *Klinik fu¨r Anästhesiologie und Intensivmedizin, Universität Duisburg-Essen, Universitätsklinikum Essen, Essen, Germany; and †Klinik fu¨r Anästhesiologie, Universität Du¨sseldorf, Universitätsklinikum Du¨sseldorf, Du¨sseldorf, Germany. Accepted for publication January 7, 2009. Reprints will not be available from the author. Supported Solely by Institutional Funding. Address correspondence to Hartmann M, Klinik fu¨r Ana¨sthesiologie und Intensivmedizin, Universita¨tsklinikum Essen, Hufelandstr. 55, D-45122 Essen, Germany. Address e-mail to matthias. [email protected]. Copyright © 2009 International Anesthesia Research Society DOI: 10.1213/ane.0b013e3181a27cfb

Vol. 109, No. 1, July 2009

various animal models of sepsis.1 Furthermore, the results of the PROWESS study demonstrate that the latter agent also reduced the mortality from sepsis in humans.4 A main cause of DIC in sepsis is tissue factor, the principal activator of coagulation in vivo. Coagulation is initiated by binding and activation of factor VII to tissue factor, activation of the coagulation cascade and formation of fibrin.5 Under physiological conditions, tissue factor is expressed exclusively on the surface of extravascular cell types but not in the intravascular space to avoid intravascular thrombus formation.6 In patients with Gram-negative sepsis, however, bacterial lipopolysaccharide (LPS) stimulates toll-like receptor 4, activates nuclear factor kappa B, and induces the expression of tissue factor on the surface of monocytes and endothelial cells, leading to DIC.7,8 We recently demonstrated that LPS-induced tissue factor expression can be reduced by heat shock treatment under in vitro conditions.9 No information, however, is available on the effect of drugs, which are commonly used perioperatively, on tissue factor expression. Thiopental has been demonstrated to affect the immune response leading to an increased rate of nosocomial infections when used for long-term sedation. On the cellular level, thiopental has been described to inhibit immunological cell functions of T lymphocytes via the transcription factor nuclear factor 109

kappa B and the suppression of I␬B kinase.10,11 The effects of the barbiturate on another key immune cell type, the monocyte, which is responsible for the pathophysiological important DIC in sepsis, is not known. Therefore, it was the aim of the present study to investigate whether thiopental can affect the LPSinduced activation of coagulation in whole blood samples, which is mediated by monocytes expressing tissue factor.

METHODS Blood Sampling After receiving approval by the local ethics committee and written informed consent, venous blood was drawn from the antecubital vein of six healthy volunteers. After discarding the first 2 mL, blood was collected in one-tenth volume of citrate (3.8%, Becton Dickinson Vacutainer) and samples were immediately used for the experiments.

Incubation of Whole Blood Samples (i) Blood samples were incubated with thiopental (250, 500, 1000 ␮g/mL final concentration), which was dissoved in NaCl 0.9% or vehicle for 1 h. Thereafter, LPS (final concentration 100 ␮g/mL) was added for a further 4 h. All incubations were performed at 37°C. (ii) To determine the mechanism of the LPS effect on whole blood clotting time, samples were incubated with the tissue factor inhibitor active site-inhibited factor VIIa (ASIS, added 10 min before measurement) or the protein synthesis inhibitor cycloheximide (added 30 min before LPS) under otherwise identical conditions.12 (iii) To study the tissue factor concentrations necessary to explain the observed LPS-induced decrease in clotting time and the modulatory effects of thiopental, whole blood samples were spiked with various amounts of tissue factor standards to obtain a dose-response curve. (iv) In a series of control experiments, the effects of thiopental on clotting time induced by recalcification in the presence or absence of tissue factor standard was determined. (v) Control experiments were performed to exclude the effect of thiopental and LPS on pH and cell viability. Neither thiopental nor LPS or a combination of both agents affected the pH under the experimental conditions described above. Furthermore, no effect on cell viability, as determined by white blood cell count or trypan blue exclusion, which was performed as recently described, was observed.9

Determination of Clotting Time and Clot Firmness Whole blood clotting time and maximal clot firmness were determined as described.12 In short, citrated whole blood samples containing LPS and thiopental, as stated in the respective experiments, were recalcified with calcium chloride and subjected to rotational thrombelastometry (Roteg 5™, Pentapharm, Munich, Germany), a modification of the original thrombelastography method.13 We recently demonstrated that 110

Thiopental and Tissue Factor Expression

Figure 1. Effects of thiopental on whole blood clotting time in the absence and presence of thiopental (1 mg/mL). After incubation of whole blood samples with thiopental, samples were incubated with lipopolysaccharide (LPS) (100 ␮g/mL) or vehicle for 4 h at 37°C. Thereafter, clotting time was determined by thrombelastometry. LPS evoked a shortening of clotting time, which was markedly reduced by thiopental. In the absence of LPS, thiopental did not affect clotting time. Mean ⫾ sd, n ⫽ 6.

storage of citrate blood samples for up to 5 h does not affect coagulation characteristics determined by thrombelastometry.12,14

Materials LPS (Escherichia coli; serotype 0.111:B4) was obtained from Sigma-Aldrich, Germany. ASIS was a generous gift from Novo Nordisk, Denmark. All other reagents were of analytical grade.

Statistical Analysis Data are presented as mean and standard deviation. For statistical evaluation, the t-test for independent measurements was used, normal distribution was demonstrated by use of the Shapiro-Wilk-test (SPSS, Chicago). Bonferroni correction was used when appropriate. Statistical significance was assumed with an ␣-error of ⬍0.05. For the determination of tissue factor concentrations from clotting time (Fig. 3), a four-parameter Hill function was used for the curve fit (SigmaPlot Software, San Jose).

RESULTS As shown in Figure 1, incubation of whole blood samples with LPS (100 ␮g/mL) at 37°C for 4 h markedly shortened the clotting time from 618 ⫾ 122 s to 192 ⫾ 33 s (n ⫽ 6; P ⬍ 0.05). The effect of LPS on coagulation was markedly reduced in the presence of thiopental (1 mg/mL): the clotting time shortened to only 372 ⫾ 86 s. In contrast, thiopental had no effect on clotting time in the absence of LPS (653 ⫾ 111 s; n ⫽ 6). The inhibitory effect of thiopental on the LPS-induced clotting time was concentration-dependent. Even at the lowest concentration of thiopental (0.25 mg/mL), the LPS effect was markedly reduced (Fig. 2). In contrast to the marked effects of thiopental on clotting time, the barbiturate did not affect maximal clot ANESTHESIA & ANALGESIA

Figure 2. Effects of thiopental (0, 250, 500, 1000 ␮g/mL) on clotting time in whole blood samples incubated with lipopolysaccharide (LPS) (100 ␮g/mL). Clotting time obtained in the absence of LPS was 618 ⫾ 122 s. Thiopental markedly inhibited the LPS-induced shortening of clotting time in a dosedependent manner. Means ⫾ sd, n ⫽ 6.

Figure 3. Influence of protein synthesis inhibition (cycloheximide 35 ␮M) and tissue factor blockade (active site-inhibited factor VIIa [ASIS] 50 ␮g/mL) on LPS (1 mg/mL)-induced shortening of clotting time. The LPS-induced shortening of clotting time was abolished by both inhibition of protein synthesis and tissue factor effects. In the absence of LPS, cycloheximide, and ASIS had no effect on clotting time. Means ⫾ sd, n ⫽ 6. firmness (control: 58 ⫾ 6 mm, 1 mg/mL thiopental: 60 ⫾ 5 mm), which is dependent on fibrinogen polymerization and platelet number and function. To elucidate the mechanism of the LPS-induced shortening of clotting time, experiments with the tissue factor blocker ASIS and the protein synthesis inhibitor cycloheximide were performed (Fig. 3). In the presence of the tissue factor inhibitor ASIS (50 ␮M), the LPS-induced shortening of clotting time was abolished, thus demonstrating that the LPS-effect is mediated exclusively via tissue factor. The fact that ASIS did not affect clotting time in the absence of LPS indicated that no relevant tissue factor was present in the blood of healthy volunteers. Inhibition of protein synthesis with cycloheximide (35 ␮M) also abolished the LPS effect on clotting time suggesting the de novo synthesis of tissue factor in response to LPS stimulation. In the absence of LPS, Vol. 109, No. 1, July 2009

Figure 4. Concentration response curve showing the effects of tissue factor standards on whole blood clotting time. Tissue factor shortened clotting time in a concentrationdependent fashion. Means ⫾ sd, n ⫽ 6.

Figure 5. Effect of thiopental (0, 250, 500, 1000 ␮g/mL) on the lipopolysaccharide (LPS)-induced tissue factor concentration in whole blood. LPS-induced tissue factor expression was markedly inhibited in a dose-dependent manner. Means ⫾ sd, n ⫽ 6. cycloheximide did not affect clotting time, demonstrating that neither cycloheximide nor protein synthesis are involved in coagulation under control conditions. To exclude that thiopental directly inhibits tissue factor activity or other components of the coagulation system, the effect of thiopental on the clotting time of whole blood samples spiked with exogenously applied tissue factor was determined. In this series, clotting time was not affected by thiopental (259 ⫾ 11 s vs 265 ⫾ 37 s; n ⫽ 6; n.s.). In a further series, tissue factor standards were added to whole blood samples and clotting time was determined (Fig. 4). Using this standard curve, we calculated the effect of thiopental on the LPS-induced tissue factor concentration using a four-parameter Hill plot (Fig. 5). Tissue factor concentration in whole blood samples, which amounted to 5.9 ⫾ 2.6 pM, was © 2009 International Anesthesia Research Society

111

inhibited from 64% to 86% by thiopental (0.25–1 mg/mL).

DISCUSSION The results of the present study demonstrate that thiopental inhibits LPS-induced tissue factor expression in whole blood, suggesting a possible role of a sedation regimen in DIC caused by sepsis. For the determination of tissue factor in whole blood samples, we used thrombelastometry as the clotting assay as reported previously.12 The sensitivity of this method is high, as shown by the concentration response curve for tissue factor, because exogenously applied tissue factor shortened clotting time at concentrations as low as 100 fM. This concentration is 60 times smaller than the tissue factor concentration evoked by LPS in the present study. Another advantage of the clotting assay is the determination of tissue factor in intact cells instead of homogenates allowing the measurement of the biologically relevant extracellular tissue factor fraction.6 Furthermore, it is important to state that clotting assays, in contrast to enzymelinked immunosorbent assay or chromogenic assays, are not affected by biologically inactive soluble tissue factor.15 The specificity of thrombelastometry for tissue factor was demonstrated by use of ASIS.16 In the presence of the tissue factor inhibitor, LPS-induced shortening of clotting time was abolished. The fact that ASIS did not prolong the clotting time in whole blood samples without LPS demonstrates that tissue factor is absent in whole blood under physiological conditions and suggests that coagulation of whole blood samples after recalcification is mediated via a surface-induced activation of the intrinsic coagulation system. During Gram-negative sepsis LPS activates both the immune system and coagulation. Stimulation of tolllike receptor 4 induced by LPS results in the expression of tissue factor on the surface of monocytes leading to DIC, which is associated with poor prognosis.3 In the present study, thiopental inhibited the LPS-induced tissue factor activity by about 70%, even at the lowest barbiturate concentration (0.25 mg/mL). At 1 mg/mL, thiopental the activity was reduced by 86%. The concentrations used in the present study are in the clinical range. During induction of anesthesia with 50 –100 mg thiopental, blood concentrations up to 142 ␮g/mL have been reported.17 In view of the findings of Loop et al.,11 who demonstrated that the inhibitory effect of thiopental on NF-␬B activation persisted several hours beyond exposure, it cannot be excluded that even small doses of the barbiturate might reduce tissue factor expression under clinical conditions. During long-term treatment of intracranial hypertension, plasma concentrations are even higher; thiopental concentrations of 580 ␮g/mL have been described.18 Moreover, it has been demonstrated that thiopental in the brain and thymus exceeded plasma concentrations up to 20-fold.17 112

Thiopental and Tissue Factor Expression

Concerning the mechanism of thiopental’s effect on coagulation, our results demonstrate that the barbiturate affected neither clotting time nor clot firmness in the absence of LPS when coagulation was induced by the artificial surface during thrombelastometry. Thus, a direct activation of coagulation factors, which would have resulted in a shortening of clotting time, can be excluded. Furthermore, an effect on fibrin formation and platelets can be excluded as clot firmness was not altered. Moreover, a direct effect of thiopental on tissue factor, e.g., a change in conformation of the molecule or a blockade of the factor 7 binding site, and an inhibitory effect on coagulation factors was excluded because clotting time induced by exogenously applied tissue factor was not affected by thiopental. In addition, the result of our experiments with the protein synthesis inhibitor cycloheximide demonstrate that LPS induces the de novo protein synthesis of tissue factor under our experimental conditions and it is obvious that the inhibitory effects of the barbiturate are mediated via an effect on protein expression. Concerning the mechanism of the thiopental-induced reduction of LPS-induced tissue factor, it is interesting to note that the barbiturate has marked effects on the immune system. Treatment with thiopental increases the rate of infectious diseases when used for the treatment of brain edema.19 Furthermore, experimental studies demonstrate that the barbiturate inhibits the phagocytic activity of rat peritoneal macrophages and the antigen-specific lymphocyte proliferation and interleukin 2 production.20,21 As possible mechanisms of the immune suppressive action, an inhibition of NF-␬B via IkB, a calmodulin/calcineurin-dependent mechanism, an effect on mitogen activated protein kinases and the small G-proteins Ras and Rac-1, and the induction of apoptosis in lymphocytes, have been described.11,22–24 As the immune system and coagulation system are tightly interrelated and tissue factor expression is modulated by NF-␬B, the molecular mechanisms involved in the immunosuppressive action of thiopental may possibly be identical to those responsible for tissue factor suppression as observed in the present study. It is important to state that the results of the present study were obtained under in vitro conditions, which cannot be substituted for studies in human sepsis. Although we investigated the effect of LPS on tissue factor formation in human whole blood, which is an important source of tissue factor in sepsis, it is important to state that the endothelium is also involved in the pathophysiology of DIC. Despite these limitations, it is an important result of the present study that sedation might affect DIC in sepsis or systemic inflammation. In our in vitro study, we used LPS at a concentration of 100 ␮g/mL to stimulate tissue factor. This concentration was chosen according to a concentration response curve which we constructed under identical experimental conditions.12 The 50% effective concentration (EC50) of the LPS effect in that study ANESTHESIA & ANALGESIA

was 18 ␮g/mL and we decided to maximally stimulate tissue factor using the fivefold concentration in the present study. The relevance of the chosen LPS concentration is underlined by the fact that the LPS content of erythrocytes from septic patients has been demonstrated to be 77 ⫾ 26 ␮g/mL.25 In general, it is important to state that the LPS concentrations used in whole blood samples are typically three orders of magnitude higher than those used in cell culture experiments with low protein content. This difference can be explained by the fact that LPS always binds to peptides and proteins in biological fluids.26 Moreover, human plasma has been described to neutralize up to 6 ␮g/mL LPS.27 In conclusion, thiopental, in clinically relevant concentrations, inhibits the effects of LPS on tissue factor expression. As tissue factor formation in the vascular space is a key molecular event leading to DIC, sedation with thiopental might reduce this disturbance of the coagulation system during systemic inflammation. REFERENCES 1. Opal SM. Interactions between coagulation and inflammation. Scand J Infect Dis 2003;35:545–54 2. Dempfle CE. Coagulopathy of sepsis. Thromb Haemost 2004; 91:213–24 3. Bakhtiari K, Meijers JC, de Jonge E, Levi M. Prospective validation of the International Society of Thrombosis and Haemostasis scoring system for disseminated intravascular coagulation. Crit Care Med 2004;32:2416 –21 4. Bernard GR, Vincent JL, Laterre PF, LaRosa SP, Dhainaut JF, Lopez-Rodriguez A, Steingrub JS, Garber GE, Helterbrand JD, Ely EW, Fisher CJ Jr. Recombinant human protein C worldwide evaluation in severe sepsis (PROWESS) study group. Efficacy and safety of recombinant human activated protein C for severe sepsis. N Engl J Med 2001;344:699 –709 5. Hoffman M, Monroe DM III. A cell-based model of hemostasis. Thromb Haemost 2001;85:958 – 65 6. Butenas S, Bouchard BA, Brummel-Ziedins KE, Parhami-Seren B, Mann KG. Tissue factor activity in whole blood. Blood 2005;105:2764 –70 7. Hotchkiss RS, Swanson PE, Freeman BD, Tinsley KW, Cobb JP, Matuschak GM, Buchman TG, Karl IE. Apoptotic cell death in patients with sepsis, shock, and multiple organ dysfunction. Crit Care Med 1999;27:1230 –51 8. Levi M, Ten Cate H. Disseminated intravascular coagulation. N Engl J Med 1999;341:586 –92 9. Sucker C, Zacharowski K, Thielmann M, Hartmann M. Heat shock inhibits lipopolysaccharide-induced tissue factor activity in human whole blood. Thromb J 2007;24:5–13 10. Loop T, Liu Z, Humar M, Hoetzel A, Benzing A, Pahl HL, Geiger KK, J Pannen BH. Thiopental inhibits the activation of nuclear factor kappaB. Anesthesiology 2002;96:1202–13

Vol. 109, No. 1, July 2009

11. Loop T, Humar M, Pischke S, Hoetzel A, Schmidt R, Pahl HL, Geiger KK, Pannen BH. Thiopental inhibits tumor necrosis factor alpha-induced activation of nuclear factor kappaB through suppression of IkappaB kinase activity. Anesthesiology 2003;99:360 –7 12. Zacharowski K, Sucker C, Zacharowski P, Hartmann M. Thrombelastography for the monitoring of lipopolysaccharide induced activation of coagulation. Thromb Haemost 2006;95:557– 61 13. Ganter MT, Hofer CK. Coagulation monitoring: current techniques and clinical use of viscoelastic point-of-care coagulation devices. Anesth Analg 2008;106:1366 –75 14. Sucker C, Zotz RB, Go¨rlinger K, Hartmann M. Rotational thrombelastometry for the bedside monitoring of recombinant hirudin. Acta Anaesthesiol Scand 2008;52:358 – 62 15. Censarek P, Bobbe A, Grandoch M, Schro¨r K, Weber AA. Alternatively spliced human tissue factor (asHTF) is not procoagulant. Thromb Haemost 2007;97:11– 4 16. Petersen LC. Active-site inhibited seven: mechanism of action including signal transduction. Semin Hematol 2001;38:39 – 42 17. Yasuda T, Yamaba T, Sawazaki K, Masuyama F, Nadano D, Takeshita H, Kishi K. Postmortem concentrations of thiopental in tissues: a sudden death case. Int J Legal Med 1993;105:239 – 41 18. Neuwelt EA, Kikuchi K, Hill SA, Lipsky P, Frenkel E. Barbiturate inhibition of lymphocyte function. Differing effects of various barbiturates used to induce coma. J Neuro Surg 1982;56:254 –9 19. Eberhardt KE, Thimm BM, Spring A, Maskos WR. Dosedependent rate of nosocomial pulmonary infection in mechanically ventilated patients with brain oedema receiving barbiturates: a prospective case study. Infection 1992;20:12– 8 20. Salman H, Bergman M, Bessler H, Alexandrova S, Beilin B, Djaldetti M. Effect of sodium thiopentone anesthesia on the phagocytic activity of rat peritoneal macrophages. Life Sci 1998;63:2221– 6 21. Correâ-Sales C, Tosta CE, Rizzo LV. The effects of anesthesia with thiopental on T lymphocyte responses to antigen and mitogens in vivo and in vitro. Int J Immunopharmacol 1997;19:117–28 22. Humar M, Pischke SE, Loop T, Hoetzel A, Schmidt R, Klaas C, Pahl HL, Geiger KK, Pannen BH. Barbiturates directly inhibit the calmodulin/calcineurin complex: a novel mechanism of inhibition of nuclear factor of activated T cells. Mol Pharmacol 2004;65:350 – 61 23. Humar M, Andriopoulos N, Pischke SE, Loop T, Schmidt R, Hoetzel A, Roesslein M, Pahl HL, Geiger KK, Pannen BH. Inhibition of activator protein 1 by barbiturates is mediated by differential effects on mitogen-activated protein kinases and the small G proteins ras and rac-1. J Pharmacol Exp Ther 2004;311:1232– 40 24. Keel M, Mica L, Stover J, Stocker R, Trentz O, Ha¨rter L. Thiopental-induced apoptosis in lymphocytes is independent of CD95 activation. Anesthesiology 2005;103:576 – 84 25. Po¨schl JM, Leray C, Ruef P, Cazenave JP, Linderkamp O. Endotoxin binding to erythrocyte membrane and erythrocyte deformability in human sepsis and in vitro. Crit Care Med 2003;31:924 – 8 26. Chaby R. Lipopolysaccharide-binding molecules: transporters, blockers and sensors. Cell Mol Life Sci 2004;61:1697–713 27. Warren HS, Novitsky TJ, Ketchum PA, Roslansky PF, Kania S, Siber GR. Neutralization of bacterial lipopolysaccharides by human plasma. J Clin Microbiol 1985;22:590 –5


113

Technology, Computing, and Simulation Section Editor: Dwayne Westenskow

Pulse Contour Analysis and Transesophageal Echocardiography: A Comparison of Measurements of Cardiac Output During Laparoscopic Colon Surgery Mario R. Concha, MD Verońica F. Mertz, MD Luis I. Cortínez, MD Katya A. Gonza´lez, MD Jean M. Butte, MD

BACKGROUND: Pulse wave analysis (PWA) allows cardiac output (CO) measurement after calibration by transpulmonary thermodilution. A PWA system that does not require previous calibration, the FloTrac/Vigileo (FTV), has been recently developed. We compared determinations of CO made with the FTV to simultaneous measurements using transesophageal echocardiography (TEE). METHOD: Ten ASA I-II patients scheduled for laparoscopic colorectal surgery were studied. A radial 20-gauge cannula was inserted and connected to a hemodynamic monitor and a FTV system for PWA and determination of CO (COPWA). TEE CO (COTEE) was determined as previously described. Measurements were made after intubation, 5 min after establishing the lithotomy position, 5 min after establishing pneumoperitoneum, every 30 min, or each time mean arterial blood pressure decreased below basal values. Statistical analysis was made with the Bland and Altman method. RESULTS: Eighty-eight measurements were compared. The COTEE values ranged from 3.23 to 12 Lt/min (mean 6.21 ⫾ 1.85). Values for COPWA ranged from 2.9 to 8.5 Lt/min (mean 4.84 ⫾ 1.14). Bias was 1.17 and limits of agreement ⫺2.02 and 4.37. The percentage error between all COTEE and COPWA measurements was 40% (27%–50%) mean (range). CONCLUSION: During laparoscopic colon surgery, clinically important differences were observed between CO determinations made with TEE and FTV. (Anesth Analg 2009;109:114 –8)

T

he “gold standard” method for measuring cardiac output (CO) in the clinical setting is thermodilution using a pulmonary artery catheter (PAC) (COPACT).1–3 However, because the risk of PAC insertion cannot be justified in routine cases, a less invasive method for obtaining this information is desirable. Pulse wave analysis of CO (COPWA) has been established as a valid and cost-effective method for CO measurement after calibration by transpulmonary thermodilution or COPACT to compensate for interindividual differences in arterial compliance.4 –7 A device that does not require a previous calibration, the FloTrac/Vigileo (FTV), has been introduced by Edwards Lifesciences LLC, Irvine, CA. The system uses the pressure sensor attached to arterial pressure tubing to derive continuous CO measurements from the arterial pressure wave. Conflicting results have been reported and its clinical From the Departments of Anesthesiology and Digestive Surgery, Escuela de Medicina, Pontificia Universidad Cato´lica de Chile, Santiago, Chile. Accepted for publication January 18, 2009. Address correspondence and reprint requests to Mario Concha, MD, Departamento de Anestesiología, Hospital Clínico Universidad Cato´lica de Chile, Santiago, Chile. Address e-mail to [email protected]. Copyright © 2009 International Anesthesia Research Society DOI: 10.1213/ane.0b013e3181a491b8

114

utility and reliability must be tested in different clinical settings and under changing hemodynamic conditions.8 –15 Considering that in most clinical situations COPACT has no indication, it would be useful to compare FTV to a less invasive measurement of CO, such as transesophageal echocardiography (TEE). Good agreement between determinations made with COPACT and TEE CO measurements (COTEE) has been demonstrated.16,17 This provides an alternative method for the intraoperative measurement of CO, allowing the comparison of measurements made with FTV with this validated method. The objective of this study was to compare simultaneous determinations of CO made with COPWA using FTV and COTEE during laparoscopic colon surgery.

METHODS After institutional ethics committee approval (Facultad de Medicina, Pontificia Universidad Cato´lica, Santiago, Chile), written informed consent was obtained from 10 patients scheduled for laparoscopic colorectal surgery. All patients were ASA physical status I-II and had no contraindication for the use of TEE. After standard monitoring (continuous electrocardiogram, noninvasive arterial blood pressure, and pulse oximetry), anesthesia was induced with fentanyl 2– 4, thiopental 5– 6, and vecuronium 0.1 mg/kg. The Vol. 109, No. 1, July 2009

Stroke volume ⫽ k ⫻ Pulsatility where k is a constant quantifying arterial compliance and vascular resistance, and pulsatility is proportional to the standard deviation of the arterial pressure wave over a 20-s interval. k is derived from patient characteristics (gender, age, weight, and height) according to the method described by Langewouters et al.18 and wave form characteristics (e.g., skewness and kurtosis Vol. 109, No. 1, July 2009

Table 1. Demographic Data 59.2 ⫾ 11.5 77.2 ⫾ 24.6 164 ⫾ 0.08 5/5

Age (yr) Weight (kg) Height (cm) Gender (M/F) Data are mean ⫾

SD

or n.

Table 2. Type of Surgery Sigmoidectomy Right hemicolectomy Left hemicolectomy Total colectomy Anterior resection Total

6 1 1 1 1 10

5 4

TEECO - VigileoCO

trachea was then intubated and ventilation was adjusted to an end-tidal carbon dioxide tension of 30 –35 mm Hg. A 20-gauge cannula was inserted into the radial artery and connected to a hemodynamic monitor and a FTV system. All pressure transducers were zeroed to the midchest level and the pressure wave form was not dampened at the time of measurements. The software used was version 1.07. During surgery, anesthesia was maintained with isoflurane 1–1.5 minimum alveolar anesthetic concentration in combination with nitrous oxide (50%), in oxygen. Supplemental bolus doses of fentanyl 1–2 ␮g/kg were administered to maintain mean arterial blood pressure (MAP) and heart rate (HR) within 20% of basal values (mean value of three determinations made at rest the day before surgery). Muscle relaxation was maintained with supplemental 1 mg doses of vecuronium throughout surgery. Nasopharyngeal temperature and urine output were also monitored. Patients were actively warmed with a forced air warming system to maintain temperatures above 36°C. Before TEE probe insertion, a nasogastric tube was used to empty air from the stomach and then removed. All TEE determinations were performed by the same certified anesthesiologist (VM, American National Board of Echocardiography) with a Philips Envisor C and an Omniplane III (2–7 MHz) TEE probe. CO by TEE was determined according to the method described by Perrino et al.,16 the left ventricular outflow tract (LVOT) diameter was determined in a midesophageal aortic long-axis view to allow the automatic calculation of the cross-sectional area of the LVOT (CSALVOT) (CSALVOT ⫽ ␲ (LVOT diameter/2)2). Then, to obtain measurements of aortic blood flow, the transesophageal ultrasound probe was positioned in a transgastric shortaxis view of the left ventricle at the midpapillary level. By rotating the imaging array to approximately 120°, the LVOT and ascending aorta were imaged lying parallel to the ultrasound beam. Aortic blood flow velocities were measured by a continuous wave Doppler at the level of the aortic valve. At end expiration, two consecutive velocity wave forms were recorded and then the velocity time integral was traced. Doppler CO is calculated as the product of velocity time integral, CSALVOT, and HR. Simultaneously, CO was determined with the FTV system that bases its calculations on arterial wave form characteristics in conjunction with demographic data. The system calculates the arterial pressure using arterial pulsatility (standard deviation of the pressure wave over a 20-s interval), resistance, and compliance, according to the following general equation:

3 2 1 0 -1 -2 -3

2

4

6

8

10

12

14

16

(TEECO + VigileoCO) / 2

Figure 1. Bland Altman for repeated measurements plot.

TEECO ⫽ cardiac output determined by transesophageal echocardiography; VigileoCO ⫽ cardiac output determined with Flo Trac Vigileo system. Solid lines represent bias (mean difference between COTEE measurements and COPWA). Dashed lines are limits of agreement.

of individual waves). The exact calculation method and the relative weight given each factor are proprietary information. k is recalculated every 10 min and CO every 20 s on the basis of the last 20-s interval of arterial wave form analysis.13,19,20 Measurements were performed by an investigator blinded to FTV data at the following times: After tracheal intubation, in supine position (T1), and after a period of at least 5 min of MAP and HR within ⫾20% of basal values. Five minutes after installing the patient in the lithotomy position and with no degree of inclination of the operating table (T2). Five minutes after establishing pneumoperitoneum (T3). Every 30 min or each time MAP decreased more than 20% of basal values. Some of these measurements were made in Trendelenburg (T4) and some in reverse Trendelenburg position (T5). During colonic anastomosis by a Pfannenstil incision, in steep Trendelenburg position (T6). © 2009 International Anesthesia Research Society

115

Table 3. Cardiac Output Measurements T1 T2 T3 T4 T5 T6 T7

COTEE

COPWA

Bias

CI 95%

PE

5.7 ⫾ 2.1 5.3 ⫾ 1.5 5.6 ⫾ 1.1 6.3 ⫾ 1.1 6.4 ⫾ 1.9 6.7 ⫾ 1.9 7 ⫾ 2.6

4.7 ⫾ 1.1 4.7 ⫾ 1 4.7 ⫾ 0.8 4.5 ⫾ 0.9 5 ⫾ 0.3 5.1 ⫾ 1.5 5 ⫾ 1.3

1.02 0.54 0.91 1.94 1.35 1.61 2.04

4.4–2.5 3.6–2.5 2.7–0.9 4.7–0.8 3.8–1.1 4.2–1.0 5.1–1.1

71 (39–86) 64 (43–84) 37 (27–45) 53 (37–62) 47 (26–56) 47 (28–72) 57 (33–78)

Data are mean and range. Cardiac output data are mean ⫾ SD. COTEE ⫽ cardiac output determined by transesophageal echocardiography; COPWA ⫽ cardiac output determined by pulse wave analysis; CI ⫽ confidence interval; PE ⫽ percentage error.

At the end of surgery, without pneumoperitoneum, and with no inclination of the operating table (T7). During surgery intraabdominal pressure was maintained at a maximum of 15 mm Hg. Agreement between methods was analyzed with the method described by Bland and Altman21 for repeated measurements. Limits of agreement (LOA) were defined as the range in which 95% of the differences between both methods were expected to lie. Bias was calculated as the mean difference between COTEE measurements and COPWA. Because Bland and Altman’s method does not compensate for the relationship between the magnitude of CO measurements and the size of the error, the percentage 6 4 2

TEECO-VigileoCO

0 -2 -4

T1 0

2

4

6

8

10

12

RESULTS Demographic and surgical data are shown in Tables 1 and 2. Eighty-eight measurements were analyzed. Figure 1 shows the Bland Altman plot. The COTEE values ranged from 3.23 to 12 Lt/min (mean 6.21 ⫾ 1.85), whereas the values for COPWA ranged from 2.9 to 8.5 (mean 4.84 ⫾ 1.14). Bias was 1.17 and LOA ⫺2.02 and 4.37. The percentage error between all COTEE and COPWA measurements was 40% (27%–50%) mean (range). Comparisons of data points in each situation when CO determinations were made are shown in Table 3 and Figure 2.

DISCUSSION The most important finding of this study was that CO determined with the FTV system showed important differences with COTEE measurements and that these differences were independent of the surgical conditions at the time of CO determinations. Even though the clinical setting and comparison method of this study are very different to the ones described in

6

6

4

4

2

2

0

0

-2

-2

-4

-6

error (⫾2 sd/␮) was calculated for each set of data according to Critchley and Critchley22 suggestions. A percentage of error of up to ⫾30% was considered acceptable for this new technique.22

0

2

4

6

8

10

12

6

4

4

4

2

2

2

0

0

0

-2

-2

-2

T5

-6 0

2

4

6

8

10

6

-4 -6 12

2 0 -2

T3

-4

-6

6

-4

4

-4

T2

-6

6

2

4

6

8

10

2

4

6

8

-4

T6 0

0

12

10

12

T4

-6 0

2

4

6

8

10

12

T7

-6 0

2

4

6

8

10

12

(TEECO+VigileoCO)/2 Figure 2. Comparison of data points in each situation studied. TEECO ⫽ cardiac output determined by transesophageal echocardiography; VigileoCO ⫽ cardiac output determined with Flo Trac Vigileo system; T1 ⫽ after intubation, in supine position, and after a period of at least 5 min of MAP and HR within ⫾20% basal values; T2 ⫽ 5 min after installing the patient in lithotomy position and with no degree of inclination of operating table; T3 ⫽ 5 min after establishing pneumoperitoneum; T4 ⫽ every 30 min, or each time MAP decrease more than 20% of basal value, in Trendelenburg position; T5 ⫽ every 30 min, or each time MAP decrease more than 20% of basal value, in reverse Trendelenburg position; T6 ⫽ during colonic anastomosis by a Pfannestil incision, in steep Trendelenburg position; T7 ⫽ at the end of surgery, without pneumoperitoneum, and with no inclination of the operating table. 116

Pressure Wave Form and Cardiac Output


other studies,8 –10,14 results are similar in the sense of an important bias and a wide range of LOA when measurements were made with a PAC. The observed differences may be explained by the characteristics of the algorithm used by the FTV system to determine the CO, the clinical situation in which comparison was performed, and the temporal relationship between the time in which k was calculated and COTEE determinations were made. As explained by Mayer et al.,8 algorithms of arterial wave form analysis are based on properties of the arterial system, such as impedance, peripheral vascular resistance, and compliance. Aortic impedance is required to calculate stroke volume, and it has important patient-to-patient variations. Former approaches to arterial wave analysis excluded this source of error by performing an initial invasive calibration.4,7 Interindividual differences in aortic impedance may contribute to inaccuracies in calculating the CO when calibration is performed based solely upon demographic data. Pneumoperitoneum and the important changes in patient position required by laparoscopic colon surgery may influence afterload and aortic diameter.23–25 Both of these factors are directly involved in the algorithm used by the FTV system to determine CO; however, these changes are not considered in the derivation of the k value. This fact can be a serious limitation for the use of FTV in this kind of surgery or in other situations that are associated with changes in peripheral vascular resistance. Finally, and as occurs in laparoscopic surgery, in situations of frequent hemodynamic changes, the k value may be determined at a time in which arterial compliance, vascular peripheral resistance, or impedance may not be the same as the time of measurement. Furthermore, potential weaknesses of the system include artifacts or alterations of the arterial pressure wave form, as in aortic valve disease, use of an intraaortic balloon pump, or important reductions of systemic vascular resistance.14,20,26 However, none of these conditions were present in the patients studied. All these factors may be clinically relevant in situations of hemodynamic instability because it can expose patients to incorrect therapeutic decisions. A new software (version 1.10, introduced in spring 2006), that updates variables that account for dynamic changes in vascular tone every 60 s as opposed to every 10 min with the previous version, has been developed and could help reduce this problem. However, conflicting results have been reported.12,15 A possible limitation of our study is that there was no comparison with COPACT and that echocardiographic measurement of CO is an operator-dependent technique. To reduce these problems, all determinations were performed by the same certified anesthesiologist (VM). However, when a good determination of aortic valve area and proper alignment of the ultrasound beam and the LVOT are obtained, there is agreement between TEECO and PACCO.16,17,26 Another possible limitation of this study is that, although not Vol. 109, No. 1, July 2009

evident from the visual inspection of the Bland Altman plots, our results might have been influenced by outlier data because we studied 88 CO measurements coming from only 10 patients. Even though our clinical setting was different from those used in other studies,8,10,14,15 we agree with their conclusions in the sense that determinations of COPWA were not comparable with those obtained using other validated methods. The FTV must be evaluated in different clinical situations and compared with a standard method of CO measurement to avoid exposing patients to ineffective or harmful therapeutic decisions. In summary, this study shows that during laparoscopic colon surgery, there is an important difference between CO determinations made with TEE and FTV. Hemodynamic changes determined by laparoscopic surgery, that are not considered in the algorithm used by FTV, may explain these differences and suggest that more evaluation in other situations with frequent and fast hemodynamic changes is necessary. REFERENCES 1. Swan HJ, Ganz W, Forrester J, Marcus H, Diamond G, Chonette D. Catheterization of the heart in man with use of a flowdirected balloon-tipped catheter. N Engl J Med 1970;283:447–51 2. Branthwaite MA, Bradley RD. Measurement of cardiac output by thermal dilution in man. J Appl Physiol 1968;24:434 – 8 3. Gomez CM, Palazzo MG. Pulmonary artery catheterization in anaesthesia and intensive care. Br J Anaesth 1998;81:945–56 4. Rodig G, Prasser C, Keyl C, Liebold A, Hobbhahn J. Continuous cardiac output measurement: pulse contour analysis vs thermodilution technique in cardiac surgical patients. Br J Anaesth 1999;82:525–30 5. Go¨dje O, Friedl R, Hannekum A. Accuracy of beat-to-beat cardiac output monitoring by pulse contour analysis in hemodynamical unstable patients. Med Sci Monit 2001;7:1344 –50 6. Penttila J, Snapir A, Kentala E, Koskenvuo J, Posti J, Scheinin M, Scheinin H, Kuusela T. Estimation of cardiac output in a pharmacological trial using a simple method based on arterial blood pressure signal waveform: a comparison with pulmonary thermodilution and echocardiographic methods. Eur J Clin Pharmacol 2006;62:401–7 7. Yamashita K, Nishiyama T, Yokoyama T, Abe H, Manabe M. Cardiac output by PulseCO is not interchangeable with thermodilution in patients undergoing OPCAB. Can J Anaesth 2005;52:530 – 4 8. Mayer J, Boldt J, Schollhorn T, Rohm KD, Mengistu AM, Suttner S. Semi-invasive monitoring of cardiac output by a new device using arterial pressure waveform analysis: a comparison with intermittent pulmonary artery thermodilution in patients undergoing cardiac surgery. Br J Anaesth 2007;98:176 – 82 9. Opdam HI, Wan L, Bellomo R. A pilot assessment of the FloTrac cardiac output monitoring system. Intensive Care Med 2007;33:344 –9 10. Sander M, Spies CD, Grubitzsch H, Foer A, Muller M, von Heymann C. Comparison of uncalibrated arterial waveform analysis in cardiac surgery patients with thermodilution cardiac output measurements. Crit Care 2006;10:R164 11. Sakka SG, Kozieras J, Thuemer O, van Hout N. Measurement of cardiac output: a comparison between transpulmonary thermodilution and uncalibrated pulse contour analysis. Br J Anaesth 2007;99:337– 42 12. Button D, Weibel L, Reuthebuch O, Genoni M, Zollinger A, Hofer CK. Clinical evaluation of the FloTrac/Vigileo system and two established continuous cardiac output monitoring devices in patients undergoing cardiac surgery. Br J Anaesth 2007;99:329 –36 13. Manecke GR Jr, Auger WR. Cardiac output determination from the arterial pressure wave: clinical testing of a novel algorithm that does not require calibration. J Cardiothorac Vasc Anesth 2007;21:3–7 © 2009 International Anesthesia Research Society

117

14. Biais M, Nouette-Gaulain K, Cottenceau V, Vallet A, Cochard JF, Revel P, Sztark F. Cardiac output measurement in patients undergoing liver transplantation: pulmonary artery catheter versus uncalibrated arterial pressure waveform analysis. Anesth Analg 2008;106:1480 – 6, table of contents 15. Compton FD, Zukunft B, Hoffmann C, Zidek W, Schaefer JH. Performance of a minimally invasive uncalibrated cardiac output monitoring system (Flotrac/Vigileo) in haemodynamically unstable patients. Br J Anaesth 2008;100:451– 6 16. Perrino AC Jr, Harris SN, Luther MA. Intraoperative determination of cardiac output using multiplane transesophageal echocardiography: a comparison to thermodilution. Anesthesiology 1998;89:350 –7 17. Darmon PL, Hillel Z, Mogtader A, Mindich B, Thys D. Cardiac output by transesophageal echocardiography using continuouswave Doppler across the aortic valve. Anesthesiology 1994;80: 796 – 805; discussion 25A 18. Langewouters GJ, Wesseling KH, Goedhard WJ. The static elastic properties of 45 human thoracic and 20 abdominal aortas in vitro and the parameters of a new model. J Biomech 1984;17:425–35 19. Manecke GR. Edwards FloTrac sensor and Vigileo monitor: easy, accurate, reliable cardiac output assessment using the arterial pulse wave. Expert Rev Med Devices 2005;2:523–7

118

Pressure Wave Form and Cardiac Output

20. Manecke GR Jr. Cardiac output from the arterial catheter: deceptively simple. J Cardiothorac Vasc Anesth 2007;21:629 –31 21. Bland JM, Altman DG. Statistical methods for assessing agreement between two methods of clinical measurement. Lancet 1986;1:307–10 22. Critchley LA, Critchley JA. A meta-analysis of studies using bias and precision statistics to compare cardiac output measurements techniques. J Clin Monit Comput 1999;15:85–91 23. Ivankovich AD, Miletich DJ, Albrecht RF, Heyman HJ, Bonnet RF. Cardiovascular effects of intraperitoneal insufflation with carbon dioxide and nitrous oxide in the dog. Anesthesiology 1975;42:281–7 24. Johannsen G, Andersen M, Juhl B. The effect of general anaesthesia on the haemodynamic events during laparoscopy with CO2-insufflation. Acta Anaesthesiol Scand 1989;33:132– 6 25. Torrielli R, Cesarini M, Winnock S, Cabiro C, Mene JM. [Hemodynamic changes during celioscopy: a study carried out using thoracic electric bioimpedance]. Can J Anaesth 1990;37:46 –51 26. Lorsomradee S, Lorsomradee S, Cromheecke S, De Hert SG. Uncalibrated arterial pulse contour analysis versus continuous thermodilution technique: effects of alterations in arterial waveform. J Cardiothorac Vasc Anesth 2007;21:636 – 43


Technical Communication

Can Mixed Venous Hemoglobin Oxygen Saturation Be Estimated Using a NICO Monitor? Yoshifumi Kotake, MD, PhD* Takashige Yamada, MD† Hiromasa Nagata, MD† Takeshi Suzuki, MD, PhD† Junzo Takeda, MD, PhD†

BACKGROUND: We hypothesized that mixed venous hemoglobin oxygen saturation (SvO2) can be estimated by calculation from CO2 production, cardiac output, and arterial oxygen saturation measured using a noninvasive cardiac output (NICO) monitor (Novametrix-Respironics, Wallingford, CT). METHODS: Twenty-three patients undergoing aortic aneurysm repair underwent SvO2 monitoring using a pulmonary artery catheter and cardiac output monitoring using a NICO monitor. The estimated SvO2 value calculated from NICO monitorderived values was compared with the SvO2 value measured using a pulmonary artery catheter. The accuracy of this estimation was analyzed with Bland-Altman method. The ability of this estimation to track the change of SvO2 was also evaluated using correlation analysis to compare the changes of estimated SvO2 and measured SvO2. RESULTS: The bias ⫾ limits of agreement of the estimated SvO2 against measured SvO2 was ⫺2.1% ⫾ 11.2%. The change of estimated SvO2 was modestly correlated with the change of measured SvO2. CONCLUSIONS: SvO2 derived from the values measured by the NICO monitor cannot be used interchangeably with the values measured spectrophotometrically using the pulmonary artery catheter. More refinement is required to obtain more reliable estimate of SvO2 less invasively. However, large changes of SvO2 may be detected with this method and can be used as a precautionary sign when the balance between oxygen supply and demand is compromised without inserting a central venous catheter. (Anesth Analg 2009;109:119 –23)

M

ixed venous oxygen saturation (SvO2) or its surrogate, central venous oxygen saturation (ScvO2), is used to assess the oxygen supply-demand relationship.1–3 However, central venous access is required to measure these values, which carries some risk to the patient. Theoretically, SvO2 can be calculated using cardiac output (CO), the concentration of oxygen saturated hemoglobin (Hb), and oxygen consumption (VO2).4 If the relationship between VO2 and CO2 production (VCO2) remains stable during anesthesia, SvO2 can be estimated with CO, VCO2, and Hb concentration and its arterial oxygen saturation. Because CO, VCO2,

From the *Department of Anesthesiology, Toho University Medical Center Ohmori Hospital; and †Department of Anesthesiology, School of Medicine, Keio University, Tokyo, Japan. Accepted for publication January 21, 2009. Supported by Intramural Department Sources Only. Yoshifumi Kotake is a paid consultant of Edwards Lifesciences. All others authors have no conflict of interest to disclose. Reprints will not be available from the author. Address correspondence to Yoshifumi Kotake, MD, PhD, Department of Anesthesiology, Toho University Medical Center Ohmori Hospital, 6-11-1, Ohmori-nishi, Ohta, Tokyo 143-8541, Japan. Address e-mail to [email protected]. Copyright © 2009 International Anesthesia Research Society DOI: 10.1213/ane.0b013e3181a85c22

Vol. 109, No. 1, July 2009

and Spo2 are continuously monitored with a noninvasive cardiac output (NICO) monitor (RespironicsNovametrics, Wallingford, CT), we hypothesized that SvO2 could be mathematically estimated using these NICO-derived parameters and the Hb concentration.5,6 The purpose of this prospective study was to assess the accuracy of estimating SvO2 using the NICO monitor in anesthetized, ventilated patients.

METHODS The study protocol was approved by the IRB of Keio University, and written informed consent was obtained from the participants. Twenty-three patients undergoing elective infrarenal aortic reconstruction were enrolled in this prospective, observational study. Patients’ lungs were mechanically ventilated with a tidal volume of 10 mL/kg, and the respiratory rate was adjusted to maintain normocapnia. CO, VCO2, and Spo2 were continuously monitored with a NICO monitor (Ver. 5.2, Novametrix-Respironics, Wallingford, CT). The NICO sensor was placed between the heat and moisture exchanger and the Y-piece of the respiratory circuit, and the length of the rebreathing circuit was adjusted to maintain optimal condition. NICO-derived CO, VCO2, and Spo2 values were downloaded to a personal computer and later used for calculation. For CO, results of average mode (CO-a) 119

were used. An 8Fr pulmonary artery catheter (PAC) equipped with continuous CO measurement and venous oximetry functions (746HF8, Edwards Lifesciences, Irvine, CA) was inserted via the right internal jugular vein. CO and SvO2 were continuously monitored after the SvO2 was calibrated using an in vivo method of measuring mixed venous blood with a CO-oximeter (Stat M Profile, Nova Biochemicals, Waltham, MA). The averaged SvO2 during a 3-min rebreathing cycle of the NICO monitor was used as a representative measured SvO2 value. Hb was measured every 60 min by the same blood gas analyzer. Between these measurements, Hb concentration was assumed to change linearly between these measurements, and these interpolated values were used for subsequent calculations. Estimated SvO2 is defined by following formula (see appendix):

Figure 1. Distribution of bias between estimated mixed venous hemoglobin oxygen saturation (SvO2) and measured SvO2 from an individual subject. x axis denotes ranges of the bias from an individual subject. y axis denotes number of subjects. Number of the measurement in each subject varies between 45 and 82.

Estimated SvO2 ⫽ SpO2 ⫺

VCO2 (mL/min)/0.85 1.36 ⫻ Hb (g/dL) ⫻ CO-a (L/min) ⫻ 0.1

Data were excluded from the analysis when the signal quality indicator was reported as 3 or 4 by the Vigilance II monitor. Data obtained during the first 20 min after cross-clamp release were also excluded from the analysis, because venous blood returned from infrarenal lesions upon the release of cross-clamp was significantly hypoxemic and caused independent effects on SvO2. Additionally, CO was measured several times during the surgery with intermittent bolus thermodilution method (ICO). The estimated SvO2 value with ICO was then calculated by substituting the ICO value in the above formula to analyze the possible contribution of measurement error with NICOderived CO. Data are expressed as mean ⫾ SD and were statistically analyzed using the Prism software (ver 4, Graphpad, San Diego, CA). Bland-Altman analysis was initially applied to the individual data to describe the agreement between estimated SvO2 and measured SvO2. Thereafter, the data were pooled to derive overall bias ⫾ limits of agreement (2SD of difference) in all the measurement pairs. A correction was made according to the literature because multiple observations from individuals were used in this analysis.7,8 To assess how the estimated SvO2 tracks the changes of measured SvO2, the relationship between the change of estimated SvO2 between the two consecutive measurement against the incidental change of measured SvO2 was analyzed with Pearson’s correlation coefficient and regression analysis according to a previous report.9

RESULTS Twenty male and three female patients participated in this study. The age, height, and weight of these 120

SvO2 Estimation by NICO Monitor

Figure 2. Distribution of precision between estimated mixed venous hemoglobin oxygen saturation (SvO2) and measured SvO2 from an individual subject. x axis denotes ranges of the limits of agreement from an individual subject. y axis denotes number of subjects. Number of the measurement in each subject varies between 45 and 82.

participants were 72 ⫾ 8 yr old, 164 ⫾ 6 cm, and 62 ⫾ 13 kg, respectively. The duration of surgery and of cross-clamp were 266 ⫾ 54 min and 59 ⫾ 18 min, respectively. The mean (range) of measured SvO2 was 78% (44%–94%). One thousand three hundred thirtythree pairs of estimated SvO2 and measured SvO2 were included in the analysis. Between 45 and 82 pairs of measurements from each individual were included. The bias and the limits of agreement in each study subjects are summarized in Figures 1 and 2. Figure 3 demonstrates the Bland-Altman plot of all the measurements. The bias ⫾ limits of agreement of estimated SvO2 by NICO monitor was ⫺2.1% ⫾ 11.2% against the spectrophotometrically measured SvO2. Even when the analysis was limited to the hemodynamically stable period, which is arbitrarily defined as CO change was within 2% between 3-min interval, the bias did not significantly change (1172 data points, ANESTHESIA & ANALGESIA

Figure 3. Bland-Altman plot of 1333 pairs of estimated and measured mixed venous hemoglobin oxygen saturation (SvO2).

The bias is demonstrated as a solid line, and the limits of agreement (2SD of the difference) are also demonstrated as a dashed line in this figure.

Figure 4. Changes of measured mixed venous hemoglobin oxygen saturation (SvO2) and estimated SvO2 between two consecutive measurement cycles. n ⫽ 1265, solid line denotes regression line: r2 ⫽ 0.145, P ⬍ 0.001, the change of estimated SvO2 ⫽ 0.39 ⫻ the change of measured SvO2 ⫺ 0.06. bias ⫾ limits of agreement: ⫺2.1% ⫾ 11.0%). CO was determined with bolus thermodilution 120 times in these subjects. The bias ⫾ limits of agreement of ICO-derived SvO2 estimation against measured SvO2 was ⫺4.4% ⫾ 13.8% and no clear improvement of agreement was demonstrated. The relationship between the change of estimated SvO2 and the change of measured SvO2 is summarized in Figure 4. The change of estimated SvO2 was modestly but significantly correlated with the change of measured SvO2 between the measurement cycle (r2 ⫽ 0.145, P ⬍ 0.001).

DISCUSSION This study demonstrated that the bias ⫾ limits of agreement between the estimated SvO2 and the measured SvO2 were ⫺2.1% ⫾ 11.2%. This study confirmed that SvO2 could be reasonably estimated by the Vol. 109, No. 1, July 2009

NICO-derived parameters with several assumptions. However, these data indicate that the NICO-derived SvO2 value cannot be used interchangeably against the SvO2 value obtained with PAC equipped with oximetry function. It is intuitively advantageous to objectively assess hemodynamic status with CO. However, the goal of hemodynamic optimization should be to achieve wellbalanced oxygen supply and VO2 instead of a predetermined CO. Based on this perspective, the clinical importance of venous oximetry as a tool to evaluate oxygen supply-demand balance is well established.1,2 However, measuring SvO2 requires PAC insertion, which carries some additional risk over central venous catheterization. Alternatively, ScvO2 has been used for the same purpose.3,10 However, previous reports demonstrated that there was a significant variability between SvO2 and ScvO2 and concluded that these two parameters are not interchangeable.9,11–13 The differences in regional oxygen extraction, especially the hepatosplanchnic region and central nervous system have been attributed to this significant variability. To assess the balance of oxygen supply and demand less invasively, several attempts have been made to estimate the SvO2 value without inserting a central venous catheter.14 –16 Alternatively, SvO2 can be estimated if CO, Hb, SaO2, and VO2 are known. As NICO measures CO, Spo2, and VCO2, we speculated that SvO2 could be calculated with reasonable accuracy if Hb and the relationship between VO2 and VCO2 were adequately estimated. We found considerably wide limits of agreement between the estimated SvO2 and the measured SvO2. This finding may be attributed to several factors, including the inaccuracy or the temporal delay of one of the measured values incorporated into the formula and several assumptions. First, the accuracy of CO measurement may contribute to a relatively large variation. The NICO monitor has been known to © 2009 International Anesthesia Research Society

121

underestimate the CO in general.17,18 Our unpublished analysis demonstrated that the bias ⫾ limits of agreement of the current version of the NICO monitor was ⫺0.18 ⫾ 1.70 L/min against standard ICO. However, when CO values derived from ICO were used in the same calculation, the limits of agreement remained large. Thus, using a presumably more accurate and instantaneous CO data from ICO failed to increase the accuracy of the estimation of SvO2. These findings suggest that the inaccuracy of CO data from the NICO monitor cannot fully account for the variability between the estimated and measured SvO2. Furthermore, the temporal delay of NICO-derived CO after hemodynamic change may also contribute to the wide variability. Although the response time of the NICO monitor has not yet been determined, SvO2 changes are generally believed to change more rapidly than continuous CO and NICO. Alternatively, the response of SvO2 to the actual hemodynamic change may be delayed, because the physiological process, such as increasing oxygen extraction in tissues, is involved in this process. These possibilities remain to be evaluated. Second, several assumptions used in this study may contribute to the observed variability. For example, Hb concentration may be a source of error, because it is only determined intermittently and the interpolated value between these measurement was used in this study. However, bias ⫾ limits of agreement of paired measurements was not significantly improved when the Hb was measured (data not shown). Similarly, using SaO2 data obtained from blood gas analysis did not improve the accuracy either. Therefore, the Hb concentration or Spo2 may not be a major source of error. Recently, a device that continuously and percutaneously measures Hb has become available. Thus, with the advance of technology, this issue may be resolved in the future. Using VCO2 to estimate VO2 is another potential source of error in this estimation. We used a fixed 0.85 value of the VCO2/VO2 ratio to convert VCO2 to VO2 throughout this study to make monitoring settings as simple as possible. Although it is intuitively obvious that the ratio may significantly vary in anesthetized patients and critically ill patients, this variability may primarily contribute to the large limits of agreement found in this study. Although measuring VO2 instead of VCO2 should improve the accuracy, the precise and continuous determination of VO2 requires specialized equipment and may substantially compromise the applicability of this method. It is debatable whether the estimated value of SvO2 has reasonable accuracy for clinical use. Previous discussions on SvO2 and its surrogate, ScvO2, can be used for this purpose. Dueck et al.9 investigated the bias and limits of agreement between ScvO2 and SvO2 in neurosurgical patients and found the largest 95% limits of agreement to be ⫾9.3%. They concluded that ScvO2 values are not interchangeable with SvO2, but they reported that the trend of the two parameters 122

SvO2 Estimation by NICO Monitor

moved in the same direction. Reinhart et al.11 also reported the limits of agreement between ScvO2 and SvO2 as 7.95% in critically ill patients. In our study, the 95% limits of agreement was 11.2% and was larger than the limits of agreement found in these previous studies. This finding suggests that the estimated SvO2 values are not interchangeable with measured SvO2. However, our results indicate that estimated SvO2 demonstrates large changes when actual SvO2 significantly changed. Rivers emphasized the importance of demonstrating trend of oxygen supply-demand relationship19 and we believe that our methods also enable physicians to detect deteriorated oxygen supplydemand balance without a central venous catheter. Although this technique can only be applied to tracheally intubated and mechanically ventilated patients, we believe this possibility is especially useful for the anesthetic management of patients with limited cardiac reserve or significant intraoperative blood loss when central venous catheterization is not indicated. In conclusion, we demonstrated that SvO2 can be estimated from the data obtained by the NICO monitor, Hb concentration, and assumption of VCO2/VO2 ratio in mechanically ventilated patients. The data suggest that this estimation is not interchangeable with the SvO2 value obtained from a PAC, but it can detect large changes of SvO2. This method may provide an estimate of balance of oxygen supply and demand without inserting a central venous catheter. APPENDIX: THE FORMULA TO ESTIMATE SVO2 BY CO, VCO2, AND HEMOGLOBIN CONCENTRATION VO2 (mL/min) ⫽ DO2 (mL/min) ⫺ 1.36 ⫻ Hb (g/dL)

⫻ CO (L/min) ⫻ SvO2 (%) ⫻ 0.1 VO2 (mL/min) ⫽ 1.36 ⫻ Hb (g/dL) ⫻ CO (L/min)

⫻ SaO2 (%) ⫻ 0.1 ⫺ 1.36 ⫻ Hb (g/dL) ⫻ CO (L/min) ⫻ SvO2 (%) ⫻ 0.1 If SaO2 ⫽ SpO2 and VO2 (mL/min) ⫽ VCO2 (mL/min)/R R: respiratory quotient

VCO2 (mL/min)/R ⫽ 1.36 ⫻ Hb (g/dL) ⫻ CO (L/min) ⫻ SpO2 (%) ⫻ 0.1 ⫺ 1.36 ⫻ Hb (g/dL) ⫻ CO (L/min) ⫻ SvO2 (%) ⫻ 0.1 SvO2 ⫽ SpO2 ⫺

VCO2 (mL/min)/R 1.36 ⫻ Hb (g/dL) ⫻ CO (L/min) ⫻ 0.1

If R can be estimated as 0.85,

SvO2 ⫽ SpO2 ⫺

VCO2 (mL/min)/0.85 1.36 ⫻ Hb (g/dL) ⫻ CO (L/min) ⫻ 0.1

Adapted from Mark and Slaughter.4 ANESTHESIA & ANALGESIA

REFERENCES 1. Vedrinne C, Bastien O, De Varax R, Blanc P, Durand PG, Du Gres B, Bouvier H, Saroul C, Lehot JJ. Predictive factors for usefulness of fiberoptic pulmonary artery catheter for continuous oxygen saturation in mixed venous blood monitoring in cardiac surgery. Anesth Analg 1997;85:2–10 2. Marx G, Reinhart K. Venous oximetry. Curr Opin Crit Care 2006;12:263– 8 3. Rivers E, Nguyen B, Havstad S, Ressler J, Muzzin A, Knoblich B, Peterson E, Tomlanovich M. Early goal-directed therapy in the treatment of severe sepsis and septic shock. N Engl J Med 2001;345:1368 –77 4. Mark JB, Slaughter TF. Cardiovascular monitoring. In: Miller RD, ed. Miller’s anesthesia. 6th ed. Philadelphia: Elsevier Churchill Livingstone, 2005:1265–362 5. Levy RJ, Chiavacci RM, Nicolson SC, Rome JJ, Lin RJ, Helfaer MA, Nadkarni VM. An evaluation of a noninvasive cardiac output measurement using partial carbon dioxide rebreathing in children. Anesth Analg 2004;99:1642–7 6. Suzuki M, Koda S, Nakamura Y, Kawamura N, Shimada Y. The relationship between cardiac output measured by the thermodilution method and that measured by the carbon dioxide rebreathing technique during laparoscopic surgery. Anesth Analg 2005;100:1381–3 7. Bland JM, Altman DG. Measuring agreement in method comparison studies. Stat Methods Med Res 1999;8:135– 60 8. Bland JM, Altman DG. Agreement between methods of measurement with multiple observations per individual. J Biopharm Stat 2007;17:571– 82 9. Dueck MH, Klimek M, Appenrodt S, Weigand C, Boerner U. Trends but not individual values of central venous oxygen saturation agree with mixed venous oxygen saturation during varying hemodynamic conditions. Anesthesiology 2005;103:249 –57

Vol. 109, No. 1, July 2009

10. Pearse R, Dawson D, Fawcett J, Rhodes A, Grounds RM, Bennett ED. Changes in central venous saturation after major surgery, and association with outcome. Crit Care 2005;9:R694 –R699 11. Reinhart K, Kuhn HJ, Hartog C, Bredle DL. Continuous central venous and pulmonary artery oxygen saturation monitoring in the critically ill. Intensive Care Med 2004;30:1572– 8 12. Varpula M, Karlsson S, Ruokonen E, Pettila V. Mixed venous oxygen saturation cannot be estimated by central venous oxygen saturation in septic shock. Intensive Care Med 2006;32:1336 – 43 13. Sander M, Spies CD, Foer A, Weymann L, Braun J, Volk T, Grubitzsch H, von Heymann C. Agreement of central venous saturation and mixed venous saturation in cardiac surgery patients. Intensive Care Med 2007;33:1719 –25 14. Levy RJ, Stern WB, Minger KI, Montenegro LM, Ravishankar C, Rome JJ, Nicolson SC, Jobes DR. Evaluation of tissue saturation as a noninvasive measure of mixed venous saturation in children. Pediatr Crit Care Med 2005;6:671–5 15. Wei W, Zhu Z, Liu L, Zuo Y, Gong M, Xue F, Liu J. A pilot study of continuous transtracheal mixed venous oxygen saturation monitoring. Anesth Analg 2005;101:440 –3 16. Jones AE, Kuehne K, Steuerwald M, Kline JA. End expiratory oxygen concentrations to predict central venous oxygen saturation: an observational pilot study. BMC Emerg Med 2006;6:9 17. Tachibana K, Imanaka H, Miyano H, Takeuchi M, Kumon K, Nishimura M. Effect of ventilatory settings on accuracy of cardiac output measurement using partial CO2 rebreathing. Anesthesiology 2002;96:96 –102 18. Kotake Y, Moriyama K, Innami Y, Shimizu H, Ueda T, Morisaki H, Takeda J. Performance of noninvasive partial CO2 rebreathing cardiac output and continuous thermodilution cardiac output in patients undergoing aortic reconstruction surgery. Anesthesiology 2003;99:283– 8 19. Rivers E. Mixed vs central venous oxygen saturation may be not numerically equal, but both are still clinically useful. Chest 2006;129:507– 8


123

Patient Safety Section Editor: Sorin J. Brull

Liability Related to Peripheral Venous and Arterial Catheterization: A Closed Claims Analysis Sanjay M. Bhananker, MD, FRCA Derek W. Liau, MD Preetma K. Kooner, BA, BS Karen L. Posner, PhD Robert A. Caplan, MD Karen B. Domino, MD, MPH

BACKGROUND: Serious complications after peripheral IV and arterial vascular cannulations have been reported. To assess liability associated with these peripheral vascular catheters for anesthesiologists, we reviewed claims in the American Society of Anesthesiologists Closed Claims database. METHODS: Claims related to peripheral vascular catheterization were categorized as related to IV or arterial catheters. Complications related to IV catheters were categorized as to type of complication. Patient and case characteristics, severity of injury, and payments were compared between claims related to IV catheters and all other (nonperipheral catheter) claims in the database. Payment amounts were adjusted to 2007-dollar amounts using the consumer price index. RESULTS: Claims related to peripheral vascular catheterization accounted for 2% of claims in the database (n ⫽ 140 of 6894 claims), most (91%) associated with IV catheters (n ⫽ 127). The most common complications were skin slough (28%), swelling/infection (17%), nerve damage (17%), fasciotomy scars (16%), and air embolism (8%). Approximately half of these complications (55%) occurred after extravasation of drugs or fluids. Compared with other claims, IV claims involved a larger proportion of cardiac surgery (25% vs 2% for other, P ⬍ 0.001) and smaller proportion of emergency procedures (8% vs 22% for other, P ⬍ 0.001). Claims related to arterial catheters were few (n ⫽ 13, 8%), with only seven associated with radial artery catheterization. CONCLUSIONS: Claims related to IV catheters were an important source of liability for anesthesiologists, approximately half of which resulted from extravasation of drugs or fluid. Claims related to radial arterial catheterization were uncommon. (Anesth Analg 2009;109:124 –9)

P

eripheral IV and arterial vascular cannulations are relatively straightforward procedures. However, significant complications after each of these have been reported1–5 and may be a source of liability for anesthesiologists. Extravasation injuries,1 local or systemic infection,3 air embolism,6 thrombophlebitis,7 and vascular insufficiency from arterial occlusion due to spasm or thrombosis,4,5 have all been reported. We analyzed the claims related to peripheral IV or arterial (peripheral vascular) catheterizations from the American Society of Anesthesiologists (ASA) Closed Claims database to assess the anesthesia liability associated with performing and using peripheral vascular catheters. We hypothesized that the liability profile of claims related to From the Department of Anesthesiology, University of Washington School of Medicine, Seattle, Washington. Accepted for publication September 18, 2008. Supported in part by the American Society of Anesthesiologists (ASA), Park Ridge, IL as part of the Closed Claims Project. All opinions expressed are those of the authors and do not necessarily reflect those of the ASA. Reprints will not be available from the authors. Address correspondence to Sanjay M. Bhananker, MD, FRCA, Department of Anesthesiology, Harborview Medical Center, 325 Ninth Avenue, Box 359724, Seattle, WA 98104-2499. Address e-mail to [email protected]. Copyright © 2009 International Anesthesia Research Society DOI: 10.1213/ane.0b013e31818f87c8

124

peripheral catheterization would differ from other anesthesia malpractice claims.

METHODS The ASA Closed Claims database is a structured evaluation of adverse anesthetic outcomes obtained from the closed claim files of 35 United States professional liability insurance companies. The data collection process has been described in detail.8,9 Briefly, each closed claim file was reviewed by a practicing anesthesiologist and each review typically consisted of relevant hospital and medical records; narrative statements from involved health care personnel; expert and peer reviews; summaries of depositions from plaintiffs, defendants, and expert witnesses; outcome reports; and the cost of settlement or jury award. The reviewer completed a standardized form that recorded information about patient characteristics, surgical procedures, sequence and location of events, critical incidents, clinical manifestations of injury, standard of care, and outcome. Anesthesia care was addressed as appropriate, less than appropriate (substandard), or impossible to judge based on the legal concept of reasonable and prudent care at the time of the event.8,10 Reliability of reviewer judgments has been found to be acceptable.10 Vol. 109, No. 1, July 2009

Figure 1. Injuries related to IV catheters (n ⫽ 127).

The physical or psychological injury for which the patient was seeking compensation was recorded in each claim. Each claim was assigned a severity of injury score that was designated by the on-site reviewer using the insurance industry’s 10-point scale that rates severity of injury from 0 (no injury) to 9 (death).8 For purposes of analysis, injuries were grouped into three categories: temporary or nondisabling (score ⫽ 0 –5), disabling and permanent (score ⫽ 6 – 8), and death (score ⫽ 9). This study focused on the analysis of claims related to peripheral IV and arterial vascular catheters. Complications associated with IV catheters were classified as to type of complication and whether or not it was related to extravasation of drugs or fluids. Patient demographics, severity of injury, and frequency and amount of payment to the plaintiff of these claims were compared with all other claims in the database. Differences between proportions were evaluated using the ␹2 analysis, the Fisher’s exact test, and the z test. Payments for settlement and jury award were expressed in dollar amounts adjusted to 2007 dollars using the Consumer Price Index.* Because payment did not exhibit a normal distribution, the median and range were used for descriptive purposes. Statistical comparisons of payment distributions were made using the Kolmogorov-Smirnov test. P ⬍ 0.001 was required for statistical significance.

RESULTS There were 140 claims for injuries between 1975 and 2000 related to peripheral vascular catheterization (2.1% of 6894 total claims in the database), of which 127 (91%) were related to IV catheters, whereas 13 (9%) claims were related to arterial catheters.

Claims Related to IV Catheters The most common complications related to IV catheters were skin slough or necrosis (n ⫽ 35) followed by swelling/inflammation/infection (n ⫽ 22), *Consumer Price Index Inflation Calculator. U.S. Department of Labor, Bureau of Labor Statistics. Available at: http://www.bls.gov/ data/home.htm. Accessed June 2, 2008. Vol. 109, No. 1, July 2009

nerve damage (n ⫽ 22), fasciotomy scars from compartment syndrome (n ⫽ 20), and air embolism (n ⫽ 10, Fig. 1). Most claims for air embolism resulted from air in blood bags from cell savers. Burns due to heat compresses used to treat IV infiltrations accounted for 3% of claims (n ⫽ 4, Fig. 1). Approximately half (55%) (95% confidence interval 46%– 64%) of peripheral IV complications were related to the extravasation of drugs or fluids. The most commonly reported drugs causing skin slough were thiopental (n ⫽ 15), vasopressors (dopamine [n ⫽ 2], dobutamine [n ⫽ 1], epinephrine [n ⫽ 1]), and calcium chloride (n ⫽ 4). Compartment syndromes accounted for 22% of all IV-related nerve damage cases. Interestingly, there were no claims for complications of IV catheters in patients who had previously had an axillary node dissection on the ipsilateral arm. Miscellaneous claims (n ⫽ 7) involved a rash from taping of the catheter, a sheared-off catheter, patient concern over placement of an IV catheter in the same arm as an arteriovenous fistula, a metal stylette from a catheter imbedded in the patient’s thigh, ecchymosis from an IV, and accidental placement in a radial artery. One patient with severe vasculitis developed hand ischemia after administration of cold blood through the IV catheter. She required amputation of several fingers. Frivolous claims (n ⫽ 7) involved allegations of pain after difficult IV cannulation without demonstration of any pathology or claims in which there was no relationship between the site of IV cannulation and the alleged area of injury. The death of a patient with a skin slough was unrelated to the IV catheter complication. Claims related to IV catheters involve a larger proportion of cardiac surgery (25%) and a smaller proportion of emergency procedures (8%) than other claims in the database (P ⬍ 0.001, Fig. 2). There were no statistically significant differences in age, ASA status, and body habitus (obesity) between claims related to IV catheters compared with other claims (Table 1). Claims for IV catheter complications were more likely to involve temporary nondisabling injury than other claims (P ⬍ 0.001, Fig. 2). © 2009 International Anesthesia Research Society

125

Figure 2. Patient and case characteristics associated with claims related to IV catheters. Claims with missing data were excluded. *P ⬍ 0.001.

Table 1. Case Characteristicsa Intravenous catheters (n ⫽ 127)

Arterial catheters (n ⫽ 13)

All other claims (n ⫽ 6754)

78 (63%) 46 (37%)

8 (62%) 5 (38%)

3946 (59%) 2758 (41%)

7 (6%) 103 (88%) 7 (6%) 45 ⫾ 18

2 (15%) 8 (62%) 3 (23%) 46 ⫾ 26

550 (9%) 5321 (82%) 598 (9%) 43 ⫾ 20

49 (58%) 36 (42%)

4 (44%) 5 (56%)

3246 (68%) 1558 (32%)

19 (37%) 32 (63%)

3 (43%) 4 (57%)

1430 (43%) 1916 (57%)

Gender, n (%) Female Male Age (yr) 0–16 17–69 70⫹ Mean ⫾ sd ASA status, n (%) 1–2 3–5 Obese, n (%) Yes No

No statistically significant differences were found between groups. SD ⫽ standard deviation; ASA ⫽ American Society of Anesthesiologists. a Table excludes missing data.

Claims Related to Arterial Catheters There were 13 arterial catheter claims, with seven involving the radial artery (Table 2). Two claims with severe injuries (death and permanent brain damage) involved iliac artery puncture from a femoral arterial catheter that caused a retroperitoneal hemorrhage. Arterial thrombosis and limb ischemia after femoral and brachial artery cannulations occurred in infants.

Liability Approximately half (54%) of all claims related to peripheral vascular catheters resulted in payment for injury (Table 3), similar to the proportion of all other claims (57%). Monetary compensation for claims related to peripheral catheters ranged from $342 to $12,525,000 (median $47,700; Table 3). The size of payments for claims related to peripheral catheters was smaller compared with all other claims (median $215,000; P ⬍ 0.001). Claims related to air embolism 126

Intravenous/Arterial Catheter Liability

Table 2. Vessel and Complication in Arterial Catheter Claims (n⫽13) Complication

No. of claims

Radial artery Retained wire/catheter Radial nerve damage Arterial thrombosis/ischemia Hematoma/carpal tunnel syndrome Femoral artery Arterial thrombosis/ischemia Iliac artery puncture/ retroperitoneal hemorrhage Brachial artery Arterial thrombosis/ischemia

7 2 2 2 1 5 2 3 1 1

had the highest median payment and a rate of 100% payment-per-claim (Table 3). Although the overall size of payments was smaller in peripheral catheter claims compared with other ANESTHESIA & ANALGESIA

Table 3. Payment by Peripheral Catheter Complication No. of Substandard Claims resulting in $ 2007 median paymentb deaths care, n (%) paymenta, n (%)

$ 2007 payment rangeb

Complications

n

Arterial catheter Intravenous catheter Skin slough or necrosis Swelling/inflammation/ infection Nerve damage Fasciotomy scar Air embolism Burn from treatment of IV infiltration Frivolous Miscellaneous Total

13 127 35 22

1 5 1 0

2 (15%) 37 (29%) 9 (26%) 4 (18%)

8 (62%) 64 (53%) 19 (58%) 8 (38%)

$49,000 $47,475 $66,270 $11,580

$9,675–$12,525,000 $342–$11,550,000 $2,104–$154,476 $342–$43,932

22 20 10 4

0 0 4 0

4 (18%) 7 (35%) 9 (90%) 3 (75%)

12 (55%) 9 (47%) 8 (100%) 3 (75%)

$49,575 $42,750 $325,000 $75,600

$3,860–$1,215,500 $14,000–$140,000 $25,800–$4,120,200 $20,000–$210,000

7 7 140

0 0 6

0 (0%) 1 (14%) 39 (28%)

1 (14%) 4 (57%) 72 (54%)

$10,000 $69,500 $47,700

$10,000–$10,000 $1,032–$11,550,000 $342–$12,525,000

IV ⫽ intravenous catheter. a Claims with missing payment data were excluded from the calculation of percentages. b Payment amounts were adjusted to 2007 dollars using the Consumer Price Index.

claims in the database, there were seven claims with payments in excess of $1,000,000 (in 2007 dollar amounts) involving compensation for permanent brain damage, amputations, and permanent disability due to nerve injury. Three of these claims involved air embolism and permanent brain damage in young children. A fourth involved permanent brain damage in a 32 yr old as a result of an iliac artery injury and cardiac arrest upon removal of an arterial line postoperatively. This patient received a compensation of $12,525,000. Two other outliers involved amputations, after arterial occlusion following arterial cannulation: in one case amputation of four fingers in a middle-aged patient (payment $11,550,000) and the other a below-knee amputation in an infant. The final payment ⬎$1 million involved a young adult patient with multiple IV placement attempts followed by diazepam phlebitis. This patient had thrombosed hand veins excised and developed reflex sympathetic dystrophy, resulting in the loss of hand and arm function.

DISCUSSION Complications secondary to peripheral vascular catheters contribute to injuries to patients and financial liability to anesthesiologists. This analysis of peripheral vascular catheter-related complications has identified important mechanisms, demographics, and types of complications. As expected, approximately half of the claims related to IV catheters involved the extravasation of drugs or fluids outside the intended vascular structure. Claims for peripheral vascular catheter complications comprise 2% of all claims in the ASA Closed Claims database. Although this proportion may seem small, it is similar in magnitude to those arising from gas delivery equipment11 and central venous catheters12 and more than from warming devices.13 Skin slough cases comprised the highest percentage of IV catheter claims. Thiopental was the most commonly reported drug in these claims. As the use of Vol. 109, No. 1, July 2009

thiopental as an induction drug declines, claims related to skin slough may decrease. A decreasing frequency of local complications of thiopental was reported even before the introduction of propofol.14 Extravasation and gangrene after peripheral administration of dopamine have been reported.15,16 If it is necessary to administer vasopressors into a peripheral vein in the urgent setting, the infusion pump should be set to detect small changes in infusion pressure to detect extravasation early.17,18 A simple clinical test to detect extravasation involves inflation of a blood pressure cuff proximal to the IV site. This impairs flow of a gravity-dependent infusion when the cannula is intravascular, but has no effect in an extravascular location.19 Successful use of transdermal nitroglycerin for prevention and treatment of phlebitis and extravasation has been described20,21 and may be considered when the intravascular location of an IV catheter is in doubt. Also, an easily avoidable complication of thrombophlebitis/swelling is burn injury due to heat compresses. Cheney et al.13 reported burn injuries due to direct application of warmed IV fluid bags and bottles to skin and cautioned that these devices should not be used for patient warming. Claims due to air embolism had the highest median compensation and a 100% rate of payment-per-claim. Several of these resulted from air in the blood bag from the shed red cell recovery devices. Air embolism from IV infusion may be potentially preventable by meticulous attention to de-airing the infusion set and IV/blood bags and incorporating autoventing filters in pressure infusion devices.22,23 Cardiac surgery cases represented the largest single surgical case type among peripheral catheter malpractice claims. Although the database lacks specific data on arm tucking in each case, we speculate that this finding is the result of the common practice of arm tucking during cardiac surgery, which results in the inability to monitor IV catheters intraoperatively. The Australian Incident Monitoring study also observed © 2009 International Anesthesia Research Society

127

33 cases involving peripheral venous access, including 14 cases of extravasation in their analysis of 2000 incidents.24 There were compartment syndrome claims in which there were warning signs of infiltration such as a slowly dripping IV. A low threshold to visually check the site of peripheral IV when the IV is dripping slower than normal may help detect an IV infiltration early on, and thus prevent the development of a compartment syndrome. There were surprisingly few claims related to arterial vascular catheterization. Our findings of limited liability associated with radial arterial cannulation are consistent with prospective reports of the safety of radial artery cannulation.25,26 Although partial or complete radial artery occlusion after decannulation occurred in a quarter of almost 1700 patients, no ischemic damage to the hand or disability occurred in any of the patients.25 In contrast, liability associated with femoral arterial catheters may be greater. Cases of intraperitoneal hemorrhage from laceration of the external iliac artery27 and retroperitoneal hematoma28 from femoral artery catheterization have been reported. As a retroperitoneal hematoma may be concealed, anesthesiologists should be alert to this possibility. Other preventive measures may include puncturing the femoral artery below the inguinal ligament rather than above it; avoiding transfixion and thus puncture of the posterior wall; and adequate compression in case of unsuccessful cannulation.29 The limitations of interpreting the data gathered from the ASA Closed Claims Project Database have been described.8,9 The database does not have data on the total number of adverse outcomes (the numerator) or the total number of anesthetics performed (the denominator), thus making it impossible to provide any numerical estimates of the risks associated with peripheral vascular catheterizations. Our data are retrospective, gathered over a time span of more than three decades and were collected in a nonrandom manner from direct participants. Finally, the database has only that information which the reviewer could obtain from the insurance company files. Incompleteness of specific detailed information regarding the sequence of events or mechanism of injury makes closed claims analysis weaker than prospectively collected data. Although our data cannot be used for establishing cause-and-effect relations, patterns of injury in this study of peripheral vascular catheter complications have identified important preventable patient complications, such as air embolism, burn injuries due to heat compresses for infiltrations/thrombophlebitis, and compartment syndromes. In summary, claims related to IV catheters were an important source of liability for anesthesiologists. Approximately half of these complications involved the extravasation of drugs or fluids. Claims related to radial artery catheterization were uncommon. 128

Intravenous/Arterial Catheter Liability

ACKNOWLEDGMENTS The authors acknowledge the contributions of John Campos, MS for data management and Lynn Akerlund for secretarial assistance. They are members of the Closed Claims Project research staff in the Department of Anesthesiology at the University of Washington, Seattle, Washington. The authors also acknowledge the closed claims reviewers from the American Society of Anesthesiologists and participation of the following liability insurance companies who have given permission to be acknowledged: Anesthesia Service Medical Group, Inc., San Diego, California; Armed Forces Institute of Pathology, Silver Spring, Maryland; Ascension Health, St. Louis, Missouri; COPIC Insurance Company, Denver, Colorado; Department of Veterans Affairs, Washington, DC; The Doctors’ Company, Napa, California; ISMIE Mutual Insurance Company, Chicago, Illinois; MAG Mutual Insurance Company, Atlanta, Georgia; Medical Liability Mutual Insurance Company, New York, New York; Midwest Medical Insurance Company, Minneapolis, Minnesota; Mutual Insurance Company of Arizona, Phoenix, Arizona; NORCAL Mutual Insurance Company, San Francisco, California; Pennsylvania Medical Society Liability Insurance Company, Mechanicsburg, Pennsylvania; Physicians Insurance A Mutual Company, Seattle, Washington; Physicians Insurance Company of Wisconsin, Madison, Wisconsin; Preferred Physicians Medical Risk Retention Group, Shawnee Mission, Kansas; ProMutual (Medical Professional Mutual Insurance Company), Boston, Massachusetts; Risk Management Foundation, Cambridge, Massachusetts; State Volunteer Mutual Insurance Company, Brentwood, Tennessee; The University of Texas Medical System, Austin, Texas; Utah Medical Insurance Association, Salt Lake City, Utah. REFERENCES 1. Steinmann G, Charpentier C, O’Neill TM, Bouaziz H, Mertes PM. Liposuction and extravasation injuries in ICU. Br J Anaesth 2005;95:355–7 2. Frezza EE, Mezghebe H. Indications and complications of arterial catheter use in surgical or medical intensive care units: analysis of 4932 patients. Am Surg 1998;64:127–31 3. Hampton AA, Sherertz RJ. Vascular-access infections in hospitalized patients. Surg Clin North Am 1988;68:57–71 4. Mangar D, Laborde RS, Vu DN. Delayed ischaemia of the hand necessitating amputation after radial artery cannulation. Can J Anaesth 1993;40:247–50 5. Scheer B, Perel A, Pfeiffer UJ. Clinical review: complications and risk factors of peripheral arterial catheters used for haemodynamic monitoring in anaesthesia and intensive care medicine. Crit Care 2002;6:199 –204 6. Adhikary GS, Massey SR. Massive air embolism: a case report. J Clin Anesth 1998;10:70 –2 7. Barbut F, Pistone T, Guiguet M, Gaspard R, Rocher M, Dousset C, Meynard JL, Carbonell N, Maury E, Offenstadt G, Poupon R, Frottier J, Valleron AJ, Petit JC. Complications due to peripheral venous catheterization. Prospective study [in French]. Presse Med 2003;32:450 – 6 8. Cheney FW, Posner K, Caplan RA, Ward RJ. Standard of care and anesthesia liability. JAMA 1989;261:1599 – 603 9. Cheney FW. The American Society of Anesthesiologists Closed Claims Project: what have we learned, how has it affected practice, and how will it affect practice in the future? Anesthesiology 1999;91:552– 6 10. Posner KL, Sampson PD, Caplan RA, Ward RJ, Cheney FW. Measuring interrater reliability among multiple raters: An example of methods for nominal data. Stat Med 1990;9:1103–15 (Erratum: Stat Med 1992;11:1401)


11. Caplan RA, Vistica MF, Posner KL, Cheney FW. Adverse anesthetic outcomes arising from gas delivery equipment: a closed claims analysis. Anesthesiology 1997;87:741– 8 12. Domino KB, Bowdle TA, Posner KL, Spitellie PH, Lee LA, Cheney FW. Injuries and liability related to central vascular catheters: a closed claims analysis. Anesthesiology 2004;100: 1411– 8 13. Cheney FW, Posner KL, Caplan RA, Gild WM. Burns from warming devices in anesthesia. A closed claims analysis. Anesthesiology 1994;80:806 –10 14. Davies DD. Local complications of thiopentone injection. A further report. Br J Anaesth 1979;51:1147–9 15. Siwy BK, Sadove AM. Acute management of dopamine infiltration injury with Regitine. Plast Reconstr Surg 1987;80:610 –2 16. Bhatt-Mehta V, Nahata MC. Dopamine and dobutamine in pediatric therapy. Pharmacotherapy 1989;9:303–14 17. Philip JH. Model for the physics and physiology of fluid administration. J Clin Monit 1989;5:123–34 18. Scott DA, Fox JA, Philip BK, Lind LJ, Cnaan A, Palleiko MA, Stelling JM, Philip JH. Detection of intravenous fluid extravasation using resistance measurements. J Clin Monit 1996;12:325–30 19. Goodie DB, Philip JH. Is the i.v. obstructed or infiltrated? A simple clinical test. J Clin Monit 1995;11:47–50 20. Denkler KA, Cohen BE. Reversal of dopamine extravasation injury with topical nitroglycerin ointment. Plast Reconstr Surg 1989;84:811–3

Vol. 109, No. 1, July 2009

21. Tjon JA, Ansani NT. Transdermal nitroglycerin for the prevention of intravenous infusion failure due to phlebitis and extravasation. Ann Pharmacother 2000;34:1189 –92 22. Rieger A, Philippi W, Spies C, Eyrich K. Safe and normothermic massive transfusions by modification of an infusion warming and pressure device. J Trauma 1995;39:686 – 8 23. Pedersen NT, Hessov I. Venous air embolism through infusion sets. Theoretical considerations, model experiments and prevention. Acta Anaesthesiol Scand 1978;22:117–22 24. Singleton RJ, Webb RK, Ludbrook GL, Fox MA. The Australian Incident Monitoring Study. Problems associated with vascular access: an analysis of 2000 incident reports. Anaesth Intensive Care 1993;21:664 –9 25. Slogoff S, Keats AS, Arlund C. On the safety of radial artery cannulation. Anesthesiology 1983;59:42–7 26. Martin C, Saux P, Papazian L, Gouin F. Long-term arterial cannulation in ICU patients using the radial artery or dorsalis pedis artery. Chest 2001;119:901– 6 27. Zavela NG, Gravlee GP, Benckart DH, Park SB, Gahtan V. Case 3–1996. Unusual cause of hypotension after cardiopulmonary bypass. J Cardiothorac Vasc Anesth 1996;10:553– 6 28. Illescas FF, Baker ME, McCann R, Cohan RH, Silverman PM, Dunnick NR. CT evaluation of retroperitoneal hemorrhage associated with femoral arteriography. AJR Am J Roentgenol 1986;146:1289 –92 29. Muralidhar K. Complication of femoral artery pressure monitoring. J Cardiothorac Vasc Anesth 1998;12:128 –9


129

Eliminating Arterial Injury During Central Venous Catheterization Using Manometry Catalin S. Ezaru, MD Michael P. Mangione, MD Todd M. Oravitz, MD James W. Ibinson, MD Richard J. Bjerke, MD

BACKGROUND: Unintended arterial puncture occurs in 2%– 4.5% of central venous catheterizations, resulting in arterial injury in 0.1%– 0.5% of patients. Routine performance of manometry during catheterization may successfully identify unintended arterial puncture and avoid arterial cannulation and injury. METHODS: We conducted a retrospective review of all cases of central venous catheter placement during a 15-yr period after implementation of a safety program requiring mandatory use of manometry to verify venous access. Arterial injuries were defined as unintended arterial cannulations with a 7-French or larger catheter or dilator. Arterial punctures were defined as the unintended placement of an 18-gauge catheter or needle into the artery. Data were reviewed for all arterial injuries during the entire 15-yr period. In addition, data on both arterial puncture and subsequent arterial injury were evaluated during the final year of analysis. RESULTS: A total of 9348 central venous catheters were placed during the observation period. During the full 15 yr of observation, there were no cases of arterial injury. During the final year of assessment, 511 central venous catheters were placed, with arterial punctures in 28 patients (5%). Arterial puncture was recognized without manometry in 24 cases. Arterial puncture was identified only with manometry in 4 cases, with no incidents of arterial injury. CONCLUSIONS: Consistent use of manometry, to verify venous placement, during central venous catheterization effectively eliminated arterial injury from unintended arterial cannulation during the 15-yr assessment. (Anesth Analg 2009;109:130 –4)

I

n the United States, more than 5 million central venous catheters are inserted annually for monitoring, fluid resuscitation, drug administration, dialysis and diagnostic studies, with complications occurring in 3%–25% of patients.1,2 Large-scale retrospective reviews of central venous catheterizations identified unintended arterial puncture in 2%– 4.5% of patients and large-bore catheter cannulation in 0.1%– 0.5% of patients.3,4 Arterial injuries result from the placement of a large bore catheter or a dilator into an artery, risking stroke or death, even if immediately recognized. Although the occurrence of arterial injury is relatively infrequent, the potential severity of outcome in these patients makes eliminating this risk a worthwhile goal. Performing central venous catheterization may be hindered by the inability to adequately assess the From the Department of Anesthesiology, Veterans Affairs Medical Center, University of Pittsburgh School of Medicine, Pittsburgh, Pennsylvania. Accepted for publication September 17, 2008. Address correspondence and reprint requests to Catalin S. Ezaru, MD, Department of Anesthesiology, 3A101, Veterans Affairs Medical Center, University Dr. C, Pittsburgh, PA 15240. Address e-mail to [email protected]. Copyright © 2009 International Anesthesia Research Society DOI: 10.1213/ane.0b013e31818f87e9

130

vasculature. The standard approach to differentiating venous from arterial cannulation is to assess the color and flow of the blood in an 18-gauge (or smaller) catheter or needle.5 Color and flow may be unreliable indicators due to profound hypotension, arterial desaturation, or catheter kinking or occlusion. Ultrasound guidance aids in the localization of the vessels and effectively reduces catheterization complications.6 Ultrasound has been consistently promoted as the primary technique for safely performing central venous catheterization6,7; however, it is infrequently used in clinical settings, even when the technology is available. A recent survey of the Society of Cardiovascular Anesthesiologists members revealed that 67% never or almost never used ultrasound when performing central venous catheterization, with only 15% always or almost always using ultrasound.8 Interestingly, most of those surveyed had experienced vascular complications during central venous catheterization, including carotid puncture (75%), carotid injury (3%), stroke (1%), and hemothorax (4%). The most common reasons for not using ultrasound were a perception that it was unnecessary (46%), lack of availability (18%), and time delay with addition of ultrasound (16%). Manometry is a simple and efficient technique to verify venous placement during catheterization. A 50 cm extension set is attached to an 18-gauge Vol. 109, No. 1, July 2009

Figure 1. Manometry verification of central venous location is demonstrated, including syringe aspiration into a 20-in. IV tubing (A) and visualization of a descending column of blood (B) to verify venous placement.

catheter or needle and the blood column in the elevated tubing is assessed. Performing manometry avoided carotid injury in a series of ⬎3000 patients undergoing internal jugular cannulations.9 This technique provides an effective assessment tool that could be readily used at the bedside without special equipment or training. We describe a 15-yr experience using manometry as part of a mandatory safety protocol to identify unintended arterial punctures in patients undergoing central venous catheterization at a university hospital. These data expand on results from an earlier report describing 11 yr of data, published as an abstract.10

METHODS A retrospective review of anesthesia department quality assurance and workload databases at a university-affiliated Veterans’ Administration hospital was conducted from January 1, 1992 to December 31, 2006. In 1991, unintended arterial cannulation during central venous catheterization resulted in patient stroke and subsequent death. This led to the implementation of a policy mandating manometry to verify venous placement before vessel cannulation. The hospital’s Quality Improvement Department made central venous catheterization safety a performance measure for the anesthesia department and also monitored the outcomes. During the catheterization procedure, the vein is cannulated with an 18-gauge catheter over the needle. After catheter placement, a 20-in. sterile IV tubing extension set with a syringe at one end is connected to the intravascular catheter and filled by aspiration of the syringe (Fig. 1A). The syringe is then disconnected, the IV tubing held vertically, and the blood column observed. Visualization of a descending column of blood indicates the catheter placement within a vein (Fig. 1B). With venous placement, the column of Vol. 109, No. 1, July 2009

blood moves with respiration, rising with coughing and descending with inspiration during spontaneous ventilation. Alternatively, during mechanical ventilation, the column of blood rises with inspiration and descends during exhalation if the catheter is IV. Regardless of the mode of ventilation, the overall movement of the blood column with IV placement is downward. Visualization of an ascending column of blood regardless of respiratory cycle indicates that the catheter is placed in an artery. Visualization of a static column of blood suggests catheter kinking, vessel wall impingement or partial catheter placement in the vessel. In this case the catheter is adjusted until continuous flow is demonstrated in the tubing and movement of the column of blood occurs. If this cannot be achieved, the catheter is removed and insertion reattempted. Guidewires are inserted only after a descending column of blood in the extension tubing is demonstrated. At this point, central venous catheterization is completed. Manometry adds ⬍1 min of additional time to the catheterization procedure. After January 1, 1992, all central venous catheters, by policy, were to be placed with this verification technique. Physician staff members were informed that procedural violations would not be tolerated. Compliance was monitored by direct observation by the department leadership, which remained constant over the entire 15-yr period. Also, certified registered nurse anesthetists and anesthesia technicians assisting with these procedures were expected to remind physicians of the requirement if necessary and/or report any noncompliance. All available data for central venous catheterizations were reviewed. Collected data included catheter location, catheter size, and experience level of staff performing catheterization. Staff members were divided into attending anesthesiologists and trainees, including medical students, residents, and fellows. All © 2009 International Anesthesia Research Society

131

Predictive analyses could not be performed since arterial injury did not occur.

Table 1. Characteristics of Central Venous Catheterization Procedures Characteristic

Number (%) total n ⫽ 9348

Location Internal jugular Subclavian Femoral Catheter size (French) 7 8.5–9 18 Staff performing procedure Attending anesthesiologist Supervised trainee

9144 (98) 87 (1) 117 (1) 4793 (51) 4403 (47) 152 (2) 6291 (67) 3057 (33)

procedures performed by trainees were supervised by an attending anesthesiologist. Arterial injury with a large-bore catheter and subsequent morbidity and mortality data were also collected. Arterial injury was defined as the occurrence of unintended cannulation of an artery with a large-bore catheter (7 French or larger) or a dilator. A secondary analysis was performed for central venous catheters placed in 2006. Additional database information was available on the occurrence of arterial puncture and subsequent complications from arterial puncture for this 1-yr period. Arterial punctures were defined as the unintended placement of an 18-gauge catheter or needle into the artery. These data were not available for the previous years of assessment. During this 1-yr assessment, data on arterial puncture with or without subsequent arterial injury and the incidence of arterial injury were both recorded.

RESULTS A total of 9348 central venous catheterizations were performed with manometry during the 15-yr assessment period. Large-bore catheters used for these procedures included 7 French double-lumen and triple-lumen catheters, 8.5 and 9 French sheath introducers for pulmonary artery catheterization (Arrow International, Reading, PA), and 18 French cannulas used for veno-veno bypass in liver transplantation (Edwards Lifesciences LLC, Irvine, CA). Catheterization details are shown in Table 1. The vast majority of catheters were placed in the internal jugular vein. Most catheters were 7–9 French, with ⬍2% using 18 French catheters. Two-thirds of catheters were placed by attending anesthesiologists. Thirty-eight individual anesthesiologists were involved in performing and/or supervising the placement of central venous catheters.

Arterial Injury There were no arterial injuries with any of the central venous catheterizations over this 15-yr period. 132

Safety of Central Venous Catheterization

Arterial Puncture From January 1, 2006 to December 31, 2006, 511 central venous catheters were placed. Similar to data for the full study, central venous catheters placed in 2006 were primarily 7 French (n ⫽ 159, 31%) and 9 French (n ⫽ 297, 58%), with a minority 18 French (n ⫽ 55, 11%). Arterial punctures occurred with the 18gauge catheter in 28 patients (5%). Catheter location and staff performing the procedures were similar for those cases resulting in arterial puncture (Table 2). There were no clinical sequelae associated with arterial puncture in any of the patients. Arterial puncture was easily identified by bright red blood spurting from the catheter in 24 patients. In 4 of the 28 patients, arterial puncture was not identified based on blood flow and color but only after performing manometry. Two of these 4 cases were performed by attending anesthesiologists and 2 by supervised trainees.

DISCUSSION This report provides 15 yr of data from more than 9000 patients who underwent central venous catheterization guided by manometry to verify venous, rather than arterial, needle placement before vessel dilation. Previously published reports of 0.1%– 0.5% incidence of arterial injury with central venous catheterization would predict between 9 and 47 cases of arterial injury in this sample. Indeed, a subanalysis of patients with recorded arterial puncture during the final year of the review predicted a potential incidence of 0.8% arterial injuries, for as many as 56 possible cases for the entire 15 yr period. The use of manometry resulted in no cases of arterial injury, regardless of catheter location and training level of the staff member performing the catheterization. This excellent outcome occurred despite the variations in catheter location and size used in this study, as well as the presence of multiple practitioners of varying levels of training. The occurrence of arterial puncture and predicted potential arterial injury without manometry in the current report were similar to reports of mechanical complications with catheterization.4 These data support and extend the data previously published on manometry verification of venous placement, by providing a larger sample size and a variety of catheterization procedures and providers. Data reported by Fabian and Jesudian9 included only internal jugular catheterizations. Although the vast majority of procedures in the current study likewise involved internal jugular catheterization, more than 200 procedures with subclavian or femoral sites were also reported, with similarly good outcome at these sites. Data on arterial puncture from the single year of observation reported similar catheter location and staff performing procedures that did and did not result in arterial ANESTHESIA & ANALGESIA

Table 2. Central Venous Catheterizations in 2006 Location (%) All catheterizations n ⫽ 511 Catheterizations resulting in arterial puncture n ⫽ 28

Internal jugular

Subclavian

Femoral

Attending

Trainee

504 (98) 28 (100)

3 (1) 0

4 (1) 0

327 (64) 20 (71)

184 (36) 8 (29)

puncture. These data suggest that vigilance in preventing arterial injury is needed for all attempts at central venous catheterization, regardless of location or experience of staff performing the procedure. The dilators used to aid in the insertion of the large bore catheters are sometimes blamed when arterial injuries occur. Some authors believe that after the guidewire is placed in the vein, the stiff dilator will bend the wire, go through the vein and into the artery.11,12 Oropello et al.11 have advocated changes in the design of the dilators in order to make them shorter and possibly reduce the incidence of arterial injury. Our results argue against this hypothesis of the dilator “jumping” from the vein into the artery since verifying venous placement with manometry would not prevent this mechanism of arterial injury. In fact, no other venous verification technique (transduction, real-time ultrasound, fluoroscopy) would prevent the dilator from exiting the vein and causing an arterial injury since that would occur after the venous verification takes place. Previous reports showed that venous verification before dilation is paramount in avoiding arterial injury4,6,9,10 and that the potential role played by the length or size of the dilator is over-stated. Despite the success of manometry for avoiding arterial injury in the current report, it is important to recognize that this technique does not prevent the occurrence of arterial puncture. Ready identification of arterial punctures and needle removal, however, typically prevents the development of complications. It is also important to recognize that manometry does not diminish the value of ultrasound. Ultrasound evaluation provides a comprehensive assessment of venous structures and allows for a more efficient procedure for practitioners with limited experience or in patients with difficult anatomy.6 However, the poor adoption of ultrasound reported by Bailey et al.8 suggests a need to promote a more user-friendly and efficient alternative. The ability to routinely use manometry for all central venous catheter placements makes manometry an attractive technique to prevent serious and potentially life-threatening complications, such as arterial injury. Data from this study support the benefit of consistent use of manometry for all central venous catheter placements. Inclusion of this simple and efficient technique into a standard safety protocol for central venous catheterization may increase the likelihood of compliance by practitioners. Vol. 109, No. 1, July 2009

Staff (%)

This report is limited by the inherent flaws of observational studies, including limited access to patient variables and outcomes and inability to verify accuracy of data. Although only a limited number of outcome measures was available from the accessed database for this study, the current analysis provides a long experience with routine use of manometry in all patients receiving central venous catheters. It has now been more than 20 yr since manometry was first described,9 but neither manometry nor other venous verification techniques such as real-time ultrasound are widely used in clinical practice.8 This is mainly due to additional time required to perform them and to the perception that arterial injuries are rare complications of central line insertions that do not justify additional steps for safety. Our experience is the largest published to date that shows that arterial injuries are completely preventable despite the involvement of multiple practitioners and trainees, but only if the verification procedure is simple, efficient and performed without fail. In the current maturing health care environment that emphasizes patient safety and zero tolerance for preventable errors, even the low risk of arterial injury during central venous catheterizations should no longer be accepted. In summary, manometry is a simple technique that rapidly and effectively identifies venous access during central venous catheterization. Standard use of manometry during catheterization can eliminate the risk of arterial injury from unrecognized arterial cannulation. Manometry can be successfully used by practitioners with different levels of experience and for all catheterization locations. REFERENCES 1. McGee DC, Gould MK. Preventing complications of central venous catheterization. N Engl J Med 2003;348:1123–33 2. Domino KB, Bowdle TA, Posner KL, Spitellie PH, Lee LA, Cheney FW. Injuries and liability related to central vascular catheters: a closed claim analysis. Anesthesiology 2004;100:1411– 8 3. Golden LR. Incidence and management of large-bore introducer sheath puncture of the carotid artery. J Cardiothorac Vasc Anesth 1995;9:425– 8 4. Jobes DR, Schwartz AJ, Greenhow DE, Stephenson LW, Ellison N. Safer jugular vein cannulation: recognition of arterial puncture and preferential use of external jugular route. Anesthesiology 1983;59:353–5 5. Mark JB, Slaughter TF. Cardiovascular monitoring. In: Miller RD, ed. Miller’s Anesthesia. Philadelphia: Churchill Livingstone, 2005:1265–362 6. Hind D, Calvert N, McWilliams R, Davidson A, Paisley S, Beverly C, Thomas S. Ultrasonic devices for central venous cannulation: meta-analysis. BMJ 2003;327:361–7 © 2009 International Anesthesia Research Society

133

7. Feller-Kopman D. Ultrasound-guided central venous catheter placement: the new standard of care? Crit Care Med 2005;33:1875–7 8. Bailey PL, Glance LG, Eaton MP, Parshall B, McIntosh S. A survey of the use of ultrasound during central venous catheterization. Anesth Analg 2007;104:491–7 9. Fabian JA, Jesudian MC. A simple method for improving the safety of percutaneous cannulation of the internal jugular vein. Anesth Analg 1985;64:1032–3 10. Bjerke R, Mangione M, Oravitz T. Major arterial injury need not be a risk of central venous catheterization (abstract). Anesth Analg 2004;98:SCA 98

134

Safety of Central Venous Catheterization

11. Oropello JM, Leibowitz AB, Manasia A, Del Guidice R, Benjamin E. Dilator-associated complications of central vein catheter insertion: possible mechanisms of injury and suggestions for prevention. J Cardiothor Vasc Anesth 1996;10:634 –7 12. Kulvatunyou N, Heard SO, Bankey PE. A subclavian artery injury, secondary to internal jugular vein cannulation, is a predictable right-sided phenomenon. Anesth Analg 2002;95:564 – 6


Case Report

Carotid Dissection: A Complication of Internal Jugular Vein Cannulation with the Use of Ultrasound Andrea J. Parsons, MD John Alfa, MBBS, DA, FRCA

Central venous catheters (CVCs) are often used in intensive care units and operating rooms. They facilitate hemodynamic monitoring, administration of fluids and medications, transvenous pacing and renal fluid replacement therapy. Severe complications can arise from inserting CVCs, some of which may be life threatening. A safe insertion technique with confirmation of correct placement of these catheters is of utmost importance. We present an obese 66-yr-old man who had carotid artery dissection with compromised cerebral circulation after CVC insertion under ultrasound guidance. The dissection was immediately repaired with no neurological sequelae to the patient. (Anesth Analg 2009;109:135–6)

CASE REPORT A morbidly obese 66-yr-old man (Body MA Index 56 kg/m2) with a thoracoabdominal aortic aneurysm and Type B dissection was scheduled for elective endovascular repair. He had a history of aortic dissection and total aortic arch repair 2 yr earlier. His other co-morbidities included diabetes mellitus, hypertension, chronic obstructive pulmonary disease, a history of congestive heart failure, and depression. The anesthetic plan consisted of general anesthesia and tracheal intubation. Preoperatively, a radial arterial catheter and spinal drain were placed. After induction of general anesthesia, a central venous catheter (CVC) was introduced for hemodynamic monitoring and possible intravascular volume resuscitation. Ultrasound (Site Rite II, Bard Access Systems, Salt Lake City, UT) was used to define the neck anatomy for the placement of a CVC in the right internal jugular (RIJ) vein. The patient was placed in Trendelenburg position, and the neck was prepared and draped for CVC insertion. Anatomic landmarks were difficult to define due to the patient’s short, thick neck. The ultrasound was used to confirm the location of the RIJ vein and the carotid artery. A 22-gauge needle was used as a finder needle before the larger, thin-walled needle, and a syringe (Arrow International, Reading, PA) was used to locate the RIJ vein. A 9-French introducer sheath was inserted into the neck using a wire and dilator (Seldinger technique). The ultrasound was not used in real-time during this procedure. The introducer was placed with a second provider holding traction on the neck to assist catheter insertion. Blood aspirated from the side port of the catheter appeared dark in color, and this was thought to be consistent with venous blood. Additionally, the side port flushed easily. The catheter was sutured in place, and dressing was applied. A bag of lactated Ringer’s solution was connected to the side arm of the introducer, but the fluid did not flow freely. From the Department of Anesthesiology, University of Michigan Hospital, Ann Arbor, Michigan. Accepted for publication March 9, 2009. Address correspondence and reprint requests to Andrea Parsons, MD, Department of Anesthesiology, 1H247 UH, 1500 E. Medical Center Dr., Ann Arbor, MI 48109. Address e-mail to [email protected]. Copyright © 2009 International Anesthesia Research Society DOI: 10.1213/ane.0b013e3181a7f5a4

Vol. 109, No. 1, July 2009

At this point, we suspected that the CVC may be in the carotid artery. A transducer was connected to the CVC, but a central venous wave form was not present. The transduced pressure was 90 mm Hg, but the tracing lacked a classical pulsatile wave form. This pressure correlated with the mean arterial blood pressure measured at the left radial artery. Blood was drawn from the CVC and the radial arterial catheter. The comparison of the two blood gas samples confirmed that the blood obtained from the CVC was arterial (CVC blood: pH 7.41, Pco2 50, Po2 243 and radial arterial catheter blood: pH 7.41, Pco2 50, Po2 250). The surgeons were immediately informed of possible carotid artery cannulation. Contrast-enhanced neck radiograph further confirmed common carotid arterial cannulation. The aortic endovascular repair was postponed, and the surgeons decided to explore and remove the catheter that had dissected the common carotid artery. They noted that the carotid artery superior to the point of entry of the catheter was pulseless with little to no blood flow into the cerebral circulation. The 9-Fr catheter had traversed the posterior wall of the internal jugular vein (IJV), and entered the common carotid artery just distal to the carotid bulb (Fig. 1). The catheter had dissected the posterior wall of the carotid artery, with the tip lodged in the dissection flap. Color flow ultrasound confirmed a moderate dissection plane. The RIJ vein and the carotid artery were then repaired surgically. Good blood flow through the artery was reestablished. The patient had no neurological or further vascular sequelae, and he returned to the operating room 3 days later to have the aortic endovascular repair performed. He did well postoperatively and was discharged from the hospital 5 days later.

DISCUSSION CVCs are often used in intensive care units and operating rooms but are not without complications. Most of these complications are mechanical and occur during insertion of the catheter. The incidence of carotid artery cannulation with a large-bore catheter is reported as 1–7 cases per 1000 attempts of line placement in the IJV.1,2 Severe complications from iatrogenic carotid artery injury include life-threatening hemorrhage, stroke, 135

Figure 1. Surgical dissection reveals the central venous catheter (CVC) traversing through the right internal jugular vein (RIJ) into the right common carotid artery (RCC).

pseudoaneurysm, arteriovenous fistula, arterial dissection, and airway compromise.3 Thus, vascular surgeons should be informed of any possible cases of carotid cannulation with large-bore catheters. Surgical management of this complication often depends on whether the catheter is obstructing blood flow through the carotid artery.3 This, in turn, depends on the size of the catheter used and whether the artery is diseased and narrowed. In this case, the introducer likely caused a dissection and was large enough to obstruct blood flow. With the catheter tip lodged in the dissection flap in the false lumen, transducing the distal lumen of the CVC showed a high pressure without a pulsatile wave form. The likely reason for this is the interruption to blood flow created by the catheter in the dissection flap. With no continuous column of blood flow through the false lumen and the catheter, there was no notable pulsatile wave form, despite a high pressure reading. Additionally, this patient was known to have vaso-occlusive disease, and his carotid arteries may well have been narrowed by this disease. As soon as reduced flow through the carotid artery was noted, the anesthetic management was adjusted to increase mean arterial blood pressures to promote collateral flow in the cerebral vasculature and help preserve blood flow to the brain during the repair of the dissection. Many reports have shown that the use of ultrasound before vessel puncture reduced the number of needle passes and complications.4 Ultrasound has been shown to decrease the incidence of arterial puncture and increase the success rate of CVC insertion, particularly in patients with poor anatomic landmarks.5 Also, using ultrasound in real-time is becoming a popular technique because many studies have shown that real-time ultrasonography further increased the success 136

Case Report

rate and reduced complications (defined as carotid artery puncture, hematoma, pneumothorax, hemothorax, and catheter malposition).6,7 When ultrasound is used in real-time, the depth and angle of the needle in relation to the carotid artery can be assessed. This can help map the course of the vein and reduce the likelihood of traversing the vein and puncturing the carotid artery. Additionally, the American College of Surgeons recommends real-time ultrasonography for insertion of CVCs into the IJV in adults and children in elective circumstances.8 When a practitioner is uncomfortable holding an ultrasound probe while attempting cannulation, a second provider who is experienced with ultrasonography can help with this process.9 A second provider can prove to be helpful, especially in obese patients with short, thick necks. Ultrasound can be used to confirm guidewire position in the IJV before catheter insertion. Also, ultrasound has been used to assist guidewire placement when resistance was encountered during insertion.10 This case demonstrates that serious complications with potentially fatal outcomes can occur with CVC placement despite ultrasound guidance. Using ultrasound to assess anatomy before venous puncture and continuing to use ultrasound throughout the entire procedure, including real-time ultrasonography, can reduce the risk of devastating complications with CVC placement. REFERENCES 1. Shah KB, Rao TL, Laughlin S, El-Etr AA. A review of pulmonary artery catheterization in 6,245 patients. Anesthesiology 1984;61: 271–5 2. Golden LR. Incidence and management of large-bore introducer sheath puncture of the carotid artery. J Cardiothorac Vasc Anesth 1995;9:425– 8 3. Mussa FF, Towfigh S, Rowe VL, Major K, Hood DB, Weaver FA. Current trends in the management of iatrogenic cervical carotid artery injuries. Vasc Endovascular Surg 2006;40:354 – 61 4. Keenan SP. Use of ultrasound to place central lines. J Crit Care 2002;17:126 –37 5. Hayashi H, Amano M. Does ultrasound imaging before puncture facilitate internal jugular vein cannulation? Prospective randomized comparison with landmark-guided puncture in ventilated patients. J Cardiothorac Vasc Anesth 2002;16:572–5 6. Karakitsos D, Labropoulos N, De Groot E, Patrianakos A, Kouraklis G, Poularas J, Samonis G, Tsoutsos D, Konstadoulakis M, Karabinis A. Real-time ultrasound-guided catheterization of the internal jugular vein: a prospective comparison with the landmark technique in critical care patients. Crit Care 2006; 10:R162 7. Leung J, Duffy M, Finckh A. Real-time ultrasonographicallyguided internal jugular vein catheterization in the emergency department increases success rates and reduces complications: a randomized, prospective study. Ann Emerg Med 2006;48:540 –7 8. American College of Surgeons. Statement on recommendations for uniform use of real-time ultrasound guidance for placement of central venous catheters. Bull Am Coll Surg 2008;93:35– 6 9. Mey U, Glasmacher A, Hahn C, Gorschluter M, Ziske C, Mergelsberg M, Sauerbruch T, Schmidt-Wolf IG. Evaluation of an ultrasound-guided technique for central venous access via the internal jugular vein in 493 patients. Support Care Cancer 2003;11:148 –55 10. Stone M. Identification and correction of guide wire malposition during internal jugular cannulation with ultrasound. CJEM 2007;9:131–2


Case Report

Refractory Anaphylactic Cardiac Arrest After Succinylcholine Administration Antoine Baumann* Daniela Studnicska, MD* Ge´rard Audibert, MD, PhD*

Refractory shock from anaphylaxis can occur after induction of general anesthesia. We report two cases of fatal cardiac arrest with increased blood tryptase and immunoglobulin E values after succinylcholine administration. Tryptase and immunoglobulin E assays may help to identify anaphylactic reactions when cardiac arrest occurs at induction of anesthesia. (Anesth Analg 2009;109:137–40)

Attila Bondar, MD, FCARCSI* Yannick Fuhrer, MD* Jean-Pierre Carteaux, MD, PhD† Paul M. Mertes, MD, PhD*

A

naphylactic reactions occurring during anesthesia can be life-threatening and require urgent therapy to avoid a fatal outcome. Anaphylactic reactions presenting as cardiac arrest resistant to cardiopulmonary resuscitation (CPR) despite prompt diagnosis and appropriate treatment occur but are rarely described in medical literature. We report two cases of anaphylactic reactions with fatal cardiac arrest after succinylcholine administration.

CASE DESCRIPTION Case 1 A 74-yr-old woman (1.70 m, 79 kg) was admitted to the hospital with acute appendicitis and scheduled for emergency surgery. Her medical history revealed dyslipidemia treated with fenofibrate and hypertension well controlled with diuretics (spironolactone and furosemide). Her surgical history consisted of two previous interventions 29 and 15 yr ago without any complications according to the patient. No atopic history or preexisting allergy was reported. General anesthesia with rapid sequence induction was planned. Anesthesia was induced with propofol (2 mg/kg), cricoid pressure applied, and succinylcholine (1 mg/kg) administered. Tracheal intubation was performed 1 min after succinylcholine injection and correct placement of the endotracheal From the *De´partement d’Anesthe´sie-Reánimation, Hoˆpital Central, Centre Hospitalier Universitaire, 29, avenue du Marećhal de Lattre de Tassigny, Nancy; and †Service de Chirurgie Cardiovasculaire et Transplantation, Hoˆpitaux de Brabois, Centre Hospitalier Universitaire de Nancy, 54500 Vandoeuvre-les-Nancy, France. Accepted for publication March 5, 2009. Address correspondence and reprint requests to Antoine Baumann, MD, DESA, De´partement d’Anesthe´sie-Reánimation, Hoˆpital Central, Centre Hospitalier Universitaire, 29, avenue du Marećhal de Lattre de Tassigny, 54000 Nancy, France. Address e-mail to [email protected]. Copyright © 2009 International Anesthesia Research Society DOI: 10.1213/ane.0b013e3181a775b2

Vol. 109, No. 1, July 2009

tube was confirmed. Four minutes after succinylcholine administration, severe bronchospasm with high airway pressure and decreased end-tidal CO2 (ETco2) (from 32 to 15 mm Hg) was observed along with undetectable oxygen saturation, accompanied by hypotension (40/28 mm Hg), bradycardia (40 bpm), and no palpable carotid pulse. Despite the absence of any cutaneous signs, anaphylactic shock was immediately suspected and CPR was started without delay in accordance with international guidelines (Fig. 1)1: 1 mg boluses of IV epinephrine were given three times at 3-min intervals and intravascular volume administration with Ringer’s lactate solution (total 1500 mL over 90 min) was administered. This was followed by 2 mg and then by 5 mg boluses of epinephrine (total dose 25 mg). Chest compression was started immediately and maintained throughout CPR. Ventricular ectopy preceded ventricular fibrillation. Defibrillation using an external defibrillator was applied and repeated eight times for recurrent episodes of ventricular fibrillation. Ventricular fibrillation alternated with asystole throughout the CPR without return of spontaneous circulation. Two grams of calcium chloride and 250 mL sodium bicarbonate 8.4% were also administered. After 60 min of unsuccessful CPR, two boluses of 1 mg terlipressin at 5-min intervals were administered without any benefit. Despite 90 min continuous CPR, efficient cardiac activity was never restored, pulmonary edema occurred, and the patient died. Blood tests performed 70 min after the onset of the reaction showed a tryptase level of more than 200 ␮g/L (threshold: 13.5 ␮g/L) and a specific immunoglobulin E (IgE) for succinylcholine confirming the diagnosis of anaphylactic shock: quaternary ammonium (QA) fixation of 4.44% (⬎3% confirm allergic reaction) with succinylcholine inhibition of 92% (⬎20% confirm allergic reaction). Hemolysis of the blood sample prevented histamine measurements.

Case 2 A 49-yr-old man (1.72 m, 98 kg) was scheduled for emergency appendectomy. He had a medical history of hypertension treated with spironolactone, altizide (thiazide diuretic), and rilmenidine (␣2 blocking drug) and a congenital single kidney without renal function impairment. He had 137

Figure 1. Advanced life support cardiac arrest algorithm.1

no surgical history and no allergic reactions. General anesthesia was induced with propofol (2 mg/kg), cricoid pressure applied, and succinylcholine (1 mg/kg) was given to facilitate tracheal intubation. Endotracheal tube position was confirmed. A low ETco2 (25 mm Hg) was noticed immediately and extensive erythema with piloerection appeared concomitantly. Anaphylactic shock was immediately suspected. Epinephrine was administered as a 0.2 mg IV bolus, his arterial blood pressure collapsed (50/35 mm Hg), and Spo2 became undetectable. Pulseless electrical activity cardiac arrest occurred without any palpable carotid pulse. CPR with chest compression, rapidly increasing doses of IV epinephrine boluses (three boluses of 1 mg followed by 2 mg and then three boluses of 5 mg), and fluid administration with saline was administered1 (Fig. 1). Two minutes later ventricular fibrillation occurred. After 15 min of CPR, including cardioversion and IV amiodarone 300 mg, two boluses of 1 mg terlipressin at 5-min intervals were administered. Twenty milligram epinephrine, 2000 mL saline, 500 mL sodium bicarbonate 8.4%, and 2 g calcium chloride were administered. Emergency circulatory support was instituted after 75 min of CPR, using extracorporeal membrane oxygenation (ECMO). The ECMO circuit consisted of a centrifugal pump (Rota Flow RF-32, Jostra Medizintechnik AG, Hirrlingen, Germany) and a hollow-fiber membrane oxygenator (Quadrox D, Jostra Medizintechnik AG). His systolic blood 138

Case Report

pressure remained low (60 mm Hg) and pulmonary edema occurred. The patient died 12 h later from refractory shock and multiple organ failure, including acidosis, renal failure, rhabdomyolysis, and coagulopathy. Blood tests performed 50 min after the onset of the reaction showed a tryptase level over 200 ␮g/L and specific IgE against succinylcholine confirming the diagnosis of anaphylactic shock: QA fixation of 29.96% and succinylcholine inhibition of 36%.

DISCUSSION The risk of immediate hypersensitivity reaction during anesthesia is estimated to be between 1/10,000 and 1/20,000 anesthesia,2,3 and neuromuscular blocking drugs are the most frequently involved substances (50%–70% of cases), succinylcholine being involved in the majority of cases.4 The mortality of anesthesiarelated anaphylaxis ranges from 5% to 10%.5,6 Our two cases document the potential acuteness, suddenness, and severity of cardiovascular events in anaphylaxis, which can be refractory to epinephrine and vasopressin analogs even with prompt treatment. In our two cases ETco2 collapse was the first sign ANESTHESIA & ANALGESIA

preceding cardiac arrest. Several authors have underlined the value of this early clinical sign for the suspicion of anaphylactic shock, even in the absence of any cutaneous manifestations.7 Epinephrine-resistant anaphylactic shock is a strong incentive to search for possible alternative treatments.8 Methylene blue,9 glucagon,10 and ␣-agonists11 have recently been proposed as alternative therapeutic options. Successful results have also been reported using arginine-vasopressin12 or vasopressin analogues (terlipressin)13 in epinephrineresistant anaphylactic shock resuscitation. Argininevasopressin is not available in France, therefore we used terlipressin. Cardiopulmonary bypass and extracorporeal support have also been successfully used in one case after 60 min of futile resuscitation.14 However, prompt resuscitation, use of terlipressin, and circulatory support were ineffective in our cases. Tryptase is a marker of mast cell activation,15 with a specificity exceeding 93% for the diagnosis of anaphylaxis during anesthesia when a 20 ␮g/L threshold is used.16 Its use for the postmortem diagnosis of anaphylaxis has been questioned because of the possibility of nonspecific mast cell activation during resuscitation.17 However, one should emphasize that this nonspecific increase is very limited during resuscitation. Therefore, vastly increased serum tryptase levels as observed in our cases support the strong argument in favor of an allergic reaction.18,19 Identification of the causal agent is possible using specific IgE measurement, even when these tests are performed at the time of reaction.20 In our cases, the presence of specific QA IgE was investigated by the University Laboratory of Medical Biochemistry using radioimmunoassay based on coupling of a choline analog to Sepharose (quaternary ammonium Sepharoseradioimmunoassay, positive threshold 1.5%). The sensitivity and specificity of these assays have been estimated to be 88% and 97%, respectively.21 A combination of high tryptase and high IgE values allowed us to conclude definitively the diagnosis of anaphylaxis to succinylcholine. In anaphylaxis with early cardiac arrest, direct myocardial injury has been suggested.22 Several clinical observations and experimental studies support this hypothesis.23,24 Myocardial injury could be related to a high myocardial mast cell infiltration in these patients.25 Unfortunately, in our patients the degree of cardiac mast cell infiltration could not be investigated because of the refusal of autopsy. In conclusion, despite the immediate diagnoses and prompt therapy, anaphylaxis can be lethal. Anaphylactic reactions to succinylcholine have been described. We postulate that a direct deleterious effect of the various mediators on myocardial function is a potential cause of the refractory cardiac arrest.25 The failure of promptly instituted epinephrine, vasopressin analog, and even ECMO to treat these patients once again emphasizes the need for further studies on the mechanism of Vol. 109, No. 1, July 2009

severe anaphylaxis to identify possible alternative treatments to epinephrine. In case of unexpected death or cardiac arrest at induction of general anesthesia, blood tryptase and IgE measurements may help to identify possible anaphylactic reactions to neuromuscular blocking drugs. REFERENCES 1. Nolan JP, Deakin CD, Soar J, Bottiger BW, Smith G. European Resuscitation Council guidelines for resuscitation 2005. Section 4. Adult advanced life support. Resuscitation 2005;67:S39 –S86 2. Fisher MM, Baldo BA. The incidence and clinical features of anaphylactic reactions during anesthesia in Australia. Ann Fr Anesth Reanim 1993;12:97–104 3. Mitsuhata H, Hasegawa J, Matsumoto S, Ogawa R. [The epidemiology and clinical features of anaphylactic and anaphylactoid reactions in the perioperative period in Japan: a survey with a questionnaire of 529 hospitals approved by Japan Society of Anesthesiology]. Masui 1992;41:1825–31 4. Mertes PM, Laxenaire MC. Adverse reactions to neuromuscular blocking agents. Curr Allergy Asthma Rep 2004;4:7–16 5. Axon AD, Hunter JM. Editorial III: anaphylaxis and anaesthesia—all clear now? Br J Anaesth 2004;93:501– 4 6. Light KP, Lovell AT, Butt H, Fauvel NJ, Holdcroft A. Adverse effects of neuromuscular blocking agents based on yellow card reporting in the U.K.: are there differences between males and females? Pharmacoepidemiol Drug Saf 2006;15:151– 60 7. Mertes PM, Laxenaire MC. Allergic reactions occurring during anaesthesia. Eur J Anaesthesiol 2002;19:240 – 62 8. Dunser MW, Torgersen C, Wenzel V. Treatment of anaphylactic shock: where is the evidence? Anesth Analg 2008; 107:359 – 61 9. Buzato MA, Viaro F, Piccinato CE, Evora PR. The use of methylene blue in the treatment of anaphylactic shock induced by compound 48/80: experimental studies in rabbits. Shock 2005;23:582–7 10. Thomas M, Crawford I. Best evidence topic report. Glucagon infusion in refractory anaphylactic shock in patients on betablockers. Emerg Med J 2005;22:272–3 11. Klouche K, Weil MH, Sun S, Tang W, Zhao DH. A comparison of alpha-methylnorepinephrine, vasopressin and epinephrine for cardiac resuscitation. Resuscitation 2003;57:93–100 12. Schummer W, Schummer C, Wippermann J, Fuchs J. Anaphylactic shock: is vasopressin the drug of choice? Anesthesiology 2004;101:1025–7 13. Rocq N, Favier JC, Plancade D, Steiner T, Mertes PM. Successful use of terlipressin in post-cardiac arrest resuscitation after an epinephrine-resistant anaphylactic shock to suxamethonium. Anesthesiology 2007;107:166 –7 14. Lafforgue E, Sleth JC, Pluskwa F, Saizy C. [Successful extracorporeal resuscitation of a probable perioperative anaphylactic shock due to atracurium]. Ann Fr Anesth Reanim 2005;24: 551–5 15. Mertes PM, Laxenaire MC, Alla F. Anaphylactic and anaphylactoid reactions occurring during anesthesia in France in 1999 –2000. Anesthesiology 2003;99:536 – 45 16. Edston E, van Hage-Hamsten M. Beta-Tryptase measurements post-mortem in anaphylactic deaths and in controls. Forensic Sci Int 1998;93:135– 42 17. Randall B, Butts J, Halsey JF. Elevated postmortem tryptase in the absence of anaphylaxis. J Forensic Sci 1995;40:208 –11 18. Pumphrey RS, Roberts IS. Postmortem findings after fatal anaphylactic reactions. J Clin Pathol 2000;53:273– 6 19. Malinovsky JM, Decagny S, Wessel F, Guilloux L, Mertes PM. Systematic follow-up increases incidence of anaphylaxis during adverse reactions in anesthetized patients. Acta Anaesthesiol Scand 2008;52:175– 81 20. Guttormsen AB, Johansson SG, Oman H, Wilhelmsen V, Nopp A. No consumption of IgE antibody in serum during allergic drug anaphylaxis. Allergy 2007;62:1326 –30 21. Gueant JL, Mata E, Monin B, Moneret-Vautrin DA, Kamel L, Nicolas JP, Laxenaire MC. Evaluation of a new reactive solid phase for radioimmunoassay of serum specific IgE against muscle relaxant drugs. Allergy 1991;46:452– 8 © 2009 International Anesthesia Research Society

139

22. Mertes PM, Pinaud M. [What are the physiopathological mechanisms? How can severe anaphylactoid reactions be explained]. Ann Fr Anesth Reanim 2002;21:55s–72s 23. Assem ES, Ezeamuzie IC. Suxamethonium-induced histamine release from the heart of naive and suxamethonium-sensitized guinea-pigs: evidence suggesting spontaneous sensitization in naive animals, and relevance to anaphylactoid reactions in man. Agents Actions 1989;27:146 –9

140

Case Report

24. Porter JM, Magner J, Phelan D. Anaphylaxis due to suxamethonium—manifested at induction of anaesthesia by bradycardia and cardiac arrest. Ir J Med Sci 1999;168:99 –101 25. Marone G, Bova M, Detoraki A, Onorati AM, Rossi FW, Spadaro G. The human heart as a shock organ in anaphylaxis. Novartis Found Symp 2004;257:133– 49; discussion 149 – 60, 276 – 85


Critical Care and Trauma Section Editor: Jukka Takala

The Effects of Fenoterol Inhalation After Acid Aspiration-Induced Lung Injury Michael T. Pawlik, MD* Thomas Schubert, MD† Susanne Hopf, MD* Matthias Lubnow, MD‡ Michael Gruber, PhD* Christoph Selig, MD§ Kai Taeger, MD, PhD* Karl P. Ittner, MD*

BACKGROUND: Acid aspiration is a serious complication that can occur during general anesthesia. Studies show that ␤-agonists have beneficial effects on lung injury. Therefore, we tested the effect of the nebulized ␤-agonist fenoterol on lung variables in a rodent model of acid-induced lung injury. METHODS: In a prospective, randomized, and controlled study, we evaluated the effects of fenoterol inhalation on lung oxygenation, inflammation, and pulmonary histology in a rat model of acid-induced lung injury. Sprague-Dawley rats underwent sevoflurane anesthesia with tracheotomy and carotid catheter insertion. Lung injury was induced by instillation of 0.4 mL/kg 0.1 M hydrochloric acid. The lungs were ventilated for 6 h and randomized to receive either fenoterol inhalation 10 ␮g or saline inhalation, both at 15 and 180 min after acid aspiration. Mean arterial blood pressures and peak airway pressures were documented, arterial blood gases were determined at 30, 90, 180, 270, and 360 min, and postmortem histology was subsequently examined. Additionally, fenoterol concentrations in bronchoalveolar lavage fluid (BALF) and plasma were determined by liquid chromatography/tandem mass spectroscopy. After 360 min tumor necrosis factor (TNF)-␣ and interleukin (IL)-6 were determined in the BALF, and lungs were dried for determination of the wet/dry ratio. RESULTS: Inhalation treatment with 10 ␮g fenoterol significantly increased oxygenation after 270 and 360 min when compared with placebo. Fenoterol-treated rats showed a significant decrease in IL-6 and TNF-␣ levels and in the wet/dry weight ratio of the lungs. The histologic appearance showed significantly less interstitial edema and leukocyte infiltration in the fenoterol group. The concentration of fenoterol was 10.3 ␮g/L (median) in the BALF and ⬍1 ␮g/L in the plasma. CONCLUSIONS: Fenoterol inhalation improved oxygenation after 270 and 360 min, attenuated the release of TNF-␣ and IL-6, and diminished the lung edema and infiltration of polymorphonuclear leukocytes. (Anesth Analg 2009;109:143–50)

T

herapy for acute lung injury (ALI) after acid aspiration is still an important issue during anesthesia in the perioperative period. Although the incidence has decreased from 1 in 2.131 in 1986 to 1 in 3.216 today, it is still a potentially lethal complication.1 Despite intensive research, pharmacological treatment options are merely supportive, and the mortality rate is high in patients developing acute respiratory distress syndrome (ARDS). One pharmacological target is to reduce the inflammatory response after acid aspiration, which is marked by acute infiltration of macrophages and

From the Departments of *Anesthesiology, †Pathology, ‡Cardiology, Pulmonology and Intensive Care, University Hospital, Regensburg, Germany; and §Department of Anesthesiology, University Hospital Ulm, Ulm, Germany. Accepted for publication February 23, 2009. Supported exclusively by the Department of Anesthesiology, University Hospital, Regensburg, Germany. Address correspondence and reprint requests to Michael Pawlik, MD, DEAA, Department of Anesthesiology, University of Regensburg, Franz-Josef-Strauss-Allee 11, D-93053 Regensburg, Germany. Address e-mail to [email protected]. Copyright © 2009 International Anesthesia Research Society DOI: 10.1213/ane.0b013e3181a2a85d

Vol. 109, No. 1, July 2009

neutrophils into the alveolar space. Activation of neutrophils results in local production of cytokines and release of reactive oxygen species, which results in damage to pulmonary cells in the affected lung. Several studies have demonstrated that ␤2-agonists have beneficial effects on a variety of factors that contribute to the severity of ARDS.2,3 In several murine models, pretreatment with ␤ adrenergics reduced the number of neutrophils and inflammatory cytokines, such as interleukin (IL)-6 in the bronchoalveolar fluid, and stabilized epithelial and endothelial functions.4 Furthermore, ␤ agonists induce bronchodilation and have, therefore, been mainstay medications in the treatment of asthma and chronic obstructive pulmonary disease for many years, and there is increasing interest in their potential role for pharmacological intervention in patients with ARDS.5 Previous in vitro studies found that administration of ␤2receptor agonists led to increased intracellular levels of cyclic adenosine monophosphate, which resulted in a dose-related reduction in neutrophil-endothelial cell adhesion.6 Although many mechanisms of the pharmacological effects of ␤2-agonists have been elucidated, many questions still remain concerning the 143

optimal drug, route, and dose-related effects in patients with ALI. One study showed that IV administered salbutamol reduced lung water in patients with ALI and ARDS, but caused undesired side effects, such as cardiac arrhythmias.7 Administration of ␤agonists by inhalation seems to be advantageous, because higher concentrations can be achieved in the lung tissue and simultaneously minimizes systemic adverse effects. Besides long-acting ␤-agonists like formoterol or salmeterol, short-acting adrenoceptor agonists, such as terbutaline or salbutamol are characterized by a rapid onset but relatively short duration of action. Fenoterol is another selective, short-acting ␤2-agonist with high intrinsic efficacy that is frequently used for inhalation therapy in ventilated patients with chronic airflow obstruction.8 We hypothesized that the short-acting ␤agonist fenoterol, when nebulized immediately after acid aspiration, would improve gas exchange and reduce pulmonary damage by attenuating lung inflammation. An additional aim of this study was to provide information about the concentration of fenoterol in the bronchoalveolar lavage fluid (BALF) and rat plasma and show its effect on histology and oxygenation.

METHODS Animal Preparation The study was conducted after obtaining written approval from the local Bioethics Committee and in accordance with guidelines of the German National Institute of Health and the institutional Animal Care and Use Committee. The study was performed with 20 male Sprague-Dawley rats (Charles River, Sulzfeld, Germany) weighing between 280 and 320 g. The rats were kept on a 12-h light-dark cycle and had free access to food and water. Anesthesia was induced by inhalation of 4.0% sevoflurane in air in a 2 L induction container. After preparation of the trachea, a 14-G IV cannula with a mark indicating the distance from tracheostoma to a position approximately 0.5 cm above the carina was advanced for ventilation and maintenance of anesthesia. The distance was measured in preliminary experiments. The right carotid artery and a hindpaw vein were exposed and cannulated using a 24-G and 26-G needle, respectively. To maintain fluid balance, we administered 5 mL 䡠 kg⫺1 䡠 h⫺1 saline solution via the venous line. Normothermia (mean 37.5°C ⫾ 0.5 sd) was maintained by positioning the animal supine on an electric heat plate. Each rat was then continuously monitored using an electrocardiogram, and invasive mean arterial blood pressure, oxygen saturation, and rectal temperature (Siemens 9000C, Erlangen, Germany) were continuously measured. All animals received volume-controlled mechanical ventilation (Animal Ventilator, Inspira AV, Harvard 144

Fenoterol Inhalation After Acid Aspiration

Apparatus, March-Hugstetten, Germany). A tidal volume of 8 mL/kg and a positive end-expiratory pressure of 3 cm were considered to be appropriate. The respiratory rate was adjusted to maintain arterial Pco2 between 35 and 45 mm Hg. Sevoflurane was administered at a concentration of 3.0% using an anesthesia circuit system (Sulla 303V, Draeger, Lu¨beck, Germany). Peak pressure was recorded, and a fraction of inspired oxygen (Fio2) of 0.35 was maintained throughout the experiment. After completion of the invasive procedures, a period of 30 min of hemodynamic stabilization was allowed for two baseline measurements of MABP, heart rate, and blood gas samples (Bayer 850, blood gas analyzer, Bayer Diagnostics, Leverkusen, Germany). Arrhythmic events during the observation time were recorded and documented. ALI was then induced by instillation of 0.4 mL/kg 0.1 M HCl (pH approximately 1.0) into both lungs, as previously described.9 The animals were randomly assigned to one of two groups: fenoterol group, inhalation of 10 ␮g fenoterol in 2 mL saline, or placebo group, inhalation of 2 mL normal saline. Both groups received inhaled fluids 15 and 180 min after aspiration for 15 min. Inhalation was repeated after 180 min based on the clinical half-life of fenoterol. Inhalation was performed and blood gas and electrolyte samples (each 0.3 mL) were taken 30, 90, 180, 270, and 360 min after acid instillation and replaced by 0.6 mL of saline. The end point of the experiment was survival after 6 h. Surviving animals were killed by an overdose of thiopental. Cytokine levels were determined in the BALF of the right lung, and the left lung was removed for histopathological examination and determination of the wet/dry ratio.

Bronchoalveolar Lavage Fluid At the end of the experiment, thoracotomy was performed and the left main bronchus was ligated proximally. BALF was prepared by carefully instilling only 2 mL of normal saline into the right lung to avoid overdistension. The fluid was withdrawn and the process repeated three times until 80% of the instilled fluid was obtained in all animals. After the third lavage, the BALF was collected and placed on ice. The samples were centrifuged at 250g for 10 min and the BALF supernatants were stored at ⫺70°C.

Determination of Fenoterol in BALF and Plasma To determine the concentration of fenoterol in the BALF, we anesthetized eight additional rats in a second experiment. After induction of anesthesia with sevoflurane, rats were tracheally intubated with an 18-G IV catheter under direct laryngoscopy and inhalation of 10 ␮g fenoterol was performed for 15 min as previously mentioned. Fifteen minutes after inhalation, we instilled 1.5 mL BALF using physiological sodium chloride at 37°C. We sampled at least 80% of the instilled BALF and took a retrobulbar blood sample of 1 mL. All samples were placed on ice and centrifuged at 250g for 10 min, ANESTHESIA & ANALGESIA

and the BAL supernatants were stored at ⫺20°C. Levels of fenoterol were quantified in rat plasma and lavage by the Institute for Biomonitoring (Bayer Industry Services, Leverkusen, Germany), with liquid chromatography/ mass spectrometry-mass spectrometry (Waters/ Micromass Quattro Ultima, Milford, MA) with a limitof-detection of 1 ␮g/L (S/N 3:1) for both matrices. Briefly, labetalol hydrochloride (Sigma, St. Louis, MO) was added before preparation as an internal standard to each sample. Sample preparation differed depending on the origin of the probes. Fifty microliter lavage was diluted with methanol/ water (50/50 v/v) and filtered through MINI UNI PREP centrifugation vessels (Chromacol, Herts, UK). Plasma was extracted using Strata XC-Cartridges (Phenomenex, Aschaffenburg, Germany). Ten microliter of the extracts was separated on an EC 50/3 nucleodur C18, ISIS, 1.8 ␮m column (Macherey & Nagel) with an ammonium acetate-water/methanol gradient. The mass spectrometer conditions were: ESI interface in positive ion mode; ion transitions in MRM-mode fenoterol: 304.1 m/z ⬎135.12 m/z; and labetalol: 329 m/z ⬎311 m/z. The intraseries recovery rate was 107% based on the internal standard at a spike level of 10 ng/mL with a coefficient of variation of 5.3%.

Cytokine Assays The tumor necrosis factor (TNF)-␣ and IL-6 concentrations in rat BALF were measured quantitatively at the end of the experiment using solid-phase sandwich ELISA kits (Biosource International, Camarillo, CA). Samples, including standards of known cytokine concentrations, control specimens, and unknowns, were analyzed according to the manufacturer’s protocol. Samples with cytokine concentrations larger than the test range were diluted with standard diluent buffer. After addition of 100 ␮L stop solution, samples were read at 450 nm with a microplate reader (Molecular Devices, Sunnyvale, CA).

Histological Studies Left lungs were inflated to a pressure of 20 cm H2O twice and then to 10 cm H2O and fixed in 4% neutral-buffered formalin for 48 h for histological examination. The lungs were cut in three planes, so that the basal, middle, and apical portions could be analyzed by histology. After embedding in paraffin, the tissue was sectioned with a microtome at 4 ␮m before hematoxylin and eosin staining. Slides were read by a pathologist blinded to the study groups, and scored using a scoring system developed by Simons et al.10 for ranking the degree of lung injury. Briefly, 10 fields at each of the three levels were examined in each lung at a magnification of 200⫻ and scored from 0 to 5 using the following criteria: septal thickening, alveolar and interstitial edema, hyaline membranes, inflammatory cell infiltration, and small airway epithelial Vol. 109, No. 1, July 2009

injury (with 0 indicating normal and 5 the most severe injury).

Wet-to-Dry Weight Ratio At the end of each experiment, total lung water was assessed using the lung wet-to-dry weight ratio. Onehalf of the upper lobe of the left lung was excised and weighed (wet weight) using a precision balance (BP 221 S, Sartorius, Go¨ttingen, Germany). Samples were then heated (WTB Binder, Labortechnik, Tutlingen, Germany) at 70°C for 72 h and weighed again (dry weight) to obtain the wet/dry lung weight ratio as an indicator of lung edema.

Statistical Analysis All data were analyzed using the statistical software SPSS 15.0 (SPSS, Chicago, IL). Data were examined for Gaussian distribution with the Kolmogorov-Smirnov test. Blood gas analysis and peak pressures with repeated measurements were compared between fenoterol and placebo by using the one-way analysis of variance procedure at each measurement time point. For each variable, the Bonferroni multiple comparison adjustment was applied across all the time points. Nonparametric histological data and nonnormally distributed cytokine data were examined with the two-sided Mann-Whitney U-test for between-group differences. Statistical significance was assumed at P ⬍ 0.05. All data presented are expressed as mean ⫾ sd (sd).

RESULTS Arterial Blood Gases As an index of pulmonary function, the Pao2/Fio2 was not different between the groups before treatment, but it decreased after instillation of acid in the placebo group to ⬍300, indicating ALI (Fig. 1). The oxygenation index was significantly better in the fenoterol group after 180 min (380 ⫾ 48, P ⫽ 0.032), 270 min (388 ⫾ 27, P ⫽ 0.001), and 360 min (403 ⫾ 51, P ⫽ 0.001) than in the placebo group (335 ⫾ 35, 298 ⫾ 42, and 248 ⫾ 22, respectively), as shown in Figure 1. The Paco2 and pH values, hemoglobin, and serum electrolytes were comparable in all groups and showed no differences among surviving animals during the time course of the experiments (Table 1). Only base excess values tended to be more negative, indicating metabolic acidosis in the placebo group, although without reaching significance (Table 1). MABP did not differ between the groups (Table 1).

Peak Pressure Although not statistically significant at every time point, there was a tendency toward lower peak pressure values in the fenoterol 10 ␮g group during the entire observation period. Comparing the fenoterol group with the placebo group at corresponding time points, we observed significant differences of the means from t30, t45, and t60 (Fig. 2). © 2009 International Anesthesia Research Society

145

Figure 1. Effect of fenoterol inhalation on Po2 (mean ⫾ sd)

after acid-induced acute lung injury in rats. Sprague-Dawley rats were randomized into two groups (n ⫽ 10 per group): inhalation of placebo, inhalation of fenoterol 10 ␮g, (*P ⬍ 0.05).

Determination of Fenoterol Levels in the BALF Fenoterol concentrations in the BALF of the fenoterol group had a median of 10.3 ␮g/L (4.7–13.6 ␮g/L). Fenoterol was not detectable in plasma (⬍1 ␮g/L).

TNF-␣ and IL-6 in the BALF

By 6 h after acid aspiration, TNF-␣ values were significantly lower in fenoterol-treated animals than in those treated with placebo (91.2 ⫾ 48.7 pg/mL vs 255.9 ⫾ 136.8 pg/mL; P ⫽ 0.038, Fig. 3a). IL-6 levels were significantly lower in the fenoterol group

Figure 2. Peak airway pressure as a variable of airway resistance. Pmax (mean ⫾ sd) showed significantly decreased airway resistance in the group treated with fenoterol inhalation only at corresponding time points from t30 until t60, although the overall decrease was not significant. (222.9 ⫾ 144.3 pg/mL) than in the placebo group (563.6 ⫾ 191.2 pg/mL; P ⫽ 0.008, Fig. 3b).

Drop-Out and Survival Rates Experiments were started with n ⫽ 20 animals, each group containing n ⫽ 10 rats. Only five animals reached the end of the observation period at t360 in the placebo group. Kaplan—Meyer analysis revealed a significantly better survival rate in the fenoterol group than in the placebo group (100% vs 50%, log-rank test, P ⫽ 0.011).

Table 1. Blood Gas, Hemoglobin, and Mean Arterial Blood Pressure (MABP) During Experiment Parameter⶿ time (min) Pao2/Fio2 CO2(fenoterol) mm Hg CO2(placebo) mm Hg pH(fenoterol) pH(placebo) BE(fenoterol) mmol/L BE(placebo) mmol/L Hb(fenoterol) mg/dL Hb(placebo) mg/dL MABP mm Hg(fenoterol) MABP mm Hg(placebo)

⫺15

0

40.2 ⫾ 9.2

41.4 ⫾ 6.4

30

90

180

270

360

43.3 ⫾ 7.6

41.5 ⫾ 5.9

37.0 ⫾ 14.2

39.6 ⫾ 10.9 38.4 ⫾ 10.1 51.0 ⫾ 12.4 50.6 ⫾ 11.6

39.3 ⫾ 10.6

38.4 ⫾ 11.9

42.3 ⫾ 14.2

7.47 ⫾ 0.8 7.45 ⫾ 0.1 2.5 ⫾ 1.3

7.42 ⫾ 0.3 7.45 ⫾ 0.1 2.6 ⫾ 3.3

7.35 ⫾ 0.7 7.36 ⫾ 0.7 2.9 ⫾ 2.9

7.37 ⫾ 0.1 7.37 ⫾ 0.2 1.7 ⫾ 5.4

7.40 ⫾ 0.1 7.42 ⫾ 0.1 2.1 ⫾ 4.3

7.41 ⫾ 0.8 7.35 ⫾ 0.2 1.6 ⫾ 3.3

2.8 ⫾ 2.9

2.3 ⫾ 2.8

1.5 ⫾ 4.8

0.6 ⫾ 5.0

⫺7.1 ⫾ 4.2

13.0 ⫾ 0.8

12.4 ⫾ 0.8

11.9 ⫾ 1.0

11.6 ⫾ 1.2

11.6 ⫾ 1.4

11.3 ⫾ 1.4

11.1 ⫾ 1.4

13.2 ⫾ 1.0

12.9 ⫾ 0.8

12.8 ⫾ 0.8

12.4 ⫾ 0.7

12.2 ⫾ 0.6

12.2 ⫾ 1.4

11.2 ⫾ 1.1

108 ⫾ 11

102 ⫾ 13

91 ⫾ 13

94 ⫾ 14

82 ⫾ 12

80 ⫾ 23

70 ⫾ 15

107 ⫾ 12

105 ⫾ 16

93 ⫾ 4

86 ⫾ 15

79 ⫾ 21

83 ⫾ 14

54.5 ⫾ 12.1 47.0 ⫾ 8.2

102 ⫾ 2

Drop-out placebo animals at time of death 282 ⫾ 48 33.5 ⫾ 5.3

7.40 ⫾ 0.8 7.38 ⫾ 0.7 ⫺2.4 ⫾ 4.4

7.20 ⫾ 0.2

⫺3.9 ⫾ 10.9 ⫺4.0 ⫾ 4.0

⫺13.9 ⫾ 8.4

12.1 ⫾ 0.7

31 ⫾ 5

Analysis and results of blood gas analysis, hemoglobin concentration, and MABP in the groups before and after acid aspiration throughout the experiment. There were no significant differences between the groups. Last column shows average values of placebo animals at time of death. MABP ⫽ mean arterial blood pressure.

146



Wet/Dry Ratio The wet-to-dry weight ratio was significantly higher in placebo rats than in the fenoterol-treated rats, indicating a higher lung water content (7.32 ⫾ 0.4 vs 6.39 ⫾ 0.4, respectively), P ⫽ 0.04.

Histology Histological examination of the lungs demonstrated differences in edema and in lung injury between the groups (Fig. 4). Nine animals in the placebo group (one died after 90 min, another four after 300 min) and 10 animals in the fenoterol group were examined. For each group, 240 fields were scored by a pulmonary pathologist (TS) blinded to the treatment groups. The median histological lung injury score was 32– 4 for the placebo group, and two1– 4 for the fenoterol 10 ␮g group. The fenoterol 10 ␮g group showed less interstitial edema and infiltration by polymorphonuclear leukocytes than the placebo group (P ⫽ 0.023, Mann-Whitney U-test).

DISCUSSION

Figure 3. Effects of fenoterol inhalation or placebo inhalation on the concentration of tumor necrosis factor (TNF)-␣ (a) and interleukin (IL)-6 (b) in the bronchoalveolar lavage of rat lungs. Figure shows mean ⫾ sd. Table 1 presents the oxygenation index, Pco2, MABP, and acid-base status of five placebo animals at time of death. Severe hypotension and metabolic acidosis were found before death.

In a model of ALI caused by acid instillation, the ␤-agonist fenoterol showed positive effects on arterial Po2, survival and a reduction in cytokine release. We found improved lung histology and decreased lung water content after the immediate inhalation of fenoterol. Beneficial effects of ␤-agonists in the therapy of ALI have been discussed and shown in several animal models.4,11,12 Beta agonists can upregulate alveolar fluid clearance and improve outcome in patients with ALI.13 Furthermore, the breakdown of the pulmonary capillary integrity due to noxious stimuli may be reduced by stabilization of endothelial permeability.14 Several mechanisms of these beneficial effects are understood, but many factors are still unknown, including the optimal drug and the optimal dosing of the drug.15,16 Additionally, most studies have been done with IV administration of ␤ agonists, although only a few studies have been performed by inhalation.

Figure 4. (a) Placebo-treated rat lung 6 h after acid-induced lung injury (n ⫽ 9). Section shows marked interstitial edema (#) and infiltration by polymorphonuclear leukocytes (J) (original magnification, ⫻20; hematoxylin-eosin). (b) Fenoterol 10 ␮g treated rat lung 6 h after acid-induced lung injury (n ⫽ 10). Edema and infiltration are less pronounced when compared with (a). Lungs of placebo rats contain significantly more lung water (c). Figure shows mean ⫾ sd. Vol. 109, No. 1, July 2009


147

We present data on the short-acting ␤-agonist fenoterol, which is in clinical use for both chronic obstructive pulmonary disease and patients in intensive care units. This investigation is of special interest as it provides additional information regarding the National Heart Lung and Blood Institute’s continuing drug study “albuterol to treat ALI,” which is currently recruiting patients in the United States. Oxygenation is compromised in the early stages of ALI because of acid aspiration. Lung edema and bronchoconstriction may be factors leading to a decreased Pao2/Fio2. We found improvement in Po2 in the fenoterol group after 180, 270, and 360 min. This result contrasts with the findings of other researchers17 who found no differences in oxygenation after 270 min. One likely explanation for this difference may be the 10-fold higher aspiration volume that was used in their model, although we used an aspiration volume that was adjusted to the clinical situation most investigators consider to be the threshold for inducing inflammation resulting in lung injury.18 Despite the observed initial effect on pulmonary mechanics, with a decrease in airway pressure after inhalation of fenoterol, there was no improvement in Pao2/Fio2. Beta agonists increase pulmonary blood flow by antagonizing the initial hypoxic vasoconstriction.19 The ␤2-agonist fenoterol is vessel-relaxing and increases the perfusion of nonventilated lung regions, and thus increases the shunt fraction by antagonizing hypoxic pulmonary vasoconstriction. However, the main effect of fenoterol on oxygenation in our study is the reduction of inflammation and lung edema that results in a better oxygenation index 5 h after the insult on the lung. Lung edema resolution from the alveolar space is a critical issue in the recovery from direct and indirect injuries of the lung. Although clearance of edema fluid depends on the balance between edema formation and reabsorption processes, the reabsorption of fluid is dependent on the active transport of sodium and electrolytes. Moreover, increased alveolar fluid clearance results in rapid improvement of lung injury, reduced duration of ventilation, and improved survival.13 Inhalation of ␤-agonists leads to an improvement of alveolar edema by increasing alveolar fluid clearance.11,17 We performed wet-lung weight measurements in our experiments and found decreased lung water content, which is in agreement with other previously published investigations.7,11 Additionally, ␤2-agonists may antagonize initial bronchoconstriction because of acid aspiration.20 Moreover, inhalational ␤-agonists are successfully used to decrease airway pressures in anesthetized patients suffering from bronchoconstriction.21 High airway pressure associated with lung injury worsens lung injury; therefore, it is reasonable to reduce high airway pressure with pharmacological intervention.22 148


Increased airway resistance after aspiration is reflected by increased airway pressures in volumecontrolled, ventilated animals23; one would expect inhalation of fenoterol to reverse bronchoconstriction. We observed a significant bronchodilating effect of fenoterol only in the first hour after acid aspiration, although the response after the second inhalation failed to demonstrate statistical significance. This may have been influenced by several factors. One possible explanation is that bronchoconstriction after acid aspiration may play a role in the first hour after such an event and decrease over the remaining course. Another reason may have been the lack of statistical power in our study. However, larger airway pressures in the placebo group after the acid injury may have been a contributing factor as a potential factor of worsening structural damage. ␤2-agonists are reported to reduce inflammation after respiratory injury, which is one of the major targets of pharmacotherapy in ALI. TNF-␣ plays an essential role in the cytokine cascade and is considered to be one of the initiators of lung inflammation after acid-induced lung injury.24 It is responsible for recruiting neutrophil granulocytes to the site of injury, expression of endothelial adhesion molecules, and the production of other relevant ILs, such as IL-6 and IL-8.25,26 We found a significant reduction in proinflammatory cytokines, such as TNF-␣ and IL-6, in the fenoterol group in contrast to placebo, consistent with the results of in vitro studies.27 Interestingly, we observed differences in the survival of animals in our study. The animals that died in our study showed metabolic acidosis, a finding that has been described in the literature.28 Furthermore, it has been found that rats are prone to develop right heart failure due to atelectasis, which results in a mortality of more than 50%. The accumulation of fluid in the alveolar space after aspiration leads to atelectasis of rat lungs.9 Fenoterol-mediated reduction of lung water content is probably one factor responsible for diminishing atelectasis. It is notable that all animals in the fenoterol group survived for a period of 6 h, although half the animals in the placebo group died from systemic hypotension and metabolic acidosis. We observed a reduction in TNF-␣ and IL-6 levels in our study, which has been reported in allergic airway diseases and is because of elevated intracellular cyclic adenosine monophosphate levels.29 Our result confirms an observation by Wu et al.,30 who reported that terbutaline, another ␤2-agonist, reduced TNF-␣ production and led to better survival in a mouse model. Moreover, in addition to the reduction of the proinflammatory cytokines after fenoterol inhalation in our study, we found a significantly improved histological pattern, with less septal thickening, fewer hyaline membranes, and reduced alveolar and interstitial edema as signs of decreased inflammation in the animals that inhaled fenoterol. ANESTHESIA & ANALGESIA

Our results have several implications for the use of inhaled ␤2-agonists. First, inhalation of fenoterol immediately after acid aspiration leads to an improvement in oxygenation. Even if Po2 is only a surrogate variable for lung function, which is influenced by several factors, this helps to maintain proper oxygenation although avoiding harmful oxygen concentrations after lung injury.31 Furthermore, inhalation of a ␤-agonist in the delivered concentration had no adverse effect on the cardiovascular system (e.g., arrhythmias). IV ␤-agonist administration is frequently followed by arrhythmic events, which may cause additional problems in the intensive care patient.7 The inhalation route of administering ␤-agonists in lung injury may help avoid cardiac arrhythmias by reaching higher concentrations in the affected side of the lung than in the systemic circulation. Our findings support previous work showing that inhaled ␤2agonists can have positive effects on inflammation and oxygenation, and they suggest clinical relevance for the management of patients with pulmonary edema. However, there are some limitations to our study. The beneficial effects shown in this model demonstrate the potential of ␤2-agonists in an animal model with mild to moderate ALI. It remains to be seen whether the severely injured alveolar epithelium (as in ARDS patients) also responds to the administration of ␤-agonists. Nebulized drugs are more difficult to place at their target than IV drugs. The concentration of a given dose may vary significantly in the alveolar space, which makes it difficult to establish a clear correlation between the nebulized dose and the reduction of inflammation. However, measurements of fenoterol concentration in the alveolar space demonstrated the feasibility of nebulization and revealed beneficial effects.

CONCLUSIONS Fenoterol inhalation improved oxygenation after 270 and 360 min, attenuated the release of TNF-␣ and IL-6, and diminished the lung edema and the infiltration of polymorphonuclear leukocytes. The positive effects of the ␤2-agonist fenoterol on these variables in rodents will have to be established in human studies. ACKNOWLEDGMENTS The authors would like to thank David Tracey, PhD, Department of Anatomy and Developmental Biology, University College London, for his helpful comments regarding this manuscript. REFERENCES 1. Warner MA, Warner ME, Weber JG. Clinical significance of pulmonary aspiration during the perioperative period. Anesthesiology 1993;78:56 – 62 2. Sakuma T, Hida M, Nambu Y, Osanai K, Toga H, Takahashi K, Ohya N, Inoue M, Watanabe Y. Beta1-adrenergic agonist is a potent stimulator of alveolar fluid clearance in hyperoxic rat lungs. Jpn J Pharmacol 2001;85:161– 6 Vol. 109, No. 1, July 2009

3. Saldias FJ, Lecuona E, Comellas AP, Ridge KM, Sznajder JI. Dopamine restores lung ability to clear edema in rats exposed to hyperoxia. Am J Respir Crit Care Med 1999;159:626 –33 4. Dhingra VK, Uusaro A, Holmes CL, Walley KR. Attenuation of lung inflammation by adrenergic agonists in murine acute lung injury. Anesthesiology 2001;95:947–53 5. Sears MR, Lotvall J. Past, present and future– beta2-adrenoceptor agonists in asthma management. Respir Med 2005;99:152–70 6. Bloemen PG, van den Tweel MC, Henricks PA, Englels F, Kester MH, van de Loo PG, Blomjous FJ, Nijkamp FP. Increased cAMP levels in stimulated neutrophils inhibit their adhesion to human bronchial epithelial cells. Am J Physiol 1997;272:L580 –7 7. Perkins GD, McAuley DF, Thickett DR, Gao F. The beta-agonist lung injury trial (BALTI): a randomized placebo-controlled clinical trial. Am J Respir Crit Care Med 2006;173:281–7 8. Bernasconi M, Brandolese R, Poggi R, Manzin E, Rossi A. Dose-response effects and time course of effects of inhaled fenoterol on respiratory mechanics and arterial oxygen tension in mechanically ventilated patients with chronic airflow obstruction. Intensive Care Med 1990;16:108 –14 9. Pawlik MT, Schreyer AG, Ittner KP, Selig C, Gruber M, Feuerbach S, Taeger K. Early treatment with pentoxifylline reduces lung injury induced by acid aspiration in rats. Chest 2005;127:613–21 10. Simons RK, Maier RV, Chi EY. Pulmonary effects of continuous endotoxin infusion in the rat. Circ Shock 1991;33:233– 43 11. Palmieri TL, Enkhbaatar P, Bayliss R, Traber LD, Cox RA, Hawkins HK, Herndon DN, Greenhalgh DG, Traber DL. Continuous nebulized albuterol attenuates acute lung injury in an ovine model of combined burn and smoke inhalation. Crit Care Med 2006;34:1719 –24 12. Saldias FJ, Comellas A, Ridge KM, Lecuona E, Sznajder JI. Isoproterenol improves ability of lung to clear edema in rats exposed to hyperoxia. J Appl Physiol 1999;87:30 –5 13. Ware LB, Matthay MA. Alveolar fluid clearance is impaired in the majority of patients with acute lung injury and the acute respiratory distress syndrome. Am J Respir Crit Care Med 2001;163:1376 – 83 14. Khimenko PL, Barnard JW, Moore TM, Wilson PS, Ballard ST, Taylor AE. Vascular permeability and epithelial transport effects on lung edema formation in ischemia and reperfusion. J Appl Physiol 1994;77:1116 –21 15. Sekut L, Champion BR, Page K, Menius JA Jr, Connolly KM. Anti-inflammatory activity of salmeterol: down-regulation of cytokine production. Clin Exp Immunol 1995;99:461– 6 16. van der Poll T, Jansen J, Endert E, Sauerwein HP, van Deventer SJ. Noradrenaline inhibits lipopolysaccharide-induced tumor necrosis factor and interleukin 6 production in human whole blood. Infect Immun 1994;62:2046 –50 17. McAuley DF, Frank JA, Fang X, Matthay MA. Clinically relevant concentrations of beta2-adrenergic agonists stimulate maximal cyclic adenosine monophosphate-dependent airspace fluid clearance and decrease pulmonary edema in experimental acid-induced lung injury. Crit Care Med 2004;32:1470 – 6 18. Ng A, Smith G. Gastroesophageal reflux and aspiration of gastric contents in anesthetic practice. Anesth Analg 2001;93:494 –513 19. Berthiaume Y, Staub NC, Matthay MA. Beta-adrenergic agonists increase lung liquid clearance in anesthetized sheep. J Clin Invest 1987;79:335– 43 20. Ishikawa T, Sekizawa SI, Sant’Ambrogio FB, Sant’Ambrogio G. Larynx vs. esophagus as reflexogenic sites for acid-induced bronchoconstriction in dogs. J Appl Physiol 1999;86:1226 –30 21. Fine GF, Motoyama EK, Brandom BW, Fertal KM, Mutich R, Davis PJ. The effect on lung mechanics in anesthetized children with rapacuronium: a comparative study with mivacurium. Anesth Analg 2002;95:56 – 61, table of contents 22. Gattinoni L, Chiumello D, Russo R. Reduced tidal volumes and lung protective ventilatory strategies: where do we go from here? Curr Opin Crit Care 2002;8:45–50 23. Westervelt CL, Choe EU, Arya J, Lippton HL, Flint LM, Ferrara JJ. Effects of anti-inflammatory agents on hydrochloric acidinduced pulmonary injury. J Invest Surg 1996;9:283–91 24. Goldman G, Welbourn R, Kobzik L, Valeri CR, Shepro D, Hechtman HB. Tumor necrosis factor-alpha mediates acid aspiration-induced systemic organ injury. Ann Surg 1990;212: 513–19, discussion 519 –20 © 2009 International Anesthesia Research Society

149

25. Goodman RB, Pugin J, Lee JS, Matthay MA. Cytokine-mediated inflammation in acute lung injury. Cytokine Growth Factor Rev 2003;14:523–35 26. Mulligan MS, Vaporciyan AA, Miyasaka M, Tamatani T, Ward PA. Tumor necrosis factor alpha regulates in vivo intrapulmonary expression of ICAM-1. Am J Pathol 1993;142:1739 – 49 27. Gu Y, Seidel A. Influence of salbutamol and isoproterenol on the production of TNF and reactive oxygen species by bovine alveolar macrophages and calcitriol differentiated HL-60 cells. Immunopharmacol Immunotoxicol 1996;18:115–28 28. Duggan M, McCaul CL, McNamara PJ, Englelberts D, Ackerley C, Kavanagh BP. Atelectasis causes vascular leak and lethal right ventricular failure in uninjured rat lungs. Am J Respir Crit Care Med 2003;167:1633– 40

150


29. Yoshimura T, Kurita C, Nagao T, Usami E, Nakao T, Watanabe S, Kobayashi J, Yamazaki F, Tanaka H, Inagaki N, Nagai H. Inhibition of tumor necrosis factor-alpha and interleukin-1-beta production by beta-adrenoceptor agonists from lipopolysaccharide-stimulated human peripheral blood mononuclear cells. Pharmacology 1997;54:144 –52 30. Wu CC, Liao MH, Chen SJ, Chou TC, Chen A, Yen MH. Terbutaline prevents circulatory failure and mitigates mortality in rodents with endotoxemia. Shock 2000;14:60 –7 31. Nader-Djalal N, Knight PR, Davidson BA, Johnson K. Hyperoxia exacerbates microvascular lung injury following acid aspiration. Chest 1997;112:1607–14


The Effects of Lung Recruitment on the Phase III Slope of Volumetric Capnography in Morbidly Obese Patients Stephan H. Bo¨hm, MD* Stefan Maisch, MD* Alexandra von Sandersleben, MD* Oliver Thamm, MD*† Isabel Passoni, PhD‡ Jorge Martinez Arca, MSc‡ Gerardo Tusman, MD§

BACKGROUND: In this study, we analyzed the effect of the alveolar recruitment strategy (ARS) and positive end-expiratory pressure (PEEP) titration on Phase III slope (SIII) of volumetric capnography (VC) in morbidly obese patients. METHODS: Eleven anesthetized morbidly obese patients were studied. Lungs were ventilated with tidal volumes of 10 mL 䡠 kg⫺1, respiratory rates of 12–14 bpm, inspiration:expiration ratio of 1:2, and Fio2 of 0.4. ARS was performed by increasing PEEP in steps of five from 0 end-expiratory pressure to 15 cm H2O. During lung recruitment, plateau pressure was limited to 50 cm H2O, whereas tidal volume was increased to the ventilator’s maximum value of 1400 mL, and PEEP was increased to 20 cm H2O for 2 min. Thereafter, PEEP was reduced in steps of 5 cm H2O, from 15 to 0. VC, arterial blood gases, and lung mechanics data were determined for each PEEP step. RESULTS: SIII decreased from 0.014 ⫾ 0.006 to 0.005 ⫾ 0.005 mm Hg/mL when 0 end-expiratory pressure was compared against 15 cm H2O of PEEP after ARS (15ARS, P ⬍ 0.05). This decrement in SIII was accompanied by increases in Pao2 (27%, P ⬍ 0.002) and compliance (32%, P ⬍ 0.001), whereas Paco2 decreased by 8% (P ⬍ 0.038) when comparing values before and after ARS. A good prediction of the lung recruitment effect by SIII was derived from the receiver operating characteristic curve analysis (area under the curve of 0.81, sensitivity of 0.75, and specificity of 0.74; P ⬍ 0.001). CONCLUSION: The SIII in VC was useful to detect the optimal level of PEEP after lung recruitment in anesthetized morbidly obese patients. (Anesth Analg 2009;109:151–9)

D

uring general anesthesia, morbidly obese patients develop considerable amounts of atelectasis in dependent lung areas, much more than normal weight patients.1–5 Such “compression atelectasis” has a known negative effect on gas exchange and lung mechanics.5–7 Lung recruitment maneuvers (RMs), one of which is the alveolar recruitment strategy (ARS), are ventilatory strategies which were developed to reverse anesthesia-induced lung collapse.7,8 The positive effects of lung recruitment on pulmonary physiology have been demonstrated in anesthetized patients of varying body masses.7,9,10

From the *Clinic of Anesthesiology, University Hospital, Hamburg-Eppendorf, Hamburg, Germany; †currently at Clinic of Plastic and Reconstructive Surgery, Burn Care Center, Hospital Cologne-Merheim, University of Witten/Herdecke, Germany; ‡Department of Bioengineering, University of Mar del Plata, Argentina; and §Department of Anesthesiology, Hospital Privado de Comunidad, Mar del Plata, Argentina. Accepted for publication November 13, 2008. Supported by the Clinic of Anesthesiology, University Hospital Hamburg-Eppendorf, Germany. Address correspondence to Stephan H. Bo¨hm, MD, CSEM Centre Suisse d’Electronique et de Microtechnique SA, Research Centre for Nanomedicine, Medical Sensors, Schulstr. 1, CH-7302 Landquart, Switzerland. Address e-mail to [email protected]. Copyright © 2009 International Anesthesia Research Society DOI: 10.1213/ane.0b013e31819bcbb5

Vol. 109, No. 1, July 2009

Noninvasive means of monitoring the effects of ARS and positive end-expiratory pressure (PEEP) on lung function are needed at the bedside. In theory, real-time and breath-by-breath respiratory variables related to the collapse-recruitment physiology of the lungs could help in performing RMs more safely and in determining the optimal level of PEEP for each patient. Our group described the effect of lung recruitment on volumetric capnography (VC) curves (the curve formed by the expired CO2 concentration plotted against expired tidal volume [VT]) not only in anesthetized patients11,12 but also in an experimental model of acute lung injury.13 Dead space and several VC-derived variables changed favorably after lung recruitment and at adequate levels of PEEP, as could be documented by improved arterial oxygenation, improved respiratory mechanics, and reduced areas of collapse on computed tomography images. Of all VC-derived variables, the slope of Phase III (SIII) is of particular physiological interest because it is an indicator of both CO2 transport within the airways (ventilation) and the CO2 delivered to the pulmonary capillaries (perfusion).14 –18 Thus, this variable has also been associated with the ventilation/perfusion (V/Q) relationship in the lungs.19 –21 There is clear evidence suggesting that any increment in SIII is a reflection of impaired V/Q matching, whereas decrements indicate its improvement.22–26 Considering the above concepts, 151

it seems that SIII is an on-line breath-by-breath variable that could provide valuable information about the effects of RMs and PEEP on lung function in anesthetized patients. The goal of this observational study was twofold: 1) to test the role of SIII as a noninvasive real-time VC-derived variable for monitoring the effects of ARS and PEEP during a titration process and 2) to identify other variables derived from VC that could describe the lung collapse-recruitment phenomenon in a noninvasive way. This protocol was performed in morbidly obese patients, a good clinical model of anesthesia-induced atelectasis,4 –5 which magnifies the typical changes in pulmonary physiology because of lung collapse routinely seen in normal weight patients.

METHODS The study was approved by the local ethics committee of the University Hospital Hamburg-Eppendorf, Germany. We examined 11 patients with body mass indexes (BMI, weight/height2) ⬎40 kg 䡠 m⫺2 undergoing bariatric laparoscopic surgery. After obtaining written informed consent, we enrolled only patients without known cardiovascular and pulmonary diseases. In the induction room, patients lay in the supine position. We inserted a 20 G radial artery catheter under local anesthesia. After induction of anesthesia, a central venous catheter (Certofix-Trio-Set 5730, B. Braun, Melsungen, Germany) and a pulmonary artery catheter (Swan-Ganz, Baxter, Irvine, CA) were introduced into the same internal jugular vein. Cardiac output was measured in triplicate using the thermodilution method. Pulmonary vascular pressures and all other measurements were performed at the end of expiration. The catheters for rather extensive hemodynamic monitoring were inserted immediately after induction of anesthesia but served primarily the needs of a subsequent intraoperative study in the same patients, investigating the effects of capnoperitoneum and PEEP on heart and lung function. In the current study, we analyzed the data of only a subset of 11 of 19 morbidly obese patients enrolled in the original study (in preparation for publication) who had complete sets of CO2 and reference data. The current study protocol was completed before moving the patient into the operating room where the surgery took place. One liter of colloidal solution (RheoHAES, B. Braun, Melsungen, Germany) was administered for intravascular volume expansion before induction of anesthesia. IV saline solution infusion was kept at 5 mL 䡠 kg⫺1 (lean body weight) during the study. This preload optimization was necessary to minimize potential hemodynamic impairment because of high airway pressures if patients presented with a relative or inadvertent hypovolemia.27 Oxygen (100%) was administered for 3 min. Thereafter, anesthesia was induced with etomidate 0.15– 0.3 152

Volumetric Capnography in Obese Patients

mg 䡠 kg⫺1, sufentanyl 0.1– 0.5 ␮g 䡠 kg⫺1, and succinylcholine 1 mg 䡠 kg⫺1. Balanced anesthesia was maintained with sevoflurane plus IV boluses of 0.1– 0.5 ␮g 䡠 kg⫺1 sufentanyl and 0.075– 0.15 mg 䡠 kg⫺1 rocuronium as needed. After tracheal intubation, the lungs were ventilated using a Cicero EM (Dra¨ger, Lu¨beck, Germany) in a volume-controlled mode with a VT of 10 mL 䡠 kg⫺1 lean body weight, a respiratory rate of 10 –14 bpm, an inspiratory-to-expiratory ratio of 1:2, Fio2 of 0.4, and initially without PEEP (ZEEP). Minute ventilation was adjusted by changing the respiratory rate to maintain Paco2 within the normal range.

VC and Lung Mechanics Variables For VC and respiratory mechanics measurements, we used the respiratory profile monitor COSMOplus (Respironics, Wallingford, CT). Its Capnostat威 sensor was inserted between the endotracheal tube and Y-piece of the ventilator. Data were continuously recorded using the software Aplus (Novametrix, Wallinford, CT). The VC is traditionally divided into three phases.21 Phase I represents expired gases free of CO2. Phase II is formed by the rapid increase in expired CO2 coming from lung acini, whereas Phase III consists of gas resident in the alveolar space. The maximum slope of Phase II (SII) was calculated by using linear regression from the data points between the start of Phase II and 40% of the expired volume. SIII was calculated by linear regression, considering all data points between 40% and 80% of VC. Both slopes are expressed as mm Hg/mL. VDaw is the airway dead space as determined by Fowler method.28 Physiologic dead space (VDphys) was calculated using Enghoff modification of Bohr formula29 as:

VDphys ⴝ (Paco2 ⴚ Peco2)/Paco2 ⴛ VT where PeCO2 is the mean expired partial pressure of CO2. Next, the ratio of physiological VD/VT was calculated as:

VD/VT ⴝ VDphys/VT VDalv was computed by subtracting VDaw from VDphys. The ratio of alveolar dead space to alveolar VT (VDalv/VTalv) was then obtained by dividing VDalv by the alveolar part of the tidal volume (VTalv ⫽ VT ⫺ VDaw). The arterial to end-tidal partial pressure of CO2 (Pa-ETco2) was also calculated. Dynamic compliance (CDYN) was calculated as VT divided by ␦ pressure (plateau pressure minus total PEEP). Expiratory airway resistance (RAW) was computed as ␦ pressure divided by expiratory flow. Peak expiratory flow (PEF) was defined as the maximum value of expiratory flow of a breath. The expiratory time constant (ETC) was calculated as the product of CDYN and RAW. Two new variables derived from the VC and lung mechanics were developed and tested. These variables ANESTHESIA & ANALGESIA

are related to the CO2 elimination kinetic during expiration: 1. The time-constant for eliminating CO2 (Tau-CO2), represented as the amount of CO2 eliminated during 1 expiratory time-constant.30 It is calculated as:

Tau-CO2 ⴝ C DYN ⴛ R AW ⴛ VTCO2,br where VTCO2,br is the amount of CO2 eliminated in 1 breath as derived from an integration of the area under the CO2 versus volume curve. 2. Because of the fact that ARS, in conjunction with adequate levels of PEEP, increased CDYN, while at the same time decreasing RAW and having some transient effects on VTCO2,br (at least if observation intervals are shorter than the time for complete equilibration, as in this study), we modified the initial Tau-CO2 formula by placing RAW in the denominator of the above equation. This new variable derived from the “modified” Tau-CO2 formula was referred to as CO2flow:

CO2flow ⴝ C DYN ⴛ VTCO2,br/R AW

Protocol PEEP titration and alveolar recruitment were initiated before the start of surgery. ARS was slightly modified from our previous publications7,9,11 to meet the higher pressure requirements of this study population. Eichenberger et al.4 showed that morbidly obese patients developed more atelectatic areas in the perioperative period when compared with patients of normal weight. Additionally, Rothen et al.8 observed that 1 of 16 patients still had atelectasis after lung recruitment with 40 cm H2O of Paw. These authors found a close correlation between the Paw needed to reexpand the atelectatic lungs and the BMI. Thus, we reasoned that the pressures needed to open all collapsed lung units in morbidly obese patients with otherwise healthy lungs would be higher than those in normal weight patients because of the decreased transpulmonary pressures caused by additional abdominal and thoracic amounts of fatty tissue.3,4,9 The lung’s opening pressures were assumed to be around 50 cm H2O of plateau pressure. Therefore, airway pressures were limited to 50 cm H2O, whereas both PEEP and VTs were increased to the machine’s maximum values of 20 cm H2O and 1400 mL, respectively. However, because of the pressure limitation the latter value would never be reached. The high-flow rate together with the pressure limitation resulted in a decelerating flow pattern even with the ventilator’s volumecontrolled mode of ventilation. As shown in Figure 1, PEEP was initially increased from 0 to 15 cm H2O in steps of 5 cm H2O. After 2 min at maximal airway pressures, previous ventilator settings were applied, starting at a PEEP of 15 cm H2O (15ARS), which was then decreased in steps of 5 cm H2O to 10 (10ARS), 5 (5ARS), and 0 (0ARS) cm Vol. 109, No. 1, July 2009

Figure 1. Symmetrical study protocol showing phases of increasing and decreasing levels of positive end-expiratory pressure (PEEP) separated by an alveolar recruitment strategy (ARS). Numbers on x axis show the level of PEEP applied during each 3-min period, as depicted by a rectangle. IndexARS ⫽ level of PEEP after an alveolar recruitment strategy. * ⫽ time points at which data were collected. H2O. Each level of PEEP before and after ARS was maintained for exactly 3 min. Data for analyzing VC and respiratory mechanics were taken starting at the third minute, and a mean value for each variable was calculated from approximately 12 breaths. Hemodynamic values and arterial blood for blood gas analysis were taken at the end of the third minute, just before changing PEEP. Samples were processed without delay by the blood gas analyzer ABL System 615 (Radiometer, Copenhagen, Denmark). From previous protocols, it was known that during the actual RM no steady-state condition would be reached. Therefore, we elected not to take blood samples for gas analysis at this step.

Analysis of the Recruitment Effect Pao2, shunt, CDYN, and the study of CO2 kinetics (Paco2 or dead space) are known markers of the effects of lung recruitment.13,31,32 Therefore, we used them to determine the physiological efficacy of PEEP applied before and after the ARS. Shunt was calculated according to the following formula and assumptions: Because the CvO2 needed for the formula CO ⫽ VO2/CaO2 ⫺ CvO2 could not be obtained during our short study periods, the assumption that VO2 ⫽ VCO2 ⫻ RQ (0.85) was used instead. Thus, CvO2 ⫽ VO2/CO ⫺ CaO2, now allowing the classic shunt formula to be applied as: Shunt ⫽ CcO2 ⫺ CaO2/CcO2 ⫺ CvO2.

Statistical Analysis Statistical analysis was performed using the program SPSS version 15.0 (SPSS, Chicago, IL). Variables © 2009 International Anesthesia Research Society

153

were analyzed by one-way analysis of variance and Tukey honestly significant difference criterion. Results are presented as mean and standard deviation. Sensitivity and specificity of a 20% decrease in SIII to detect a lung recruitment effect were evaluated by constructing the receiver operating characteristics curve (ROC)33 (Area Under the Curve, AUC ⫽ 0.7, P ⬍ 0.05). An AUC of 0.5 would mean no predictive value at all (pure chance), whereas an AUC of 1 signifies the best possible prediction. A P ⬍ 0.05 was considered significant. We considered changes in Pao2 ⱖ 20%, in CDYN ⱖ 40%, in VDalv/VTalv ⱕ20%, and in Paco2 ⱕ 5% as cutoffs for defining a recruitment effect.

RESULTS Nine men and two women morbidly obese patients undergoing anesthesia for laparoscopic gastric banding were studied. The mean patient age was 38 ⫾ 6 yr. Mean height was 1.68 ⫾ 0.1 m and mean weight was 144 ⫾ 29 kg, corresponding to a mean BMI (weight/height2) of 51 ⫾ 10 kg/m2. The effect of lung recruitment on SIII is presented in Figure 2. SIII showed significantly higher values before than after lung recruitment at 15ARS. SIII increased again when PEEP decreased to values below 15 cm H2O. Regression analysis showed that SIII together with SII and VTCO2,br predict the real Paco2 value according to the following equation: Paco2 ⫽ 719.2 ⫻ SIII ⫹ 23.74 ⫻ SII ⫹ 1.039 ⫻ VTCO2,br ⫹ 4.107 (R2 ⫽ 0.84, P ⬍ 0.05). No good correlations were found between SIII and Pao2, VDalv/VTalv, or CDYN at a high variability among these 11 patients (Fig. 2). However, the receiver operating characteristic analysis showed good sensitivity and specificity of SIII to detect lung recruitment effects (Table 1). Arterial oxygenation improved with lung recruitment (Fig. 2). Pao2 values after lung recruitment were higher than those at the same level of PEEP before ARS. The highest Pao2 value was observed at 15ARS and was statistically different from baseline (ZEEP). Dead space variables during the protocol are shown in Table 2. In general, the inefficiency of ventilation decreased with lung recruitment and ventilation at 15 cm H2O of PEEP. VDalv, VDalv/VTalv, and Pa-ETco2 showed significant changes after ARS compared with ZEEP ventilation before recruitment. Both Pao2 and dead space variables declined when PEEP was decreased stepwise from 15ARS to 0ARS, suggesting a derecruitment of previously recruited lung areas. Data on ventilation and lung mechanics are shown in Table 3. Lung mechanics improved after ARS. CDYN increased with PEEP and lung recruitment, reaching the highest values at 15ARS. RAW decreased with increasing PEEP, reaching the lowest values at 15 cm H2O both before and after recruitment. Similar to gas exchange, CDYN and RAW showed a progressive deterioration when PEEP went from 15ARS to 0ARS. ETC 154


was prolonged and reached significant differences after ARS when compared with ventilation without recruitment. All values of PEF after lung recruitment were lower than those before ARS. Despite the fact that VTCO2,br did not change significantly throughout the entire protocol, Tau-CO2 showed consistent increments after lung recruitment, describing a combined positive effect of ARS on lung mechanics and CO2 elimination (Fig. 3). CO2flow, however, described the impact of lung recruitment in a more pronounced way, showing a pattern similar to Pao2 and CDYN in Figure 2. Hemodynamic data are provided in Table 4. Mean systemic and pulmonary arterial pressures, cardiac filling pressures, and cardiac output remained stable and within normal ranges throughout the entire protocol while shunt presented an almost exact mirror image of Pao2. No complications occurred.

DISCUSSION This study in morbidly obese anesthetized patients showed that an ARS in conjunction with adequate levels of PEEP decreased the SIII of the VC curve, reflecting improved global lung function. This suspected improvement was confirmed by simultaneous and positive changes in arterial oxygenation and CO2, shunt, dead space, and respiratory mechanics, which were most pronounced after the lung recruitment.7,9,13,31,32 Two main clinical implications can be deduced from this study. First, the association between VCderived variables and the collapse-recruitment physiology suggests that monitoring these variables on a breath-by-breath basis could guide RMs and identify the appropriate level of PEEP in mechanically ventilated patients. Second, lung recruitment together with a PEEP of 15 cm H2O showed the best effect on respiratory physiology in these morbidly obese patients undergoing general anesthesia. Thus, it can be reasoned that it was the active lung recruitment that induced these improvements in lung function while PEEP maintained them. This same level of PEEP without lung recruitment did not show the same physiological effect in morbidly obese patients, as also observed by Santesson34 and Erlandsson et al.35 Their results showed some similarities with our data, and 15 cm H2O of PEEP without prior lung recruitment had an inconsistent and limited impact on cardiopulmonary physiology.

Effects of Lung Recruitment on SIII The origin of SIII of the expired CO2-volume curve has been investigated over the last decades.12–17 From the ventilatory point of view, SIII is caused by two kinds of inhomogeneities in the transport of gas within the airways: “convection-dependent inhomogeneity,” which is a large-scale inhomogeneity within and between different ventilated lung zones mediated by a convective transport of CO2, and “diffusion-convection-dependent ANESTHESIA & ANALGESIA

Figure 2. Individual responses to the lung recruitment maneuver for Pao2, Paco2, dead space (VD)alv/tidal volume (VT)alv, the slope of Phase III (SIII), dynamic compliance (CDYN) and CO2flow. ⌬ ⫽ differences in the mean values summarizing all conditions before and after the alveolar recruitment strategy (ARS) are presented as percentage values above each graphic. A P value ⬍0.05 was considered significant. * ⫽ significant difference between 0 positive end-expiratory pressure (PEEP) and 15ARS.

inhomogeneity,” which is caused by the interaction of both convective and diffusive CO2 transport within asymmetric small airways.36 As an example, asthmatic and emphysematous patients show high SIII because of the known increases in RAW, which also affect the CO2 transport within the lungs.22,23 Vol. 109, No. 1, July 2009

From the perfusion point of view, SIII is caused by the continuous elimination of CO2 molecules through the alveolar-capillary membrane, as this gas is delivered to the lungs by the pulmonary blood flow. It was postulated that the effect of lung perfusion on SIII is responsible for approximately 10% of the sloping.18 © 2009 International Anesthesia Research Society

155

Table 1. Receiver Operating Characteristic (ROC) Analysis Variable

Cutoff (%)

AUC

Sensitivity

Specificity

P

Pao2 CDYN VDalv/VTalv Paco2

⬎20 ⬎40 ⬍20 ⬍5

0.81 0.84 0.67 0.81

0.75 0.78 0.65 0.67

0.74 0.76 0.62 0.66

⬍0.001 ⬍0.001 ⬍0.001 ⬍0.001

Prediction of the recruitment effect by a decrease in the slope of Phase III (SIII) was tested by calculating the area under the curve (AUC) of the ROC curve comparing zero positive end-expiratory pressure (PEEP) (ZEEP) before lung recruitment against the value at 15 cm H2O after it. We used a % change in SIII ⱕ20% and a cutoff in the partial pressure of oxygen (PaO2) ⱖ20%, in dynamic compliance (CDYN) ⬎40%, in the ratio of alveolar dead space to alveolar tidal volume (VDalv/VTalv) ⱕ20% and in the partial pressure of carbon dioxide (PaCO2) ⱕ5%. A P value ⬍0.05 was considered statistically significant.

Table 2. Data on Dead Space and Other Volumetric Capnography Variables During the Protocol ARS

PEEP (cm H2O) 0 VDaw (mL) VDalv (mL) VDphys (mL) VD/VT Pa-ETco2 (mm Hg) VTCO2,br (mL) SII (mm Hg/mL)

5

10

15

15ARS

10ARS

5ARS

0ARS

113 ⫾ 13* 116 ⫾ 13 124 ⫾ 12 135 ⫾ 15 117 ⫾ 16 107 ⫾ 15* 105 ⫾ 15* 107 ⫾ 13* 131 ⫾ 46† 122 ⫾ 43 110 ⫾ 38 101 ⫾ 44 84 ⫾ 46 99 ⫾ 45 106 ⫾ 42 110 ⫾ 48 244 ⫾ 48 239 ⫾ 43 234 ⫾ 42 236 ⫾ 48 201 ⫾ 48 206 ⫾ 50 211 ⫾ 46 216 ⫾ 51 0.39 ⫾ 0.08 0.39 ⫾ 0.07 0.38 ⫾ 0.06 0.38 ⫾ 0.07 0.32 ⫾ 0.06 0.33 ⫾ 0.07 0.34 ⫾ 0.07 0.35 ⫾ 0.08 6.3 ⫾ 0.3† 5.7 ⫾ 0.3 5.6 ⫾ 0.5 5.6 ⫾ 0.6 2.9 ⫾ 0.4 4.4 ⫾ 0.4 5.5 ⫾ 0.3 6.5 ⫾ 0.3 16.6 ⫾ 2.3 16.6 ⫾ 2.5 16.2 ⫾ 2.5 16.6 ⫾ 3.4 16.8 ⫾ 3.3 17.3 ⫾ 2.8 17.3 ⫾ 3.1 17.2 ⫾ 2.9 0.39 ⫾ 0.08 0.41 ⫾ 0.09 0.39 ⫾ 0.13 0.36 ⫾ 0.07 0.34 ⫾ 0.07 0.35 ⫾ 0.07 0.36 ⫾ 0.07 0.38 ⫾ 0.08

Data are presented as mean ⫾ SD. IndexARS ⫽ level of positive end-expiratory pressure (PEEP) applied after an alveolar recruitment strategy (ARS); VDaw ⫽ airway dead space; VDalv ⫽ alveolar dead space; VDphys ⫽ physiological dead space; VD/VT ⫽ ratio of physiological dead space to tidal volume; Pa-ETCO2 ⫽ arterial to end-tidal partial pressure differences of CO2; VTCO2,br ⫽ carbon dioxide elimination per breath; and SII ⫽ slope of Phase II. * Vs 15, P ⬍ 0.05. † Vs 15ARS, P ⬍ 0.05.

Table 3. Data on Ventilation and Lung Mechanics ARS

PEEP (cm H2O) 0

5

10

15

20ARS

15ARS

10ARS

5ARS

0ARS

PIP (cm H2O) 27 ⫾ 4* 28 ⫾ 4* 30 ⫾ 5 34 ⫾ 4 50 ⫾ 1 30 ⫾ 3 27 ⫾ 5* 26 ⫾ 6* 25 ⫾ 5* PLP(cm H2O) 21 ⫾ 6 23 ⫾ 5 24 ⫾ 6 27 ⫾ 7 50 ⫾ 1 24 ⫾ 4 20 ⫾ 6 20 ⫾ 4 20 ⫾ 5 Paw (cm H2O) 11 ⫾ 2*†‡§ 13 ⫾ 2†‡ 17 ⫾ 3† 21 ⫾ 3‡ 29 ⫾ 3 20 ⫾ 2 16 ⫾ 2 13 ⫾ 3 11 ⫾ 3 VT (mL) 613 ⫾ 72 614 ⫾ 69 609 ⫾ 80 618 ⫾ 62 1007 ⫾ 132 614 ⫾ 68 617 ⫾ 62 611 ⫾ 65 616 ⫾ 62 RR (bpm) 12 ⫾ 1.3 12 ⫾ 1.2 12 ⫾ 1.1 11 ⫾ 1.5 11 ⫾ 1.4 11 ⫾ 1.2 11 ⫾ 1.2 11 ⫾ 1.0 11 ⫾ 1.1 CDYN (mL/cm 30 ⫾ 6*†‡ 36 ⫾ 9†‡ 42 ⫾ 10†‡ 48 ⫾ 10† 39 ⫾ 15 68 ⫾ 13 57 ⫾ 16 44 ⫾ 12 33 ⫾ 8 H2O) RAW (mL 䡠 cm 18 ⫾ 3.9 17 ⫾ 3.2 16 ⫾ 2.0 15 ⫾ 1.8 17 ⫾ 4.1 15 ⫾ 3.7 16 ⫾ 3.6 19 ⫾ 4.3 19 ⫾ 3.7 H2O⫺1 䡠 s⫺1) PEF (L/s) 36 ⫾ 6 37 ⫾ 6㛳 39 ⫾ 5‡㛳 40 ⫾ 5†‡ 45 ⫾ 6 32 ⫾ 6 31 ⫾ 5 31 ⫾ 5 32 ⫾ 5* ETC (s) 0.98 ⫾ 0.2†‡ 0.96 ⫾ 0.1†‡ 0.92 ⫾ 0.2†‡ 0.91 ⫾ 0.1†‡ 0.81 ⫾ 0.1 1.27 ⫾ 0.3 1.25 ⫾ 0.3 1.20 ⫾ 0.2 1.16 ⫾ 0.2 Data are presented as mean ⫾ SD. IndexARS ⫽ level of positive end-expiratory pressure (PEEP) after an alveolar recruitment strategy; PIP ⫽ peak airway pressure; PLP ⫽ plateau pressure; Paw ⫽ mean airway pressure; VT ⫽ tidal volume; RR ⫽ respiratory rate; CDYN ⫽ respiratory dynamic compliance; RAW ⫽ expiratory airway resistance; PEF ⫽ expiratory peak flow; ETC ⫽ expiratory time constant. * Vs 15, P ⬍ 0.05. † Vs 15ARS, P ⬍ 0.05. ‡ Vs 10ARS, P ⬍ 0.05. § Vs 10, P ⬍ 0.05. 㛳 Vs 5ARS, P ⬍ 0.05.

Our group has confirmed the perfusion-dependent mechanism in the genesis of SIII in anesthetized patients.19 Therefore, SIII depends highly on the spatial and temporal distribution of ventilation and perfusion and can therefore be considered a general index of V/Q matching.18 –20 Considering the above concepts, changes in SIII can be explained best by the effect that lung recruitment has on the transport of CO2 from the capillary blood 156


into the airway opening. First, increasing the area for gas exchange by actively recruiting previously collapsed lung acini will naturally decrease the resistance to CO2 diffusion through the alveolar-capillary membrane. Second, increasing the cross-sectional area of the airways will result in decreased resistance to intrapulmonary CO2 transport by both diffusion and convection. The latter mechanism seems to be the most relevant for decreasing SIII. This hypothesis has ANESTHESIA & ANALGESIA

on the mechanical properties of the respiratory system. Thus, any change in respiratory mechanics induced by lung recruitment will play a major role in the genesis of SIII because the expiratory flow patterns change accordingly. As opposed to CO2 kinetics, the effect of lung collapse and recruitment on respiratory mechanics is well known.7,9 –11 Our results fit well with these studies showing that the elastic properties of the respiratory system increase and airway flow resistance decreases when lung volumes are restored by RMs and PEEP. These changes in lung mechanics can be explained by the concept of expiratory time constant (ETC). The ETC describes how fast the passive respiratory system responds to an external mechanical perturbation during expiration. Classically, a short ETC goes along with a fast response, whereas a long ETC indicates a delayed attainment of a new equilibrium within the respiratory system. At ZEEP, the lower CDYN caused by atelectasis allows the equilibrium to be reached very quickly while the VT is distributed within a smaller lung. This leads to decreased global compliance and increased PEF. The opposite mechanism is observed after lung recruitment: higher CDYN allows a slower and more homogeneous expiratory flow. ETC increases despite a decrement in RAW mainly because of a more-than-proportional increase in CDYN. VT is now distributed within more functional lung tissue, the characteristic of an “open-lung” condition. Increased compliance and decreased PEF are the consequences. The relationship between the effects of ARS and PEEP on lung mechanics and CO2 kinetics is represented by the Tau-CO2 concept (Fig. 3). Despite stable hemodynamics (Table 4) and strict protocol timing, this composite variable showed that the amount of CO2 eliminated within 1 ETC was highest at 15ARS, a condition related to the most favorable lung condition, as reflected by oxygenation and dead space. The CO2flow variable showed an even better profile than the Tau-CO2 concept for monitoring the recruitment effect, because the lower RAW seen after recruitment, which is now in the denominator, will augment the

Figure 3. Time constants for CO2 elimination throughout the study protocol. Tau-CO2 ⫽ CDYN ⫻ RAW ⫻ VTCO2,br, where CDYN is the dynamic compliance, RAW is the expiratory airway resistance, and VTCO2,br is the amount of CO2 eliminated in one breath. CO2flow ⫽ CDYN ⫻ VTCO2,br/RAW, using the same variables as above but in a modified arrangement, with RAW in the denominator. IndexARS ⫽ level of positive end-expiratory pressure (PEEP) after an alveolar recruitment strategy. Data are presented as mean ⫾ sd. * Vs 15ARS, P ⬍ 0.05. already been shown in normal weight patients with anesthesia-induced atelectasis11,12 and is now confirmed in these morbidly obese patients. In contrast to these results, Blanch et al.37 found no differences in VC indices when PEEP levels of 0, 5, 10, and 15 cm H2O were applied; however, this was without prior recruitment in patients with normal lungs or in patients with acute lung injury or acute respiratory distress syndrome. Their results are almost identical to our data obtained before the lung RM, in which changes were small and nonsignificant. The physiologic differences between the effect of PEEP with and without lung recruitment are well known7,8,13,32 and can easily explain such contradictory results.

Mechanism Behind Change in SIII Induced by Lung Recruitment As we have already pointed out, mechanisms affecting the CO2 transport within the lungs during expiration are the main determinants of SIII. Expiratory flow is a passive process that depends exclusively

Table 4. Hemodynamic Data ARS

PEEP (cm H2O)

HR (beats/min) MAP (mm Hg) MPAP (mm Hg) PCWP (mm Hg) CVP (mm Hg) CO (L/min) Shunt (%)

0

5

10

15

20ARS

15ARS

10ARS

5ARS

0ARS

69 ⫾ 4 81 ⫾ 14 23 ⫾ 7 14 ⫾ 3 14 ⫾ 4 6.7 ⫾ 1.5 0.14 ⫾ 0.07*†‡

67 ⫾ 5 75 ⫾ 10 24 ⫾ 6 14 ⫾ 2 14 ⫾ 4 6.5 ⫾ 1.5 0.14 ⫾ 0.06*†‡

66 ⫾ 6 78 ⫾ 11 26 ⫾ 6 16 ⫾ 3 13 ⫾ 3 6.4 ⫾ 1.6 0.13 ⫾ 0.06*†

64 ⫾ 6 79 ⫾ 10 26 ⫾ 5 16 ⫾ 4 15 ⫾ 4 6.3 ⫾ 1.6 0.11 ⫾ 0.05*

78 ⫾ 11 80 ⫾ 19 28 ⫾ 4 19 ⫾ 4 18 ⫾ 3 5.4 ⫾ 1.1 —

68 ⫾ 7 80 ⫾ 13 25 ⫾ 5 17 ⫾ 3 15 ⫾ 3 6.1 ⫾ 1.6 0.06 ⫾ 0.03

65 ⫾ 8 83 ⫾ 14 24 ⫾ 6 16 ⫾ 3 14 ⫾ 4 6.1 ⫾ 1.5 0.08 ⫾ 0.04

67 ⫾ 8 85 ⫾ 15 25 ⫾ 7 16 ⫾ 3 14 ⫾ 3 6.4 ⫾ 1.6 0.10 ⫾ 0.04

68 ⫾ 8 84 ⫾ 16 25 ⫾ 5 15 ⫾ 2 14 ⫾ 4 6.5 ⫾ 1.7 0.10 ⫾ 0.03

Data are presented as mean ⫾ SD. IndexARS ⫽ level of positive end-expiratory pressure (PEEP) after an alveolar recruitment strategy; MAP ⫽ mean systemic arterial blood pressure; MPAP ⫽ mean pulmonary arterial pressure; PCWP ⫽ pulmonary capillary wedge pressure; CVP ⫽ central venous pressure; CO ⫽ cardiac output; HR ⫽ heart rate. * Vs 15ARS, P ⬍ 0.05. † Vs 10ARS, P ⬍ 0.05. ‡ Vs 5ARS, P ⬍ 0.05.

Vol. 109, No. 1, July 2009


157

signal of lung improvement (Fig. 3). The physiological rationale of CO2flow goes beyond this simple mathematical association and is related to the above explanations. RAW to expiratory flow has an indirect correlation with CO2 transport within the lungs, thereby explaining why a decrement in RAW also decreases SIII.38 The main limitation of our study, beyond the small number of patients, is the fact that for methodological and ethical reasons we could not apply the different levels of PEEP in random order. The physiologic response to PEEP is totally different if it is used alone or in conjunction with a prior recruitment. Lung recollapse after lung recruitment is a pressure-dependent mechanism that is affected by the level of PEEP applied. Therefore, a perfectly designed study protocol would have applied all pressure levels in random order, with disconnections between any one of the PEEP steps and additional recruitment interventions before applying each one of the postrecruitment PEEPARS levels. Such a pure but rather radical protocol would not only have led to surfactant liberation39 by extensive lung RMs but also to yet another prolongation of the study time. Thus, this kind of study can only be performed under experimental conditions. Similarly, because of organizational limitations, the time at each PEEP level had to be limited to only 3 min, and therefore, the impact of longer time periods on the variables of interest was not assessed. Additionally, ETco2 and VTCO2,br values stabilized within the first minute at any PEEP level, supporting the fact that the 3-min time frame chosen was enough to obtain reliable data. Cardiovascular stability during the entire protocol excluded major hemodynamic interference with SIII. Despite general hemodynamic stability, we could not measure the distribution of pulmonary blood flow within the lungs. In theory, an uneven distribution of blood flow through different lung regions with different efficiencies for gas exchange could have increased SIII regardless of constant total amounts of such flow. However, the magnitude of the effect of an uneven distribution of pulmonary blood flow was estimated to be around 10% of total SIII and thus should not have influenced our results in an undue manner.18,19 In conclusion, the SIII of the VC curve decreased after lung recruitment and adequate levels of PEEP because of a decreased resistance to CO2 elimination within alveoli and airways. This slope provides aggregate information about gas exchange at the alveolarcapillary membrane, gas transport within airways, and respiratory mechanics. Because it can be measured noninvasively at the bedside on a breath-bybreath basis, this variable may be useful for guiding recruitment and for identifying appropriate levels of PEEP in anesthetized patients with impaired lung function, such as those in this study. 158


Note added in proof: See also Böhm et al.40 in this issue which reports different findings in the same patients. REFERENCES 1. Brismar B, Hedenstierna G, Lundquist H, Strandberg Å, Svensson L, Tokics L. Pulmonary densities during anaesthesia with muscular relaxation—a proposal of atelectasis. Anesthesiology 1985;62:422– 8 2. Strandberg A, Tokics L, Brismar B, Lundquist H, Hedenstierna G. Constitutional factors promoting development of atelectasis during anesthesia. Acta Anaesthesiol Scand 1987;31:21– 4 3. Damia G, Mascheroni D, Croci M, Tarenzi L. Perioperative changes in functional residual capacity in morbidly obese patients. Br J Anaesth 1988;60:574 – 8 4. Eichenberger AS, Proietti S, Wicky S, Frascarolo P, Suter M, Spahn DR, Magnusson L. Morbid obesity and postoperative pulmonary atelectasis: an underestimated problem. Anesth Analg 2002;95:1788 –92 5. Pelosi P, Croci M, Ravagnan I, Tredici S, Pedoto A, Lissoni A, Gattinoni L. The effects of body mass on lung volumes, respiratory mechanics and gas exchange during general anesthesia. Anaesth Analg 1998;87:654 – 60 6. Hedenstierna G, Tokics L, Strandberg A, Lundquist H, Brismar B. Correlation of gas exchange impairment to development of atelectasis during anaesthesia and muscle paralysis. Acta Anaesthesiol Scand 1986;30:183–91 7. Tusman G, Bo¨hm SH, Vazquez de Anda GF, do Campo JL, Lachmann B. Alveolar recruitment strategy improves arterial oxygenation during general anaesthesia. Br J Anaesth 1999; 82:8 –13 8. Rothen HU, Sporre B, Wegenius G, Hedenstierna G. Reexpansion of atelectasis during general anaesthesia: a computed tomography study. Br J Anaesth 1993;71:788 –95 9. Tusman G, Bo¨hm SH, Melkun F, Nador CR, Staltari D, Rodriguez A, Turchetto E. Efectos de la maniobra de reclutamiento alveolar y la PEEP sobre la oxigenacioń arterial en pacientes obesos anestesiados. Rev Esp Anestesiol Reanim 2002;49:177– 83 10. Whalen FX, Gajic O, Thompson GB, Kendrick ML, Que FL, Williams BA, Joyner MJ, Hubmayr RD, Wagner DO, Sprung J. The effects of the alveolar recruitment maneuver and positive end-expiratory pressure on arterial oxygenation during laparoscopic bariatric surgery. Anesth Analg 2006;102:298 –305 11. Tusman G, Bo¨hm SH, Sua´rez Sipmann F, Turchetto E. Alveolar recruitment improves ventilatory efficiency of the lungs during anesthesia. Can J Anaesth 2004;51:723–7 12. Tusman G, Bo¨hm SH, Sua´rez Sipmann F, Maisch S. Lung recruitment improves the efficiency of ventilation and gas exchange during one-lung ventilation anesthesia. Anesth Analg 2004;98:1604 –9 13. Tusman G, Suarez Sipmann F, Bo¨hm SH, Pech T, Reissmann H, Meschino G, Scandurra A, Hedenstierna G. Monitoring dead space during recruitment and PEEP titration in an experimental model. Intensive Care Med 2006;32:1863–71 14. Dutrieue B, Vanholsbeeck F, Verbank S, Paiva M. A human acinar structure for simulation of realistic alveolar plateau slopes. J Appl Physiol 2000;89:1859 – 67 15. Schwardt JD, Gobran SR, Neufeld GR, Aukburg SJ, Scherer PW. Sensitivity of CO2 washout to changes in acinar structure in a single-path model of lung airways. Ann Biomed Engl 1991;19:679 –97 16. Verbank S, Paiva M. Model simulations of gas mixing and ventilation distribution in the human lung. J Appl Physiol 1990;69:2269 –79 17. Glenny RW, Lamm WJ, Albert RK, Robertson HT. Gravity is a minor determinant of pulmonary blood flow distribution. J Appl Physiol 1991;71:620 –9 18. Prisk GK, Guy HJ, Elliot AR, West JB. Inhomogeneity of pulmonary perfusion during sustained microgravity on SLS-1. J Appl Physiol 1994;76:1730 – 8 19. Tusman G, Areta M, Climente C, Plit R, Suarez-Sipmann F, Rodríguez-Nieto MJ, Peces-Barba G, Turchetto E, Bo¨hm SH. Effect of pulmonary perfusion on the slopes of single-breath test of CO2. J Appl Physiol 2005;99:650 –5 20. Hofbrand BI. The expiratory capnogram: a measure of ventilation-perfusion inequalities. Thorax 1966;21:518 –24


21. Fletcher R, Jonson B. The concept of deadspace with special reference to the single breath test for carbon dioxide. Br J Anaesth 1981;53:77– 88 22. Fletcher R, Jonson B. Deadspace and the single breath test for carbon dioxide during anesthesia and artificial ventilation. Br J Anaesth 1984;56:109 –19 23. Englel LA. Gas mixing within acinus of the lung. J Appl Physiol 1983;54:609 –18 24. You B, Peslin R, Duvivier C, Vu VD, Grilliat JP. Expiratory capnography in asthma: evaluation of various shape indices. Eur Respir J 1994;7:318 –23 25. Schwardt JD, Neufeld GR, Baumgardner JE, Scherer PW. Noninvasive recovery of acinar anatomic information from CO2 expirograms. Ann Biomed Engl 1994;22:293–306 26. Kars AH, Bogaard JM, Stijnen T, de Vries J, Verbraak AF, Hilvering C. Deadspace and slope indices from the expiratory carbon dioxide-tension volume curve. Eur Respir J 1997;10:1829 –36 27. Jellinek H, Krafft P, Fitzgerald RD, Schwarz S, Pinsky MR. Right atrial pressure predicts hemodynamic response to apneic positive airway pressure. Crit Care Med 2000;28:672– 8 28. Fowler WS. Lung function studies. II. The respiratory dead space. Am J Physiol 1948;154:405–16 29. Englhoff H. Volumen inefficax. Bemerkungen zur Frage des scha¨dlichen Raumes. Uppsala La¨kareforen Forhandl 1938;44: 191–218 30. Suarez Sipmann F, Bo¨hm SH, Tusman G, Borges JB, Hedenstierna G. Tau-CO2: a novel variable to help optimizing PEEP. Intensive Care Med 2007;33(suppl 2):S143, PS550 31. Lachmann B, Jonson B, Lindroth M, Robertson B. Modes of artificial ventilation in severe respiratory distress syndrome. Lung function and morphology in rabbits after wash-out of alveolar surfactant. Crit Care Med 1982;10:724 –32 32. Suarez Sipmann F, Bo¨hm SH, Tusman G, Pesch T, Thamm O, Reissmann H, Reske A, Magnusson A, Hedenstierna G. Use of dynamic compliance for open lung positive end-expiratory pressure titration in an experimental study. Crit Care Med 2007;35:214 –21

Vol. 109, No. 1, July 2009

33. Zweig MH, Campbell G. Receiver-operating characteristic (ROC) plots: a fundamental evaluation tool in clinical medicine. Clin Chem 1993;39:561–77 34. Santesson J. Oxygen transport and venous admixture in the extremely obese. Influence of anaesthesia and artificial ventilation with and without positive end-expiratory pressure. Acta Anaestesiol Scand 1976;20:387–94 35. Erlandsson K, Odenstedt H, Lundin S, Stenqvist O. Positive end-expiratory pressure optimization using electric impedance tomography in morbidly obese patients during laparoscopic gastric bypass surgery. Acta Anaestesiol Scand 2006;50:833–9 36. Crawford ABH, Makowska M, Paiva M, Englel LA. Convectionand diffusion-dependent ventilation misdistribution in normal subjects. J Appl Physiol 1985;59:838 – 46 37. Blanch LL, Lucangelo U, Lopez-Aguilar J, Fernandez R, Romero PV. Volumetric capnography in patients with acute lung injury: effect of positive end-expiratory pressure. Chest 1994;105:219 –23 38. Blanch LL, Fernandez R, Saura P, Baigorri F, Artigas A. Relationship between expired capnogram and respiratory system resistance in critically ill patients during total ventilatory support. Eur Respir J 1999;13:1048 –54 39. Wirtz HRW, Dobbs LG. Calcium mobilization and exocytosis after one mechanical stretch of lung epithelial cells. Science 1990;250:1266 –9 40. Böhm SH, Thamm OC, von Sandersleben A, Bangert K, Langwieler TE, Tusman G, Strate TG, Standl TG. Alveolar recruitment strategy and high-positive end-expiratory pressure levels do not affect hemodynamics in morbidly obese intravascular volume-loaded patients. Anesth Analg 2009;109:160 –3


159

Brief Report

Alveolar Recruitment Strategy and High Positive EndExpiratory Pressure Levels Do Not Affect Hemodynamics in Morbidly Obese Intravascular Volume-Loaded Patients Stephan H. Bohm, MD* Oliver C. Thamm, MD* Alexandra von Sandersleben, MD* Katrin Bangert, MD* Thomas E. Langwieler, MD†

We evaluated the effect of the alveolar recruitment strategy and high positive end-expiratory pressure (PEEP) on hemodynamics in 20 morbidly obese (body mass index 50 ⫾ 9 kg/m2), intravascular volume-loaded patients undergoing laparoscopic surgery. The alveolar recruitment strategy was sequentially performed with and without capnoperitoneum and consisted of an upward PEEP trial, recruitment with 50 – 60 cm H2O of plateau pressure for 10 breaths, and a downward PEEP trial. Recruitment and high PEEP did not cause significant disturbances in any hemodynamic variable measured by systemic and pulmonary artery catheters. Transesophageal echocardiography revealed no differences in end-diastolic areas or evidence of segmental abnormalities in wall motion. (Anesth Analg 2009;109:160 –3)

Gerardo Tusman, MD‡ Tim G. Strate, MD, PhD† Thomas G. Standl, MD, PhD*

A

nesthesia-induced atelectasis is more pronounced in morbidly obese patients.1–5 The alveolar recruitment strategy (ARS) eliminates such atelectasis, reinstating normal lung function without collapse.6 However, the level of positive end-expiratory pressure (PEEP) needed to keep the lungs open is expected to be higher than in patients of normal weight.7–9 The high intrathoracic

From the Clinics of *Anesthesiology, †General Visceral and Thoracic Surgery, University Hospital Hamburg-Eppendorf, Germany; and ‡Department of Anesthesiology, Hospital Privado de Comunidad, Mar del Plata, Argentina. Accepted for publication March 19, 2009. Supported by the Clinic of Anesthesiology, University Hospital Hamburg-Eppendorf, Germany. Address correspondence to Stephan H. Bo¨hm, MD, CSEM Centre Suisse d’Electronique et de Microtechnique SA, Research Centre for Nanomedicine, Medical Sensors, Schulstr. 1, CH-7302 Landquart, Switzerland. Address e-mail to [email protected]. Oliver C. Thamm is currently at Clinic of Plastic and Reconstructive Surgery, Handsurgery, Burn Care Center, Hospital CologneMerheim, University of Witten/Herdecke, Germany. Katrin Bangert is currently at Department of Anaesthesiology, Israelitisches Krankenhaus Hamburg, Germany. Thomas G. Standl is currently at Clinic of Anesthesia, Critical Care and Palliative Care Medicine, Academic Hospital Solingen, Germany. Thomas E. Langwieler is currently at Department of Surgery, Evangelisches Amalie Sieveking-Krankenhaus, Hamburg, Germany. Tim G. Strate is currently at Clinic of Surgery, KrankenhausReinbek, Hamburg, Germany. Copyright © 2009 International Anesthesia Research Society DOI: 10.1213/ane.0b013e3181a801a3

160

pressures reached during the maneuver could potentially compromise hemodynamics, especially in morbidly obese patients, because of their predisposition for cardiovascular diseases.10 –13 The objective of this prospective study was to analyze the hemodynamic consequences of ARS and PEEP before and during capnoperitoneum in anesthetized morbidly obese patients.

METHODS After obtaining ethics committee approval and written informed consent from patients, we examined 20 patients with a body mass index (BMI) above 40 kg/m2 undergoing laparoscopic gastric banding. Patients with known cardiopulmonary diseases were excluded. Selected data from 11 of these 20 patients appear in this issue of the Journal.8 Anesthesia was induced with etomidate 0.15– 0.3 mg/kg, sufentanyl 0.1– 0.5 ␮g/kg, and succinylcholine 1 mg/kg of lean body weight and maintained with sevoflurane and boluses of 0.1– 0.5 ␮g/kg sufentanyl. After tracheal intubation, lungs were ventilated using a Cicero EM (Dra¨ger, Lu¨beck, Germany) with a tidal volume of 10 mL/kg of lean body weight, a respiratory rate of 10 –12 bpm, an I:E ratio of 1:1, and an Fio2 of 0.4. One liter of colloidal solution RheoHAES 70/0.4 6% (Braun, Melsungen, Germany) was administered IV for intravascular volume expansion before induction of anesthesia, whereas the preload status was assessed by measuring the end-diastolic area (EDA), Vol. 109, No. 1, July 2009

Figure 1. Dynamic pressure-flow plots describing changes in pulmonary vascular resistance caused by the alveolar recruitment strategy (ARS) with and without capnoperitoneum (CP). The pressure decrease across the pulmonary circulation as (mean pulmonary arterial pressure [MPAP] minus pulmonary capillary wedge pressure [PCWP]) is displayed on the vertical axis, and CO is displayed on the horizontal axis. Baseline values for CO and MPAP-PCWP at 0 cm H2O of positive end-expiratory pressure (PEEP) before the ARS intervention are found at the intersection of the axes. Differences from baseline for all patients and all PEEP levels are plotted as individual points (defined by ⌬CO on the x axis and ⌬MPAP-PCWP on the y axis) for the two studies and the two ventilatory conditions (with/without capnoperitoneum and before/after ARS). Shaded triangles represent zones of uncertainty, because either the slopes or the pressure intercepts of the lines defined by (MPAP-PCWP/CO) are physiologically implausible. The area above the shaded zones represents states of vasoconstriction relative to the baseline condition, whereas the area below them represents vasodilatation. The amount of measurements within each area of certainty is presented as a percentage of the total count, whereas the remaining ones (not shown) decrease within the shaded zones of uncertainty.15,16

with values between 16.0 and 31.2 cm2 defining “euvolemia.”14 A continuous infusion of saline solution was run at 5 mL per kg lean body weight and hour during anesthesia. Electrocardiogram, pulse oximetry, and invasive arterial blood pressure measurement in a radial artery were monitored. A pulmonary artery catheter (SwanGanz, Baxter, Irvine, CA) was placed through the right internal jugular vein and cardiac output (CO) was measured using triplicate thermodilutions. Systemic vascular resistance (SVR) was calculated using a standard formula with the mean systemic arterial blood pressure (MAP), the central venous pressure (CVP), and the CO values as follows:

SVR ⫽ MAP ⫺ CVP/CO ⫻ 80 (dyn 䡠 s/cm5) Pulmonary vascular resistance (PVR) was calculated using a standard formula with the mean pulmonary arterial pressure (MPAP), the pulmonary capillary Vol. 109, No. 1, July 2009

wedge pressure (PCWP), and the CO values as follows:

PVR ⫽ MPAP ⫺ PCWP/CO ⫻ 80 (dyn 䡠 s/cm5) Transesophageal echocardiography (TEE) was conducted using the Sonos 5500威 (Phillips Medical Systems, Bo¨bligen, Germany). The echocardiographic probe was initially inserted into the stomach to examine the morphology and function of the heart. Thereafter, the probe was withdrawn into the esophagus and had to remain there during the study to facilitate the surgery in the upper gastric region. We measured EDA of the left ventricle by planimetry using the leading edge method.14 Protocol: Patients were placed in a reverse Trendelenburg position (approximately 30°). ARS and PEEP titrations were sequentially performed before (without capnoperitoneum) and after (with capnoperitoneum) the start of surgery. The ARS was performed © 2009 International Anesthesia Research Society

161

Table 1. Hemodynamic Data PEEP (cm H2O) Without CP 0 5 10 15 ARS 15 10 5 0 With CP 0 5 10 15 20 ARS 20 15 10 5 0

HR MAP MPAP CVP PcwP CO PVR SVR (bpm) (mm Hg) (mm Hg) (mm Hg) (mm Hg) (L/min) (dyn 䡠 s/cm5) (dyn 䡠 s/cm5)

EDA (cm2)

67 ⫾ 9 66 ⫾ 8 67 ⫾ 9 65 ⫾ 9 66 ⫾ 13 65 ⫾ 11 65 ⫾ 10 67 ⫾ 11 68 ⫾ 12

74 ⫾ 8 77 ⫾ 9 78 ⫾ 9 79 ⫾ 9 78 ⫾ 17 82 ⫾ 12 82 ⫾ 13 84 ⫾ 14 86 ⫾ 16

24 ⫾ 7 26 ⫾ 6 26 ⫾ 7 28 ⫾ 7 28 ⫾ 5 26 ⫾ 5 25 ⫾ 6 26 ⫾ 7 25 ⫾ 7

14 ⫾ 4 15 ⫾ 4 15 ⫾ 4 18 ⫾ 4 15 ⫾ 4 15 ⫾ 4 14 ⫾ 4 15 ⫾ 4 15 ⫾ 4

16 ⫾ 4 17 ⫾ 4 18 ⫾ 4 18 ⫾ 4 19 ⫾ 4 18 ⫾ 3 18 ⫾ 4 18 ⫾ 4 17 ⫾ 4

6.6 ⫾ 1.4 6.5 ⫾ 1.3 6.4 ⫾ 1.3 6.3 ⫾ 1.3 5.5 ⫾ 1.4 5.9 ⫾ 1.5 6.1 ⫾ 1.5 6.4 ⫾ 1.6 6.4 ⫾ 1.6

104 ⫾ 38 115 ⫾ 32 114 ⫾ 45 118 ⫾ 53 142 ⫾ 64 107 ⫾ 57 97 ⫾ 42 114 ⫾ 59 101 ⫾ 41

773 ⫾ 190 804 ⫾ 195 819 ⫾ 201 851 ⫾ 220 938 ⫾ 290 941 ⫾ 238 924 ⫾ 222 893 ⫾ 218 894 ⫾ 255

26 ⫾ 4 27 ⫾ 4 27 ⫾ 4 27 ⫾ 3 25 ⫾ 3 26 ⫾ 3 28 ⫾ 4 29 ⫾ 5 30 ⫾ 4*

73 ⫾ 11 77 ⫾ 13 80 ⫾ 15 78 ⫾ 16 79 ⫾ 16 78 ⫾ 18 77 ⫾ 16 78 ⫾ 15 82 ⫾ 13 81 ⫾ 12 82 ⫾ 13

94 ⫾ 16 95 ⫾ 18 87 ⫾ 16 84 ⫾ 15 84 ⫾ 16 74 ⫾ 20 80 ⫾ 16 85 ⫾ 12 88 ⫾ 14 86 ⫾ 13 87 ⫾ 15

24 ⫾ 6 25 ⫾ 5 25 ⫾ 6 25 ⫾ 5 25 ⫾ 4 24 ⫾ 7 22 ⫾ 5 23 ⫾ 5 24 ⫾ 5 23 ⫾ 5 23 ⫾ 4

14 ⫾ 4 14 ⫾ 4 14 ⫾ 5 14 ⫾ 4 14 ⫾ 4 14 ⫾ 6 14 ⫾ 4 13 ⫾ 4 13 ⫾ 4 12 ⫾ 4 11 ⫾ 4*

16 ⫾ 5 15 ⫾ 5 16 ⫾ 5 16 ⫾ 5 15 ⫾ 3 15 ⫾ 5 14 ⫾ 5 15 ⫾ 5 14 ⫾ 5 13 ⫾ 4 11 ⫾ 4*

6.7 ⫾ 1.1 6.8 ⫾ 1.1 6.8 ⫾ 1.4 6.5 ⫾ 1.3 6.4 ⫾ 1.4 5.6 ⫾ 1.3 5.5 ⫾ 1.2* 6.1 ⫾ 1.5 6.8 ⫾ 1.1 7.3 ⫾ 1.2 7.4 ⫾ 1.5

103 ⫾ 35 118 ⫾ 38 128 ⫾ 42 113 ⫾ 28 132 ⫾ 32 149 ⫾ 63 111 ⫾ 53 90 ⫾ 21 101 ⫾ 34 102 ⫾ 26 116 ⫾ 33

1004 ⫾ 201 1034 ⫾ 215 926 ⫾ 207 847 ⫾ 168 887 ⫾ 177 931 ⫾ 262 1023 ⫾ 248 949 ⫾ 209 885 ⫾ 200 803 ⫾ 218* 825 ⫾ 205

29 ⫾ 3 29 ⫾ 4 28 ⫾ 3 29 ⫾ 4 27 ⫾ 4 25 ⫾ 5 27 ⫾ 4 28 ⫾ 4 29 ⫾ 3 28 ⫾ 4 30 ⫾ 3

Data are presented as mean ⫾ SD. PEEP ⫽ positive end-expiratory pressure; HR ⫽ heart rate; MAP ⫽ mean systemic arterial blood pressure; MPAP ⫽ mean pulmonary arterial pressure; CVP ⫽ central venous pressure; PCWP ⫽ pulmonary capillary wedge pressure; CO ⫽ cardiac output; PVR ⫽ pulmonary vascular resistance; SVR ⫽ systemic vascular resistance; EDA ⫽ end-diastolic area measured by transesophageal echocardiography; ARS ⫽ alveolar recruitment strategy. * Comparison of values obtained at the same level of PEEP before and after ARS; P ⬍ 0.05.

by increasing PEEP in increments of 5 cm H2O from 0 to 15 (without capnoperitoneum) and to 20 cm H2O (with capnoperitoneum). The patients’ lungs’ opening pressures were assumed to be around 50 cm H2O of airway pressure without capnoperitoneum and 60 cm H2O with capnoperitoneum. Therefore, plateau pressures of 50 and 60 cm H2O were applied for about 10 breaths to actively recruit collapsed lung tissue before and during capnoperitoneum, respectively. After approximately 1 min at maximal airway pressures, PEEP was decreased in steps of 5 cm H2O down to 0 cm H2O. Each level of PEEP before and after ARS was maintained for 3 min. Hemodynamic measurements and arterial blood samples were taken during the 3rd min of each PEEP period. Intraabdominal pressure was around 20 mm Hg as measured by the insufflator UHI-2 (Olympus, Tokyo, Japan). Statistical analysis was performed using SPSS version 13.0 (SPSS, Chicago, IL). Descriptive analysis and the Wilcoxon’s test were applied. Variables are presented as mean ⫾ sd, and a value of P ⬍ 0.05 was considered significant.

RESULTS Morbidly obese patients aged 39 ⫾ 6 yr with a BMI of 50 ⫾ 9 kg/m2 were enrolled in the study. All patients finished the protocol without complications. At the highest airway pressures, CO decreased from 6.3 ⫾ 1.3 to 5.5 ⫾ 1.4 L/min before capnoperitoneum and from 6.5 ⫾ 1.3 to 5.6 ⫾ 1.3 L/min during capnoperitoneum, returning to baseline values within 162

Brief Report

the next PEEP step. The absolute values of SVR and PVR did not show significant differences during the protocol. However, PVR values were slightly lower immediately after ARS than before ARS (Fig. 1).15,16 The remaining variables were stable during the entire protocol (Table 1). TEE did not reveal previous heart disease, evidence of segmental wall motion abnormalities, or significant differences in EDA during ARS (Table 1).

DISCUSSION The ARS and high levels of PEEP were hemodynamically well tolerated in intravascular volume-loaded morbidly obese patients undergoing laparoscopic surgery. This hemodynamic stability was observed both before and during capnoperitoneum. The potential hemodynamic repercussions of a reverse Trendelenburg position, capnoperitoneum, and mechanical ventilation depend mainly on their negative effect on venous return.17–21 Thus, the intravascular volume status plays a crucial role in any patient undergoing bariatric surgery, regardless of BMI. Jellinek et al.22 demonstrated the absence of any hemodynamic compromise at high levels of PEEP if CVPs were kept higher than 10 mm Hg. Our data confirm these results: no hemodynamic consequences were observed at higher airway pressures, provided that preload was kept within a normal range, as documented by the unremarkable EDAs and filling pressures. This is not too surprising, because the significantly elevated intraabdominal pressures of morbidly obese patients, particularly during the ANESTHESIA & ANALGESIA

surgery, reduced the transmural pressure acting on the hemodynamics. The intravascular volume loading with colloid (15 mL/kg of lean body weight) before anesthesia obviously prevented any hemodynamic disturbances in these fasted morbidly obese patients. Our results are also in agreement with those of Erlandsson et al.,17 who found that infusion of 1 L of intravascular volume expanders before applying high levels of PEEP avoided any negative effects in hemodynamics in morbidly obese patients. Limitations: Because of the particular surgical procedure, we could not assume the intragastric TEE position typically needed for optimal CO measurement. Therefore, we decided to report only values for EDA and the presence or absence of segmental wall motion abnormalities.

CONCLUSIONS After optimization of preload, lung recruitment and high positive airway pressures were hemodynamically well tolerated in morbidly obese patients with or without capnoperitoneum. ACKNOWLEDGMENTS The authors thank Stefan Maisch, Clinic of Anesthesiology, University Hospital Hamburg-Eppendorf, Germany, for his help during the conduct of this study and for valuable input during the revision process. REFERENCES 1. Brismar B, Hedenstierna G, Lundquist H, Strandberg A, Svensson L, Tokics L. Pulmonary densities during anaesthesia with muscular relaxation—a proposal of atelectasis. Anesthesiology 1985;62:422– 8 2. Damia G, Mascheroni D, Croci M, Tarenzi L. Perioperative changes in functional residual capacity in morbidly obese patients. Br J Anaesth 1988;60:574 – 8 3. Pelosi P, Croci M, Ravagnan I, Tredici S, Pedoto A, Lissoni A, Gattinoni L. The effects of body mass on lung volumes, respiratory mechanics and gas exchange during general anaesthesia. Anaesth Analg 1998;87:654 – 60 4. Strandberg A, Tokics L, Brismar B, Lundquist H, Hedenstierna G. Constitutional factors promoting development of atelectasis during anaesthesia. Acta Anaesthesiol Scand 1987;31:21– 4 5. Eichemberger AS, Proietti S, Wicky S, Frascarolo P, Suter M, Spahn DR, Magnusson L. Morbid obesity and postoperative pulmonary atelectasis: an underestimated problem. Anesth Analg 2002;95:1788 –92

Vol. 109, No. 1, July 2009

6. Tusman G, Bo¨hm SH, Vazquez de Anda GF, do Campo JL, Lachmann B. Alveolar recruitment strategy improves arterial oxygenation during general anaesthesia. Br J Anaesth 1999;82:8 –13 7. Whalen FX, Gajic O, Thompson GB, Kendrick ML, Que FL, Williams BA, Joyner MJ, Hubmayr RD, Warner DO, Sprung J. The effect of the alveolar recruitment maneuver and positive end-expiratory pressure on arterial oxygenation during laparoscopic bariatric surgery. Anesth Analg 2006;102:298 –305 8. Bo¨hm SH, Maisch S, von Sandersleben A, Thamm O, Passoni I, Martinez Arca J, Tusman G. The effects of lung recruitment on the phase III slope of volumetric capnography in morbidly obese patients. Anesth Analg 2009;109:151–9 9. Tusman G, Bo¨hm SH, Melkun F, Nador CR, Staltari D, Rodriguez A, Turchetto E. Efectos de la maniobra de reclutamiento alveolar y la PEEP sobre la oxigenacioń arterial en pacientes obesos anestesiados. Rev Esp Anestesiol Reanim 2002;49:177– 83 10. Manson JE, Colditz GA, Stampfer MJ, Willett WC, Rosner B, Monson RR, Speizer FE, Hennekens CH. A prospective study of obesity and risk of coronary heart disease in women. New Engl J Med 1990;322:882–9 11. Abdollahi M, Cushman M, Rosendaal FR. Obesity: risk of venous thrombosis and the interaction with coagulation factor levels and oral contraceptive use. Thromb Haemost 2003;89:493– 8 12. Valensi P, Thi BN, Lormeau B, Paries J, Attali JR. Cardiac autonomic function in obese patients. Int J Obes Relat Metab Disord 1995;19:113–18 13. Kenchaiah S, Evans JC, Levy D, Wilson PW, Benjamin EJ, Larson MG, Kannel WB, Vasan RS. Obesity and the risk of heart failure. N Engl J Med 2002;347:305–13 14. Cahalan MK, Ionescu P, Melton HE Jr, Adler S, Kee LL, Schiller NB. Automated real-time analysis of intraoperative transesophageal echocardiograms. Anesthesiology 1993;78:477– 85 15. Naeije R. Pulmonary vascular resistance. A meaningless variable? Intensive Care Med 2003;29:526 –9 16. Versprille A. Pulmonary vascular resistance. A meaningless variable. Intensive Care Med 1984;10:51–3 17. Erlandsson K, Odenstedt H, Lundin S, Stenqvist O. Positive end-expiratory pressure optimization using electric impedance tomography in morbidly obese patients during laparoscopic gastric bypass surgery. Acta Anaestesiol Scand 2006;50:833–9 18. Artuso D, Wayne M, Cassaro S, Cerabona T, Teixeira J, Rossi R. Hemodynamic changes during laparoscopic gastric bypass procedures. Arch Surg 2005;140:289 –92 19. Perilli V, Sollazzi L, Bozza P, Modesti C, Chierichini A, Tacchino RM, Ranieri R. The effects of the reverse trendelenburg position on respiratory mechanics and blood gases in morbidly obese patients during bariatric surgery. Anesth Analg 2000;91:1520 –5 20. Goodale RL, Beebe DS, McNevin MP, Boyle M, Letourneau JG, Abrams JH, Cerra FB. Hemodynamic, respiratory, and metabolic effects of laparoscopic cholecystectomy. Am J Surg 1993;166:533–7 21. Pinsky MR. Cardiovascular issues in respiratory care. Chest 2005;128:592S–597S 22. Jellinek H, Krafft P, Fitzgerald RD, Schwarz S, Pinsky MR. Right atrial pressure predicts hemodynamic response to apneic positive airway pressure. Crit Care Med 2000;28:672– 8


163

Case Report

Helmet Ventilation for Acute Respiratory Failure and Nasal Skin Breakdown in Neuromuscular Disorders Fabrizio Racca, MD* Lorenzo Appendini, MD† Giacomo Berta, MD* Luigi Barberis, MD* Ferdinando Vittone, MD*

Noninvasive ventilation (NIV) has been widely used to decrease the complications associated with tracheal intubation in mechanically ventilated patients with neuromuscular diseases in acute respiratory failure. However, nasal ulcerations might occur when masks are used as an interface. Helmet ventilation is a possible option in this case. We describe two patients with acute respiratory failure due to Duchenne muscular dystrophy who developed nasal bridge skin necrosis during NIV. Helmet pressure support ventilation caused significant patient-ventilator asynchrony, leading to NIV intolerance. Thus, biphasic positive airway pressure delivered by helmet was applied, which improved gas exchange and patientventilator interaction, allowing successful NIV. (Anesth Analg 2009;109:164 –7)

Cesare Gregoretti, MD‡ Gabriela Ferreyra, RT* Rosario Urbino, MD* V. Marco Ranieri, MD*

N

oninvasive ventilation (NIV) has been successfully used to support patients with Duchenne-type muscular dystrophy (DMD).1 However, NIV applied by mask often fails when high levels of support and/or long periods of ventilation are required to manage acute respiratory failure (ARF).2,3 A helmet may be an effective interface to deliver NIV in patients with ARF, avoiding skin breakdown and optimizing comfort.4 However, data suggest that noninvasive pressure support ventilation (N-PSV) delivered by helmet may worsen the patient–ventilator interaction compared with standard masks, eventually leading to NIV failure.5 We describe the drawbacks and a possible solution for two DMD patients with ARF and nasal ulceration treated with helmet ventilation. From the *Dipartimento di Anestesia e Rianimazione, Universita` di Torino, Ospedale S. Giovanni Battista-Molinette, Torino; †Divisione di Pneumologia, Fondazione Salvatore Maugeri, IRCCS, Istituto Scientifico di Veruno, Veruno (No); and ‡Servizio di Anestesia e Rianimazione, Azienda Ospedaliera CTO-CRF-Maria Adelaide, Torino Italy. Accepted for publication January 2, 2009. This work was performed at Universita` di Torino, Ospedale S. Giovanni Battista-Molinette, Torino, Italy. Address correspondence and reprint requests to Fabrizio Racca, MD, Dipartimento di discipline Medico-Chirurgiche—Sezione di Anestesiologia e Rianimazione, Ospedale S. Giovanni Battista Corso Dogliotti 14, 10126 Torino. Address e-mail to fabrizio.racca@gmail. com. Copyright © 2009 International Anesthesia Research Society DOI: 10.1213/ane.0b013e3181a1f708

164

CASE DESCRIPTIONS Case 1 A 19-yr-old DMD patient on long-term nocturnal nasal NIV was admitted to the emergency room with severe septic shock. In a couple of hours, he was transferred to the intensive care unit (ICU), where he was tracheally intubated and his lungs mechanically ventilated (Evita4 ventilator, Draeger, Lubeck, Germany). Six days later his trachea was extubated and mask N-PSV 24 h/d was initiated, using the same ICU ventilator with the leak compensation facility switched on. His arterial blood gases (ABGs) remained satisfactory (Table 1). Two different types of nasal mask, a Profile Lite nasal gel mask (Respironics, PA) and a Mirage nasal mask (ResMed, Australia), were alternated. A barrier dressing (a hydrocolloid sheet) was used to reduce the risk of skin breakdown. The masks were secured to avoid air leaks although allowing enough space to pass two fingers beneath the head strap. Four days later the patient developed nasal ulceration. Various interfaces were tried (nasal pillows, full face mask, and mouthpiece), and a bilevel ventilator (BiPAPVision—Respironics, Murrysville, PA) replaced the ICU ventilator without any success in terms of patient tolerance to mechanical ventilation. Subsequently, N-PSV was applied via a helmet interface (CaStar“R” StarMed, Italy). With this setting (Table 1) the patient showed poor clinical tolerance to helmet N-PSV, severe dyspnea and paradoxic respiratory motion. Vol. 109, No. 1, July 2009

Table 1. Ventilator Modes, Ventilator Settings, and Arterial Blood Gases in One of the Patients (Case 1) Over Time Timing Ventilator mode Interface Settings Pinsp PEEP (cm H2O) RR (breaths/min) TI (s) Blood gases pH Pao2/Fio2 Paco2 (mm Hg)

Before extubation

After extubation

Day 4 after extubation




PSV ETT

PSV Nasal mask

PSV Helmet

BIPAP Helmet

BIPAP Helmet

PSV Nasal mask

16 4 — —

18 5 — —

26 5 — —

26 5 20 1.1

20 5 16 1.2

14 5 — —

7.42 306 47

7.40 298 50

NA NA NA

7.42 315 48

7.43 295 45

7.44 300 40

BIPAP ⫽ biphasic positive airway pressure; ETT ⫽ endotracheal tube; FIO2 ⫽ fractional inspired oxygen tension; NA ⫽ not available; PaCO2 ⫽ arterial carbon dioxide tension; PaO2 ⫽ arterial oxygen tension; PEEP ⫽ positive end-expiratory pressure; Pinsp ⫽ inspiratory pressure support above PEEP; PSV ⫽ pressure support ventilation; RR ⫽ respiratory rate; TI ⫽ inspiratory time.

Figure 1. Experimental record from patient 1. Airway opening pressure (Pao), ribcage (RC), abdomen (AB), and respiratory inductive plethysmography (RIP) tidal volume (VT) signals during helmet pressure support ventilation (PSV) and biphasic positive airway pressure (BIPAP) are shown. Respiratory pattern, ribcage–abdominal motion and major patient–ventilator asynchrony were measured by means of simultaneous recording of RIP (RespitracePlus, NIMS, FL) and Pao. Pao was measured at the helmet inspiratory port with a differential transducer (Digima-Clic, ⫾100 cm H2O, Special Instruments, Germany). Vt was reported in arbitrary units. All signals were collected on a personal computer through a 12-bit analog-to-digital converter (National Instrument DAQCard 700; TX) at a sampling rate of 200 Hz (ICU-lab, KleisTEK Engineering, Bari, Italy). The horizontal dashed line indicates the nominal inspiratory pressure level delivered by the ventilator. TI is the inspiratory time. During BIPAP, most of the mechanical breaths were not triggered by the patient (VT begins after start of mechanical assistance). The percentage of ventilator inspiratory assistance, defined as the percentage of TI spent at nominal pressure level (time elapsed between the first and the second vertical dashed lines), is greater during BIPAP than during PSV. Ineffective efforts (VT without any mechanical assistance identified by arrows), associated with paradoxical AB motion, are prevalent during PSV.

Several ineffective efforts were also disclosed by the simultaneous analysis of airway opening pressure (Pao) and respiratory inductive plethysmography (RIP) (RespitracePlus, NIMS, FL) tracings (Fig. 1). This occurred despite the fact that the ventilator (Evita4 ventilator) was set at the highest trigger sensitivity that did not induce auto-triggering. The Evita4 ventilator was switched to BIPAP set with the same Vol. 109, No. 1, July 2009

inspiratory-expiratory pressures used during N-PSV, with respiratory rate (RR) and inspiratory time (TI) set as close as possible to the patient’s RR and timing (Table 1) as measured during a helmet continuous positive airway pressure (CPAP) trial. With these ventilator settings, dyspnea, patient-ventilator synchrony (Fig. 1), and ABGs improved (Table 1). The patient was kept on helmet ventilation for 8 days until the nasal ulceration healed and the ABGs were stable. Eighteen days after admission to the ICU he was returned to nocturnal nasal NIV, and 3 days later he was discharged to the neurological ward. The patient was eventually discharged from the hospital and is living at home on nasal N-PSV.

Case 2 An 18-yr-old DMD patient on long-term nocturnal nasal NIV was admitted comatose to the emergency room with decompensated respiratory acidosis. He was immediately transferred to the ICU, where he was tracheally intubated and his lungs mechanically ventilated for 48 h. After improvement in ABGs, the patient’s trachea was extubated and he was switched to mask ventilation. Two different types of nasal mask and a hydrocolloid sheet were used to reduce the risk of skin breakdown. His ABGs remained satisfactory (pH 7.35, Pao2/fraction of inspired oxygen [Fio2] 275, Paco2 52 mm Hg). However, the patient failed N-PSV after 72 h because of nasal ulceration. A ventilator strategy similar to that used in Case 1 was attempted, without any success in terms of patient tolerance to mechanical ventilation. Next, N-PSV was applied using a helmet interface. The patient showed severe dyspnea, paradoxic respiratory motion, and several ineffective efforts. Subsequently, the ventilator was switched to BIPAP, set at the same inspiratory-expiratory pressures used during N-PSV, with a RR and TI set as close as possible to the patient’s RR and timing (RR 18 min, TI 1.3 s). With these ventilator settings, dyspnea, patient-ventilator synchrony, and ABGs (pH 7.38, Pao2/Fio2 240, Paco2 49 mm Hg) improved. © 2009 International Anesthesia Research Society

165

The patient remained on helmet ventilation for 6 days until skin breakdown healed and his ABGs were stable. Seven days after beginning helmet ventilation he was returned to nocturnal nasal NIV, and 1 wk later he was discharged to the neurological ward. The patient was eventually discharged from the hospital and lived at home on nasal N-PSV. He died 1 yr later because of a primary nonrespiratory cause (abdominal sepsis).

DISCUSSION In both clinical cases presented here, we attempted to overcome the nasal mask-related complications that precluded NIV effectiveness by using a helmet interface. The N-PSV irreversible intolerance to conventional ventilator management was successfully treated with helmet ventilation in an assist-controlled (synchronized BIPAP) mode. We confirm previous reports that the risk of facial lesions and other side effects associated with the use of a standard interface may be increased because of high ventilator dependence in DMD patients with ARF.3 Notwithstanding this risk, our DMD patients were switched early to NIV when they were still completely ventilator-dependent, according to current guidelines and clinical practice,1,6 to decrease the complications associated with tracheal intubation in mechanically ventilated patients. Until the onset of skin lesions, NIV proved to be effective in reversing ARF in both patients (Table 1). Despite efforts made to avoid it, skin breakdown on the nose bridge occurred after 72–96 h of continuous NIV. Thus alternative patient-ventilator interfaces were evaluated in an attempt to avoid invasive mechanical ventilation. Studies suggest that N-PSV delivered by helmet is better tolerated than N-PSV delivered by conventional masks and is similarly effective in reducing the need for endotracheal intubation.4 These data warranted the choice of helmet ventilation in the two DMD patients. At first, based on the literature,4 assisted mechanical ventilation in the form of PSV was used with the helmet. In this condition, both patients showed poor patient–ventilator synchrony and several unassisted inspiratory efforts (Fig. 1), despite the fact that the ventilator was set with very high trigger sensitivity. Furthermore, the percentage of inspiratory assistance, defined as % of inspiratory time spent at nominal pressure level,7 was very low, suggesting quite poor respiratory muscle unloading (Fig. 1). In both cases, ABGs could not be measured during helmet PSV because of very severe clinical intolerance to the ventilator mode. These findings confirm previous reports in the literature showing significant patient–ventilator asynchrony, discomfort and inefficacy in unloading the respiratory muscles during helmet N-PSV as a result of helmet characteristics conflicting with the algorithms governing PSV.5 In particular, the present case 166

Case Report

report confirms that patients with DMD may be exposed to severe asynchrony during N-PSV.8 Assist-controlled modalities are commonly used in the ICU to assure triggering and cycling of the ventilator in the presence of persistent patient–ventilator asynchrony.6,9 During BIPAP provided by the Evita4 ventilator, a positive pressure breath is delivered regardless of the patient’s trigger capability, TI of the mechanical breaths is preset, and spontaneous breathing is permitted in any phase of the mechanical cycle.10 In our study, to optimize patient–ventilator interactions, TI and RR during BIPAP were matched with those obtained during a helmet CPAP trial. The inspiratory triggering threshold was set at the most sensitive level that did not induce auto-triggering, although pressure rise time was set at the ventilator’s fastest ramp. With these changes, major patient-ventilator asynchrony and the ineffective efforts during synchronized BIPAP mode disappeared, and percentage of inspiratory assistance increased (Fig. 1). Some technical aspects merit discussion. First, the assessment of patient–ventilator asynchrony requires at least an esophageal catheter system in place to record inspiratory muscle efforts11 or diaphragm recordings12 coupled with flow and Pao tracings. Esophageal catheters to record pressure or diaphragm electromyography and a mouthpiece to record airflow were not used in the two patients presented here because of swallowing dysfunction and dyspnea. As an alternative, we used RIP to detect patient–ventilator asynchrony and tidal volume. According to the criteria used in sleep medicine, ineffective inspiratory efforts were defined as thoracic-abdominal displacements not assisted by the ventilator positive pressure boost (Fig. 1). Second, RIP tracings and Pao were recorded in each patient only during the first hour of helmet ventilation. Subsequently, RR was clinically assessed13 and TI extrapolated from RR assuming an inspiratoryto-expiratory ratio of 1:2. Third, the pathophysiology of ARF can change quite quickly, making NIV assistance inadequate. To avoid ventilator overassistance or under-assistance, patient RR, and TI were screened daily during a CPAP trial to reset ventilator parameters as close as possible to the patient’s own values. Moreover, a weaning protocol based on progressive airway pressure reduction was implemented.13 This approach allowed nasal lesions to heal and patients to recover quickly from ARF (8 and 6 days, respectively). In conclusion, the two clinical cases presented and discussed here show that helmet ventilation can be considered to manage intolerance to a nasal mask during NIV in patients with DMD and ARF. However, because difficult triggering and poor patient ventilator synchrony occur during helmet N-PSV, assist pressurecontrolled mode may be more indicated to provide helmet ventilation. ANESTHESIA & ANALGESIA

ACKNOWLEDGMENTS The authors would like to thank Michele Mele for his technical support in preparing the manuscript. REFERENCES 1. Vianello A, Bevilacqua M, Arcaro G, Gallan F, Serra E. Noninvasive ventilatory approach to treatment of acute respiratory failure in neuromuscular disorders: a comparison with endotracheal intubation. Intensive Care Med 2000;26:384 –90 2. Carlucci A, Richard JC, Wysocki M, Lepage E, Brochard L; SRLF Collaborative Group on Mechanical Ventilation. Noninvasive versus conventional mechanical ventilation. An epidemiologic survey. Am J Respir Crit Care Med 2001;163:874 – 80 3. Gregoretti C, Confalonieri M, Navalesi P, Squadrone V, Frigerio P, Beltrame F, Carbone G, Conti G, Gamna F, Nava S, Calderini E, Skrobik Y, Antonelli M. Evaluation of patient skin breakdown and comfort with a new face mask for non-invasive ventilation: a multi-centre study. Intensive Care Med 2002;28:278 – 84 4. Antonelli M, Conti G, Pelosi P, Gregoretti C, Pennisi MA, Costa R, Severgnini P, Chiaranda M, Proietti R. New treatment of acute hypoxemic respiratory failure: noninvasive pressure support ventilation delivered by helmet—a pilot controlled trial. Crit Care Med 2002;30:602– 8 5. Racca F, Appendini L, Gregoretti C, Stra E, Patessio A, Donner CF, Ranieri VM. Effectiveness of mask and helmet interfaces to deliver noninvasive ventilation in a human model of resistive breathing. J Appl Physiol 2005;99:1262–71

Vol. 109, No. 1, July 2009

6. British Thoracic Society Standards of Care Committee. Noninvasive ventilation in acute respiratory failure. Thorax 2002; 57:192–211 7. Racca F, Appendini L, Gregoretti C, Varese I, Berta G, Vittone F, Ferreyra G, Stra E, Ranieri VM. Helmet ventilation and carbon dioxide rebreathing: effects of adding a leak at the helmet ports. Intensive Care Med 2008;34:1461– 8 8. Fanfulla F, Delmastro M, Berardinelli A, D’Artavilla Lupo N, Nava S. Effects of different ventilator settings on sleep and inspiratory effort in patients with neuromuscular disease. Am J Respir Crit Care Med 2005;172:619 –24 9. Mehta S, Hill NS. State of the art: noninvasive ventilation. Am J Respir Crit Care Med 2001;163:540 –77 10. Putensen C, Wrigge H. Airway pressure release ventilation. In: Tobin MJ, ed. Principles and practice of mechanical ventilation. 2nd ed. New York: McGraw-Hill, 2006:327–34 11. Nava S, Bruschi C, Fracchia C, Braschi A, Rubini F. Patientventilator interaction and inspiratory effort during pressure support ventilation in patients with different pathologies. Eur Respir J 1997;10:177– 83 12. Beck J, Gottfried SB, Navalesi P, Skrobik Y, Comtois N, Rossini M, Sinderby C. Electrical activity of the diaphragm during pressure support ventilation in acute respiratory failure. Am J Respir Crit Care Med 2001;164:419 –24 13. Ely EW, Meade MO, Haponik EF, Kollef MH, Cook DJ, Guyatt GH, Stoller JK. Mechanical ventilator weaning protocols driven by nonphysician health-care professionals: evidence-based clinical practice guidelines. Chest 2001;120:454S– 463S


167

Obstetric Anesthesiology Section Editor: Cynthia A. Wong

The Interaction Between Epidural 2-Chloroprocaine and Morphine: A Randomized Controlled Trial of the Effect of Drug Administration Timing on the Efficacy of Morphine Analgesia Paloma Toledo, MD Robert J. McCarthy, PharmD Mary Jane Ebarvia, BS, RN Christopher J. Huser, MD Cynthia A. Wong, MD

BACKGROUND: The efficacy and duration of epidural morphine analgesia is diminished when administered after 2-chloroprocaine compared with lidocaine. The mechanism of the interaction between 2-chloroprocaine and morphine is unknown. Possible explanations include differences in the latency and duration of action of the two drugs or opioid receptor antagonism. We hypothesized that administration of epidural morphine 30 min before the initiation of 2-chloroprocaine anesthesia would result in postoperative analgesia of similar duration and quality to that achieved by epidural morphine after the initiation of lidocaine anesthesia in patients undergoing postpartum tubal ligation. METHODS: Subjects undergoing bilateral postpartum tubal ligation after vaginal delivery with epidural analgesia were randomized to one of three groups. Subjects received epidural morphine or saline 30 min before the initiation of analgesia with 3% 2-chloroprocaine (two groups) or 2% lidocaine (one group), and at the time of surgical incision, they received either epidural saline or morphine. The duration of analgesia was defined as the time from morphine administration until the first request for supplemental analgesia. Duration of epidural morphine analgesia was compared among groups using Kaplan–Meier survival analysis and the log-rank test. RESULTS: Administration of epidural morphine 30 min before the initiation of 2-chloroprocaine anesthesia (n ⫽ 29) resulted in a longer median duration of analgesia (28.6 h [95% CI 4.4 –52.7]) compared with the administration of morphine after 2-chloroprocaine anesthesia (n ⫽ 30) (2.2 h [95% CI 0 – 4.8]) (P ⫽ 0.006). The median duration of analgesia observed when morphine was administered before 2-chloroprocaine was similar to that observed when morphine was administered after initiation of lidocaine anesthesia (n ⫽ 28) (25.8 h [95% CI 10.7– 40.9]) (P ⫽ 0.83). Pain scores were not different in the postanesthesia care unit, but were higher on admission to the postpartum unit in the subjects receiving morphine after 2-chloroprocaine. Supplemental morphine equivalents administered in the first 48 h were similar among groups and there were no differences in opioid-related side effects. DISCUSSION: This study demonstrates that administration of epidural morphine 30 min before epidural anesthesia with 2-chloroprocaine provides a similar duration of analgesia as epidural morphine after epidural lidocaine anesthesia. This suggests that the observed interaction between epidural morphine and 2-chloroprocaine is a result of differences in latency and duration of action of the two drugs, or that the administration of morphine before 2-chloroprocaine effectively blocks a receptor site antagonism. (Anesth Analg 2009;109:168 –73)

S

tudies have shown that there is considerable pain after postpartum tubal ligation (PPTL), and this pain is often undertreated.1–3 Marcus et al.4 found that epidural morphine, as a part of a multimodal analgesic regimen, decreased the need for systemic analgesia after PPTL. However, in this study the morphine was

administered after epidural lidocaine anesthesia because of concern for decreased morphine efficacy if administered after the administration of epidural 2-chloroprocaine. Epidural 2-chloroprocaine anesthesia is often used for PPTL procedures because the rapid onset and short duration of effect better match

From the Department of Anesthesiology, Northwestern University Feinberg School of Medicine, Chicago, Illinois. Accepted for publication January 14, 2009. Cynthia A. Wong is the Section Editor of Obstetric Anesthesiology for the Journal. This manuscript was handled by Steven L. Shafer, Editor-in-Chief, and Dr. Wong was not involved in any way with the editorial process or decision.

Reprints will not be available from the author. Address correspondence to Cynthia A. Wong, MD, Department of Anesthesiology, 251 E. Huron St., F5-704, Chicago, IL 60611. Address e-mail to [email protected]. Copyright © 2009 International Anesthesia Research Society

168

DOI: 10.1213/ane.0b013e3181a40cf6

Vol. 109, No. 1, July 2009

the duration of surgery compared with epidural lidocaine. It would therefore be ideal if one could administer epidural morphine with 2-chloroprocaine and provide effective analgesia. Both the quality and duration of analgesia have been shown to be diminished when epidural morphine is administered with or shortly after the administration of 2-chloroprocaine.5,6 There is controversy regarding the mechanism of this antagonism. Suggested mechanisms include direct antagonism of the opioid receptors by 2-chloroprocaine or one of its metabolites, antagonism of an intracellular second messenger, decreased morphine availability because of a decrease in perineural pH, or a pharmacodynamic window phenomenon caused by the long latency of morphine’s analgesic effect combined with the rapid regression of the 2-chloroprocaine nerve block.7–9 We hypothesized that epidural administration of morphine before the epidural administration of 2-chloroprocaine would provide better analgesia compared with morphine administered after epidural administration of 2-chloroprocaine. The purpose of this study was to examine the effect of varying the timing of the epidural administration of morphine relative to the epidural administration of 2-chloroprocaine, when compared with epidural morphine administered with epidural lidocaine.

METHODS This double-blind, prospective study was approved by the Northwestern University IRB. Clinical trial registration for this study can be found at ClinicalTrial. gov; url: http://www.clinicaltrials.gov; registration identifier: NCT00487084. Healthy women, scheduled for bilateral PPTL after vaginal delivery with epidural analgesia were eligible to participate in the study. All parturients had maintenance labor analgesia with patient-controlled epidural analgesia consisting of bupivacaine 0.625 mg/mL with 2 ␮g/mL fentanyl. After delivery, all patients were monitored for a minimum of 1 h before starting the PPTL. Exclusion criteria included body mass index more than 40 kg/m2, history of obstructive sleep apnea, drug abuse, chronic opioid use, a nonfunctioning epidural catheter, or an allergy or sensitivity to any of the study medications. Subjects were randomized via a computer-generated random number sequence to one of three study groups. Group assignments were sealed in sequentially numbered opaque envelopes. Patients who consented to study participation were randomized after vaginal delivery. A series of three syringes was prepared for each patient by an anesthesiologist or anesthesia nurse who was not involved in patient care. The first syringe contained either 3 mg (6 mL) of preservative free morphine sulfate (morphine 0.5 mg/mL) or 6 mL of preservative free saline. The second syringe contained 30 mL of either 3% 2-chloroprocaine or 2% lidocaine with epinephrine 5 ␮g/mL and NaHCO3 Vol. 109, No. 1, July 2009

1 mEq/10 mL. The third syringe contained either 6 mL of preservative-free saline or 3 mg (6 mL) of preservative-free morphine sulfate. The first syringe was administered 30 min before the initiation of epidural anesthesia. Epidural anesthesia was initiated using 15–30 mL of the local anesthetic in syringe 2. After confirmation of a T4 sensory level, surgery commenced, and at the time of skin incision, the contents of the third syringe were administered. The sequence of syringes for each group was as follows, MCS Group, morphine followed by 2-chloroprocaine, then saline; SCM Group, saline followed by 2-chloroprocaine, then morphine; SLM Group, saline, followed by lidocaine, then morphine (Fig. 1). All subjects were premedicated with 0.3M oral sodium citrate (30 mL), IV ranitidine 50 mg, and metoclopramide 10 mg before the procedure. A periumbilical incision was made and the PPTL was performed using a modified Pomeroy technique. Intraoperatively, midazolam was administered at the discretion of the anesthesiologist. Supplemental subcuticular lidocaine infiltration by the obstetrician, used for inadequate anesthesia, was recorded. Nausea and/or vomiting were treated with ondansetron 4 mg. The epidural catheter was removed at the conclusion of the procedure. Postoperatively, ibuprofen 600 mg was administered per os every 6 h. The first dose was administered within 15 min of arrival in the postanesthesia care unit (PACU). At the first request for supplemental analgesia, patients were given acetaminophen 325 mg with hydrocodone 10 mg per os, with an additional dose allowed after 1 h if requested. Acetaminophen/hydrocodone was given every 4 h as requested. Pruritus was treated with IV nalbuphine 2.5 mg. Nausea and/or vomiting were treated with IV ondansetron 4 mg or promethazine 12.5 mg. Demographic data collected included the subject’s age, height, and weight. Vaginal delivery data included the presence and degree of perineal lacerations. Intraoperative data included the time interval from delivery to PPTL incision, the amount of local anesthetic administered to achieve surgical anesthesia, duration of the procedure and the use of midazolam or supplemental local anesthetic administered by the surgeon. The duration of the PACU stay was recorded. Pain, nausea, and pruritus scores were determined on admission to the PACU and postpartum unit and every 4 h for 12 h by nurses blinded to group assignment. Subjects rated their pain using an 11-point verbal rating scale, ranging from 0 for no pain to 10 for the worst pain imaginable. Pruritus and nausea were rated on a three-point scale (none, mild, and moderate to severe). Treatments for nausea or pruritus were recorded for the first 24 h. The duration of analgesia was defined as the time from morphine administration until the first request for supplemental analgesia. Acetaminophen/hydrocodone or other opioids administered during the first 48 h postoperatively were recorded and the hydrocodone © 2009 International Anesthesia Research Society

169

Figure 1. Study medications and postoperative analgesia administration sequence. Groups, MCS ⫽ morphine followed by 2-chloroprocaine, then saline; SCM ⫽ saline followed by 2-chloroprocaine, then morphine; SLM ⫽ saline followed by lidocaine, then morphine.

dose was converted to oral morphine equivalents (10 mg hydrocodone ⫽ 7.5 mg morphine). The primary outcome variable was time to first request for supplemental analgesia. Secondary outcome variables included the pain verbal rating scale score, total 48 h opioid consumption, pruritus, and nausea scores, and the rate of treatment for pruritus or nausea. Based on a previous study that demonstrated a decreased efficacy of morphine when administered after 2-chloroprocaine, a sample of 56 subjects was estimated to have an 80% power to detect a 30% difference in the proportion of subjects with continuing analgesia in the SLM Group versus the SCM Group when 50% of the subjects in the SCM had requested supplemental analgesia.5 Expecting a similar difference between groups given morphine before (MCS) and after 2-chloroprocaine (SCM), a third group of 28 subjects was added to compare the influence of time of morphine and 2-chloroprocaine administration. The primary outcome was compared using Kaplan–Meier survival analysis and the logrank test. Subjects who did not request supplemental analgesia were assigned a time to analgesia request of 48 h. Categorical data were compared using a ␹2 statistic and interval data were compared using Kruskal–Wallis one-way analysis of variance. Bonferroni correction was used for post hoc comparisons. P ⬍ 0.05 was used to reject the null hypothesis.

RESULTS Ninety-nine subjects were consented and randomized. Twelve subjects were excluded after randomization. The flow of subjects through the study protocol, including protocol violations, is shown in Figure 2. Eighty-seven subjects completed the protocol. There were no differences among the groups in weight, height, or the presence or degree of perineal lacerations (Table 1). The median duration of analgesia for subjects in MCS and SLM Groups was 28.6 h (95% CI 4.4 –52.7) and 25.8 h (95% CI 10.7– 40.9), respectively, compared with 2.2 h (95% CI 0 – 4.8) for SCM Group (Fig. 3). Thirty-three percent of the subjects in the SCM Group compared with none in MCS and SLM Groups requested analgesia within 90 min of morphine administration (P ⬍ 0.01). However, during the 48-h observation 170

Epidural Morphine and Chloroprocaine

period, 41% of the subjects in MCS Group compared with 46% of the subjects in SLM Group and 63% or the subjects in SCM Group required rescue analgesia (P ⫽ 0.20). Pain scores did not differ among groups on admission to the PACU. However, on admission to the postpartum unit, patients in SLM Group had less pain on the verbal rating scale than SCM and MCS Groups (Fig. 4). Forty-eight hour supplemental analgesia requirements, measured in morphine equivalents, did not differ among groups. There were no differences in the delivery to procedure time, procedure duration, intraoperative medication use, or duration of PACU stay among groups. Neither the incidence of side effects, the percentage of subjects requiring treatment for nausea, vomiting, or pruritus differed among groups at any time (Table 2).

DISCUSSION The important finding from this study was that the administration of morphine 30 min before the initiation of anesthesia with 2-chloroprocaine for PPTL resulted in a duration of analgesia similar to that of morphine administered with lidocaine. When epidural morphine was administered after 2-chloroprocaine, the interval to the first request for supplemental analgesia was decreased, compared with both of the other study groups. The reduction in opioid analgesia when morphine was administered after 2-chloroprocaine observed in this study is consistent with the findings of Grice et al.7 In their study, laboring parturients requesting epidural analgesia were randomized to receive a test dose of either 2% 2-chloroprocaine or 2% lidocaine, followed by epidural analgesia with bupivacaine, fentanyl, and epinephrine. The duration of epidural analgesia was reduced after an epidural test dose of 2-chloroprocaine. Likewise, Camann et al.10 demonstrated a reduced duration of epidural fentanyl analgesia after cesarean delivery when epidural 2-chloroprocaine was administered for anesthesia compared with lidocaine. In their study, fentanyl was not administered until the first request for analgesia. Although there was a difference in the initial duration of analgesia, the 24-h opioid requirements were not different between groups. ANESTHESIA & ANALGESIA

Figure 2. Subject flow diagram. Groups, MCS ⫽ morphine followed by 2-chloroprocaine, then saline; SCM ⫽ saline followed by 2-chloroprocaine, then morphine; SLM ⫽ saline followed by lidocaine, then morphine. Similar results have been demonstrated in other studies when morphine was administered after 2chloroprocaine. Eisenach et al.5 randomized women undergoing cesarean delivery to receive 2-chloroprocaine or lidocaine as an epidural test dose, followed by epidural anesthesia with 0.5% bupivacaine and epidural morphine administration (5 mg) after delivery of the infant. There was a reduction in the median duration of analgesia from 24 h in the lidocaine group to 16 h in the 2-chloroprocaine group. In another study, subjects who received epidural morphine after epidural anesthesia with 2-chloroprocaine required more morphine in the Vol. 109, No. 1, July 2009

first 4 h postoperatively after elective cesarean deliveries than subjects who had epidural lidocaine anesthesia.6 However, the groups did not differ in the amount of morphine consumed over 24 h. In contrast, Hess et al.9 did not find a difference in the duration of analgesia when parturients were randomized to receive 3% 2-chloroprocaine 150 mg or placebo, followed by epidural morphine 3 mg after cesarean delivery. Unlike the aforementioned studies, intraoperative anesthesia in both groups was provided with spinal bupivacaine, and a neuraxial opioid (fentanyl) was administered before © 2009 International Anesthesia Research Society

171

Table 1. Subject Characteristics

Age (yr) Body mass index (kg/m2) Perineal laceration n (%) Degree of laceration (n) 1st 2nd 3rd Delivery to procedure time (min) Intraoperative medications (n) Midazolam Lidocaine infiltration Procedure duration (min)a

MCS n ⴝ 29

SCM n ⴝ 30

SLM n ⴝ 28

P

33 (28–35) 32 (28–34) 13 (45)

31 (26–35) 31 (27–37) 7 (23)

29 (26–32) 28 (25–33) 8 (29)

0.15 0.17 0.18

7 6 0 138 (111–214)

4 3 0 152 (123–240)

3 4 1 147 (119–285)

0.77

0 6 63 (54–73)

3 3 60 (46–67)

1 7 60 (47–69)

0.25

0.49

0.62

Values are median (interquartile range) or n (percent). Groups MCS ⫽ morphine followed by 2-chloroprocaine, then saline; SCM ⫽ saline followed by 2-chloroprocaine, then morphine; SLM ⫽ saline followed by lidocaine, then morphine. a Time from local anesthetic (syringe 2) administration to admission to the PACU (postanesthesia care unit).

Figure 4. Box plots of postprocedure verbal rating scale pain Figure 3. Kaplan–Meier curves of the percentage of subjects

with continuing analgesia. Time 0 ⫽ time of epidural morphine administration. Subjects who did not request analgesia during the first 48 h were assigned a time of 48 h. The median duration for MCS Group 28.6 h (95% CI 4.4 –52.7) was different than SCM Group 2.2 h (95% CI 0 – 4.8) (P ⫽ 0.006), but not SLM Group 25.8 h (95% CI 10.7– 40.9) (P ⫽ 0.83). SCM Group was different from SLM Group (P ⫽ 0.009). Groups, MCS ⫽ morphine followed by 2-chloroprocaine, then saline; SCM ⫽ saline followed by 2-chloroprocaine, then morphine; SLM ⫽ saline followed by lidocaine, then morphine.

the administration of epidural 2-chloroprocaine and morphine. There are several proposed mechanisms for morphine antagonism by 2-chloroprocaine. Direct antagonism of the opioid receptors by 2-chloroprocaine or one of its metabolites, antagonism of an intracellular second messenger, or decreased morphine availability because of a decrease in perineural pH have all been suggested as possible mechanisms for the reduced efficacy of opioids.8 Coda et al. conducted binding studies that demonstrated a Ki of 435 ␮M for 2-chloroprocaine at mu-opioid receptor sites. They concluded that it was unlikely that a direct opioid receptor antagonism was responsible for the decreased pharmacodynamic effect, except perhaps in unusual clinical circumstances when high neuraxial 172

Epidural Morphine and Chloroprocaine

scores. The box solid line represents the median value, the boxes are the interquartile range and the whiskers are the 10th and 90th percentile range. † ⫽ SCM and MCS Groups different from SLM Group, P ⬍ 0.05. Groups, MCS ⫽ morphine followed by 2-chloroprocaine, then saline; SCM ⫽ saline followed by 2-chloroprocaine, then morphine; SLM ⫽ saline followed by lidocaine, then morphine.

concentrations of 2-chloroprocaine are used.8 Although it remains unknown whether the timing of opioid administration affects the binding, our data cannot directly support or refute either competitive inhibition of chloroprocaine by morphine, or a steric interaction at the receptor. Several investigators have suggested that the observed decrease in analgesia duration may be due to a disparity between the time the 2-chloroprocaine anesthesia resolves and the onset of epidural morphine analgesia, resulting in an analgesia window.5,6,9 The duration of action of epidural 2-chloroprocaine anesthesia is 30 – 45 min, and the onset of epidural morphine analgesia is 60 –70 min. Therefore, the regression of sensory blockade occurs before the onset of the morphine analgesia and this could result in such a window. Our data may support this hypothesis as more subjects in the SCM Group (n ⫽ 10) requested supplemental analgesia within 90 min of the procedure, compared with the MCS (n ⫽ 0) or SLM (n ⫽ 0) Groups. An ANESTHESIA & ANALGESIA

Table 2. Postprocedure Outcomes

Pruritus n (%) Nausea n (%) Vomiting n (%) PACU stay (min) Subjects requesting rescue analgesia n (%) Supplemental morphine equivalents during first 48 h postprocedure (mg)

MCS n ⴝ 29

SCM n ⴝ 30

SLM n ⴝ 28

P

7 (24) 2 (7) 3 (10) 95 (79–107) 12 (41) 22.5 (7.5–30.0)

6 (20) 2 (7) 4 (13) 95 (79–115) 19 (63) 15.0 (7.5–30.0)

11 (39) 7 (25) 6 (21) 110 (75–132) 13 (46) 22.5 (7.5–33.7)

0.28 0.09 0.49 0.24 0.20 0.91

Values are median (interquartile range) or n (percent). PACU ⫽ postanesthesia care unit; Groups, MCS ⫽ morphine followed by 2-chloroprocaine, then saline; SCM ⫽ saline followed by 2-chloroprocaine, then morphine; SLM ⫽ saline followed by lidocaine, then morphine.

alternative explanation of our data is that direct antagonism of the opioid receptor by 2-chloroprocaine is inhibited by the presence of morphine. In this study, morphine was administered before 2-chloroprocaine in MCS Group, suggesting that the presence of opioid at the receptor before 2-chloroprocaine administration may reduce the antagonistic effect of 2-chloroprocaine on the efficacy of morphine. Both mechanisms could also explain the negative results of the Hess et al. study, prolonged anesthesia/analgesia was provided in both groups with spinal bupivacaine, and opioid receptors were occupied by fentanyl, again in both groups, before the administration of 2-chloroprocaine. There are limitations to our study design and conclusions. We used a single 30-min interval between morphine and 2-chloroprocaine administration and additional investigation is warranted to determine the optimum interval for administration. Subjects were only followed for 48 h and differences among the groups may have been obscured by limiting the time of follow-up. Our finding of no difference in the need for oral analgesics to treat breakthrough pain among our study groups may have been because of our decision to use a multimodal pain management regimen which included scheduled ibuprofen.4 Treatment of breakthrough pain with oral acetaminophen/ hydrocodone required a request through the nursing personnel. The use of a patient-controlled device for analgesia administration might have allowed for more precise determination of opioid requirements. In addition, we were unable to show a difference in the duration of PACU stay as we measured the time that the patient left the PACU and not when the patient met discharge criteria. In summary, the administration of epidural morphine before 2-chloroprocaine anesthesia for PPTL allowed parturients the benefit of short surgical anesthesia accompanied by prolonged postoperative analgesia. The short duration of 2-chloroprocaine anesthesia may

Vol. 109, No. 1, July 2009

be beneficial if it is associated with shorter PACU stays, less need for one-on-one nursing care, shorter separation time from newborns, and earlier mobilization. Many clinicians, however, are hesitant to administer epidural 2-chlorprocaine if they plan to use epidural morphine anesthesia for postprocedure analgesia. The findings of this study suggest epidural morphine may be used with 2-chloroprocaine, with similar duration of action, efficacy, and incidence of side effects found with epidural lidocaine, if the morphine is administered 30 min before the administration of 2-chloroprocaine. REFERENCES 1. Campbell DC, Riben CM, Rooney ME, Crone LL, Yip RW. Intrathecal morphine for postpartum tubal ligation postoperative analgesia. Anesth Analg 2001;93:1006 –11 2. Habib AS, Muir HA, White WD, Spahn TE, Olufolabi AJ, Breen TW. Intrathecal morphine for analgesia after postpartum bilateral tubal ligation. Anesth Analg 2005;100:239 – 43 3. Wittels B, Faure EA, Chavez R, Moawad A, Ismail M, Hibbard J, Principe D, Karl L, Toledano AY. Effective analgesia after bilateral tubal ligation. Anesth Analg 1998;87:619 –23 4. Marcus RJ, Wong CA, Lehor A, McCarthy RJ, Yaghmour E, Yilmaz M. Postoperative epidural morphine for postpartum ligation analgesia. Anesth Analg 2005;101:876 – 81 5. Eisenach JC, Schlairet TJ, Dobson CE, Hood DH. Effect of prior anesthetic solution on epidural morphine analgesia. Anesth Analg 1991;73:119 –23 6. Karambelkar DJ, Ramanathan S. 2-Chloroprocaine antagonism of epidural morphine analgesia. Acta Anaesthesiol Scand 1997;41:774 – 8 7. Grice SC, Eisenach JC, Dewan DM. Labor analgesia with epidural bupivacaine plus fentanyl: enhancement with epinephrine and inhibition with 2-chloroprocaine. Anesthesiology 1990;72:623– 8 8. Coda B, Bausch S, Haas M, Chavkin C. The hypothesis that antagonism of fentanyl analgesia by 2-chloroprocaine is mediated by direct action on opioid receptors. Reg Anesth 1997;22:43–52 9. Hess PE, Snowman CE, Hahn CJ, Kunze LJ, Ingold VJ, Pratt SD. Chloroprocaine may not affect epidural morphine for postcesarean delivery analgesia. J Clin Anesth 2006;18:29 –33 10. Camann WR, Hartigan PM, Gilbertson LI, Johnson MD, Datta S. Chloroprocaine antagonism of opioid analgesia: a receptor specific phenomenon? Anesthesiology 1990;73:860 –3


173

Serotonin Receptor Antagonists for the Prevention and Treatment of Pruritus, Nausea, and Vomiting in Women Undergoing Cesarean Delivery with Intrathecal Morphine: A Systematic Review and Meta-Analysis Ronald B. George, MD, FRCPC* Terrence K. Allen, MBBS, FRCA† Ashraf S. Habib, MBBCh, MSc, FRCA†

BACKGROUND: We performed a systematic review to determine the overall efficacy of serotonin (5-HT3) receptor antagonists for the prevention and treatment of pruritus, nausea, and vomiting in women receiving spinal anesthesia with intrathecal morphine for cesarean delivery. METHODS: Reports of randomized, controlled trials that compared prophylaxis or treatment of pruritus and/or nausea, and vomiting using one of the 5-HT3 receptor antagonists or placebo in women undergoing cesarean delivery were reviewed. The articles were scored for validity and data were extracted by the authors independently and summarized using relative risks (RR) with 95% confidence intervals (CI). RESULTS: Nine randomized, controlled trials were included in the systematic review. The nine trials had a total of 1152 patients enrolled; 539 received 5-HT3 receptor antagonists, 413 received placebo, and 200 received other antiemetics and were not included in the analysis. The incidence of pruritus was not reduced with 5-HT3 receptor antagonists prophylaxis compared with placebo (80.7% vs 85.8%, RR [95% CI] ⫽ 0.94 [0.81–1.09]). However, their use reduced the incidence of severe pruritus and the need for treatment of pruritus (number-needed-to-treat ⫽ 12 and 15, respectively). Their use for the treatment of established pruritus showed improved efficacy compared with placebo with a number-needed-to-treat of three. There was a significant reduction in the incidence of postoperative nausea (22.0% vs 33.6%, RR [95% CI] ⫽ 0.75[0.58 – 0.96]) and vomiting (7.7% vs 16.8%, RR [95% CI] ⫽ 0.49 [0.30 – 0.81]), and the need for postoperative rescue antiemetic treatment with the use of 5-HT3 receptor antagonists when compared with placebo (9% vs 23%, RR [95% CI] ⫽ 0.38 [0.21– 0.68]). CONCLUSIONS: Although prophylactic 5-HT3 receptor antagonists were ineffective in reducing the incidence of pruritus, they significantly reduced the severity and the need for treatment of pruritus, the incidence of postoperative nausea and vomiting, and the need for rescue antiemetic therapy in parturients who received intrathecal morphine for cesarean delivery. They were also effective for the treatment of established pruritus. Although more studies are warranted, the current data suggest that the routine prophylactic use of those drugs should be considered in this patient population. (Anesth Analg 2009;109:174 –82)

I

ntrathecal morphine is commonly used to enhance postoperative analgesia in women undergoing cesarean delivery under spinal anesthesia. However, From the *Department of Women’s and Obstetric Anesthesia, IWK Health Centre, Dalhousie University, Halifax, Nova Scotia, Canada; and †Department of Anesthesiology, Division of Women’s Anesthesia, Duke University Medical Center, Durham, North Carolina. Accepted for publication January 22, 2009. This study was presented in part at the meeting of the American Society of Anesthesiologists, San Francisco, October 2007. Supported solely by funding from the Duke University Medical Center’s Department of Anesthesiology, Division of Women’s Anesthesia. The authors have no relevant conflicts of interest. Address correspondence and reprint requests to Ashraf S. Habib, MBBCh, MSc, FRCA, Duke University Medical System, PO Box 3094, Durham, North Carolina. Address e-mail to habib001@ mc.duke.edu. Copyright © 2009 International Anesthesia Research Society DOI: 10.1213/ane.0b013e3181a45a6b

174

its use is associated with a frequent incidence of side effects, including pruritus, nausea, and vomiting.1 Serotonin (5-HT3) receptor antagonists were specifically developed for the management of nausea and vomiting. Their favorable side effect profile, in particular, lack of sedation, confers an advantage over older generation antiemetics. The efficacy of those drugs for the prophylaxis of postoperative nausea and vomiting in the general surgical population has been established.2–5 Some studies also suggested that these drugs might be effective for the prophylaxis and treatment of opioid-induced pruritus.6 We therefore performed this systematic review to assess the efficacy of 5-HT3 receptor antagonists for the prevention and treatment of intrathecal morphineinduced pruritus, nausea, and vomiting in women undergoing cesarean delivery under spinal anesthesia. Vol. 109, No. 1, July 2009

METHODS The current meta-analysis adhered to the QUOROM guidelines for reporting meta-analyses.7 A systematic search was performed for full reports of randomized, controlled trials that compared prophylaxis or treatment of pruritus and/or nausea, and vomiting using one of the 5-HT3 receptor antagonists (ondansetron, granisetron, tropisetron, and dolasetron) versus placebo in women undergoing cesarean delivery under spinal anesthesia. Relevant trials had to report the main end points, namely the incidence of pruritus, and/or nausea, and/or vomiting in all study groups. The spinal anesthetic technique in both the treatment and control groups had to be standardized and include the administration of intrathecal morphine. The databases of MEDLINE (1966 –2008), the Cochrane Central Register of Controlled Trials, EMBASE, Web of Science, and CINAHL were searched without language restriction. Furthermore, the reference lists of retrieved reports and review articles were screened. Data from abstracts, letters, retrospective trials, case reports, and unpublished data were not considered. Keywords used in the search were “ondansetron,” “granisetron,” “tropisetron,” “dolasetron,” “pruritus,” “nausea,” “vomiting,” “postoperative nausea and vomiting,” and “cesarean.” Both medical subject headings and text words were used. The date of the last computer search was June 2008. The articles meeting the inclusion criteria were scored independently by two authors (RG and TA) for methodological validity using the 4-item, 7-point, Modified Oxford scale (Appendix).8 Any discrepancies were resolved by discussion with the third author (AH). All three authors extracted data independently. A data collection form was used to collect the following data: i) surgical procedure, ii) anesthesia technique, iii) intrathecal opioid, iv) therapeutic allocation, v) outcome measures including the incidence and severity of postoperative pruritus, incidence and severity of nausea and vomiting, and need for rescue treatment of pruritus, nausea, and vomiting, vi) side effects, and vii) for treatment studies, success of treatment. When event rates were reported over multiple time intervals and not over the entire duration of the study, the authors were contacted and asked to provide such information. If the authors did not respond, the highest recorded incidence over the duration of the study was used in the analysis. If any other additional data were required, the authors were contacted and asked to provide the additional information. If the studies included three groups and the third group did not include a placebo or a 5-HT3 receptor antagonist, data from this third group were not included in the analysis. If two 5-HT3 receptor antagonist groups were included in addition to placebo, data from all the groups were extracted. Drug specific and pooled analyses combining all the 5-HT3 receptor antagonists were then performed. Vol. 109, No. 1, July 2009

Dichotomous data were extracted and summarized using relative risks (RR) with 95% confidence intervals (CI). If the 95% CI included a value of 1, it was assumed that there was no significant difference between the 5-HT3 receptor antagonist and placebo. A fixed effects model was used by default. If the statistical test for heterogeneity was significant (P ⬍ 0.1), a random effects model was used and the reason for heterogeneity was explored. Analyses were performed using the Review Manager Software (Review Manager [Revman] Version 5.0. Copenhagen: The Nordic Cochrane Centre, The Cochrane Collaboration, 2008). Forrest plots were used to graphically represent and evaluate treatment effects. The number-neededto-treat (NNT) was calculated to estimate the overall clinical impact of statistically significant interventions. We performed a subgroup analysis of the incidence of postoperative pruritus by including only studies that used intrathecal morphine without lipophilic opioids. Dose responsiveness was assessed using assumptions previously published in similar analyses.3,9,10 First, if the 95% CI of the RR of one dose did not overlap with the 95% CI of another dose, we assumed that these two doses were significantly different. Second, if the lower dose was not significantly different than control but the higher doses were more effective than control and the NNT decreased by more than 20%, this was considered dose responsive.

RESULTS Twenty-four potentially relevant articles were identified (Fig. 1). Fifteen were excluded11–25 leaving nine randomized, controlled trials included in the systematic review. Six of these trials contained results regarding prophylaxis against pruritus,26 –31 six contained data regarding prophylaxis against nausea and vomiting,26 –29,32,33 and one presented results for the treatment of pruritus.34 The trials’ details are outlined in Table 1. All studies investigating prophylaxis against pruritus and/or nausea, and vomiting had an observation period of 24 h26 –30,33 except for one study which lasted for 28 h.31 The dose of intrathecal morphine ranged from 0.1 to 0.2 mg. There were no studies investigating the efficacy of 5-HT3 receptor antagonists for the treatment of established nausea and vomiting. Methodological validity scores ranged from 3 to 7. The nine trials had a total of 1152 patients enrolled, 539 received 5-HT3 receptor antagonists, 413 received placebo, and 200 patients received non-5-HT3 receptor antagonist drugs. These data were not included in this review. Tropisetron,28 granisetron,29,32 and ondansetron26 –31,33 were the 5-HT3 receptor antagonists used. All trials used a fixed dose of 5-HT3 receptor drugs, except a single trial which used ondansetron 0.1 mg/kg.31 For analysis purposes, this trial was included in the 8 mg category. © 2009 International Anesthesia Research Society

175

pruritus in either group (84.3% vs 85.9%, RR [95% CI] ⫽ 0.97 [0.90 –1.05]). Need for Treatment Six of the nine studies published data regarding the need for treatment of pruritus.26 –31 Treatment of pruritus consisted of propofol,26,31 diphenhydramine,29,30 naloxone,28 nalbuphine,27 and hydroxyzine.28 In three studies, treatment was offered to patients with “severe” pruritus,26,30,31 in two studies pruritus was treated on patients’ request,28,29 and in one study, the trigger for the treatment of pruritus was not defined.27 There was a reduction in the need for treatment of pruritus with ondansetron 4 mg (NNT ⫽ 7) and when all doses of ondansetron were combined (NNT ⫽ 17). There was no evidence of dose responsiveness for prophylactic ondansetron when need for treatment of pruritus was the outcome. Tropisetron and granisetron, each used in one trial, were not more effective than placebo at reducing the need for treatment of pruritus. When all trials which reported the need for treatment of pruritus were combined, 5-HT3 receptor antagonist prophylaxis was significantly more effective than placebo at reducing the need for treatment of pruritus (Table 2).

Figure 1. Flow diagram of screened, excluded, and analyzed studies.

Pruritus Incidence of Pruritus The effect of prophylaxis with 5-HT3 receptor antagonists on the incidence of pruritus was reported in five studies.26,28 –31 Ondansetron 4 mg was tested in one study,26 ondansetron 8 mg in five studies,26,28 –31 tropisetron 5 mg in one study,28 and granisetron 3 mg in one study.29 Results are presented in Table 2. There was no reduction in the incidence of pruritus with any of the 5-HT3 receptor antagonists compared with placebo, nor was there a difference when all trials reporting pruritus were combined (Table 2 and Fig. 2). Ondansetron was the only drug used in more than one dose. Data did not suggest that there was dose responsiveness when ondansetron was used for the prophylaxis of pruritus. Similarly, there was no difference in the incidence of pruritus when the analysis was limited to the trials using intrathecal morphine only without lipophilic drugs (81.2% vs 89.6%, RR [95% CI] ⫽ 0.86 [0.66 –1.12]). Statistical heterogeneity was observed. Excluding the study by Yeh et al.31 eliminated the statistical heterogeneity (P ⫽ 0.63, I2 ⫽ 0%) but did not significantly change the overall incidence of 176

Side Effects of Intrathecal Morphine

Severity of Pruritus The severity of pruritus was reported in six studies.26,28 –31,33 Sarvela et al.28 graded pruritus with a numeric rating scale between 0 and 10 and chose to treat subjects who requested treatment for their pruritus. Data were reported as median and range and showed no significant differences between the treatment and placebo groups in the severity of pruritus. Data from this study were not included in the pooled analysis. One trial used a 4-point scale (0 ⫽ no pruritus, 1 ⫽ perioral, 2 ⫽ diffuse moderate, 3 ⫽ diffuse intense).33 Results were reported as mean with standard deviation and were not included in the pooled analysis. Two trials used a 4-point scale (1 ⫽ no pruritus, 2 ⫽ mild pruritus, 3 ⫽ moderate pruritus, 4 ⫽ severe pruritus) with subjects receiving treatment on request29 or moderate to severe pruritus.26 Two trials used a 3-point scale (0 ⫽ no pruritus, 1 ⫽ mild/mild to moderate pruritus, 2 ⫽ severe pruritus) and treated patients who had severe pruritus.30,31 For analysis, we converted the 4-point scale into a 3-point scale by combining Grades 3 and 4 into one grade (severe pruritus) and compared the incidence of severe pruritus between the placebo and 5-HT3 receptor antagonists groups in these four studies.26,29 –31 Combined data showed that ondansetron 4 mg and combined ondansetron doses were effective in reducing the incidence of severe pruritus (Table 2). There was no evidence of dose responsiveness in the effect of ondansetron on the severity of pruritus. Granisetron and tropisetron were not effective in reducing the severity of pruritus. When combining all drugs and doses, 5-HT3 receptor antagonists were effective in reducing the incidence of severe pruritus compared with placebo. ANESTHESIA & ANALGESIA

Table 1. Trials of 5-HT3 Receptor Antagonists for Prophylaxis and Treatment of Intrathecal Morphine-Induced Pruritus, Nausea, and Vomiting Oxford scale (R/C/B/F)a

Intrathecal morphine

Charuluxananan et al.26

7 (2/1/2/2)

0.2 mg

Harnett et al.27

4 (2/0/2/0)

0.2 mg (⫹10 ␮g ITFc)

Sarvela et al.28

6 (2/0/2/2)

0.16 mg (⫹15 ␮g ITFc)

Siddik-Sayyid et al.29

5 (2/0/1/2)

0.2 mg

Yazigi et al.30

4 (2/0/2/0) 3 (1/0/2/0)

0.1 mg (⫹2.5 ␮g ITSd) 0.15 mg

Peixoto et al.33

7 (2/1/2/2)

0.1 mg (⫹20 ␮g ITFc)

Balki et al.32

6 (2/0/2/2) 7 (2/1/2/2)

0.1 mg (⫹10 ␮g ITFc) 0.2 mg

Ref

Yeh et al.31

Charuluxananan et al.34

Randomized groups

n

Normal saline (4 mL) Nalbuphine (4 mg) Ondansetron (4 mg) Ondansetron (8 mg) Normal saline (10 mL) Ondansetron (4 mg) Scopolamine (1.5 mg) Normal saline (5 mL) Ondansetron (8 mg) Tropisetron (5 mg) Normal saline Granisetron (3 mg) Ondansetron (8 mg) Normal saline (5 mL) Ondansetron (8 mg) Normal saline Ondansetron (0.1 mg/kg) Diphenhydramine (30 mg) Normal saline Ondansetron (4 mg) Droperidol (1.25 mg) Normal saline (1 mL) Granisetron (1 mg) Normal saline (2 mL) Ondansetron (4 mg)

60 60 60 60 81 79 80 29 30 28 45 42 42 50 50 20 20 20 40 40 40 88 88 39 41

Outcomes (P/N/V)b

Observation period

P/N/V

24 h

P/N/V

24 h

P

24 h

P/N/V

24 h

P

24 h

P

28 h

N/V

24 h

N/V

PACU

P

4h

PACU ⫽ postanesthesia care unit. a R/C/B/F—randomization/concealment allocation/blinding/flow of subjects. Numbers represent points allocated for each quality indicator (see Appendix). b P/N/V—pruritus/nausea/vomiting. c ITF—intrathecal fentanyl. d ITS—intrathecal sufentanil.

Table 2. The Effect of 5-HT3 Receptor Antagonists Prophylaxis on the Incidence, Need for Treatment and Severity of Pruritus

Incidence of pruritus Ondansetron 4 mg26 Ondansetron 8 mg26,28–31a Combined ondansetron doses Tropisetron 5 mg28a Granisetron 3 mg29 All studies combined Need for treatment of pruritus Ondansetron 4 mg26,27 Ondansetron 8 mg26,28–31a Combined ondansetron doses Tropisetron 5 mg28a Granisetron 3 mg29 All studies combined Incidence of severe pruritus Ondansetron 4 mg26 Ondansetron 8 mg26,29–31 Combined ondansetron doses Granisetron 3 mg29 All studies combined

Risk of treatment group

Risk of control group

RR (95% CI)

52/60 (86.7%) 157/202 (77.7%) 209/262 (79.8%) 22/28 (78.6%) 37/42 (88.1%) 268/332 (80.7%)

56/60 (93.3%) 175/204 (85.8%) 175/204 (85.8%) 22/29 (75.9%) 39/45 (86.7%) 175/204 (85.8%)

0.93 (0.82–1.05) 0.94 (0.79–1.11) 0.93 (0.80–1.10) 1.04 (0.78–1.37) 1.02 (0.87–1.19) 0.94 (0.81–1.09)

52/139 (37.4%) 77/202 (38.1%) 129/341 (37.8%) 11/28 (39.3%) 12/42 (28.6%) 152/411 (38.0%)

74/141 (52.5%) 94/204 (46.1%) 125/285 (43.9%) 9/29 (31.0%) 18/45 (40.0%) 125/285 (43.9%)

0.71 (0.55–0.92) 0.83 (0.66–1.03) 0.79 (0.66–0.95) 1.27 (0.62–2.58) 0.71 (0.39–1.30) 0.80 (0.64–0.96)

28/60 (46.7%) 70/172 (40.7%) 98/232 (42.2%) 14/42 (33.3%) 112/274 (40.9%)

43/60 (71.7%) 85/175 (48.6%) 85/175 (48.6%) 18/45 (40.0%) 85/175 (48.6%)

0.65 (0.48–0.89) 0.84 (0.67–1.05) 0.80 (0.65–0.98) 0.83 (0.48–1.46) 0.79 (0.65–0.97)

NNT (95% CI)

7 (4–29) 17 (8 to ⫺59) 15 (7 to ⫺186) 4 (3–13) 16 (6 to ⫺29) 13 (6 to ⫺58)

RR ⫽ relative risk; CI ⫽ confidence interval; NNT ⫽ number needed to treat. a Sarvela et al.28—The highest recorded incidence over the duration of the study was used as the 24 h incidence.

Time to Onset of Pruritus Only a single prophylaxis trial published the interval from morphine administration until the onset of pruritus after prophylaxis with ondansetron 0.1 mg/kg.31 The Vol. 109, No. 1, July 2009

average time to the onset of pruritus was not significantly different in the placebo compared with the ondansetron group (183 vs 187 min, weighted mean difference [95% CI] ⫽ 4.00 [⫺28.87 to 36.87]). © 2009 International Anesthesia Research Society

177

Figure 2. Incidence of pruritus. A relative risk (RR) less than one indicates less pruritus with 5-HT3 receptor antagonists compared with control. When the 95% confidential interval (CI) does not include 1, the difference is considered statistically significant. oˆ2, ␹2, and I2 refer to the tests for statistical heterogeneity, M–H ⫽ Mantel–Haenszel test; 5-HT3RA ⫽ 5-HT3 receptor antagonist.

Treatment of Established Pruritus Only one study compared ondansetron 4 mg with placebo for the treatment of pruritus after cesarean delivery using 0.2 mg intrathecal morphine.34 A 4-point scale to assess the severity of pruritus was used. Patients who had a pruritus score of 3 or 4 were randomized and treated; treatment success was achieved if the pruritus score was decreased to 1 or 2 after treatment. Ondansetron was significantly more effective than placebo in successfully treating pruritus (80% vs 36%, RR [95% CI] ⫽ 0.30 [0.16 – 0.59], NNT ⫽ 3).

Nausea and Vomiting Intraoperative Nausea and Vomiting Data on intraoperative nausea and vomiting were reported in four studies.27,28,32,33 Two reported nausea and vomiting separately,27,32 whereas the remainder reported nausea and vomiting collectively.28,33 Of the former studies, one reported the incidence over the whole intraoperative period,27 whereas the other reported predelivery and postdelivery data separately.32 Therefore, we were unable to combine data on the incidence of intraoperative nausea and vomiting quantitatively and these data were not included in the review. Postoperative Nausea Four studies reported the incidence of postoperative nausea after prophylactic 5-HT3 receptor antagonists compared with placebo.26,27,29,33 The results are summarized in Table 3. Ondansetron 8 mg reduced postoperative nausea when compared with placebo. This reduction was also significant when these data were combined with data from the three trials investigating ondansetron 4 mg. There was no evidence of dose responsiveness for ondansetron. The study investigating granisetron 3 mg did not demonstrate a significant reduction in the incidence of postoperative nausea.29 Overall, when all drugs and doses of 5-HT3 receptor antagonists were combined, there was a significant reduction in the incidence of postoperative nausea when compared with placebo (Fig. 3). 178


Postoperative Vomiting The four studies investigating postoperative nausea also reported the incidence of postoperative vomiting (Table 3).26,27,29,33 Ondansetron 4 mg significantly reduced postoperative vomiting when compared with placebo. When combined with the studies investigating ondansetron 8 mg, the reduction was still significant. There was no evidence of dose responsiveness for ondansetron. The only study investigating the use of granisetron 3 mg was unable to demonstrate any efficacy in preventing postoperative vomiting.29 When all drugs and doses were combined, the 5-HT3 receptor antagonists significantly reduced the incidence of postoperative vomiting when compared with placebo (Fig. 4). Need for Postoperative Rescue Antiemetic Treatment Three studies reported the need for postoperative rescue antiemetic therapy.28,29,33 Metoclopramide or naloxone,33 droperidol or metoclopramide,28 or metoclopramide only29 were the rescue antiemetics of choice. Results are summarized in Table 3. A significant reduction in the need for rescue occurred with ondansetron 4 mg, and with all ondansetron doses combined. When all drugs and doses were combined, the need for postoperative rescue antiemetic was significantly reduced when 5-HT3 receptor antagonists were compared with placebo. Severity of Nausea/Vomiting Five studies assessed the severity of nausea and vomiting in patients receiving 5-HT3 receptor antagonists.26–30 In one study, nausea severity was assessed postoperatively using a 100 mm unlabeled visual analog scale.27 No difference in nausea scores was reported between the placebo and 5-HT3 receptor antagonist groups; data from this study were not included in the pooled analysis. Two studies used a 4-point scale (1 ⫽ absent nausea, 2 ⫽ queasy, 3 ⫽ severe nausea, 4 ⫽ vomiting) for assessing the severity of postoperative nausea and vomiting.26,29 Yazigi et al.30 used a 3-point scale (0 ⫽ no nausea and vomiting, 1 ⫽ mild to moderate nausea or vomiting not needing treatment, and 2 ⫽ severe nausea or vomiting needing treatment) as did Sarvela et al.28 (0 ⫽ none, 1 ⫽ nausea, 2 ⫽ disturbing nausea or vomiting). For the purposes of comparison, we converted the 4-point scales to a 3-point scale by combining Grades 3 and 4 into a single severe group and compared ANESTHESIA & ANALGESIA

Table 3. The Effect of 5-HT3 Receptor Antagonists on the Incidence of Postoperative Nausea and Vomiting and the Need for Rescue Antiemetic Therapy

Incidence of postoperative nausea Ondansetron 4 mg26,27,33a Ondansetron 8 mg26,29 Combined ondansetron doses Granisetron 3 mg29 All studies combined Incidence of postoperative vomiting Ondansetron 4 mg26,27,33a Ondansetron 8 mg26,29 Combined ondansetron doses Granisetron 3 mg29 All studies combined Need for postoperative rescue antiemetic therapy Ondansetron 4 mg33 Ondansetron 8 mg28,29 Combined ondansetron doses Granisetron 3 mg29 Tropisetron 5 mg28 All studies combined Incidence of severe postoperative nausea and vomiting Ondansetron 4 mg26 Ondansetron 8 mg26,28–30 Combined ondansetron doses Granisetron 3 mg29 Tropisetron 5 mg28 All studies combined

Risk of treatment group

Risk of control group

RR (95% CI)

50/179 (27.9%) 12/102 (11.8%) 62/281 (22.1%) 9/42 (21.4%) 71/323 (22.0%)

67/181 (37.0%) 25/105 (23.8%) 92/286 (32.2%) 9/45 (20.0%) 76/226 (33.6%)

0.76 (0.58–1.00) 0.49 (0.26–0.93) 0.69(0.53–0.89) 1.07 (0.47–2.44) 0.75 (0.58–0.96)

36/179 (20.1%) 9/102 (8.8%) 21/281 (7.4%) 4/42 (9.5%) 25/323 (7.7%)

63/181 (34.8%) 11/105 (10.5%) 42/286 (14.5%) 7/45 (15.6%) 38/226 (16.8%)

0.59 (0.43–0.80) 0.85 (0.37–1.96) 0.52 (0.32–0.85) 0.61 (0.19–1.94) 0.49 (0.30–0.81)

2/40 (5.0%) 8/72 (11.1%) 10/112 (8.9%) 5/42 (11.9%) 1/28 (3.6%) 16/182 (8.8%)

10/40 (25.0%) 16/74 (21.6%) 26/114 (22.8%) 13/45 (28.9%) 3/29 (10.3%) 26/114 (22.8%)

0.16 (0.03–0.78) 0.46 (0.19–1.15) 0.39(0.20–0.78) 0.33 (0.11–1.03) 0.32 (0.03–3.29) 0.38 (0.21–0.68)

5 (3–20)

5/60 (8.3%) 23/182 (12.6%) 28/242 (11.6%) 8/42 (19.0)% 1/28 (3.5%) 37/315 (11.7%)

14/60 (23.3%) 43/184 (23.4%) 57/244 (23.4%) 10/45 (22.2%) 4/29 (13.8%) 43/184 (23.4%)

0.36 (0.14–0.93) 0.54 (0.34–0.87) 0.50 (0.30–0.75) 0.86 (0.37–1.96) 0.26 (0.03–2.18) 0.55 (0.33–0.76)

7 (4–45) 10 (5–34) 9 (5–20)

NNT (95% CI)

8 (4–56) 10 (6–35) 9 (5–25) 7 (4–18) 14 (8–48) 12 (7–30)

7 (4–22) 7 (4–19)

9 (5–22)

RR ⫽ relative risk; CI ⫽ confidence interval; NNT ⫽ number-needed-to-treat. a Harnett et al. 27—The author provided unpublished data regarding the overall 24 h incidence of postoperative nausea and vomiting.

Figure 3. Incidence of postoperative nausea. A relative risk (RR) less than one indicates less postoperative nausea with 5-HT3 receptor

antagonists compared with control. When the 95% confidential interval (CI) does not include 1, the difference is considered statistically significant. oˆ2, ␹2, and I2 refer to the tests for statistical heterogeneity; M–H ⫽ Mantel–Haenszel test; 5-HT3RA ⫽ 5-HT3 receptor antagonist.

Figure 4. Incidence of postoperative vomiting. A relative risk (RR) less than one indicates less postoperative vomiting with 5-HT3 receptor

antagonists compared with control. When the 95% confidential interval (CI) does not include 1, the difference is considered statistically significant. oˆ2, ␹2, and I2 refer to the tests for statistical heterogeneity; M–H ⫽ Mantel–Haenszel test; 5-HT3RA ⫽ 5-HT3 receptor antagonist. Vol. 109, No. 1, July 2009


179

the incidence of severe postoperative nausea and vomiting between the 5-HT3 receptor antagonists and placebo groups. Combined data showed that ondansetron was effective in reducing the incidence of severe postoperative nausea and vomiting when compared with placebo (Table 3). Overall, when combining all drugs and doses, 5-HT3 receptor antagonists were still effective in reducing the incidence of severe postoperative nausea and vomiting when compared with placebo.

Side Effects Headache, cardiac arrhythmias, and extrapyramidal side effects were the most commonly investigated side effects of 5-HT3 receptor antagonists. Four studies evaluated patients for headache after the administration of 5-HT3 receptor antagonists.29 –32 In one study, there were no reported cases of headache directly related to the administration of 5-HT3 receptor antagonists.30 Yeh et al.31 did not report the actual incidence of headaches but reported no difference between the treatment and placebo groups. In the two remaining studies, the incidence of headaches were reported quantitatively.29,32 Both studies reported no difference between the treatment and placebo groups. Combined data also showed no difference in the risk of headache with the 5-HT3 receptor antagonists compared with placebo (6% vs 2%, RR [95% CI] ⫽ 2.78 [0.83–9.29]). Five studies evaluated patients for cardiac dysrrhythmias.26,29 –32 The single study reporting quantitative differences found no difference in the incidence of dysrhythmias (3.4% vs 4.6%), tachycardia (0% vs 1.1%), or bradycardia (6.7% vs 10.2%) between the treatment and control groups.32 Yeh et al.31 reported no significant difference in the incidence of dysrhythmias between the treatment and the placebo group. In the remaining three studies, cardiac dysrhythmias were not observed in any patient.26,29,30 Four studies evaluated patients for extrapyramidal side effects26,29 –31 and reported that these side effects were not observed in any patient.

DISCUSSION The results of this systematic review indicate that the 5-HT3 receptor antagonists do not significantly reduce the incidence of pruritus in women undergoing spinal anesthesia with intrathecal morphine. Their use, however, was associated with a significant reduction in the severity and need for treatment of pruritus. Ondansetron was effective for the treatment of established pruritus. 5-HT3 receptor antagonists were effective for the prophylaxis against postoperative nausea and vomiting in this patient population. There was no dose responsiveness with ondansetron 4 or 8 mg when used for the prophylaxis against nausea, vomiting, and pruritus. 5-HT3 receptor antagonists were not associated with a higher incidence of side effects when compared with placebo. Neuraxial opioid-induced pruritus is likely due to cephalad migration of neuraxial opioids to the medulla where the “itch center” is thought to be located 180


and where they interact with the trigeminal nucleus.35 Interactions between the serotonin and opioid receptors in the central nervous system have been suggested as a mechanism of opioid-induced pruritus.36 Specifically, the 5-HT3 receptor has been implicated and this stimulated interest in investigating the potential for the 5-HT3 receptor antagonists to reduce the incidence of this bothersome complication of intrathecal morphine. Parturients appear to be more susceptible to neuraxial opioid-induced pruritus compared with the general surgical population with a reported incidence of 60%–100%.37 An interaction of estrogen with the opioid receptors has been suggested as a reason for this increased sensitivity.38 Although our analysis demonstrated the prophylactic use of 5-HT3 receptor antagonists was associated with a reduction in the severity and the need for treatment of pruritus, this reduction was modest with a NNT of 12 and 15, respectively. Only one study investigated the use of ondansetron for the treatment of established pruritus and reported that it was more effective than placebo.34 The incidence and severity of pruritus increases with increasing doses of intrathecal morphine.39 In this review, the average incidence of pruritus was 90%, 82%, and 76% with the use of 0.2, 0.1, and 0.16 mg intrathecal morphine, respectively. It has been suggested that the combination of intrathecal morphine with a lipophilic opioid may decrease the efficacy of the 5-HT3 receptor antagonist in reducing the incidence of pruritus, due to activation of the serotonin receptors by the lipophilic opioid before being blocked by the 5-HT3 receptor antagonist.24 However, our study found no efficacy in reducing the incidence of pruritus, even when we limited our analysis to studies in which patients did not receive lipid soluble opioids. Our results regarding pruritus differ from those of Bonnet et al.40 who recently published a quantitative systematic review of the efficacy of 5-HT3 receptor antagonists for the prophylaxis of neuraxial opioids (morphine, fentanyl, and sufentanil)-induced pruritus in patients undergoing a wide variety of surgical procedures and labor. They concluded that 5-HT3 receptor antagonists were effective in reducing the incidence of pruritus. This agrees with a previous report showing a benefit of those drugs for the prophylaxis against neuraxial opioid-induced pruritus in the general surgical population.41 They also performed a subgroup analysis of patients receiving neuraxial opioids for cesarean delivery and concluded that the incidence of pruritus was reduced with 5-HT3 receptor antagonists’ prophylaxis in this patient population. However, they included a trial in which fentanyl alone was used16 in addition to trials using intrathecal morphine. Our analysis only included patients who had intrathecal morphine, because we believe that this results in a more clinically homogeneous patient population. We also included the highest incidence of pruritus recorded during the 24 h duration of the study in one of the included trials,28 whereas Bonnet et al. used the incidence of pruritus reported at 4–12 h after surgery (69%), which was lower in the 5-HT3 group than the incidence that we included (83%). ANESTHESIA & ANALGESIA

Our systematic review had several limitations. There were only a limited number of studies available for review that investigated pruritus, nausea, and vomiting in the obstetric population. Several of these studies had small sample sizes. In addition, these studies used different scoring systems for reporting the severity of pruritus and nausea. The trigger for treating pruritus was also different in the included studies. Therefore, data on the severity of pruritus and severity of postoperative nausea and vomiting, and the need for treatment of pruritus, should be interpreted with caution. Also, the duration of pruritus and nausea and vomiting episodes was not assessed in several studies. None of the studies reported complete response to antiemetic prophylaxis. We combined data on all drugs and doses because it was previously suggested that the antiemetic effect of the 5-HT3 receptor antagonists is similar when adequate doses are used.42 However, it is not clear if this also applies to their antipruritic effect. Publication bias cannot be excluded. However, funnel plots and statistical tests for detection of publication bias are unreliable in the presence of a small number of studies as is the case in our review and therefore were not performed.43–45 A strong point of our analysis was the consistent anesthetic technique in the included studies and comparable periods of follow-up. Larger studies with adequate power are required to further investigate the use of the 5-HT3 receptor antagonists for the prophylaxis against pruritus, intraoperative, and postoperative nausea and vomiting in the obstetric population. Specifically, future studies need to use validated and consistent scoring systems for assessing the severity of pruritus and nausea. These studies also need to have clearly defined consistent end points for the treatment of pruritus. Nausea and vomiting also need to be reported separately rather than collectively. Reporting of intraoperative nausea and vomiting also needs to be improved. Reporting should differentiate events that occur before delivery or after delivery because the etiology might be different at those different stages of the procedure. The usefulness and efficacy of the 5-HT3 receptor antagonists for the treatment of established pruritus also requires further investigation. In conclusion, the 5-HT3 receptor antagonists significantly reduce the severity and need for treatment of pruritus, the incidence and severity of postoperative nausea and vomiting and the need for postoperative rescue antiemetic therapy in patients who have received intrathecal morphine as part of spinal anesthesia for cesarean delivery. However, they did not reduce the overall incidence of neuraxial opioidinduced pruritus but were effective for the treatment of established pruritus. They also had a favorable side effects profile, and therefore the results of this review suggest that the prophylactic use of 5-HT3 receptor antagonists in this patient population should be considered. However, the current studies had limitations, and therefore further studies are warranted. Vol. 109, No. 1, July 2009

APPENDIX Modified Oxford score8 Validity Score (0–7) Randomization 0–None 1–Mentioned 2–Described and adequate Concealment of allocation 0–None 1–Yes Double blinding 0–None 1–Mentioned 2–Described and adequate Flow of patients 0–None 1–Described but incomplete 2–Described and adequate R/C/B/F–randomization/concealment allocation/ blinding/flow of subjects (Table 1)

REFERENCES 1. Yang T, Breen TW, Archer D, Fick G. Comparison of 0.25 mg and 0.1 mg intrathecal morphine for analgesia after Cesarean section. Can J Anaesth 1999;46:856 – 60 2. Habib AS, Gan TJ. Evidence-based management of postoperative nausea and vomiting: a review. Can J Anaesth 2004;51:326 – 41 3. Tramer MR, Reynolds DJ, Moore RA, McQuay HJ. Efficacy, dose-response, and safety of ondansetron in prevention of postoperative nausea and vomiting: a quantitative systematic review of randomized placebo-controlled trials. Anesthesiology 1997;87:1277– 89 4. Kranke P, Eberhart LH, Apfel CC, Broscheit J, Geldner G, Roewer N. [Tropisetron for prevention of postoperative nausea and vomiting: a quantitative systematic review]. Anaesthesist 2002;51:805–14 5. Loewen PS, Marra CA, Zed PJ. 5-HT3 receptor antagonists vs traditional agents for the prophylaxis of postoperative nausea and vomiting. Can J Anaesth 2000;47:1008 –18 6. Kjellberg F, Tramer MR. Pharmacological control of opioidinduced pruritus: a quantitative systematic review of randomized trials. Eur J Anaesthesiol 2001;18:346 –57 7. Moher D, Cook DJ, Eastwood S, Olkin I, Rennie D, Stroup DF. Improving the quality of reports of meta-analyses of randomised controlled trials: the QUOROM statement. Quality of Reporting of Meta-analyses. Lancet 1999;354:1896 –900 8. Jadad AR, Moore RA, Carroll D, Jenkinson C, Reynolds DJ, Gavaghan DJ, McQuay HJ. Assessing the quality of reports of randomized clinical trials: is blinding necessary? Control Clin Trials 1996;17:1–12 9. Bu¨ttner M, Walder B, von Elm E, Trame`r MR. Is low-dose haloperidol a useful antiemetic?: a meta-analysis of published and unpublished randomized trials. Anesthesiology 2004;101:1454 – 63 10. Henzi I, Sonderegger J, Tramer MR. Efficacy, dose-response, and adverse effects of droperidol for prevention of postoperative nausea and vomiting. Can J Anaesth 2000;47:537–51 11. Abouleish EI, Rashid S, Haque S, Giezentanner A, Joynton P, Chuang AZ. Ondansetron versus placebo for the control of nausea and vomiting during Caesarean section under spinal anaesthesia. Anaesthesia 1999;54:479 – 82 12. Cherian VT, Smith I. Prophylactic ondansetron does not improve patient satisfaction in women using PCA after Caesarean section. Br J Anaesth 2001;87:502– 4 13. Fujii Y, Saitoh Y, Tanaka H, Toyooka H. Granisetron/ dexamethasone combination for reducing nausea and vomiting during and after spinal anesthesia for cesarean section. Anesth Analg 1999;88:1346 –50 14. Fujii Y, Tanaka H, Toyooka H. Prevention of nausea and vomiting with granisetron, droperidol and metoclopramide during and after spinal anaesthesia for caesarean section: a randomized, double-blind, placebo-controlled trial. Acta Anaesthesiol Scand 1998;42:921–5 © 2009 International Anesthesia Research Society

181

15. Fujii Y, Tanaka H, Toyooka H. Granisetron prevents nausea and vomiting during spinal anaesthesia for caesarean section. Acta Anaesthesiol Scand 1998;42:312–5 16. Gulhas N, Erdil FA, Sagir O, Gedik E, Togal T, Begec Z, Ersoy MO. Lornoxicam and ondansetron for the prevention of intrathecal fentanyl-induced pruritus. J Anesth 2007;21:159 – 63 17. Han DW, Hong SW, Kwon JY, Lee JW, Kim KJ. Epidural ondansetron is more effective to prevent postoperative pruritus and nausea than intravenous ondansetron in elective cesarean delivery. Acta Obstet Gynecol Scand 2007;86:683–7 18. Kocamanoglu IS, Baris S, Karakaya D, Sener B, Tur A, Cetinkaya M. Effects of granisetron with droperidol or dexamethasone on prevention of postoperative nausea and vomiting after general anesthesia for cesarean section. Methods Find Exp Clin Pharmacol 2005;27:489 –93 19. Manullang TR, Viscomi CM, Pace NL. Intrathecal fentanyl is superior to intravenous ondansetron for the prevention of perioperative nausea during cesarean delivery with spinal anesthesia. Anesth Analg 2000;90:1162– 6 20. Pan PH, Moore CH. Intraoperative antiemetic efficacy of prophylactic ondansetron versus droperidol for cesarean section patients under epidural anesthesia. Anesth Analg 1996;83:982– 6 21. Pan PH, Moore CH. Comparing the efficacy of prophylactic metoclopramide, ondansetron, and placebo in cesarean section patients given epidural anesthesia. J Clin Anesth 2001;13:430 –5 22. Sanai L. Ondansetron after caesarian section. Br J Anaesth 2001;87:942–3 23. Schumann R, Hudcova J. Cholestasis of pregnancy, pruritus and 5-hydroxytryptamine 3 receptor antagonists. Acta Obstet Gynecol Scand 2004;83:861–2 24. Yazigi A, Chalhoub V, Madi-Jebara S, Haddad F. Ondansetron for prevention of intrathecal opioids-induced pruritus, nausea and vomiting after cesarean delivery. Anesth Analg 2004;98:264 25. Crighton IM, Hobbs GJ, Reid MF. Ondansetron for the treatment of pruritus after spinal opioids. Anaesthesia 1996;51:199 –200 26. Charuluxananan S, Kyokong O, Somboonviboon W, Narasethakamol A, Promlok P. Nalbuphine versus ondansetron for prevention of intrathecal morphine-induced pruritus after cesarean delivery. Anesth Analg 2003;96:1789 –93 27. Harnett MJ, O’Rourke N, Walsh M, Carabuena JM, Segal S. Transdermal scopolamine for prevention of intrathecal morphine-induced nausea and vomiting after cesarean delivery. Anesth Analg 2007;105:764 –9 28. Sarvela PJ, Halonen PM, Soikkeli AI, Kainu JP, Korttila KT. Ondansetron and tropisetron do not prevent intraspinal morphine- and fentanyl-induced pruritus in elective cesarean delivery. Acta Anaesthesiol Scand 2006;50:239 – 44 29. Siddik-Sayyid SM, Aouad MT, Taha SK, Azar MS, Hakki MA, Kaddoum RN, Nasr VG, Yazbek VG, Baraka AS. Does ondansetron or granisetron prevent subarachnoid morphine-induced pruritus after cesarean delivery? Anesth Analg 2007;104:421– 4

182


30. Yazigi A, Chalhoub V, Madi-Jebara S, Haddad F, Hayek G. Prophylactic ondansetron is effective in the treatment of nausea and vomiting but not on pruritus after cesarean delivery with intrathecal sufentanil-morphine. J Clin Anesth 2002;14:183– 6 31. Yeh HM, Chen LK, Lin CJ, Chan WH, Chen YP, Lin CS, Sun WZ, Wang MJ, Tsai SK. Prophylactic intravenous ondansetron reduces the incidence of intrathecal morphine-induced pruritus in patients undergoing cesarean delivery. Anesth Analg 2000;91:172–5 32. Balki M, Kasodekar S, Dhumne S, Carvalho JC. Prophylactic granisetron does not prevent postdelivery nausea and vomiting during elective cesarean delivery under spinal anesthesia. Anesth Analg 2007;104:679 – 83 33. Peixoto AJ, Celich MF, Zardo L, Peixoto Filho AJ. Ondansetron or droperidol for prophylaxis of nausea and vomiting after intrathecal morphine. Eur J Anaesthesiol 2006;23:670 –5 34. Charuluxananan S, Somboonviboon W, Kyokong O, Nimcharoendee K. Ondansetron for treatment of intrathecal morphineinduced pruritus after cesarean delivery. Reg Anesth Pain Med 2000;25:535–9 35. Chaney MA. Side effects of intrathecal and epidural opioids. Can J Anaesth 1995;42:891–903 36. Kyriakides K, Hussain SK, Hobbs GJ. Management of opioidinduced pruritus: a role for 5-HT3 antagonists? Br J Anaesth 1999;82:439 – 41 37. Szarvas S, Harmon D, Murphy D. Neuraxial opioid-induced pruritus: a review. J Clin Anesth 2003;15:234 –9 38. Bromage PR. The price of intraspinal narcotic analgesia: basic constraints. Anesth Analg 1981;60:461–3 39. Girgin NK, Gurbet A, Turker G, Aksu H, Gulhan N. Intrathecal morphine in anesthesia for cesarean delivery: dose-response relationship for combinations of low-dose intrathecal morphine and spinal bupivacaine. J Clin Anesth 2008;20:180 –5 40. Bonnet MP, Marret E, Josserand J, Mercier FJ. Effect of prophylactic 5-HT3 receptor antagonists on pruritus induced by neuraxial opioids: a quantitative systematic review. Br J Anaesth 2008;101:311–9 41. George RB, Allen TK, Habib AS. Prophylaxis of neuraxial opioid pruritus with 5HT3 antagonists: a systematic review. Anesthesiology 2007;107:A1039 42. Gan TJ, Meyer TA, Apfel CC, Chung F, Davis PJ, Habib AS, Hooper VD, Kovac AL, Kranke P, Myles P, Philip BK, Samsa G, Sessler DI, Temo J, Tramer MR, Vander Kolk C, Watcha M; Society for Ambulatory Anesthesia. Society for Ambulatory Anesthesia guidelines for the management of postoperative nausea and vomiting. Anesth Analg 2007;105:1615–28 43. Terrin N, Schmid CH, Lau J, Olkin I. Adjusting for publication bias in the presence of heterogeneity. Stat Med 2003;22:2113–26 44. Terrin N, Schmid CH, Lau J. In an empirical evaluation of the funnel plot, researchers could not visually identify publication bias. J Clin Epidemiol 2005;58:894 –901 45. Thornton A, Lee P. Publication bias in meta-analysis: its causes and consequences. J Clin Epidemiol 2000;53:207–16


Economics, Education, and Policy Section Editor: Franklin Dexter

Personalized Oral Debriefing Versus Standardized Multimedia Instruction After Patient Crisis Simulation Timothy M. Welke, MD* Vicki R. LeBlanc, PhD*† Georges L. Savoldelli, MD, MEd*†‡ Hwan S. Joo, MD* Deven B. Chandra, MD* Nicholas A. Crabtree, MB ChB* Viren N. Naik, MD, MEd*†

BACKGROUND: Simulation experience alone without debriefing is insufficient for learning. Standardized multimedia instruction has been shown to be useful in teaching surgical skills but has not been evaluated for use as an adjunct in crisis management training. Our primary purpose in this study was to determine whether standardized computer-based multimedia instruction is effective for learning, and whether the learning is retained 5 wk later. Our secondary purpose was to compare multimedia instruction to personalized video-assisted oral debriefing with an expert. METHODS: Thirty anesthesia residents were recruited to manage three different simulated resuscitation scenarios using a high-fidelity patient simulator. After the first scenario, subjects were randomized to either a computer-based multimedia tutorial or a personal debriefing of their performance with an expert and videotape review. After their respective teaching, subjects managed a similar posttest resuscitation scenario and a third retention test scenario 5 wk later. Performances were independently rated by two blinded expert assessors using a previously validated assessment system. RESULTS: Posttest (12.22 ⫾ 2.19, P ⫽ 0.009) and retention (12.80 ⫾ 1.77, P ⬍ 0.001) performances of nontechnical skills were significantly improved in the standardized multimedia instruction group compared with pretest (10.27 ⫾ 2.10). There were no significant differences in improvement between the two methods of instruction. CONCLUSION: Computer-based multimedia instruction is an effective method of teaching nontechnical skills in simulated crisis scenarios and may be as effective as personalized oral debriefing. Multimedia may be a valuable adjunct to centers when debriefing expertise is not available. (Anesth Analg 2009;109:183–9)

F

ull-scale mannequin simulators are increasingly recognized as useful educational adjuncts and are predominantly used to teach the management of clinical crises. Crises are clinical events that occur rarely, but are “high stakes” situations associated with significant morbidity and mortality.1 Simulation provides a patient-safe environment to learn the management of these crises.2 Nontechnical skills have been shown to be major determinants of successful anesthesia

From the *St. Michael’s Anesthesia Research into Teaching (SMART) Simulation Group, Department of Anesthesia, St. Michael’s Hospital, University of Toronto, Toronto, Ontario, Canada; †Wilson Centre for Research in Education, University Health Network, University of Toronto, Toronto, Ontario, Canada; and ‡Department of Anesthesia and Unit for Development and Research in Medical Education, University Hospitals and Faculty of Medicine of Geneva, Geneva, Switzerland. Accepted for publication January 5, 2009. Supported by a peer-reviewed grant from the Physician Services Incorporated Foundation, North York, Ontario, Canada. Reprints will not be available from the author. Address correspondence to Viren N. Naik, MD, MEd, FRCPC, Department of Anesthesia, St. Michael’s Hospital, University of Toronto, 30 Bond Street, Toronto, ON, Canada M5B 1W8. Address e-mail to [email protected]. Copyright © 2009 International Anesthesia Research Society DOI: 10.1213/ane.0b013e3181a324ab

Vol. 109, No. 1, July 2009

crisis management.3 Nontechnical skills are those that do not relate to medical knowledge or technical procedures but instead encompass cognitive skills (e.g., decision making, situation awareness) and interpersonal skills (e.g., exchanging information, assertiveness). A comprehensive and reliable assessment tool for nontechnical skills has been developed: the Anesthetists’ Nontechnical Skills (ANTS) system.4 The number of mannequin patient simulators in medical education has expanded rapidly in the last decade.5,6 Despite their popularity, their use may be limited by several factors. First, anesthesiologists have identified a fear of judgment by peers as a barrier to using simulation.7 Second, a relative paucity of simulation debriefing expertise prevents maximum implementation of simulator-based education. Finally, simulation is associated with a significant operating cost, which can strain limited education resources.8 One way to address these limitations is to explore novel adjuncts to simulator education. This study evaluates the effectiveness of computer-based multimedia instruction as an alternative to personal debriefing. A standardized computer-based multimedia 183

model offers several advantages compared with the traditional oral and video-assisted oral debriefing. First, participants control the pace of the debriefing, which engages the learner and may help to hold their attention. Second, debriefing in the absence of an instructor or peers may decrease anxiety. Finally, experts have noted anecdotally that junior trainees make similar errors in similar simulated scenarios; thus, debriefing for a particular scenario can be standardized. Computer-based multimedia instruction has been used to successfully teach a variety of medical and nonmedical technical skills, including simple surgical skills, advanced cardiac life support (ACLS), and aviation.9 –11 The question remains whether a multimedia computer program can provide the cognitive framework necessary for learning complex crisis resource management. The primary purpose of this study was to determine whether computer-based multimedia instruction is effective for learning and whether the learning is retained 5 wk later. The secondary purpose was to compare multimedia instruction to personalized video-assisted oral debriefing with an expert.

METHODS Participant Recruitment and Orientation Phase After IRB approval (St. Michael’s Hospital, University of Toronto, Ontario, Canada), 30 anesthesia residents in their first and second year of training at the University of Toronto were approached to participate as subjects. In addition to informed consent, confidentiality agreements were obtained to ensure that the content of the simulation scenarios was not disseminated before study completion. All subjects were given a structured orientation to the simulation environment. Orientation consisted of familiarization with the simulated operating room, SimMan威 mannequin (Laerdal Medical Canada), anesthetic cart, drugs and anesthetic gas machine. After orientation, participants completed a questionnaire of demographic data and previous simulator experience.

Study Design This study used a prospective randomized design with two treatment groups: traditional personalized video-assisted oral debriefing or a standardized computer-based multimedia debriefing. Randomization was performed using a computer random number generator and sealed envelope technique. Randomization was stratified for subjects’ level of training. All subjects managed three ACLS resuscitation scenarios. Subjects were aware that all scenarios could be managed with knowledge of ACLS, and they were asked to review their algorithms before their simulation sessions. Subjects’ review of ACLS was confirmed verbally by the investigators for all subjects. Anesthesiologists with simulation expertise developed scenarios using a subjective iterative process with a goal 184

Multimedia Instruction for Simulation

of equal complexity and duration. All scenarios were then piloted with attending anesthesiologists not involved in scenario design and scored on a 5-point Likert scale to confirm similar complexity and consistency in execution. One scenario consisted of bradycardia and hypotension secondary to pacemaker failure. A second scenario consisted of pulseless electrical activity secondary to hypovolemia. A third scenario consisted of ventricular fibrillation secondary to cardiac ischemia. In each scenario, participants played the role of the primary anesthesiologist. Simulation center staff and one investigator functioned as perioperative personnel in each scenario in the scripted roles of surgeon and nurse. A second investigator played the scripted role of a colleague anesthesiologist who was available to help and to perform tasks when directed but did not offer crisis management advice or differential diagnoses. All three cardiac arrest scenarios were set in the operating room and lasted 10 min. All subjects managed a first scenario as a pretest. Management was recorded with full video, audio, and a real-time monitor overlay that included heart rate, arterial blood pressure, oxygen saturation, electrocardiogram, and end-tidal carbon dioxide. After the pretest, subjects received one of two types of instruction according to their randomization. Subjects in the personalized video-assisted oral debriefing group received a constructive critique of their management of the pretest scenario focused on their crisis management and nontechnical skills from an experienced instructor.12 A videotape record of the subject’s management of the scenario was used to highlight good demonstration of skills and identifying skills that would benefit from improvement. Debriefing concluded after all the subject’s questions and instructor’s observations were addressed. Subjects in the standardized multimedia group received instruction through a standardized multimedia presentation that was learner controlled with a computer. The multimedia presentation was developed by the investigators to reflect four main objectives in crisis management: 1) planning and preparing, 2) calling for help early and announcing a crisis clearly, 3) distributing tasks to appropriate individuals, and 4) reevaluation. Subjects maintained control over the pace and review of material. The presentation used text, audio voice-over, and a set of digital videos. Unlike the personalized oral debriefing group, these videos were standardized reenactments scripted by the investigators to demonstrate poor performance in the management of a crisis. Each demonstration of a poor performance was followed by instruction, highlighting how performance could be improved by applying the four principles of crisis management stated above. The multimedia presentation would then follow with a demonstration of the same scenario with ideal management and application of the crisis ANESTHESIA & ANALGESIA

Table 1. Subject Demographics (n ⫽ 30) Personalized video-assisted oral debriefing (n ⫽ 15)

Standardized multimedia instruction (n ⫽ 15)

10:5 73 1.0 (0.25–1.75)

10:5 87 1.0 (1.0–1.75)

8:7

8:7

Male:female ratio % Previous simulator experience Previous simulator sessions of those with previous experience, median (interquartile range) Postgraduate training year (PGY) (PGY1:PGY2 ratio) Advanced cardiac life support certified (%)

management principles. Multimedia instruction concluded when subjects indicated that they had completed the module. After debriefing or multimedia instruction, all residents immediately managed a posttest crisis scenario. After a period of 5 wk, all participants returned and managed a retention test scenario.

Scoring Performances Two experienced simulator instructors/examiners independently reviewed the videotapes of the three tests: pretest, posttest, and retention test. Examiners were blinded as to the order of the performance they were scoring. Examiners were also blinded to identity and level of training of participants as well as the type of instruction they received and the sequence of the scenarios. Performances were scored using the previously validated global rating scale for nontechnical skills, the ANTS.4* The ANTS system consists of four categories: task management, team working, situation awareness, and decision making. Each category is scored out of 4, with 4 the highest score and 1 the lowest possible score. For each performance, the behaviors observed were scored at the category level and guided by descriptors of the component skill elements. Each subject also received a total ANTS score (range 4 –16) by totaling the scores of the four categories.

Statistical Analysis Statistical analysis was performed using SPSS16.0 (Chicago, IL). Descriptive statistics were used to analyze demographic data. The interrater reliability was measured both for the total ANTS score and at the category level using the Cohen kappa coefficient. Although the individual Likert scales that make up the ANTS scores are clearly ordinal in nature, these scores behave empirically as a parametric variable. Consequently, we, like other researchers, chose to use parametrical statistical analysis for our normally distributed data.13–14 The primary outcome was to detect differences in performance for the multimedia instruction group between pretest, posttest, and retention test, using the *Available at: http://www.abdn.ac.uk/iprc/ants_papers.shtml. Accessed January 2, 2009.

Vol. 109, No. 1, July 2009

12 (80)

13 (86.7)

ANTS category scores and total ANTS scores. Our secondary outcome measure was to detect whether the changes in performance were different between personalized video-assisted oral debriefing and multimedia instruction using the ANTS category scores and total ANTS scores. Our results were analyzed using a two-way mixed design analysis of variance (ANOVA). The independent variables were “type of debriefing/instruction” as the between-subjects variable and “test phase” as the within-subjects measure. A mixed ANOVA was used because our variables were a mixture of “between subjects” and “within subjects.” A Tukey post hoc test was performed for significant findings. A P value ⬍0.05 was considered significant. Sample size was calculated a priori for our primary outcome. In the field of psychology and education, an effect-size of more than 1.0 sd is considered large and acceptable for a given teaching intervention.15 Thus, assuming an effect-size of 1.0 and a power of 0.8 for a repeated measures ANOVA with three tests for within multimedia instruction, we calculated a sample size of 12 subjects (␣ ⫽ 0.05 two tailed). Conservatively, we recruited 15 subjects to account for attrition.

RESULTS Demographics All 30 participants completed the study. The videotape of one test in the multimedia instruction group was lost due to a recording malfunction. Participants’ characteristics are summarized in Table 1. There were equal numbers of first and second year residents in each group. The high ratio of male to female participants in each group reflects the demographics of our anesthesia residency. Most residents had completed ACLS training and the extent of previous simulator training was not statistically different between our treatment groups.

Interrater Reliability The overall interrater reliability for the total ANTS score level was good: Cohen ␬ ⫽ 0.68 (P ⬍ 0.001). At the category level, across the four categories, interrater reliability was also good: Cohen ␬ ⫽ 0.62 (P ⬍ 0.001).


185

Figure 1. Improvement in nontechnical skills for standardized multimedia instruction. Comparison of pretest, posttest, and retention test for the standardized multimedia instruction group for each of the four Anesthetists’ Nontechnical Skills (ANTS) categories (task management, team working, situation awareness, decision making).

Figure 2. Personalized video-assisted oral debriefing versus standardized multimedia instruction comparing pretests, posttests, and retention tests. The total ANTS score is a summation of the scores for each of the four Anesthetists’ Nontechnical Skills (ANTS) categories (task management, team working, situation awareness, decision making).

Primary Outcome Measure: ANTS Category and Total ANTS Scores for Standardized Multimedia Instruction Group In the standardized multimedia instruction group, all four ANTS categorical scores significantly improved from pretest when measured at posttest (all 186


P ⬍ 0.01) and retention test (all P ⬍ 0.01) (Fig. 1). For total ANTS scores, the posttest (P ⫽ 0.009) and retention test (P ⬍ 0.001) total ANTS scores were also significantly improved when compared with their pretest scores (Fig. 2). Therefore, the demonstrated effect of the standardized multimedia instruction was ANESTHESIA & ANALGESIA

Table 2. Anesthetists’ Nontechnical Skills (ANTS) Scores for All Three Test Phases Subject group Standardized multimedia instruction (n ⫽ 15)

Personalized video-assisted oral debriefing (n ⫽ 15)

ANTS category Task management Team working Situation awareness Decision making Total ANTS Task management Team working Situation awareness Decision making Total ANTS

Pretest score

Posttest score

Retention score

F

P

2.46 ⫾ 0.48 2.63 ⫾ 0.50 2.59 ⫾ 0.59

3.07 ⫾ 0.37 3.15 ⫾ 0.35 3.18 ⫾ 0.49

3.17 ⫾ 0.44 3.13 ⫾ 0.42 3.30 ⫾ 0.37

F(2,42) ⫽ 11.85 F(2,42) ⫽ 7.12 F(2,42) ⫽ 8.96

⬍0.001 0.002 ⬍0.001

2.61 ⫾ 0.62 10.27 ⫾ 2.10 2.27 ⫾ 0.65 2.32 ⫾ 0.53 2.23 ⫾ 0.66

3.20 ⫾ 0.43 12.22 ⫾ 2.19 2.73 ⫾ 0.65 2.77 ⫾ 0.51 2.72 ⫾ 0.61

3.30 ⫾ 0.51 12.80 ⫾ 1.77 2.90 ⫾ 0.42 2.88 ⫾ 0.39 2.82 ⫾ 0.48

F(2,42) ⫽ 7.54 F(2,42) ⫽ 9.10 F(2,42) ⫽ 4.68 F(2,42) ⫽ 5.72 F(2,42) ⫽ 4.32

0.002 ⬍0.001 0.015 0.006 0.020

2.18 ⫾ 0.64 9.00 ⫾ 2.45

2.70 ⫾ 0.61 11.30 ⫾ 2.08

2.92 ⫾ 0.36 11.62 ⫾ 1.54

F(2,42) ⫽ 7.13 F(2,42) ⫽ 9.38

0.002 ⬍0.001

0.9 standard deviations. In the multimedia group, retention test scores were not significantly different from posttest scores for all four ANTS categories (all P ⫽ NS) and total ANTS (P ⫽ 0.38) (Fig. 2, Table 2).

Secondary Outcome Measures: Comparison of Efficacy of Standardized Multimedia Instruction to Personalized Video-Assisted Oral Debriefing Our groups were homogenous before their instruction modalities when compared with our pretest total ANTS scores (both P ⫽ 0.14). The improvements in total ANTS score from pretest to posttest (P ⫽ 0.97), pretest to retention test (P ⫽ 0.94), and posttest to retention test (P ⫽ 0.84) were similar for both instruction modalities (Fig. 2). Therefore, our study demonstrated only a very small effect size of 0.02 standard deviations for instruction modality.

DISCUSSION This study examined the effectiveness of multimedia instruction for learning and demonstrating crisis resource management principles. The results of our study suggest that standardized multimedia instruction is an effective method of improving the nontechnical skills of anesthesia trainees for crisis management and that these nontechnical skills are retained 5 wk after the instruction. The use of high-fidelity patient simulators in medical training has increased significantly during the past few years.5 Simulators have several perceived advantages over traditional patient-centered didactic training. The advantages include repeatability, standardization, the ability to simulate rare events and the lack of potential patient harm.8 The strong face validity of high-fidelity simulation may explain the increasing use of simulators for training, despite the lack of clear evidence.6 Our study confirms that anesthesia nontechnical skills are not only learned by practicing crisis resource management scenarios during high-fidelity simulation but are also retained for a period of at least 5 wk. Previous studies attest to the efficacy of simulator based education on the immediate acquisition of nontechnical skills.4 However, these skills may not be Vol. 109, No. 1, July 2009

used in clinical practice for a considerable duration of time after training. We have demonstrated that, after a period of 5 wk without instruction, there is no attrition of nontechnical skills. Grober et al.16 demonstrated that simulator-acquired surgical skills are durable 4 mo later, albeit with considerable attrition. Further study is necessary to evaluate whether the nontechnical skills of crisis resource management in anesthesia are maintained for more than 5 wk. Whether these skills are retained in clinical practice is the fundamental question. It will be difficult, if not impossible, for clinical research to answer this question because of the paucity and unpredictability of clinical critical events in anesthesia. In the meantime, the simulator setting is an alternate way to test these crisis resource management skills. A secondary outcome of this study was to compare multimedia instruction to a personalized videoassisted oral debriefing. Our results suggest that multimedia instruction may be as effective as personalized video-assisted oral debriefing. Although this was not the primary purpose of this study, the small effect-size between our teaching modalities reassures us that there is no strong educational advantage of one method of instruction over the other.15 Personal video-assisted oral debriefings of simulated crises incorporate the key elements of adult learning: engagement with the learner, linking to prior knowledge, opportunity for reflection and feedback.17 Although instruction and feedback are recognized as the critical factors in simulation learning, there are few studies examining novel alternatives to the current “gold standard” of personalized video-assisted oral debriefing. Computer-based learning is useful for training a wide variety of skills from reading to piloting aircraft. In the surgical field, multimedia learning has been used to help teach the procedural skills for laparoscopic surgery.18 This study does not attempt to demonstrate why multimedia is effective, but we can speculate that multimedia-guided selfdebriefing incorporates the same principles of adult learning. © 2009 International Anesthesia Research Society

187

Managing crises whether real or simulated is a stressful experience. Debriefing may also be viewed by participants as equally stressful because of the presence of senior expert supervisors who are often responsible for academic evaluations. This may cause participants to misinterpret their debriefing as a summative evaluation. It is possible that in the absence of supervisors, such as during a multimedia debrief, the stress of debriefing is diminished and therefore allows more efficient cognitive processes to focus solely on learning. In short, anesthesiologists have identified a fear of judgment by peers as a significant barrier to simulator use in continuing education.4 Multimedia instruction may overcome this barrier. Another key difference between personalized oral debriefing and multimedia instruction is that computer-based debriefing is self-guided. Participants have control over the pace and review of the instruction and, therefore, they may choose to focus on areas of concern. This encourages an active role in the instruction session and may allow review of selfperceived weaknesses; although, this was not measured. During personalized oral debriefing with an expert, the pace and direction of the session may not be determined by the participant and individual areas of concern might not be addressed. Finally, a unique aspect of multimedia instruction is the standardization of the debrief content. All participants regardless of performance receive the same debrief information. Personalized debriefing by an instructor delivered orally is thought to be efficient because more time can be spent on areas of weakness and less on areas of strength. This approach might not always be appropriate as participants may benefit from a full demonstration of accepted good and bad performance. With multimedia instruction, participants may find it easier to model good performance rather than avoid poor performance. By depersonalizing the demonstrations, participants may feel less inclined to defend their individual simulator performances, focusing cognitive skills on learning instead. This study has key limitations that may affect the clinical applicability of the results. We chose not to include a control group of participants deprived of the benefit of any type of instruction. Instead, we decided to use the current gold standard of oral debriefing as a control group against which multimedia instruction was compared. A previous study by our group has shown that trainees who do not receive any form of instruction demonstrate no improvement in nontechnical skills. We felt that residents should not participate in a study without some educational benefit.19 All participants were orientated to the simulator before testing in an attempt to address the potential confounding influences of previous simulator experience and knowledge of the simulation environment. By assessing self-reported simulation experience, we found that there were no significant differences between the two groups. A further consideration was 188


that subjects’ performances on the retention test could have been improved simply by a further 5 wk of clinical experience, rather than retention of simulatorbased learning. We chose to measure retention at 5 wk, as opposed to a longer interval, to minimize the chance of exposure to clinical critical events and potential learning from these situations. We have demonstrated that multimedia instruction is a useful adjunct to simulator training and may offer several distinct advantages over traditional personalized video-assisted oral debriefing. Using experts to debrief participants is time consuming and costly, and indeed experts may not be readily available. Multimedia debriefing allows simulator training to be independent of these restraints, which could result in expanded access to simulator training at a reduced cost of operation. By removing the fear of judgment by peers, multimedia instruction may also help remove one of the major barriers to simulator-based continuing education in anesthesia. We recommend multimedia debriefing as an effective adjunct to the simulator-based training of nontechnical skills. ACKNOWLEDGMENTS This study would not be possible without the participation of the first and second year anesthesiology residents at the University of Toronto. The authors also thank Mr. Roger Chow, Ms. Lynn Dionne, and Ms. Erinn Macaulay for their participation and coordination of the simulations. REFERENCES 1. Gaba DM, Fish KJ, Howard SK. Crisis management. In: Anesthesiology. Philadelphia: Churchill Livingstone, 1994:5– 47 2. Issenberg SB, McGaghie WC, Hart IR, Mayer JW, Felner JM, Petrusa ER, Waugh RA, Brown DD, Safford RR, Gessner IH, Gordon DL, Ewy GA. Simulation technology for health care professional skills training and assessment. JAMA 1999; 282:861– 6 3. Fletcher GC, McGeorge P, Flin RH, Glavin RJ, Maran NJ. The role of non-technical skills in anaesthesia: a review of current literature. Br J Anaesth 2002;88:418 –29 4. Fletcher G, Flin R, McGeorge P, Glavin R, Maran N, Patey R. Anaesthetists’ Non-Technical Skills (ANTS): evaluation of a behavioural marker system. Br J Anaesth 2003;90:580 – 8 5. Morgan PJ, Cleave-Hogg D. A worldwide survey of the use of simulation in anesthesia. Can J Anesth 2002;49:659 – 62 6. Issenberg SB, McGaghie WC, Petrusa ER, Lee GD, Scalese RJ. Features and uses of high-fidelity medical simulations that lead to effective learning: a BEME systematic review. Med Teach 2005;27:10 –28 7. Savoldelli GL, Naik VN, Hamstra SJ, Morgan PJ. Barriers to use of simulation-based education. Can J Anaesth 2005; 52:944 –50 8. Wong AK. Full scale computer simulators in anesthesia training and evaluation. Can J Anaesth 2004;51:455– 64 9. Summers AN, Rinehart GC, Simpson D, Redlich PN. Acquisition of surgical skills: a randomized trial of didactic, videotape, and computer-based training. Surgery 1999;126:330 – 6 10. Roessingh JJM. Transfer of manual flying skills from PCbased simulation to actual flight-comparison of in-flight measured data and instructor ratings. Int J Aviat Psychol 2005;15:67–90 11. Schwid HA, Rooke GA, Ross BK, Sivarajan M. Use of a computerized advanced cardiac life support simulator improves retention of advanced cardiac life support guidelines better than a textbook review. Crit Care Med 1999;27:821– 4


12. Howard SK, Gaba DM, Fish KJ, Yang G, Sarnquist FH. Anesthesia crisis resource management training: teaching anesthesiologists to handle critical incidents. Aviat Space Environ Med 1992;63:763–70 13. Regehr G, MacRae H, Reznick RK, Szalay D. Comparing the psychometric properties of checklists and global rating scales for assessing performance on an OSCE-format examination. Acad Med 1998;73:993–7 14. Carifio J, Perla R. Resolving the 50-year debate around using and misusing Likert scales. Med Educ 2008;42:1150 –52 15. Cohen J. Statistical power analysis for the behavioral sciences. New York: Academic Press, 1977:24 –7 16. Grober ED, Hamstra SJ, Wanzel KR, Reznick RK, Matsumoto ED, Sidhu RS, Jarvi KA. Laboratory based training in urological microsurgery with bench model simulators: a randomized controlled trial evaluating the durability of technical skill. J Urol 2004;172:378 – 81

Vol. 109, No. 1, July 2009

17. Ostergaard HT, Ostergaard D, Lippert A. Implementation of team training in medical education in Denmark. Qual Saf Health Care 2004;13 (suppl 1):i91–i95 18. Schijven MP, Jakimowicz JJ, Broeders IA, Tseng LN. The Eindhoven laparoscopic cholecystectomy training course–improving operating room performance using virtual reality training: results from the first E.A.E.S. accredited virtual reality trainings curriculum. Surg Endosc 2005;19:1220 – 6 19. Savoldelli GL, Naik VN, Park J, Joo HS, Chow R, Hamstra SJ. Value of debriefing during simulated crisis management: oral versus video-assisted oral feedback. Anesthesiology 2006;105: 279 – 85


189

Brief Report

A Historical Perspective on Resident Evaluation, the Accreditation Council for Graduate Medical Education Outcome Project and Accreditation Council for Graduate Medical Education Duty Hour Requirement Steven H. Rose, MD Timothy R. Long, MD Beth A. Elliott, MD Michael J. Brown, MD

BACKGROUND: The Accreditation Council for Graduate Medical Education (ACGME) Outcome Project, endorsed at the 1999 ACGME annual meeting, was intended to shift the focus of residency program requirements and accreditation from processoriented assessment to an assessment of outcomes. The Outcome Project established six general competencies, each of which is supported by more specific competencies. METHODS: We compared contemporary resident evaluation based on the Outcome Project to faculty evaluation of a surgical resident at Mayo Clinic that was completed in 1917. RESULTS: The contemporary faculty assessment of resident performance was remarkably similar to the evaluation form and criteria used in 1917. All six general competencies, and nearly all of the more specific items listed under each general competency, were included in the 1917 evaluation. Duty hour data as a component of the 1917 resident evaluation included the number of hours per week spent in “practical work,” “medical library,” and “research work.” CONCLUSIONS: The remarkable similarities between the qualities assessed in the 1917 evaluation and the assessment of contemporary ACGME competencies suggest that a common set of desirable physician characteristics and behaviors can be identified and measured. (Anesth Analg 2009;109:190 –3)

M

ayo Clinic has a long history of involvement in graduate medical education (GME). The Accreditation Council for Graduate Medical Education (ACGME) Outcome Project, including formal establishment of six general competencies, was endorsed at the February 1999 ACGME annual meeting and has been progressively implemented. The six ACGME general competencies are defined in Table 1. We have previously compared the components of the ACGME Outcome Project to those of the Clinical Competence Committee Report required by the American Board of Anesthesiology to assess resident performance every 6 mo throughout the continuum of training in anesthesiology.1 In the current analysis, we compare a 1917 Mayo Clinic surgical resident evaluation to the competencies outlined in the ACGME Outcome Project. The 1917 document addressed duty From the Department of Anesthesiology, Mayo Clinic, Rochester, Minnesota. Accepted for publication December 4, 2008. Reprints will not be available from the author. Address correspondence to Steven H. Rose, MD, Department of Anesthesiology, 200 First St. SW, Rochester, MN 55905. Address e-mail to [email protected]. Copyright © 2009 International Anesthesia Research Society DOI: 10.1213/ane.0b013e3181a1653b

190

hours, allowing us to compare the duty hours recorded on this evaluation to contemporary ACGME duty hour requirements.

METHODS A colleague reviewing a historical surgical resident’s academic file for another purpose brought an assessment form from 1917 to our attention (Figs. 1 and 2) and we were impressed by the similarities to current systems of resident evaluation. The items addressed in the 1917 evaluation were compared with the contemporary evaluation of ACGME competencies assessed using the Mayo School of GME Integrated Scheduling and Evaluation System (ISES). ISES is a validated, behavioral electronic evaluation system implemented throughout the Mayo School of GME that is ACGME competency based. The authors are medical educators with significant experience in the assessment of resident performance using ACGME competency criteria. Each author independently mapped the criteria from the 1917 performance evaluation of a Mayo Clinic trainee (Figs. 1 and 2) to the ACGME competencies (Table 2). When the authors differed in the comparisons, a consensus opinion was generated through discussion. Vol. 109, No. 1, July 2009

Table 1. Accreditation Council for Graduate Medical Education (ACGME) Competencies 1. Patient care Residents must be able to provide patient care that is compassionate, appropriate, and effective for the treatment of health problems and the promotion of health. Residents are expected to: a. Communicate effectively and demonstrate caring and respectful behaviors when interacting with patients and their families b. Gather essential and accurate information about their patients c. Make informed decisions about diagnostic and therapeutic interventions based on patient information and preferences, up-to-date scientific evidence, and clinical judgment d. Develop and carry out patient management plans e. Counsel and educate patients and their families f. Use information technology to support patient care decisions and patient education g. Perform competently all medical and invasive procedures considered essential for the area of practice h. Provide health care services aimed at preventing health problems or maintaining health i. Work with health care professionals, including those from other disciplines, to provide patient-focused care 2. Medical knowledge Residents must demonstrate knowledge about established and evolving biomedical, clinical, and cognate (e.g. epidemiological and social-behavioral) sciences and the application of this knowledge to patient care. Residents are expected to: a. Demonstrate an investigatory and analytic thinking approach to clinical situations b. Know and apply the basic and clinically supportive sciences which are appropriate to their discipline 3. Practice-based learning and improvement Residents must be able to investigate and evaluate their patient care practices, appraise, and assimilate scientific evidence, and improve their patient care practices. Residents are expected to: a. Analyze practice experience and perform practice-based improvement activities using a systematic methodology b. Locate, appraise, and assimilate evidence from scientific studies related to their patients’ health problems c. Obtain and use information about their own population of patients and the larger population from which their patients are drawn d. Apply knowledge of study designs and statistical methods to the appraisal of clinical studies and other information on diagnostic and therapeutic effectiveness e. Use information technology to manage information, access on-line medical information, and support their own education f. Facilitate the learning of students and other health care professionals 4. Interpersonal and communication skills Residents must be able to demonstrate interpersonal and communication skills that result in effective information exchange and teaming with patients, their patients families, and professional associates. Residents are expected to: a. Create and sustain a therapeutic and ethically sound relationship with patients b. Use effective listening skills and elicit and provide information using effective nonverbal, explanatory, questioning, and writing skills c. Work effectively with others as a member or leader of a health care team or other professional group 5. Professionalism Residents must demonstrate a commitment to carrying out professional responsibilities, adherence to ethical principles, and sensitivity to a diverse patient population. Residents are expected to: a. Demonstrate respect, compassion, and integrity; a responsiveness to the needs of patients and society that supercedes self-interest; accountability to patients, society, and the profession; and a commitment to excellence and continuing professional development b. Demonstrate a commitment to ethical principles pertaining to provision or withholding of clinical care, confidentiality of patient information, informed consent, and business practices c. Demonstrate sensitivity and responsiveness to patients’ culture, age, gender, and disabilities 6. Systems-based practice Residents must demonstrate an awareness of and responsiveness to the larger context and system of health care and the ability to effectively call on system resources to provide care that is of optimal value. Residents are expected to: a. Understand how their patient care and other professional practices affect other health care professionals, the health care organization, and the larger society and how these elements of the system affect their own practice b. Know how types of medical practice and delivery systems differ from one another, including methods of controlling health care costs and allocating resources c. Practice cost-effective health care and resource allocation that does not compromise quality of care d. Advocate for quality patient care and assist patients in dealing with system complexities e. Know how to partner with health care managers and health care providers to assess, coordinate, and improve health care and know how these activities can affect system performance

RESULTS The contemporary assessment of resident performance was remarkably similar to the evaluation form and criteria used in 1917. All six ACGME competencies and at least 25 of the 28 items detailed under each of the competencies listed in Table 1 were included in Vol. 109, No. 1, July 2009

the 1917 evaluation. Two of the three items not included related to the use of information technology. They are inpatient care (1f) and practice-based learning and improvement (3e). The third item not included is any direct reference to preventive health in the patient care competency (1h). The 1917 evaluation © 2009 International Anesthesia Research Society

191

Table 2. Comparison of a 1917 Mayo Clinic Resident Evaluation with Contemporary Accreditation Council for Graduate Medical Education (ACGME) Competencies ACGME competencies addressed

1917 MSGME surgery resident evaluation

Figure 1. Qualities assessed in a 1917 Mayo Clinic resident evaluation form.

Figure 2. Evaluation of a 1917 Mayo Clinic resident from June 2, 1917 through July 31, 1917.

used a 5-point Likert scale (also used in our contemporary ISES). The faculty evaluator used 4 of the 5 scores within the range in this trainee performance evaluation, 1 (very good) to 4 (poor) (Fig. 2). The narrative remarks are sparse, but direct, in providing feedback regarding this trainee’s level of industry. The 1917 evaluation included three questions regarding duty hours. The questions addressed the estimated average number of hours per week spent in “practical work,” “medical library,” and “research work.” Viewed in the context of historical duty hour legends (especially in the surgical specialties), and in the context of modern duty hour requirements, the estimated duty hours were low, at only 48 h per week.

DISCUSSION The similarities between an assessment tool used to evaluate a Mayo Clinic resident more than 90 yr ago and a contemporary resident assessment are striking and reassuring. They suggest a common set of physician characteristics and behaviors can be identified and measured to provide meaningful feedback to physicians throughout the continuum of their medical education. We maintain the most important 192

Brief Report

Personal characteristics Personal appearance Manner with associates Manner with patients General demeanor Ambition Initiative Industry Persistence Promptness Honesty Loyalty Reception of criticism Training—general General intelligence Use of oral English Use of written English Chirography (handwriting) General medical training Knowledge of anatomy, physiology, and pathology Knowledge of medical ethics Comprehension of instructions Faithfulness in carrying out instructions System in work Diagnostic Skill in history taking Skill in physical examinations Thoroughness in physical examinations Judgment in referring patients for special examinations Judgment in interpreting results of special examinations Judgment in summarizing cases Operative Quickness of vision Knowledge of and faithfulness in asepsis Manual dexterity Care of postoperative cases

1a, 5a 1i, 4c 1a, 1e, 4a–b, 5a–c 1a, 1e, 4a–c, 5a–c 1b–c, 2a, 3a–e, 5a–b, 6d 1b–c, 2a, 3a–d, 5a–b, 6d 1b–c, 2a, 3a–d, 5a–b, 6d 1b–c, 2a, 3a–d, 5a–b, 6a 1b–c, 2a, 3a–d, 5a–b, 6a 1b–c, 2a, 3a–d, 5a–b, 6a 1b–c, 2a, 3a–d, 5a–b, 6a 3a 2a–b 1a, 1e, 3f, 4b–c, 6e 1a, 1e, 1i, 3f, 4b–c, 6e 1a, 1e, 4b–c 1c–d, 2a–b, 3d 1b, 1g, 2a–b, 3b 1a, 1e, 4a, 5a–c 1d–e, 4c, 5a 1d–e, 1g, 4c 1b, 1i, 4c, 6a–e 1a–b, 1e, 4a–b, 5a, 5c 1b 1b 1b–d, 1i, 4c, 5a, 6a, 6c, 6e 1b–e, 2a–b, 3b, 3d, 6c 1a–e, 1i, 2a, 3b, 3d, 3f, 4b–c, 5a–c, 6e 1g 1d, 1g, 2b, 5a, 6c 1g 1c–e, 1i, 2a–b, 3b, 3d, 4a–c, 5a–c, 6c–e

MSGME ⫽ Mayo School of Graduate Medical Education; ACGME ⫽ Accreditation Council for Graduate Medical Education.

changes in resident evaluation are not further identifying desirable characteristics of physicians but improving the instruments by which these characteristics are measured. Examples of innovations that may provide more effective, comprehensive, and sophisticated evaluation include the use of simulation (and other forms of experiential training and testing), 360° evaluations, chart-stimulated reviews, objective structured clinical ANESTHESIA & ANALGESIA

evaluations, and any of a variety of other metrics, several of which are outlined in the ACGME Outcome Project “Toolbox.”* Although the use of information technology, detailed under the general competencies of patient care (1f) and practice-based learning and improvement (3e) in Table 1, was not specifically addressed in the 1917 evaluation, the use of the *http://www.acgme.org/Outcome/, last accessed July 2008.

Vol. 109, No. 1, July 2009

medical library may addresses this competency, as this was the most advanced information technology available at the time. REFERENCE 1. Rose SH, Burkle CM. Accreditation Council for Graduate Medical Education competencies and the American Board of Anesthesiology Clinical Competence Committee: a comparison. Anesth Analg 2006;102:212–16


193

Neurosurgical Anesthesiology and Neuroscience Section Editor: Adrian W. Gelb

The Effect of Sedation on Intracranial Pressure in Patients with an Intracranial Space-Occupying Lesion: Remifentanil Versus Propofol Francois Girard, MD, FRCPC* Robert Moumdjian, MD, FRCSC† Daniel Boudreault, MD, FRCPC* Philippe Chouinard, MD, FRCPC* Alain Bouthilier, MD, FRCSC† Monique Ruel, RN*

BACKGROUND: In this study, we compared the effect of light sedation with remifentanil versus propofol on intracranial (ICP) and cerebral perfusion pressure (CPP) of patients undergoing stereotactic brain tumor biopsy under regional anesthesia. METHODS: This was a prospective, open-label, randomized, and controlled study. Forty patients undergoing stereotactic brain tumor biopsy under regional anesthesia were randomized into two groups to receive remifentanil or propofol titrated to a level of four on the modified Assessment of Alertness/Sedation Scale. ICP was measured via the biopsy needle. RESULTS: At the targeted level of sedation, the rates of infusion for remifentanil and propofol were, respectively, 4.2 ⫾ 1.8 ␮g 䡠 kg⫺1 䡠 h⫺1 and 4.3 ⫾ 2.5 mg 䡠 kg⫺1 䡠 h⫺1. At the time of ICP measurement, patients in the remifentanil group had a slower respiratory rate (11/min ⫾ 3 vs 15 per min ⫾ 3, P ⫽ 0.0001) and a higher Pco2 (48.3 ⫾ 6.2 mm Hg vs 43.1 ⫾ 5.5 mm Hg, P ⫽ 0.009) than patients in the propofol group. The mean was similar for both groups, 19.0 ⫾ 11.9 mm Hg vs 16.4 ⫾ 11.1 mm Hg for remifentanil and propofol, respectively (P ⫽ 0.48). Higher mean arterial blood pressure in the remifentanil group (101.1 ⫾ 13.7 mm Hg vs 85.8 ⫾ 12.7 mm Hg, P ⫽ 0.0008) resulted in a higher CPP than the propofol group: 82.0 ⫾ 19.0 mm Hg vs 69.5 ⫾ 17.0 ⫾ 19.0 mm Hg (P ⫽ 0.03). CONCLUSION: Light sedation with remifentanil does not result in a higher ICP than propofol in patients undergoing stereotactic brain tumor biopsy. CPP might be better preserved with remifentanil. (Anesth Analg 2009;109:194 –8)

R

emifentanil, an ultrashort-acting opioid derivative of fentanyl, has a number of properties that support its use for sedation of the neurologically impaired patient. Remifentanil’s rapid onset and offset of action allows for timely neurologic examination, even after a prolonged infusion. Its potent analgesic property helps provide stable hemodynamics during painful procedures in the intensive care unit (ICU).1,2 However, despite the increasing number of studies reporting its use as a sedative in the neuro ICU,2– 4 the effect of remifentanil on intracranial dynamics remains somewhat controversial. In mechanically ventilated patients with acute brain injury, moderately deep remifentanil sedation did not result in a higher intracranial pressure (ICP) than fentanyl, sufentanil, or morphine4 when associated From the *Department of Anesthesiology, and †Neurosurgery Division, Centre Hospitalier de l’Universite´ de Montreál, Hoˆpital Notre-Dame, Montreal, Canada. Accepted for publication January 17, 2009. Reprints will not be available from the author. Address correspondence to François Girard, MD, FRCPC, Department of Anesthesiology, CHUM, Hoˆpital Notre-Dame, 1560 Sherbrooke East, Montreal, Canada, H2L 4M1. Address e-mail to [email protected]. Copyright © 2009 International Anesthesia Research Society DOI: 10.1213/ane.0b013e3181a3ea3a

194

with propofol or midazolam. Conversely, the initiation of sedation with remifentanil as a single drug has been associated with an increase of ICP and a decrease of cerebral perfusion pressure (CPP) when compared with baseline in the same population of patients.2 Remifentanil sedation has not been directly compared with propofol in the neurosurgic setting. In addition, there is little information with regard to the effects of remifentanil on ICP and CPP in spontaneously breathing patients presenting with an intracranial mass effect. We therefore designed this prospective, open-label, randomized, and controlled study to compare the ICP and CPP of spontaneously breathing nonintubated brain tumor patients sedated with either remifentanil or propofol. We hypothesized that the use of remifentanil would result in a higher ICP than propofol because of its potential hypercapnia-induced increase of cerebral blood volume, but that CPP should be preserved in the two study groups, described later.

METHODS After institutional ethical and scientific review board approval and written informed consent, 40 consecutive adult patients undergoing stereotactic brain tumor biopsy were enrolled in this study. They Vol. 109, No. 1, July 2009

Table 1. Modified Observer’s Assessment of Alertness/ Sedation Scale (MOAAS) (Responsiveness Part)6 Responsiveness

Score level

Responds readily to name spoken in normal tone Lethargic response to name spoken in normal tone Responds only after name is called loudly and/or repeatedly Responds only after mild prodding or shaking Does not respond to mild prodding or shaking

5 (Alert) 4 3 2 1

were randomized using a computerized list to receive either remifentanil or propofol during the surgical procedure. Individual randomization assignments were kept in sealed envelopes that were opened in the operating room before the patient installation. Exclusion criteria were the presence of a ventricular drain, body mass index ⬎35 kg/m2, proven or suspected allergy to local anesthetics, propofol, or fentanyl derivatives. Tracheally intubated patients and those who had a prior craniotomy were also excluded. After the initiation of standard anesthesia monitors (electrocardiogram, noninvasive arterial blood pressure obtained from a blood pressure cuff positioned on the patient’s arm, pulse oximetry, and end-expired CO2 via a nasal oxygen canula for the assessment of respiratory rate), a scalp nerve block was performed on patients of both groups with 20 mL of a mixture of lidocaine 2% and bupivacaine 0.5% in equal amounts using a technique described by our group.5 Sedation with propofol, up to 1 mg/kg total and fentanyl (0.5–1.0 ␮g/kg), was allowed if needed for both groups during the performance of the block so that they could remain awake but calm. Patients received oxygen 2 L/min through a nasal canula during the study period. The stereotactic frame was installed and the patient brought to the radiology suite for a computed tomography scan, as part of the surgical procedure, without receiving additional sedation. In the operating room, the sedation regimen was initiated after obtaining the baseline mean arterial blood pressure (MAP). Propofol Group: the infusion was started at a rate of 3 mg 䡠 kg⫺1 䡠 h⫺1 after a 0.5 mg/kg bolus. The infusion was then adjusted by increments or decrements of 1.5 mg 䡠 kg⫺1 䡠 h⫺1 every 2 min to obtain and maintain a level of sedation of 4 on the modified observer’s assessment of alertness/sedation (MOAAS) scale (Table 1).6 At this level of sedation, the patients would have a lethargic response calling his or her name, spoken in normal tone. Remifentanil Group: the remifentanil infusion was started at a rate of 1.2 ␮g 䡠 kg⫺1 䡠 h⫺1 and then adjusted by increments or decrements of 0.6 ␮g 䡠 kg⫺1 䡠 h⫺1 to obtain and maintain a level of sedation of 4 on the MOAAS scale. Because stereotactic brain biopsies are performed in various degrees of head-up position, we measured Vol. 109, No. 1, July 2009

and noted the acute angle obtained between the horizontal plane and the back piece of the operating room table. The biopsy needle was flushed with saline and connected to a pressure transducer calibrated to atmospheric pressure at the center of the side-circle on the Cosman-Robert-Wells frame, which usually represents the target itself. We recorded the ICP obtained after a 1-min stabilization period. An acceptable ICP tracing was defined as a pulsatile wave form in synchrony, both with the arterial pulsation on the pulse oximeter and the respiratory rate of the patient. Concomitantly, the MAP, pulse oximetry (Spo2), and heart rate were noted. The MAP reading was obtained directly from the oscillotonometer display on the anesthesia monitor (Datex-Engstrom AS/3 noninvasive pressure module, Helsinki, Finland). Immediately after the ICP measurement, an arterial blood sample was obtained from a single radial artery puncture. The sample was immediately put in an ice bucket and blood gas analysis was performed at the central laboratory facility of the hospital within a few minutes after the sampling. This ended the study period. Patient’s age, gender, and the occurrence of signs and symptoms of increased ICP were recorded: alteration of the level of consciousness, headache, nausea and vomiting, and signs of papilloedema in the preoperative neurologic examination done by the attending neurosurgeon. The diameter of the lesion (or the largest one in cases of multiple lesions), the presence of perilesional edema or hydrocephalus, and the presence and degree of mid-line shift were obtained from the intraoperative computed tomography scan. Data were stored in an Excel™ database. Assuming, from our previous study,7 an average CPP of 89 mm Hg and a standard deviation (sd) of 18 mm Hg, two groups of 20 patients would be necessary to detect a difference of 15 mm Hg in CPP (17%) between the two groups, with a power of 82% and an ␣ of 0.05 for a one-tailed t-test.8 Statistical analyses were performed with the GraphPad Instat statistical software (version 3.06, 2003). Demographical and ordinal data were compared using Fisher’s exact test. Continuous data were compared using Student’s t-test or Wilcoxon’s sum-rank test and Mann–Whitney test in case of nonnormal distribution. Results are expressed as means ⫾ sd. All comparisons were two tailed and a P ⬍ 0.05 was considered significant.

RESULTS Twenty patients were recruited in each group and all the randomized patients completed the study. Demographics and preoperative data were similar in both groups (Table 2). Baseline MAP was 101.7 ⫾ 11 and 101.1 mm Hg ⫾ 15 in the remifentanil and propofol group, respectively. Intraoperatively (Table 3), patients in the remifentanil group had a slower respiratory rate (11 per min ⫾ 3 vs 15 per min ⫾ 3, P ⫽ © 2009 International Anesthesia Research Society

195

Table 2. Demographical and Preoperative Data Group remifentanil (n ⫽ 20) Demographics Age (yr) Gender (female:male) Weight (kg) ASA patient status (median, range) Preoperative data Glasgow coma scale median (range) Decreased level of consciousness (n) Headache (n) Nausea (n) Papilloedema (n) Diameter of the tumor (cm) Peritumoral edema (n) Hydrocephalus (n) Midline shift (n) Midline shift (mm) Site of the lesion Frontal Parasagittal Parietal Temporal Thalamic Multifocal Propofol sedation for scalp nerves block Fentanyl sedation for scalp nerves block Time span between scalp nerves block and intracranial pressure measurement

Group propofol (n ⫽ 20)

53 ⫾ 12 6:14 68 ⫾ 13 2 (2–3)

57 ⫾ 11 11:9 67 ⫾ 8 2 (2–3)

15 (14–15) 3 9 2 1 4.7 ⫾ 1.8 16 0 16 5.3 ⫾ 4.9

15 (14–15) 8 8 2 1 4.1 ⫾ 1.6 15 0 10 2.7 ⫾ 3.9

9 0 5 4 1 1 40.1 ⫾ 19 mg 59.8 ⫾ 18 ␮g 96 ⫾ 19 min

9 1 4 5 1 0 44 ⫾ 23 mg 52 ⫾ 21 ␮g 105 ⫾ 18

Values are mean ⫾ standard deviation. P values nonsignificant throughout. ASA ⫽ American Society of Anesthesiologists.

Table 3. Intraoperative Data Obtained During Intracranial Pressure Measurementa Group remifentanil (n ⫽ 20) Rate of infusion at the targeted level of sedation Total dosage at the end of ICP measurement MOAAS score median (range) Heart rate (bpm) Pulse oximetry (%) Respiratory rate (breath/min) Arterial blood gas pH Pco2 (mm Hg) Po2 (mm Hg) HCO3 (mEq 䡠 L⫺1) Angle from the operating table (degrees) Intracranial pressure (mm Hg) Mean arterial pressure (mm Hg) Cerebral perfusion pressure (mm Hg)

⫺1

4.2 ⫾ 1.8 ␮g 䡠 kg 224 ⫾ 114 ␮g 4 (4–4) 75 ⫾ 13 98 ⫾ 1 11 ⫾ 3

⫺1

䡠h

7.34 ⫾ 0.04 48.3 ⫾ 6.2 127 ⫾ 34 26 ⫾ 2 29 ⫾ 15 19.0 ⫾ 11.9 101.1 ⫾ 13.7 82.0 ⫾ 19.0

Group propofol (n ⫽ 20) ⫺1

4.3 ⫾ 2.5 mg 䡠 kg 208 ⫾ 78 mg 14 (3–4) 77 ⫾ 12 98 ⫾ 2 15 ⫾ 3

7.38 ⫾ 0.04 43.1 ⫾ 5.5 147 ⫾ 39 25 ⫾ 2 26 ⫾ 12 16.4 ⫾ 11.1 85.8 ⫾ 12.7 69.5 ⫾ 17.0

P ⫺1

䡠h

ns 0.0001 0.007 0.009 0.008 0.48 0.0008 0.03

Values are mean ⫾ standard deviation. MOAAS ⫽ modified observer’s assessment of alertness/sedation; ICP ⫽ intracranial pressure; ns ⫽ nonsignificant. a All values were obtained at the end of the 1 min stabilization period for intracranial pressure measurement.

0.0001) and a higher arterial Pco2 (48.3 ⫾ 6.2 mm Hg vs 43.1 ⫾ 5.5 mm Hg, P ⫽ 0.009), resulting in a lower arterial pH (7.34 ⫾ 0.04 vs 7.38 ⫾ 0.04, P ⫽ 0.007) than patients in the propofol group. The mean ICP of both groups was similar (Table 3), but higher MAP in the remifentanil group (101.1 ⫾ 13.7 mm Hg vs 85.8 ⫾ 12.7 mm Hg, P ⫽ 0.0008) resulted in a higher CPP than the propofol group: 82.0 ⫾ 19.0 mm Hg vs 196

Remifentanil Sedation in Neurosurgery

69.5 ⫾ 17.0 mm Hg (P ⫽ 0.03). No complication was encountered during this study.

DISCUSSION This study indicates that, contrary to our preliminary hypothesis, light sedation with remifentanil does not result in a higher ICP than propofol in patients ANESTHESIA & ANALGESIA

with a brain tumor undergoing stereotactic brain biopsy in a 25°–30° head-up position. The CPP was better preserved in the remifentanil group (82.0 mm Hg) because of a higher MAP, although it remained well above the expected lower threshold for autoregulation in the propofol group (69.5 mm Hg P ⬍ 0.05). Because of the respiratory-depressant effect of remifentanil, patients in this group experienced a small but significant increase in Pco2. This difference did not seem to have an impact on ICP, despite the size of the space-occupying lesions (pooled average diameter of 4.2 cm) and the signs or symptoms of intracranial hypertension presented by most patients (82.5%). In the literature, we could not find a study comparing the effect of sedation with remifentanil versus propofol as the sole drugs on intracranial dynamics. The independent use of both drugs has, however, been studied. One study showed that the initiation of sedation with remifentanil with no background sedation regimen increased the mean ICP from 17 to 19 –22 mm Hg (P ⬍ 0.05) and decreased the mean CPP from 74 to 59 – 63 mm Hg (P ⬍ 0.05) in mechanically ventilated patients with severe traumatic brain injury.2 The increase in ICP was attributed to autoregulatory vasodilation. In this report, remifentanil was given in much larger doses than our study to prevent a response to tracheal suctioning; depending on the study group, the authors administered a bolus of 1–3 ␮g/kg followed by an infusion of 15– 60 ␮g 䡠 kg⫺1 䡠 h⫺1. In contrast, our patients received an average of 4.2 ⫾ 1.8 ␮g 䡠 kg⫺1 䡠 h⫺1 and no bolus. This difference in dosage could explain why we did not encounter such a finding in our study. In addition, for obvious ethical reasons, we could not repeat ICP measurements before and after the initiation of sedation in our group of patients. In a previous study, we assessed the effect of moderately deep propofol sedation (average MOAAS score of 2) when compared with no sedation on the ICP of patients undergoing stereotactic brain tumor biopsy under regional anesthesia.7 At this level of sedation, patients in the propofol group experienced a higher arterial Pco2 than patients receiving no sedation: 48 ⫾ 8 mm Hg vs 41 ⫾ 3 mm Hg; P ⬍ 0.005. ICP was similar in both groups: 13 mm Hg (8.2–16.2 mm Hg) in the propofol group vs 15 mm Hg (8.3–21.7 mm Hg) in the no sedation group (median and 95% confidence interval, P ⫽ 0.66). CPP was lower in the propofol group: 76 ⫾ 18 mm Hg vs 89 ⫾ 18 mm Hg, P ⫽ 0.003. In this study, propofol was administered at a higher dosage (10.6 mg 䡠 kg⫺1 䡠 h⫺1), which explains the increase in Pco2. Despite this difference, CPP was in the same range in both studies: 76 mm Hg when compared with 69.5 mm Hg in this study. When added to propofol or midazolam sedation, no significant difference among remifentanil, fentanyl, and morphine could be found in mean ICP (12.0 vs Vol. 109, No. 1, July 2009

13.9 vs 10.3 mm Hg) or CPP (68.8 vs 75.6 vs 77.0 mm Hg) in mechanically ventilated ICU patients with acute brain injury or who had undergone neurosurgery.4 Remifentanil was associated with a more rapid and predictable emergence from sedation than fentanyl or morphine. Engelhard et al.3 obtained no significant effect on ICP, CPP, or cerebral blood flow velocity when remifentanil was administered as a bolus followed by an infusion in mechanically ventilated ICU patients with traumatic brain injury already deeply sedated with a propofol-sufentanil infusion. In these studies, the effect of remifentanil was evaluated in the presence of deep sedation with other drugs (propofol, sufentanil, and morphine). Therefore, meaningful conclusions on the pure effects of remifentanil on cerebral dynamics cannot be made. In this study, patients in the propofol group experienced a 14.3 mm Hg (15%) reduction in MAP (from 101.1 to 85.8 mm Hg), whereas it remained at baseline in the remifentanil group for the same level of sedation. This decrease in MAP with propofol is similar to the 13% reduction reported by Roekaerts et al.9 in the first 30 min after sedation was initiated after coronary artery bypass surgery. Similarly, Chamorro et al.10 reported a 13% decrease of MAP (from 93 to 81 mm Hg) after initiation of sedation with propofol at an average dosage of 2.8 ⫾ 1 mg 䡠 kg⫺1 䡠 h⫺1 in ICU patients, some of them with a neurologic condition. Low-dose sedation with remifentanil has been associated with a stable hemodynamic profile in ICU patients.1 Breen et al.11 reported minimal to no change of MAP with rates of sedation ranging from 7 to 10.1 ␮g 䡠 kg⫺1 䡠 h⫺1 in 40 ICU patients. Our study confirms this finding because patients in the remifentanil group experienced no reduction of MAP with an average dosage of 4.2 ␮g 䡠 kg⫺1 䡠 h⫺1. This study has some limitations. We measured intralesional (parenchymal) pressures using a fluidfilled catheter. Intraparenchymal brain pressure is usually measured with the use of a fiberoptic or a microtransducer. The ICP tracings we obtained were in all points identical to a standard ICP tracing with the classic respiratory and cardiac variations. In addition, intraparenchymal pressure is routinely measured with a fluid-filled catheter to detect abdominal12 and lower extremity compartment syndrome.13 In this study, a direct comparison between intraparenchymal brain pressure measurement and intraventricular pressure could not be performed for obvious reasons. There are also reasons to believe that intralesional pressures may not be a reflection of the pressure in the rest of the cranial vault. A study of expanding left frontal epidural mass in pigs has shown that there is a pressure gradient of brain parenchymal pressure.14 The larger gradients were reached at the maximal expansion volume of 6 mL, for a resulting ICP of 38 mm Hg, between the left frontal lobe and the cerebellum and midbrain (16 and 26 mm Hg, respectively). For the supratentorial space, a gradient was obtained © 2009 International Anesthesia Research Society

197

between the left frontal and right temporal lobe (8 mm Hg). Such gradients have also been demonstrated in a primate model of stroke.15 However, this limitation also applies to other methods of measuring ICP, and it cannot be easily overcome without measuring ICP at several different locations. We decided to initiate remifentanil sedation without a bolus, knowing that this would introduce an additional limitation to our study because propofol sedation was initiated with a bolus. We made this decision for two main reasons. First, a given level of sedation can easily and rapidly be attained with remifentanil without the use of a bolus, which is not the case with propofol. Second and more importantly, we did it for a safety issue. We are not aware of another study that used remifentanil sedation in awake nonintubated patients with a large intracranial expanding mass, whereas the experience with propofol is extensive in our institution.7 Because the timespan between scalp nerve block and ICP measurement was more than 90 min in both groups, we do not feel that the administration of small doses of fentanyl and propofol during the performance of the block could be a confounder in this study. Finally, for obvious ethical concerns, we could not perform a crossover evaluation in this study or measure ICP at several timepoints during the procedure, which would have been the best way to compare the two sedation regimens.

CONCLUSION In conclusion, this study assessed the effect of remifentanil versus propofol sedation on the ICP of spontaneously breathing, nonintubated patients with a large expanding intracranial lesion. It shows that light sedation with remifentanil does not result in a higher ICP than the use of propofol and that CPP might be better preserved with remifentanil. REFERENCES 1. Battershill AJ, Keating GM. Remifentanil. A review of its analgesic and sedative use in the intensive care unit. Drugs 2006;66:365– 85

198

Remifentanil Sedation in Neurosurgery

2. Leone M, Albane`se J, Viviand X, Garnier F, Bourgoin A, Barrau K, Martin C. The effects of remifentanil on endotracheal suctioning-induced increases in intracranial pressure in headinjured patients. Anesth Analg 2004;99:1193– 8 3. Englelhard K, Reeker W, Kochs E, Werner C. Effect of remifentanil on intracranial pressure and cerebral blood flow velocity in patients with head trauma. Acta Anaesthesiol Scand 2004;48: 396 –9 4. Karabinis A, Mandragos K, Stergiopoulos S, Komnos A, Soukup J, Speelberg B, Kirkham AJ. Safety and efficacy of analgesiabased sedation with remifentanil versus standard hypnoticbased regimens in intensive care unit patients with brain injuries: a randomized, controlled trial. Crit Care 2004;8: R268 –R80 5. Nguyen A, Girard F, Boudreault D, Fuge`re F, Ruel M, Moumdjian R, Bouthilier A, Caron JL, Bojanowski MW, Girard DC. Scalp nerve blocks decreases the severity of pain following craniotomy. Anesth Analg 2001;93:1272– 6 6. Chernik DA, Gillings D, Laine H, Hendler J, Silver JM, Davidson AB, Schwam EM, Siegel JL. Validity and reliability of the Observer’s Assessment of Alertness/Sedation Scale: study with intravenous midazolam. J Clin Psychopharmacol 1990;10:244 –51 7. Girard F, Moumdjian R, Boudreault D, Chouinard P, Bouthilier A, Sauvageau E, Ruel M, Girard DC. The effect of propofol sedation on intracranial pressure of patients with an intracranial space-occupying lesion. Anesth Analg 2004;99:573–77 8. Lenth RV. Java applets for power and sample size [Computer software]. Available at: http://www.stat.uiowa.edu/⬃rlenth/ Power/index.html. 2006 9. Roekaerts PM, Huygen FJ, de Lange S. Infusion of propofol versus midazolam for sedation in the intensive care unit following coronary artery surgery. J Cardiothorac Vasc Anesth 1993;7: 142–7 10. Chamorro C, De Latorre F, Montero A, Sanchez-Izquierdo J, Jareno A, Moreno J, Gonzalez E, Barrios M, Carpintero J, Martin-Santos F, Otero B, Ginestal R. Comparative study of propofol versus midazolam in the sedation of critically ill patients: results from a prospective, randomized, multicenter trial. Crit Care Med 1996;24:932–9 11. Breen D, Wilmer A, Bodenham A, Bach V, Bonde J, Kessler P, Albrecht S, Shaikh S. Offset of pharmacodynamic effects and safety of remifentanil in intensive care unit patients with various degrees of renal impairment. Crit Care 2004;8:R21–R30 12. Ertel W, Oberhlozer A, Patz A, Stocker R, Trentz O. Incidence and clinical pattern of the abdominal compartment syndrome after “damage-control” laparotomy in 311 patients with severe abdominal and/or pelvic trauma. Crit Care Med 2000;28: 1747–53 13. Giannotti G, Cohn SM, Brown M, Varela JE, McKenney MG, Wiseberg JA. Utility of near-infrared spectroscopy in the diagnosis of lower extremity compartment syndrome. J Trauma 2000;48:396 – 401 14. Wolfla CE, Luerssen TG, Bowman RM, Putty TK. Brain tissue pressure gradients created by expanding frontal epidural mass lesion. J Neurosurg 1996;84:642–7 15. D’Ambrosio AL, Hoh DJ, Mack WJ, Winfree CJ, Nair MN, Ducruet A, Sciacca RR, Huang J, Pinsky DJ, Connolly ES. Interhemispheric intracranial pressure gradients in nonhuman primate stroke. Surg Neurol 2002;58:295–301


The Effects of Arterial Carbon Dioxide Partial Pressure and Sevoflurane on Capillary Venous Cerebral Blood Flow and Oxygen Saturation During Craniotomy Klaus Ulrich Klein, MD* Martin Glaser, MD† Robert Reisch, MD, PhD† Achim Tresch, MSc‡ Christian Werner, MD, PhD* Kristin Engelhard, MD, PhD*

BACKGROUND: Intraoperative routine monitoring of cerebral blood flow and oxygenation remains a technological challenge. Using the physiological principle of carbon dioxide reactivity of cerebral vasculature, we investigated a recently developed neuromonitoring device (oxygen-to-see, O2C™ device) for simultaneous measurements of regional cerebral blood flow (rvCBF), blood flow velocity (rvVelo), oxygen saturation (srvO2), and hemoglobin amount (rvHb) at the capillary venous level in patients subjected to craniotomy. METHODS: Twenty-six neurosurgical patients were randomly assigned to anesthesia with 1.4% or 2.0% sevoflurane end-tidal concentration. After craniotomy, a fiberoptic probe was applied on a macroscopically healthy surface of cerebral tissue next to the site of surgery. Simultaneous measurements in 2 and 8 mm cerebral depth were performed in each patient during lower (35 mm Hg) and higher (45 mm Hg) levels (random order) of arterial carbon dioxide partial pressure (Paco2). The principle of these measurements relies on the combination of laser-Doppler flowmetry (rvCBF, rvVelo) and photo-spectrometry (srvO2, rvHb). Linear models were fitted to test changes of end points (rvCBF, rvVelo, srvO2, rvHb) in response to lower and higher levels of Paco2, 1.4% and 2.0% sevoflurane end-tidal concentration, and 2 and 8 mm cerebral depth. RESULTS: RvCBF and rvVelo were elevated by Paco2 independent of sevoflurane concentration in 2 and 8 mm depth of cerebral tissue (P ⬍ 0.001). Higher Paco2 induced an increase in mean srvO2 from 50% to 68% (P ⬍ 0.001). RvVelo (P ⬍ 0.001) and srvO2 (P ⫽ 0.007) were higher in 8 compared with 2 mm cerebral depth. RvHb was not influenced by alterations in Paco2 but positively correlated to sevoflurane concentration (P ⫽ 0.005). CONCLUSIONS: Increases in rvCBF and rvVelo by Paco2 suggest preserved hypercapnic vasodilation under anesthesia with sevoflurane 1.4% and 2.0% end-tidal concentration. A consecutive increase in srvO2 implies that cerebral arteriovenous difference in oxygen was decreased by elevated Paco2. Unchanged levels of rvHb signify that there was no blood loss during measurements. Data suggest that the device allows detection of local changes in blood flow and oxygen saturation in response to different Paco2 levels in predominant venous cerebral microvessels. (Anesth Analg 2009;109:199 –204)

M

ultiple intraoperative technologies have been investigated for the quantitative (xenon-133 clearance, thermal diffusion, double-indicator dilution technique) and qualitative (angiography, Doppler ultrasonography, laser-Doppler flowmetry) determination of cerebral blood flow (CBF). Global (jugular bulb oxymetry, nearinfrared spectroscopy), regional oxygen saturation From the Departments of *Anesthesiology, †Neurosurgery, Johannes Gutenberg-University, Mainz, Germany; and ‡Department of Chemistry and Biochemistry, Gene Center Munich, LudwigMaximilians-University, Munich, Germany. Accepted for publication January 19, 2009. Supported by departmental funding. There is no financial relationship between the authors and LEA Medizintechnik GmbH (Giessen, Germany) or any other company or organization with potential or vested interest in the outcome of the study. Parts of this work have been presented at the Annual Meeting of the American Society of Anesthesiologists, October 14, 2006, and the Society of Neurosurgical Anesthesia and Critical Care, October 13, 2006, Chicago, IL; and at the Annual Meeting of the American Vol. 109, No. 1, July 2009

(photo-spectrometry), and oxygen tension (polarographic electrodes) have also been examined.1 Electrophysiological techniques focus on the determination of brain metabolism and function (evoked potentials, electroencephalography, and bispetral index). None of these techniques, however, provide simultaneous assessment of CBF and tissue oxygenation for routine use during neurosurgery.

Society of Anesthesiologists, October 15, 2007, and the Society of Neurosurgical Anesthesia and Critical Care, October 12, 2007, San Francisco, CA. Address correspondence and reprint requests to Klaus Ulrich Klein, MD, Department of Anesthesiology, Johannes GutenbergUniversity, Langenbeckstrasse 1, 55131 Mainz, Germany. Address e-mail to [email protected]. Copyright © 2009 International Anesthesia Research Society DOI: 10.1213/ane.0b013e3181a800e5

199

This study introduces a recently developed monitoring device (oxygen-to-see, O2C™, LEA Medizintechnik, Giessen, Germany) for simultaneous measurements of regional cerebral blood flow (rvCBF), blood flow velocity (rvVelo), oxygen saturation (srvO2), and hemoglobin levels (rvHb) at the venous end of capillaries during craniotomies. The O2C device has been used in various human tissues (e.g., peripheral arterial occlusive disease,2 gastrointestinal tract,3–9 transplantation surgery10,11) except the brain. O2C parameters have been validated in the pig brain using cerebral microspheres and venous oxygen saturation.12 In this study, the physiological effect of alterations in arterial carbon dioxide partial pressure (Paco2) on cerebral vasculature was used to investigate the O2C device. Arteriolar CO2 vasoreactivity was indirectly assessed by measuring relative changes of O2C parameters at the venous end of cerebral capillaries. A fiberoptic measurement probe was applied on different areas of the human brain next to the site of surgery. It was hypothesized that increases in Paco2 would lead to regional enhancement of rvCBF, rvVelo, and srvO2 measured with the O2C device.

METHODS Preparation After approval by the Research Ethics Committee of the state of Rheinland-Palatinate (approval number 837.136.05 [4794]) and written patient informed consent, 26 ASA II–III patients scheduled for elective intracranial surgery were included in the study. Anesthesia was induced with 0.3– 0.4 ␮g 䡠 kg⫺1 䡠 min⫺1 remifentanil, 2 mg/kg propofol, and 0.5 mg/kg atracurium. Patients were tracheally intubated and mechanically ventilated pressure controlled (Fio2 ⫽ 0.5, positive end-expiratory pressure ⫽ 5 mbar). Anesthesia was continued by infusion of 0.1– 0.4 ␮g 䡠 kg⫺1 䡠 min⫺1 remifentanil, and 1.4% (n ⫽ 13) or 2.0% sevoflurane (n ⫽ 13) end-tidal concentration. All patients received a radial arterial line for continuous measurement of arterial blood pressure and intermittent blood-gas analysis.

O2C Device and Probe Description The O2C device combining laser-Doppler flowmetry (rvCBF, rvVelo) and photo-spectrometry (rvHb, srvO2) was used for cerebral measurements. The device consists of a computer with built-in laser and light-emitting diodes and maximal two fiberoptic probes (laser class 3b, protective class 1, CE-mark). The probe contains glass fibers enclosed with epoxy casting resin and is nontoxic to human tissue. An increased distance between illuminating and detecting glass fiber allows for increased measurement depth because light is backscattered to the surface in a semicircular path through tissue. Different flat probes, vertical probes, and microprobes are available providing measurements between 0.2 and 18 mm depth. In this study, one single flat probe (type LF-1, probe 200

Paco2 Effects on Cerebral Microcirculation

head: width 12 mm, height 5.5 mm, length 44.5 mm; probe length 300 cm) was applied on the cerebral cortex next to the site of surgery. This special probe has two channels and provides simultaneous measurements in 2 and 8 mm depth. In contrast to other near-infrared spectrophotometers (INVOS威, NIRO200威), cerebral measurements with the O2C device cannot be performed transcranially but must be conducted from the cortex directly.

O2C Device Measurement Principle The measurement principle of the O2C device relies on the transmission of near infrared and visible light to tissue. Continuous wave laser light and white light is scattered and collected on the tissue surface into the fiberoptic probe type LF-1 (spatial resolution 2 and 8 mm depth), where it then splits into its spectral components by charge-coupled device array (temporal resolution 500 ms). A Doppler shift of the illuminated laser light (830 nm, ⬍30 mW) caused by movements of erythrocytes is detected, analyzed, and displayed as rvVelo. The product of the amount of moving erythrocytes times the velocity of the erythrocytes is used for the calculation of rvCBF. Approximately 300 wavelengths of white light are detected simultaneously (450 – 850 nm, ⬍30 mW). The spectrum is compared with reference values of deoxygenated and oxygenated hemoglobin spectra for the determination of srvO2 and rvHb.13,14

Intraoperative Measurements Before measurements, each patient received 2 g cefuroxime as IV antibiotic. The white light source of the measurement probe was calibrated and disinfected with alcohol solution. The measurement probe was covered with sterile polyurethane protective for ultrasound transducers (Ultracover威 87110, Microtek Medical, Zutphen, The Netherlands) and flushed with warm saline solution. Patients were randomly assigned to the first measurement at lower (target Paco2 ⫽ 35) or higher (target Paco2 ⫽ 45) Paco2 levels. The probe was applied carefully without pressure to macroscopically healthy cerebral tissue next to the site of surgery. During measurements, the light of the operating microscope was switched off and a swab placed over the craniotomy to minimize surrounding light effects. Three repetitive measurements, each lasting 1 min, were performed resulting in a total of 50 –70 measurements. Thereafter, levels of initially lower Paco2 were altered to higher Paco2 or vice versa by changes in respiratory frequency. After 30 min of steady-state anesthesia, the probe was placed on the same area of cerebral tissue and measurements were repeated.

Physiological Variables Physiological variables including mean arterial blood pressure, heart rate, bladder temperature, inspired oxygen fraction, hemoglobin concentration, hematocrite level, and peripheral oxygen saturation ANESTHESIA & ANALGESIA

(SaO2) were controlled and maintained at stable levels over time. Oxygen content of blood (CaO2) was calculated using the equation: CaO2 ⫽ 1.39 ⫻ Hb ⫻ SaO2 ⫹ Pao2 ⫻ 0.003. When appropriate mean arterial blood pressure was stabilized by repetitive administration of combined IV cafedrine-Hcl 50 mg and theodrenaline-Hcl 2.5 mg (0.25 mL Akrinor威). Akrinor is a sympathomimetic agent used to counter transitory hypotension and approved in Germany, Austria, South Africa, and Eastern Europe.

Statistical Analysis Linear models were fitted to test changes of rvCBF, rvVelo, srvO2, and rvHb in response to lower (35 mm Hg) and higher (45 mm Hg) target Paco2, 1.4 and 2.0% sevoflurane end-tidal concentration, and 2 and 8 mm cerebral depth; P values of ⬍0.05 were considered significant. Data are presented as mean ⫾ sd. All analyses were performed using SPSS 13.0 (SPSS, Chicago, IL). All analyses were performed using the statistical Software R (Version 2.8.1, Lucent Technologies, Inc., Murray Hill, NY). Figures were plotted using Graph Pad Prism (Version 5.0 b, Graph Pad Software Inc., La Jolla, CA).

RESULTS Thirty-four patients were recruited and measurements were completed in 26 patients. In three patients, measurements could not be performed because of bleeding. In five patients, the probe could not be applied on the cerebral cortex because the craniotomy was too small. No tissue damage, infection, or postoperative cerebral dysfunction was attributable to the positioning of the probe. The study groups were comparable with respect to demographic characteristics and surgical covariates (Table 1). Mean changes in Paco2 levels were 9.1 ⫾ 2.8 mm Hg for 1.4% sevoflurane and 8.9 ⫾ 2.4 mm Hg for 2.0% sevoflurane end-tidal concentration. Changes in other physiological variables remained within range of 10% (Table 2). Measurements of rvCBF, rvVelo, srvO2, and rvHb are shown in Figure 1. P values of linear regression models are listed in Table 3. Higher levels of Paco2 increased rvCBF, rvVelo, and srvO2 independent of 1.4% and 2.0% sevoflurane end-tidal concentration (P ⬍ 0.001). RvVelo (P ⬍ 0.001) and srvO2 (P ⫽ 0.007) were higher in 8 mm compared with 2 mm cerebral depth. Levels of rvHb were positively correlated to end-tidal sevoflurane concentration (P ⫽ 0.005) but not dependent on Paco2 or cerebral depth.

DISCUSSION This study investigated the effects of altered Paco2 on rvCBF, rvVelo, srvO2, and rvHb assessed with the O2C device during craniotomies. Higher levels of Paco2 induced an increase in rvCBF, rvVelo, and srvO2 independent of sevoflurane concentration. Increased srvO2 represents an enhancement in oxygen availability due to increased cerebral perfusion. Levels Vol. 109, No. 1, July 2009

Table 1. Patient Demographic Characteristics, Intracranial Pathology, Cardiovascular Risk Factors, and Region of Cerebral Measurements; Groups were Comparable with Respect to Demographic Characteristics and Surgical Covariates 1.4% 2.0% Sevoflurane Sevoflurane (n ⫽ 13 (n ⫽ 13 patients) patients) Age (yr) 54 (44–68) 55 (37–71) Height (cm) 169 (165–180) 169 (161–173) Weight (kg) 70 (60–81) 70 (64–85) Body mass index (kg/m2) 23 (21–27) 27 (25–28) ASA status (0–5) 2 (2–3) 2 (2–3) Gender (M/F) 5/8 4/9 Glioblastoma (n/13) 6/13 6/13 Meningeoma (n/13) 5/13 5/13 Temporal lobe epilepsy 1/13 2/13 (n/13) Cerebral metastasis (n/13) 1/13 0/13 Arterial hypertension (n/13) 2/13 6/13 Diabetes mellitus (n/13) 2/13 1/13 Smoker (n/13) 3/13 4/13 Frontal cortex measurement 3/13 1/13 (n/13) Temporal/parietal cortex 7/13 10/13 measurement (n/13) Occipital cortex measurement 3/13 2/13 (n/13) Data are presented as median (25–75 h percentile or interquartile range) or incidence of observations.

of rvHb were unchanged, suggesting that there was no blood loss during measurements. CBF correlates positively with Paco2 at the range of 20 – 80 mm Hg and increases or decreases 2%– 4% per mm Hg change in Paco2.15 The major mechanism influencing hyper- and hypocapnic CBF response is related to modulation in extracelluar fluid pH affecting vessel resistance. All cerebral vessels respond to altered Paco2, but hypercapnia dilates smaller arterioles more than larger ones. Dilation of resistance arterioles increases the microvascular pressure gradient and thereby also increases rvVelo.16,17 This effect of altered Paco2 on venous capillaries was investigated in this study. The results show a mean of 50% increase in rvCBF and a mean of 45% increase in rvVelo when Paco2 was altered by 9 mm Hg (Fig. 1). The findings are in line with intravital microscopic observations and laser Doppler-flowmetry, where Paco2 increased capillary CBF velocity.16,18 –21 In this study, CO2 reactivity was absent in two patients anesthetized with 1.4% sevoflurane, and in three patients anesthetized with 2.0% sevoflurane end-tidal concentration. Absence of CO2 response may be physiological in some individuals but can also reflect a pathophysiological state caused by structural damage to cerebral tissue through surgery or local edema.22–25 Another explanation may be alterations in capillary perfusion because of shunting, recruitment, or increase in capillary diameter. Invalid measurements of the O2C device cannot be excluded because no “gold standard” CBF measurement was used to verify the © 2009 International Anesthesia Research Society

201

Table 2. Physiological Variables During Anesthesia with 1.4% and 2.0% Sevoflurane End-Tidal Concentration at Target PaCO2 of 35 vs (/) 45 mm Hg Target Paco2 35/45 mm Hg MAP (mm Hg) Sevoflurane, 1.4% (n ⫽ 13) Sevoflurane, 2.0% (n ⫽ 13)

83 ⫾ 9/ 81 ⫾ 8 78 ⫾ 10/ 75 ⫾ 7

HR (min⫺1)

Temp (°C)

Fio2 (%)

Pao2 (mm Hg)

Paco2 (mm Hg)

Hb (mg/dL)

Hct (%)

CaO2 (mL/dL)

SaO2 (%)

53 ⫾ 9/ 55 ⫾ 11 58 ⫾ 7/ 60 ⫾ 8

36.1 ⫾ 0.6/ 36.2 ⫾ 0.5 35.9 ⫾ 0.6/ 36.1 ⫾ 0.6

50 ⫾ 2.7/ 50 ⫾ 1.5 50 ⫾ 1.5/ 49 ⫾ 1.5

271 ⫾ 74/ 252 ⫾ 32 198 ⫾ 47/ 208 ⫾ 41

36.2 ⫾ 1.2/ 45.4 ⫾ 2.8 36.0 ⫾ 2/ 45.0 ⫾ 3

11.7 ⫾ 1.1/ 11.6 ⫾ 1.1 11.9 ⫾ 1.1/ 11.5 ⫾ 1.1

35.0 ⫾ 3.3/ 34.7 ⫾ 3.5 36.0 ⫾ 3.3/ 34.7 ⫾ 3.5

16.9 ⫾ 1.6/ 16.8 ⫾ 1.6 17.0 ⫾ 1.4/ 16.6 ⫾ 1.5

99.7 ⫾ 0.6/ 99.8 ⫾ 0.6 99.6 ⫾ 0.9/ 99.9 ⫾ 0.2

Relative changes in physiological variables remain within the range of 10%. Data are presented as mean ⫾ standard deviation. MAP ⫽ mean arterial blood pressure; HR ⫽ heart rate; Temp ⫽ bladder temperature; FIO2 ⫽ fraction of inspired oxygen; Pao2 ⫽ partial pressure of oxygen; PaCO2 ⫽ partial pressure of carbon dioxide; Hb ⫽ hemoglobin concentration; Hct ⫽ hematocrit level; CaO2 ⫽ arterial oxygen content, SaO2 ⫽ oxygen saturation of hemoglobin.

Sevoflurane 1.4%, e.t.

Sevoflurane 2.0%, e.t.

Figure 1. O2C device measured parameters during anesthesia with 1.4 and 2.0% sevoflurane end-tidal concentration in 2 and 8 mm cerebral depth at lower (35 mm Hg) and higher (45 mm Hg) levels of arterial carbon dioxide partial pressure (Paco2). Data are presented as means ⫾ sd. Higher levels of Paco2 increased regional capillary venous blood flow (rvCBF), blood flow velocity (rvVelo), and oxygen saturation (srvO2) independent of sevoflurane concentration. RvVelo and srvO2 were higher in 8 mm compared with 2 mm cerebral depth. Levels of rvHb were positively correlated to end-tidal sevoflurane concentration but not dependent on Paco2 or cerebral depth. rvCBF ⫽ regional capillary venous cerebral blood flow; rvVelo ⫽ regional capillary venous cerebral blood flow velocity; srvO2 ⫽ regional capillary venous cerebral oxygen saturation; rvHb ⫽ regional capillary venous cerebral hemoglobin amount. O2C device readings. Sevoflurane concentration had no effect on rvCBF, rvVelo, and srvO2, suggesting that CO2 reactivity was preserved even with a higher sevoflurane concentration.26 In this study, srvO2 values were between 46% and 54% at a mean Paco2 of 36.0 ⫾ 2.0 mm Hg. These findings conform to previous studies investigating human cerebral oxygen saturation.27 Intraoperative photo-spectrometry revealed a mean cortical oxygen saturation of about 50%.28,29 Measurements using polarographic oxygen electrodes reveal a ptio2 of 20 –30 mm Hg for subcortical areas and ⬎30 mm Hg for cortical areas.30 In contrast, our findings show a higher srvO2 in deeper (8 mm) compared with superficial (2 mm) cerebral depth (P ⫽ 0.007). This may be attributable to possible systematic failure of the study setup 202


(e.g., application pressure of probe, temperature, surgery, diffusion of CO2) or failure of the device. After a Paco2 mean increase of 9 mm Hg, however, srvO2 increased between 62% and 78% (Fig. 1). This increase in srvO2 may reflect the physiological decrease in arteriovenous oxygen difference (avDO2). In CO2 reactivity, vessel diameter primarily increases leading to consecutive increases in CBF and decrease in avDO2. Conversely, avDO2 essentially remains constant in both metabolic and pressure autoregulation.31 This study has several limitations. The main disadvantage of laser-Doppler flowmetry is its nonquantitative measurement. However, relative changes in flow values assessed with laser Doppler flowmetry compare with those obtained by quantitative techniques if ⬎20 repetitive measurements are performed.32 This may be ANESTHESIA & ANALGESIA

Table 3. Linear Regression Models were Used to Test the Influence of the Independent Variables Arterial Carbon Dioxide Partial Pressure (PaCO2) (35 vs 45 mm Hg), Cerebral Depth (2 vs 8 mm), and Anesthesia (1.4% vs 2.0% Sevoflurane End-Tidal Concentration) on O2C™ Device Parameters (Regional Capillary Venous Blood Flow 关rvCBF兴, Blood Flow Velocity 关rvVelo兴, Regional Capillary Venous Cerebral Oxygen Saturation 关srvO2兴 and Regional Capillary Venous Cerebral Hemoglobin Amount 关rvHb兴)a P

rvCBF (AU)

rvVelo (AU)

srvO2 (%)

rvHb (AU)

Paco2 Depth Anesthesia

⬍0.001* 0.258 0.628

⬍0.001* ⬍0.001* 0.487

⬍0.001* 0.007† 0.108

0.255 0.090 0.005†

* P ⬍ 0.001, † P ⬍ 0.05.

achieved in experimental settings but rarely in clinical scenarios. In this study, three repetitive measurements were performed to minimize laser-Doppler flowmetry associated artifacts and inaccuracies. Finding an avascular macroscopically healthy site for the probe placement is dependent on the size and location of craniotomy. In our experience, craniotomies larger than 3 cm in diameter allow for optimal placement of the LF-1 probe. Ambient light mainly emitted from the operation microscope may bias the spectrophotometric readings. To minimize ambient light effects, the probe was covered with a swab to eliminate any adverse effects of artificial light on measured hemoglobin spectra. In comparison with other techniques investigating cerebral microcirculation, the O2C device offers advantages and disadvantages. The jugular bulb oxygen catheter determines global venous oxygen saturation. The O2C device measures regionally and, therefore, can be used exclusively during craniotomies. PtiO2 (Licox威, Neurovent-PTO威) and microdialysis probes traumatize the cortex, although the risks of tissue damage, bleeding, and infections are minimal.30 PtiO2 readings are accurate and stable but measurements can be elevated if the probe is placed near an arterial vessel. Measurements using the O2C device are affected by motion artifacts. Another approach to determine cerebral oxygenation is transcranial near-infrared spectroscopy (INVOS威, NIRO-200威). Using this technique, reading of skin, subcutaneous tissue, and skull are included in the measurement of cerebral cortex. The devices usually work with only two wavelengths to determine oxygen saturation. In contrast, the O2C device determines all alterations in wavelengths within visible range. Potential error is reduced because scattering properties on tissue are considered in calculations. In conclusion, the O2C device provides real-time intraoperative measurements of cerebral microcirculation during craniotomies. Levels of rvCBF, rvVelo, and srvO2 in 2 and 8 mm cerebral depth increased with higher Paco2. These changes were independent of sevoflurane (1.4% vs 2.0%) concentration. Levels of rvHb remained unchanged, supporting that there was Vol. 109, No. 1, July 2009

no blood loss during measurements. Data suggest that the device allows detection of regional changes in blood flow, oxygen saturation, and hemoglobin amount in response to different Paco2 levels in predominant venous microvessels. According to these favorable properties of the O2C device, further studies are planned to characterize the validity of the probe to detect intraoperative cerebral ischemia and hypoxia. REFERENCES 1. Ndubuizu O, LaManna JC. Brain tissue oxygen concentration measurements. Antioxid Redox Signal 2007;8:1207–19 2. Beckert S, Witte MB, Konigsrainer A, Coerper S. The impact of the micro-lightguide O2C for the quantification of tissue ischemia in diabetic foot ulcers. Diabetes Care 2004;12:2863–7 3. Brell B, Temmesfeld-Wollbru¨ck B, Altzschner I, Frisch E, Schmeck B, Hocke AC, Suttorp N, Hippenstiel S. Adrenomedullin reduces Staphylococcus aureus alpha-toxin-induced rat ileum microcirculatory damage. Crit Care Med 2005;4:819 –26 4. Buise MP, Ince C, Tilanus HW, Klein J, Gommers D, van Bommel J. The effect of nitroglycerin on microvascular perfusion and oxygenation during gastric tube reconstruction. Anesth Analg 2005;4:1107–11 5. Wunder C, Brock RW, Krug A, Roewer N, Eichelbro¨nner O. A remission spectroscopy system for in vivo monitoring of hemoglobin oxygen saturation in murine hepatic sinusoids, in early systemic inflammation. Comp Hepatol 2005;1:1 6. Wunder C, Brock RW, Frantz S, Go¨ttsch W, Morawietz H, Roewer N, Eichelbo¨nner O. Carbon monoxide, but not endothelin-1, plays a major role for the hepatic microcirculation in a murine model of early systemic inflammation. Crit Care Med 2005;10:2323–31 7. Albuszies G, Radermacher P, Vogt J, Wachter U, Weber S, Schoaff M, Georgieff M, Barth E. Effect of increased cardiac output on hepatic and intestinal microcirculatory blood flow, oxygenation, and metabolism in hyperdynamic murine septic shock. Crit Care Med 2005;10:2332– 8 8. Schwarte LA, Fournell A, van Bommel J, Ince C. Redistribution of intestinal microcirculatory oxygenation during acute hemodilution in pigs. J Appl Physiol 2005;3:1070 –5 9. Fournell A, Scheeren TWL, Schwarte LA. Simultaneous assessment of microvascular oxygen saturation and laser-Doppler flow in gastric mucosa. Adv Exp Med Biol 2003;540:47–53 10. Knobloch K, Lichtenberg A, Pichlmaier M, Mertsching H, Krug A, Klima U, Haverich A. Microcirculation of the sternum following harvesting of the left internal mammary artery. Thorac Cardiovasc Surg 2003;5:255–9 11. Knobloch K, Lichtenberg A, Pichlmaier M, Tomaszek S, Krug A, Haverich A. Palmar microcirculation after harvesting of the radial artery in coronary revascularization. Ann Thorac Surg 2005;3:1026 –30 12. Walter B, Bauer R, Krug A, Derfuss T, Traichel F, Sommer N. Simultaneous measurement of local cortical blood flow and tissue oxygen saturation by near infra-red laser doppler flowmetry and remission spectroscopy in the pig brain. Acta Neurochir Suppl 2002;81:197–9 13. Mook GA, van Assendelft OW, Zijlstra WG. Wavelength dependency of the spectrophotometric determination of blood oxygen saturation. Clin Chim Acta 1969;1:170 –3 14. Zijlstra WG, Buursma A, Meeuwsen-van der Roest WP. Absorption spectra of human fetal and adult oxyhemoglobin, deoxyhemoglobin, carboxyhemoglobin, and methemoglobin. Clin Chem 1991;9:1633– 8 15. Kety SS, Schmidt CF. The effects of altered arterial tensions of carbon dioxide and oxygen on cerebral blood flow and cerebral oxygen consumption of normal young men. J Clin Invest 1948;4:484 –92 16. Hudetz AG. Blood flow in the cerebral capillary network: a review emphasizing observations with intravital microscopy. Microcirculation 1997;2:233–52 17. Kuschinsky W, Paulson OB. Capillary circulation in the brain. Cerebrovasc Brain Metab Rev 1992;3:261– 86 18. Hudetz AG, Biswal BB, Fehe´r G, Kampine JP. Effects of hypoxia and hypercapnia on capillary flow velocity in the rat cerebral cortex. Microvasc Res 1997;1:35– 42 © 2009 International Anesthesia Research Society

203

19. Bereczki D, Wei L, Otsuka T, Hans FJ, Acuff V, Patlak C, Fenstermacher J. Hypercapnia slightly raises blood volume and sizably elevates flow velocity in brain microvessels. Am J Physiol 1993;2:1360 –9 20. Abounader R, Vogel J, Kuschinsky W. Patterns of capillary plasma perfusion in brains in conscious rats during normocapnia and hypercapnia. Circ Res 1995;1:120 – 6 21. Duelli R, Kuschinsky W. Changes in brain capillary diameter during hypocapnia and hypercapnia. J Cereb Blood Flow Metab 1993;6:1025– 8 22. Brauer P, Kochs E, Werner C, Bloom M, Policare R, Pentheny S, Yonas H, Kofke WA, Schulte am Esch J. Correlation of transcranial Doppler sonography mean flow velocity with cerebral blood flow in patients with intracranial pathology. J Neurosurg Anesthesiol 1998;2:80 –5 23. Jackson SA, Piper I, Dunn L, Leffler C, Daley M. Assessment of the variation in cerebrovascular reactivity in head injured patients. Acta Neurochir Suppl 2000;76:445–9 24. Sahuquillo J, Munar F, Baguena M, Poca MA, Pedraza S, Rodriguez-Baeza A. Evaluation of cerebrovascular CO2reactivity and autoregulation in patients with post-traumatic diffuse brain swelling. Acta Neurochir Suppl 1998;71:233– 6 25. Schaller C, Schramm J, Haun D, Meyer B. Patterns of cortical oxygen saturation changes during CO2 reactivity testing in the vicinity of cerebral arteriovenous malformations. Stroke 2003;4:938 – 44

204


26. Engelhard K, Werner C. Inhalational or intravenous anesthetics for craniotomies? Pro inhalational. Curr Opin Anaesthesiol 2006;5:504 – 8 27. Ho¨per J, Gaab MR. Intraoperative monitoring of local Hboxygenation in human brain cortex. Adv Exp Med Biol 1994;361:483–9 28. Ho¨per J, Gaab MR. Effect of arterial PCO2 on local HbO2 and relative Hb concentration in the human brain—a study with the Erlangen micro-lightguide spectrophotometer (EMPHO). Physiol Meas 1994;2:107–13 29. Meyer B, Schaller C, Frenkel C, Ebeling B, Schramm J. Distributions of local oxygen saturation and its response to changes of mean arterial blood pressure in the cerebral cortex adjacent to arteriovenous malformations. Stroke 1999;12:2623–30 30. Dings J, Meixensberger J, Jager A, Roosen K. Clinical experience with 118 brain tissue oxygen partial pressure catheter probes. Neurosurgery 1998;5:1082–95 31. Verweij BH, Amelink GJ, Muizelaar JP. Current concepts of cerebral oxygen transport and energy metabolism after severe traumatic brain injury. Prog Brain Res 2007;161:111–24 32. Soehle M, Heimann A, Kempski O. Laser Doppler scanning: how many measurements are required to assess regional cerebral blood flow? Acta Neurochir Suppl 2000;76:181– 4


Oxygen and Glucose Deprivation in an Organotypic Hippocampal Slice Model of the Developing Rat Brain: The Effects on N-Methyl-D-Aspartate Subunit Composition Lisa Wise-Faberowski, MD* Prairie Neeley Robinson, MD* Sarah Rich, BS* David S. Warner, MD†

BACKGROUND: Oganotypic hippocampal slices (OHS) are commonly used to screen for neuroprotective effects of pharmacological agents relevant to pediatric brain injury. The importance of donor rat pup age and N-methyl-d-aspartate (NMDA) receptor subunit composition have not been addressed. In this study, we evaluated the age-dependent effect of oxygen-glucose deprivation (OGD) in the developing rat brain and determined whether OGD modulates the NMDA receptor subunit composition. METHODS: OHS were prepared from rat pups on postnatal days (PND) 4, 7, 14, and 21 and cultured 7 days in vitro. The slices were exposed to OGD for durations of 5– 60 min. After 24 and 72 h, OHS survival and NMDA subunit composition were assessed. RESULTS: Cell death was evident in OHS prepared from PND 14 and 21 rat pups (P ⬍ 0.001) with OGD durations of 5 and 10 min, respectively. In OHS prepared from PND7 rat pups, neurodegeneration was not evident until 20 min OGD (P ⬍ 0.001). Exposure to OGD in OHS prepared from PND4 and PND7 rat pups was associated with a transition in the NMDA receptor subunit composition from NR2B predominant to NR2A predominant subunit composition. CONCLUSIONS: This in vitro neonatal rat pup investigation using OHS supports both an age and an NMDA receptor subunit composition-dependent relationship between OGD and neuronal cell death. (Anesth Analg 2009;109:205–10)

N

-methyl-d-aspartate (NMDA) receptors are heterotetramers of NMDA receptor 1 (NR1) and one or more of the NR2 subunits: NR2A-D.1,2 The functional and pharmacological properties of the NMDA receptor (NR) are determined by its subunit composition, specifically the type of NR2 subunit incorporated.3,4 During development, NMDA receptors play a critical role in postsynaptic stabilization in the cortex and hippocampus.5 Age-dependent changes in the subunit composition in the developing rat brain have been determined using mRNA and NMDA receptor subunitspecific antagonists. NR2B is expressed at birth and through the first 2 wk. NR2A is expressed in the second

From the *The Children’s Hospital Pediatric Anesthesia Laboratory, Department of Anesthesiology, University of Colorado Health Sciences Center, Aurora, Colorado; and †Departments of Anesthesiology, Neurobiology, and Surgery, Multidisciplinary Neuroprotection Laboratories, Duke University Medical Center, Durham, North Carolina. Accepted for publication December 17, 2008. Supported by American Heart Association National Scientist Development Grant, Foundation for Anesthesia Education and Research Mentored Research Award, NIH Grants T32 GM08600-09 and RO1 GM067139-03. Address correspondence and reprint requests to Lisa WiseFaberowski, MD, The Children’s Hospital, Department of Anesthesiology, 13123 East 16th Ave., B090, Aurora, CO 80045. Address e-mail to [email protected]. Copyright © 2009 International Anesthesia Research Society DOI: 10.1213/ane.0b013e3181a27e37

Vol. 109, No. 1, July 2009

week and gradually increases, thereafter. Therefore, the NR2A/NR2B ratio increases with maturation.6 – 8 The developmental change in the NMDA receptor subunit composition is associated with a change in receptor properties and thus functionality.9,10 For example, different subunits are involved in activitydependent synaptic plasticity e.g., excitatory long-term enhancement and inhibitory long-term depression.11–13 In the developing hippocampus, the former is NR2Bdependent and the latter NR2A. Excitatory or inhibitory balances that are developmentally regulated can become unregulated, and thus promote susceptibility to brain injury.14 –17 NMDA-induced neurotoxicity reflects a calcium imbalance related to the developmental changes in NMDA receptor subunit composition, specifically NR2B.18,19 The NR2B predominance early in development is an adaptive response that protects the fetus during hypoxic conditions normally seen in utero.20,21 However, the response of the NMDA receptor subunit composition to acute hypoxia has not been demonstrated fully in a developing brain model.14,19,22–26 Using organotypic hippocampal slices (OHS) prepared from postnatal day (PND) 4, 7, 14, and 21 rat pups, we evaluated the effect of ontogenetic development of the NMDA receptor subunit composition in the response of neural tissue to oxygen-glucose deprivation (OGD). OHS offer a model of intact neural circuits and avoid some of the variables inherent in an in vivo investigation in small rodents.27,28 205

METHODS OHS Cultures All studies were approved by the Duke University and the University of Colorado Health Sciences Center Institutional Animal Care and Use Committees. OHS cultures were prepared according to the methods described by Stoppini et al.27 with some modification.28 PND 4, 7, 14, and 21 Sprague Dawley rat pups (Zivic Laboratories, Pittsburgh, PA) were anesthetized using an intraperitoneal injection of ketamine (10 mg/kg) and diazepam (0.2 mg/kg). The pups were decapitated and the hippocampi were removed and placed in 4°C Gey’s Balanced Solution (SigmaAldrich, St. Louis, MO) with 100 ␮M adenosine. Using a MX-TS brain slicer (Siskiyou Design Instruments, Grants Pass, OR), the hippocampi were cut transversely (300 ␮m thickness) and transferred to 30-mm diameter membrane inserts (Millicell-CM, Millipore, Bedford, MA). Approximately 3–5 slices were placed on the insert within each well of a six-well culture tray with media where they remained for 7 days. The culture media consisted of 50% Minimal Essential Media (Invitrogen, Carlsbad, CA), 25% Earle’s Balanced Salt Solution (Invitrogen, Carlsbad, CA) and 25% Hyclone Heat Inactivated Horse Serum (Perbio, Cell Culture Division, South Logan, UT) with 6.5 mg/mL glucose and 5 mM KCl. The media was exchanged after the second day in culture and twice a week thereafter. OHS were cultivated in a humidified atmosphere at 37°C and 5% CO2. No antibiotics or antimitotics were used.

Oxygen and Glucose Deprivation During OGD, the media in each cell culture well was exchanged for glucose-free Minimal Essential Media (Invitrogen, Carlsbad, CA).15 Hypoxic gas (5% O2 and 95% N2, 6 L/m) was delivered to a 3 L air-tight chamber, (Billup’s Rothenberg, San Diego, CA) housed within a water jacketed incubator, for 30 min. The hypoxic gas was heated and humidified to 37°C (Fisher-Paykel, Laguna Hills, CA). After 30 min the inserts containing OHS were placed in the hypoxic tissue culture wells above the hypoxic/glucose-free media, so as to allow exposure to media on one surface and the gas mixture on the opposing surface. The flow rate of hypoxic gas was decreased and maintained at 700 mL/min for a period of 5– 60 min, determined by random assignment. Oxygen and CO2 concentrations were continuously monitored via a sampling port connected to a medical gas analyzer (Datex Instruments Corporation, Tewksbury, MA). The pH of the media before and after exposure was 7.4 ⫾ 0.4. After exposure to OGD, the slices with inserts were removed and returned to their glucose-containing normoxic media and maintained in the incubator for 3 days and then imaged for cell death at 24 and 72 h.

Evaluation of Cell Death Sytox (Molecular Probes, Eugene, OR) is a high affinity nucleic acid stain specific for neurons with 206

Oxygen and Glucose Deprivation

compromised plasma membranes and does not penetrate live neurons.28,29 At 24 h after OGD, the media was exchanged with media containing 5 ␮M Sytox®. The OHS with Sytox were imaged using a Leica inverted microscope (2.5⫻) (Wetzlar, Germany) and fluorescent digital images were taken using a CoolSnap digital camera (Image Processing Solutions, North Reading, MA). The variables for imaging were standardized for each OHS. Both light and fluorescent microscopic images were made simultaneously for each slice to confirm region identification. Images were obtained 24 and 72 h post-OGD and stored for later analysis. Using MCID software (Imaging Research, St. Catherines, Ontario), CA1, CA2, CA3, and dentate gyrus were manually outlined on the images obtained by light microscopy of each OHS. These outlines were then superimposed on the fluorescent images of the corresponding OHS. Excitation wavelength was 490 nm and emission was 590 nm. The fluorescence (relative units) was then measured for each hippocampal region. OHS with intense fluorescence in hippocampal CA2 in baseline (control) or OGD groups were excluded from analysis. These represented nonviable slices, which constituted approximately 5%–10% of the OHS population. Analysis of the images was performed by an investigator blinded to group assignment and after completion of each experiment for consistency in measurement.

NMDA Receptor Subunit Composition To determine baseline values, OHS were maintained in culture for 7 days after which time the slices were freshly frozen without further treatment in liquid nitrogen and stored at ⫺80°C until further processing. To assess effects of OGD, OHS were maintained in culture for 7 days and exposed to OGD as described above for 10 or 30 min. The slices were then returned to their original media and 24 h later the slices were frozen in liquid nitrogen and stored at ⫺80°C until further processing.

Protein Preparations and Western Blot Analysis Protein from hippocampal slices was homogenized with PowerGen Tissue homogenizer (Fisher Scientific, Pittsburgh, PA) in buffer containing 50 mM Tris-HCl, pH 7.4, 1% SDS, 2 mM EDTA, and one Complete Protease Inhibitor Cocktail Tablet (Roche Diagnostics, Chicago, IL), briefly boiled, and centrifuged at 13,000g for 30 m. Supernatants were aliquoted and stored at ⫺80°C. Protein concentration was determined using the bicinchoninic (Smith) protein assay (Peirce, Rockford, IL). Western blot analyses were performed on protein extracts. Protein was loaded at a concentration of 50 ␮g sample/lane using Invitrogen (Carlsbad, CA) NuPage™ 4%–12% Bis-Tris prepoured gel. The protein was then transferred to polyvinylidene difluouride membranes (Immobilon™-P Transfer Membrane, Millipore, Bedford, MA) briefly soaked in methanol. Protein transfer was performed using the semidry immunoblot method (Owl Model HEP-1 Panther ANESTHESIA & ANALGESIA

Semi-Dry Electroblotter–#HEP-3). Membranes were blocked with 5% Milk PBS-tween, followed by overnight incubation with primary antibodies (anti-NMDA receptor, subunit 2A rabbit IgG Fraction; 1:200, anti-NMDA receptor, subunit 2B rabbit IgG Fraction; 1:200, Molecular Probes, Eugene, OR). After PBS-tween washes, membranes were incubated with horseradish peroxidase conjugated secondary antibodies and bands were visualized using SuperSignal® West Dura Extended Duration Substrate (Pierce Biotechnology, Rockford, IL). Imaging was performed using a UVP BioChem Imaging System and densitometry was performed using LabWorks 4.0 software. The protein loading internal control was ß-Actin Clone A-15, (Sigma, St. Louis, MO). Data are expressed as a ratio of the NMDA subunits to ß-Actin.

Statistics Experiments for neurodegeneration and NMDA receptor subunit composition were not performed concurrently and therefore were analyzed separately. Within each experiment, fluorescence intensity was compared initially with two-way analysis of variance (PND ⫻ duration of hypoxia exposure) and multivariate analysis (PND ⫻ duration of hypoxia exposure ⫻ DIV). Each hippocampal region (CA1, CA3, and dentate gyrus) was analyzed independently. Approximately 15 OHS were used for each experimental condition. The critical exposure duration for each PND was defined as the shortest OGD duration at which statistically different optical densities were observed as compared with age-matched controls. OHS with 100% neuronal cell death in CA2 were excluded from the analysis. When indicated by a significant F ratio, post hoc testing was performed using Scheffe’s test. Significance was assumed when P ⱕ 0.05.

RESULTS Neuronal Cell Death Comparison to Control Group by Region and Duration In CA1, critical exposure duration for neurodegeneration was 5 min OGD in PND21 (P ⫽ 0.005), 10 min OGD in PND14 OHS (P ⫽ 0.004) 20 min OGD in PND7 OHS (P ⫽ 0.04), and 45 min OGD in PND4 OHS (P ⬍ 0.0001) (Fig. 1). In CA3, critical exposure duration was 5 min in PND21 (P ⫽ 0.008), 10 min in PND14 (0.019), 20 min in PND7 (P ⫽ 0.04) and 45 min PND4 OHS (P ⫽ ⬍0.001) (Fig. 1). In the dentate gyrus, there were no observed differences except in the 45 min PND4 OHS (P ⬍ 0.001). The 24 and 48 h observations were not statistically different regardless of region evaluated. Comparison to PND4 by Age and Duration When compared with PND4 OHS significant differences in the ischemic duration and the resultant cell death were noted in all groups, 10 min PND21 (P ⫽ 0.0145), 20 min PND14 (P ⫽ 0.03), and 30 min PND7 OHS (P ⫽ 0.04) (Fig. 2). This effect on neuronal cell death was independent of the region and the time of observation. Vol. 109, No. 1, July 2009

Figure 1. Comparison of region and postnatal day (PND) (A) Neuronal cell death in hippocampal CA1 (upper graph) and CA3 (lower graph) at 24 h postoxygen-glucose deprivation (OGD) in PND4 organotypic hippocampal slices (OHS) exposed to 10, 30, 45, or 60 min OGD compared with PND-matched controls. Neuronal death increased with OGD duration. Forty-five minutes reflects the critical exposure duration to produce neuronal cell death that is statistically different than that observed in control OHS. (B) Neuronal cell death in CA1 (upper graph) and CA3 (lower graph) at 24 h post-OGD in PND21 OHS exposed to 5, 10, or 20 min OGD compared with PND-matched controls. Neuronal death increased with OGD duration. In both (A) and (B) statistical significance was absent between cell death in CA1 as opposed to CA3. Data are displayed as fluorescence (relative units) ⫾sd; *Denotes statistical significance P ⬍ 0.05 from control values.

Baseline NMDA Receptor Subunit Composition In OHS, in the absence of OGD, baseline NMDA receptor subunit composition changed as a function of postnatal age. In OHS the ratio of NR2B to NR2A receptor subunit composition was NR2B-predominant in PND4 and PND7 OHS but transitioned to a NR2A-predominant subunit composition by PND14 (Fig. 3). © 2009 International Anesthesia Research Society

207

Figure 3. Western blot analysis of NMDA receptor subunit composition NR2A as compared with NR2B in organotypic hippocampal slices (OHS) prepared from rat pups at postnatal day (PND) 4, 7, 14, and 21 DIV7 in the absence of oxygen-glucose deprivation. An age-related change in N-methyl-d-aspartate (NMDA) subunit composition was observed. NR2B (stippled) NR2A (solid) data expressed in densitometry units, mean ⫾ sd as a ratio to ␤–actin. PND 4 OHS showed a predominance of NR2B. PND7 OHS showed NR2B predominance with more NR2A than PND4. PND14 and PND21 OHS showed a change to NR2A predominance. Data demonstrate protein pooled from approximately 50 OHS for each PND. For each NMDA receptor subunit composition, 2–3 Western blot analyses were performed. Analysis was normalized to tissue ␤-actin.

DISCUSSION

Figure 2. Comparison of postnatal day (PND) and oxygenglucose deprivation (OGD) duration. (A) Neuronal death at 24 h post-OGD in hippocampal CA1 in PND 4, 7, 14, and 21 organotypic hippocampal slices (OHS), exposed to 10 min oxygen-glucose deprivation (OGD). Neuronal death increased as a function of increased PND. (B) CA1 neuronal death at 24 h post-OGD in PND 4, 7, 14, and 21 OHS exposed to 20 min OGD. Neuronal death increased with PND. (C) CA1 neuronal death at 24 h post-OGD in PND 4, 7, and 14 OHS exposed to 30 min OGD. Data are displayed as fluorescence (relative units) ⫾sd; *Denotes statistical significance P ⬍ 0.05 from PND4 values.

Change in NMDA Receptor Subunit Composition in Response to OGD To compare all postnatal age groups, both 10 and 30 min OGD were evaluated. These OGD durations were selected based on the critical exposure durations of each PND group, spanning 5 min for PND21 and to 45 min for PND4. Exposure to OGD for 10 or 30 min decreased the NR2B subunit composition in PND 4 and PND7 OHS and reversed the NR2B to NR2A predominance in PND4 and PND7 OHS that was observed in control (no OGD) OHS (Fig. 4). 208


This study demonstrates that OGD-induced neuronal death is dependent on the exposure duration of OGD and the age of rat pup used for OHS preparation. As shown in our work and other neonatal rat models,18,30 –33 in the absence of OGD, PND7 is the stage in rat development when a transition from NR2B to NR2A receptor subunit composition occurs. We also show that the NMDA receptor subunit composition changes in PND 4 and PND7 OHS to a NR2A predominance after OGD instead of the NR2B predominance seen in control PND 4 and 7 OHS (no OGD). When evaluating cell death 24 and 48 h after OGD in both the CA1 and CA3 regions of the hippocampus, we found that the PND 14 and 21 slices were similar for the OGD duration required to cause significant injury (10 and 5 min, respectively). In OHS derived from younger rat brains (PND 4 and 7), OHS cell death was evident only after longer durations (45 and 30 min, respectively). It is important to note that the observations made by our investigation are specific to the hippocampus. In human preterm infants developing white matter is vulnerable to hypoxic-ischemic injury. The vulnerability of gray matter, cortex, and basal ganglia increases toward term in the early postnatal period.34 Thus, responses may be different in other brain regions.9 Synaptogenesis peaks in at PND7 in the rat brain.30,35 Ontogenetically, this corresponds to 32–36 wk gestation in humans and may extend throughout the brain growth spurt (6 mo to 2-yr-of-age).36,37 Using mRNA, expression of the NMDA receptor subunit composition in the hippocampus has been determined to be a function of ANESTHESIA & ANALGESIA

Figure 4. Western blot analysis of N-methyl-daspartate (NMDA) receptor subunit composition NR2B and NR2A in organotypic hippocampal slices (OHS) prepared from rat pups at postnatal day (PND) 4, 7, 14, and 21 exposed to oxygen-glucose deprivation (OGD) for 10 or 30 min. Data shows a change in NMDA subunit composition with respect to age and OGD. NR2B (stippled) and NR2A (solid) data are expressed in densitometry units, mean ⫾ sd, as a ratio to ␤–actin. There is a receptor subunit change in response to 10 and 30 min of OGD. In PND4 and 7 OHS, the NR2B to NR2A predominance is reversed from that observed in controls not exposed to OGD (Fig. 3).

postnatal age in humans.36 The pattern of NMDA receptor subunit composition development in the human hippocampus parallels that seen in the rodent.36 We did not evaluate region-specific differences in NMDA subunit composition, but this has been described in rodents.38 The responsiveness to NMDA-induced neuronal death as a result of duration in culture in OHS has been described.39 The responsiveness to NMDAinduced cell death in OHS of the same postnatal age cultured for 1 and 2 wk differed from those cultured for 3 wk. This difference was not a function of NR2A and 2B subunit composition but the responsiveness of the receptor to NMDA. This investigation further demonstrated that OHS from PND6 pups, regardless of culture duration, did not differ in the NR2A or NR2B subunit composition. Thus, the ratio of the NR2A and NR2B subunit composition was preserved regardless of culturing conditions. However, the NR1 receptor increased after 3 wk in culture. As determined by western blot analysis, we have shown the preservation of the NR2A to NR2B ratio in our OHS prepared from rat pups of differing postnatal age. Changes in NMDA receptor subunit expression in response to hypoxia-ischemia in a neonatal piglet model have been reported.25 NR2B levels were increased at 24 h posthypoxia-ischemia. Thus, differences in our investigation could relate to difference in species, region evaluated (forebrain versus hippocampus),40 mechanism of insult, or postnatal age. Indeed, investigation in younger aged piglets found no change in the receptor subunit composition in response to 1 h of hypoxia.24 Like the rat model, the distribution of NR2A and 2B receptor subunits in neonatal piglets is greatest in the forebrain and hippocampus.24,33,40,41 Regardless, our investigation in a rat model demonstrates a postnatal age-related and OGD duration-dependent change in NMDA receptor subunit composition. Although NMDA receptor density increases with increasing postnatal age and NMDA receptor density Vol. 109, No. 1, July 2009

decreases with acute hypoxia,24,32 we normalized our results to the total ␤-actin in the sample, thus eliminating receptor density (protein) as a variable with postnatal age. It is important to note that, because of this normalization of our data to ␤-actin, the amount of NR2A is less in control PND14 and PND21 OHS as compared with control PND4 and 7 OHS. However, the ratio of NR2B to NR2A is the focus of our investigation, as detailed by the responses of the NMDA receptor subunit composition to acute hypoxia. By normalizing to ␤-actin we can further account for the total protein differences as an effect of OHS slice size which increases with increasing postnatal age. Our protein analysis did not define regional differences in CA1 as opposed to CA3 in terms of relative subunit composition.25 Tissue from those regions were pooled to provide sufficient sample for analysis. Indeed, the NMDA receptor subtype expressed varies with development. NR1 is necessary for receptor function.7,33,41 NR1, though initially paired with NR2B or NR2D early in life, becomes paired with NR2A or NR2C later. This transition parallels the decrease in hypoxia tolerance with increasing postnatal age. Transition in NMDA receptor subunit composition occurs at PND7 in rat pups with completion at PND14.6,7,41 We have demonstrated the preservation of this transition in NR2B to NR2A receptor subunit composition in our OHS. Thus, the tolerance to hypoxia as a function of postnatal age is not unique to our investigation. However, our investigation is the first to confirm this preservation of NR2A to NR2B ratio in rat OHS of differing postnatal age using protein as opposed to mRNA analysis.18,31,32,36 In summary, using an OHS model, we demonstrated an age-related change in the NMDA receptor subunit composition at baseline that parallels that seen in vivo in rodents and in humans. Age-dependent NMDA receptor subunit composition is altered by acute exposure to OGD. As a result, when using OHS to investigate NMDA © 2009 International Anesthesia Research Society

209

receptor responses to hypoxia, the stage of ontogenetic development in tissue used for analysis can be expected to have an influence on results obtained. An understanding of developmental differences and the affect of acute hypoxia on the NMDA subunit composition in the model being used may help us better understand the neurotoxic/ neuroprotective effect of a pharmacological agent in the presence of hypoxic-ischemic injury. ACKNOWLEDGMENTS The technical assistance of Ms. VA E. Beckey is greatly appreciated. REFERENCES 1. Forrest D, Yusaki M, Soares HD, Ng L, Luk DC, Sheng M, Stewart CL, Morgan JI, Conner JA, Curran T. Targeted disruption of NMDA receptor 1 gene abolishes NMDA response and results in neonatal death. Neuron 1994;13:325–38 2. Hardingham GE, Bading H. The yin and the yang of NMDA receptor signaling. Trends Neurosci 2003;26:81–9 3. Cull-Candy S, Brickley S, Farrant M. NMDA receptor subunits: diversity, development and disease. Curr Opin Neurobiol 2001;11:327–35 4. Cull-Candy S, Leszkiewicz DN. Role of distinct NMDA receptor subtypes in central synapses. Sci STKE 2004;re16 5. Durand GM, Kovalchuk Y, Konnerth A. Long-term potentiation and functional synapse induction in developing hippocampus. Nature 1996;381:71–5 6. Moyner H, Burnashev N, Laurie DJ, Sakmann B, Seeburg PH. Developmental and regional expression in the rat brain and functional properties of four NMDA receptors. Neuron 1994;12:529 – 40 7. Sheng M, Cummings J, Roldan LA, Jan YN, Jan LY. Changing subunit composition of heteromeric NMDA receptors during development of rat cortex. Nature 1994;368:144 –7 8. Groc L, Heine M, Cousins SL. NMDA receptor surface mobility depends on NR2A-2B subunits. Proc Natl Acad Sci U S A 2006;103:18769 –74 9. Massey PV, Johnson BE, Moult PR, Auberson YP, Brown MW, Molnar E, Collingridge GL, Bashir ZI. Differential roles of NR2A and NR2B-containing NMDA receptors in cortical long-term potentiation and depression. J Neurosci 2004;24:7821– 8 10. Babb TL, Mikuni N, Najm I, Wylie C, Olive M, Dollar C, MacLennan H. Pre- and postnatal expressions of NMDA receptors 1 and 2B subunit proteins in the normal rat cortex. Epilepsy Res 2005;64:23–30 11. Bear MF, Malenka RC. Synaptic plasticity: LTP and LTD. Curr Opin Neurobiol 1994;4:389 –99 12. Liu L, Wong TP, Pozza MF, Lingnehoehl K, Wang Y, Sheng M, Auberson YP, Wang YT. Role NMDA receptor subtypes in governing the direction of hippocampal synaptic plasticity. Science 2004;304:1021– 4 13. Watanabe M, Inoue Y, Sakimura K, Mishina M. Developmental changes in distribution of NMDA receptor channel subunit mRNA’s. Neuroreport 1992;3:1138 – 40 14. Gurd JW, Bissoon N, Beesley PW, Nakazawa T, Yamamoto T, Vannucci SJ. Differential effects of hypoxia-ischemia on subunit expression and tyrosine phosphorylation of the NMDA receptor in 7- and 21-day old rats. J Neurochem 2002;82:848 –56 15. Ben-Ari Y. Basic developmental rules and their implications for epilepsy in the immature brain. Epileptic Disord 2006;8:91–102 16. Ikonomidou C, Mosinger JL, Salles KS, Labruyere J, Olney JW. Sensitivity of the developing rat brain to hypobaric/ischemic damage parallels sensitivity to N-methyl-D-aspartate neurotoxicity. J Neurosci 1989;9:2809 –18 17. Zhou M, Baudry M. Developmental changes in NMDA neurotoxicity reflect developmental changes in subunit composition of NMDA receptors. J Neurosci 2006;26:2956 – 63 18. Haberny KA, Paule MG, Scallet AC, Sistare FD, Lester DS, Haniq JP, Slikker W Jr. Ontogeny of the N-methyl-D-aspartate (NMDA) receptor system and susceptibility to neurotoxicity. Toxic Sci 2002;68:9 –17

210


19. Misha OP, Fritz KI, Delivoria-Papadopoulos M. NMDA receptor and neonatal hypoxic brain injury. Ment Retard Dev Disabil Res Rev 2001;7:249 –53 20. Bickler PE, Fahlman CS, Taylor DM. Oxygen sensitivity of NMDA receptors: relationship to NR2 subunit composition and hypoxia tolerance of neonatal neurons. Neuroscience 2003;118:25–35 21. Bickler PE, Hansen BM. Hypoxia-tolerant neonatal CA1 neurons: relationship to survival to evoked glutamate release and glutamate receptor mediated calcium changes in hippocampal slices. Dev Brain Res 1998;106:57– 69 22. Grojean S, Pourie G, Vert P, Daval J. Differential neuronal fates in the CA1 hippocampus after hypoxia in newborn and 7 day-old rats: effects of pretreatment with MK-801. Hippocampus 2003;13:970 –7 23. Fernandez-Lopez D, Martinez-Orgado J, Casanova I, Bonet B, Leza JC, Lorenzo P, Moro MA, Lizasoain I. Immature rat brain slices exposed to oxygen-glucose deprivation as an in vitro model of neonatal hypoxic-ischemic encephalopathy. J Neurosci Methods 2005;30:145:205–12 24. Kim WT, Kuo MF, Mishra OP, Delivoria-Papadopoulos M. Distribution and expression of the subunits of N-methyl-Daspartate (NMDA) receptors; NR1, NR2A and NR2B in hypxic newborn piglet brains. Brain Res 1998;799:49 –54 25. Guerguerian AM, Bambrink AM, Traystman RJ, Huganir RL, Martin LJ. Altered expression and phosphorylation of N-methyl-D-aspartate receptors in piglet striatum after hypoxia-ischemia. Mol Brain Res 2002;104:66 – 80 26. Gee CE, Benquet P, Raineteau O, Rietschin L, Kirbach SW, Gerber V. NMDA receptors and the differential ischemic vulnerability of hippocampal neurons. Eur J Neurosci 2006;23:2595– 603 27. Stoppini L, Buch PA, Muller D. A simple method for organotypic cultures of nervous tissue. J Neurosci Methods 1991;37:173– 82 28. Wise-Faberowski L, Zhang H, Ing R, Pearlstein RD, Warner DS. Isoflurane-induced neuronal degeneration: an evaluation in organotypic hippocampal slice cultures. Anesth Analg 2005;101:651–7 29. Moldrich RX, Beart PM, Pascoe CJ, Cheung NS. Low-affinity kainite receptor agonists induce insult-dependent apoptosis and necrosis in cultured cortical neurons. J Neurosci Res 2000;59: 788 –96 30. Hestrin S. Developmental regulation of NMDA receptor mediated at a central synapse. Nature 1992;357:686 –9 31. Zhong J, Carrozza DP, Williams K, Britchett DB, Molshoff PB. Expression of mRNAs encoding subunits of the NMDA receptor in developing rat brain. J Neurochem 1995;64:531–9 32. Goebel DJ, Poosch MS. NMDA receptor subunit composition in the rat brain: a quantitative analysis of endogenous mRNA levels of NR1, NR2A, NR2B, NR2C, NR2D and NR3A. Mol Brain Res 1999;69:164 –70 33. Portera-Cailliau C, Price DL, Martin LJ. N-methyl-D-aspartate proteins NR2A and NR2B are differentially distributed in the developing rat central nervous system as revealed by subunitspecific antibodies. J Neurochem 1996;66:692–700 34. Barkovich AJ, Gressens P, Evrard P, Lyon G, Vrard P. Formation and maturation and developmental disorders of brain white matter and neocortex. AJNR Am J Neuroradiol 1992;13:423– 61 35. Sanchez RM, Jensen FE. Maturational aspects of epilepsy mechanisms and consequences for the immature brain. Epilepsia 2001;42:577– 85 36. Law AJ, Weickert CS, Webster MJ, Herman MM, Kleinman JE, Harrison PJ. Changes in NMDA receptor subunit mRNA’s and cyclophilin mRNA during development of the human hippocampus. Ann N Y Acad Sci 2003;1003:426 –30 37. Romijn HJ, Hoffman MA, Gramsbergen A. At what age is the developing cerebral cortex of the rat comparable to that of the full-term newborn human baby? Early Human Dev 1991;26:61–7 38. Coultrap SJ, Nixon KM, Alvestad RM, Valenzuela CF, Browning MD. Differential expression of NMDA receptor subunits and splice variants among the CA1, CA3 and dentate gyrus of the adult rat. Brain Res Mol Brain Res 2005;135:104 –11 39. Sakaguchi T, Okada M, Kuno M, Kawasaki K. Dual mode of N-methyl-D-aspartate-induced neuronal death in hippocampal slice cultures in relation to N-methyl-D-aspartate receptor properties. Neuroscience 1997;76:411–23 40. Sugimoto A, Takeda A, Kogure V, Onodera H. NMDA receptor (NMDAR1) expression in the rat hippocampus after forebrain ischemia. Neurosci Lett 1994;170:39 – 42 41. Laube B, Kuhse J, Betz H. Evidence for a tetrameric structure of recombinant NMDA receptors. J Neurosci 1998;18:2954 – 61


General Articles

The Effect of Duration of Surgery on Fluid Balance During Abdominal Surgery: A Mathematical Model Tsuneo Tatara, MD Yoshiaki Nagao, MD Chikara Tashiro, MD

BACKGROUND: There is controversy regarding which fluid management regimen provides the best postoperative outcome. Interstitial fluid accumulation may adversely affect postoperative outcome, but the effect of surgical duration on fluid balance is unknown. In this study, we used a mathematical model to describe fluid distribution. METHODS: Previously published data from bioimpedance analysis in patients undergoing abdominal surgery were used to calculate changes to interstitial volume (⌬VIT, percent change relative to baseline) in uninjured and injured tissues. Ratios of ⌬VIT in uninjured and injured tissues at the end of surgery to total fluid volume infused during surgery (VINF, mL/kg) were compared between surgeries of duration ⬍3 h (n ⫽ 5) and ⱖ3 h (n ⫽ 25). Critical values for change in plasma volume (⌬VPL, percent change relative to baseline) and ⌬VIT, which give rise to adverse outcome, were calculated from previously published data on the physiological effects of IV fluid administration in healthy volunteers. Finally, simulated abdominal surgery in a 70 kg man for 1– 8 h was used to determine the effect of crystalloid infusion rate between 2 and 30 mL 䡠 kg⫺1 䡠 h⫺1 on ⌬VPL and ⌬VIT. Fluid infusion rates that maintained ⌬VPL and ⌬VIT in uninjured tissue within critical values were then computationally determined as a function of duration of surgery. RESULTS: Bioimpedance data showed that the differences in ⌬VIT/VINF ratios between uninjured and injured tissues were significant only for surgical duration ⱖ3 h (0.30 ⫾ 0.17% 䡠 kg/mL vs 1.55 ⫾ 0.73% 䡠 kg/mL, P ⬍ 0.0001). Differences of ⌬VIT/VINF ratios between surgical durations ⬍3 and ⱖ3 h were found only for injured tissue (0.45 ⫾ 0.35% 䡠 kg/mL vs 1.55 ⫾ 0.73% 䡠 kg/mL, P ⫽ 0.003). The range of fluid infusion rates required to maintain ⌬VPL and ⌬VIT within the critical values (⬎⫺15% and ⬍20%, respectively) was wide for short-duration surgery (2–18.5 mL 䡠 kg⫺1 䡠 h⫺1 for a 2 h-surgery), whereas it was narrow for long-duration surgery (5– 8 mL 䡠 kg⫺1 䡠 h⫺1 for a 6 h-surgery). CONCLUSIONS: Based on our model, it should be possible to increase the fluid infusion rate without significant interstitial edema for abdominal surgery of ⬍3 h duration. However, our model predicts that restrictive fluid management should be used in abdominal surgery of ⬎6 h duration to avoid excessive interstitial edema. (Anesth Analg 2009;109:211–6)

T

here is controversy regarding whether liberal or restrictive fluid management provides the best postoperative outcome after abdominal surgery.1,2 Restrictive fluid management showed better postoperative outcomes than liberal administration during This article has supplementary material on the Web site: www.anesthesia-analgesia.org. From the Department of Anesthesiology, Hyogo College of Medicine, Hyogo, Japan. Accepted for publication January 19, 2009. Supported by Grants-in-Aid for Researchers, Hyogo College of Medicine, 2006 and Grants-in-Aid for Scientific Research (C) 20591846 from the Ministry of Education, Science and Culture of Japan. Address correspondence and reprint requests to Tsuneo Tatara, MD, Department of Anesthesiology, Hyogo College of Medicine, 1-1 Mukogawa-cho, Nishinomiya, Hyogo 663-8501, Japan. Address e-mail to [email protected]. Copyright © 2009 International Anesthesia Research Society DOI: 10.1213/ane.0b013e3181a3d3dc

Vol. 109, No. 1, July 2009

esophageal,3 gastrointestinal, and colon surgeries.4,5 The opposite findings were reported in patients who underwent laparoscopic cholecystectomy,6 although no differences in recovery were found in fast-track colonic surgery.7 Interstitial fluid accumulation is a possible complication of treatment with excessive perioperative fluid treatment.1,2,8 One potential variable that could account for the differences in response to fluid management is the duration of surgery. The above studies cover a range from 1 h for laparoscopic cholecystectomy to 7 h for esophagectomy. Because microvascular permeability at the surgical site peaks from 3 and 4 h after surgical injury,9 most of the operative period of a long-duration surgery may take place in a state of enhanced capillary leakage. Thus, it is important to know how surgical duration affects interstitial fluid accumulation. Given the inherent difficulties in clinical studies of controlling the duration of surgery with the same magnitude of surgical stress, we used a mathematical model describing fluid 211

distribution during abdominal surgery to study the relationship between fluid administration, interstitial fluid accumulation, and surgical duration.10 The aim of this study was to analyze how duration of surgery affects fluid balance during abdominal surgery when intraoperative fluid is administered at different rates. Previously published data from bioimpedance analysis in patients undergoing abdominal surgery11 were used to determine interstitial volume in uninjured and injured tissues in surgeries of different duration. Critical values for changes in plasma and interstitial volume that give rise to adverse outcomes in a simulated healthy subject were determined using previously published data on physiological effects of IV fluid administration in healthy volunteers.12 Finally, simulation was applied to determine fluid infusion rates required to maintain plasma volume and interstitial volume in uninjured tissue at the end of surgery within critical values as a function of duration of surgery.

METHODS

Table 1. Normal Steady-State Values for Fluid and Protein in the Compartments and Parameters Related to Capillary Exchange, Lymphatics, and Kidney in a 70 kg Male10,17,18 Variables

Values

Plasma volume, VPL (mL) Interstitial volume, VIT (mL) Plasma hydrostatic pressure (mm Hg) Plasma protein concentration (g/mL) Interstitial protein concentration (g/mL) Reflection coefficient for protein Fluid filtration coefficient, kF (mL 䡠 mm Hg⫺1 䡠 h⫺1) Permeability-surface area product for protein, PS (mL/h) Lymph flow sensitivity (mL 䡠 mm Hg⫺1 䡠 h⫺1) Lymph flow rate, JL (mL/h) Rate of urine production, JU (mL/h)

3200 8400 11 0.07 0.0298 0.875 321.4

212

Duration of Surgery and Fluid Balance

43.1 75.7 60.0

The following equations were used to calculate interstitial volume changes:

⌬VIT,UN ⫽ ⌬V IT,LM ⫻

1 ⫺ F IJ FA ⫹ FL

Interstitial Volume Changes at Different Surgical Duration To compare changes to interstitial volume in uninjured and injured tissues in surgeries of different duration, this study used clinical data of changes to interstitial volume in uninjured and injured tissues, which were obtained from previously published segmental bioimpedance analysis data from patients undergoing elective abdominal surgery.11 This previously published study was of 30 patients, aged 36 to 77 yr (mean age, 63 yr), who underwent elective surgery for abdominal cancer. During surgery (1.9 –7.7 h; mean duration, 4.1 h), isotonic acetated Ringer’s solution was infused at a rate of 7 to 15 mL 䡠 kg⫺1 䡠 h⫺1 and blood was transfused as indicated. Segmental bioimpedance analysis was conducted preoperatively (i.e., before the induction of anesthesia) and postoperatively (i.e., after recovery from anesthesia). Impedance values were obtained for the right arm, for the left side of the trunk, and for the right leg. The extracellular fluid volume in each body segment (i.e., arms, trunk, and legs) was calculated by applying the equation derived from the cell suspension theory.11 The extracellular fluid volume estimated by bioimpedance analysis is relatively insensitive to plasma volume change and thus mainly reflects interstitial fluid volume change.13 On this basis, the postoperative changes of extracellular fluid volume relative to preoperative volume estimated by bioimpedance analysis in the limbs (i.e., limbs/arms and legs, ⌬VIT,LM, expressed in mL/kg) and trunk (⌬VIT,TR, expressed in mL/kg) were converted to changes to interstitial volume in uninjured and injured tissues (⌬VIT,UN and ⌬VIT,IJ, respectively, expressed in mL/kg), assuming that fluid distribution was uniform between uninjured tissue.10

200.7

⌬VIT,IJ ⫽ ⌬V IT,TR ⫺ ⌬V IT,UN ⫻

F T ⫺ F IJ 1 ⫺ F IJ

where FA, FT, and FL are the fluid volume fraction of arms, trunk, and legs, respectively,14 and FIJ is the fractional volume of injured tissue in the whole body.10 The changes to interstitial volume (⌬VIT, expressed in percent change relative to baseline) in uninjured and injured tissues at the end of surgery were calculated from values of ⌬VIT,UN and ⌬VIT,IJ, with estimated preoperative interstitial volume (expressed in mL/kg) in uninjured and injured tissues of 8400 䡠 (1 ⫺ FIJ)/70 and 8400 䡠 FIJ/70, respectively (Table 1). We calculated the ratios of ⌬VIT at the end of surgery to total fluid volume infused during surgery (VINF, expressed in mL/kg) for surgeries that lasted ⬍3 h (n ⫽ 5) and ⱖ3 h (n ⫽ 25). All data are expressed as mean ⫾ sd. Statistical comparisons were performed using GraphPad Prism 4 (GraphPad Software, San Diego, CA). The paired t-test was used for comparison of ⌬VIT/VINF ratios between uninjured and injured tissues for groups of short (i.e., ⬍3 h) and long (i.e., ⱖ3 h) duration surgery. The comparison of ⌬VIT/VINF ratios between surgeries of duration ⬍3 and ⱖ3 h was also made for uninjured and injured tissues using unpaired t-test. Values of P ⬍ 0.05 were considered significant.

Determination of Critical Values for Changes in Plasma and Interstitial Volume To determine critical values for change in plasma volume (⌬VPL, expressed in percent change relative to baseline) and ⌬VIT that affect postoperative outcome, we compared two types of data. We considered published data on the physiological effects of IV fluid ANESTHESIA & ANALGESIA

administration in healthy volunteers.12 We also considered simulated values of ⌬VPL and ⌬VIT obtained by our mathematical model describing fluid distribution for a healthy subject. In the clinical study,12 healthy volunteers weighing approximately 70 kg received an infusion of lactated Ringer’s solution of 40 or 5 mL/kg over 3 h. Outcome assessments (pulmonary function measured by spirometry, exercise capacity, balance function, and weight) were conducted at 3, 5, 7, 11, and 27 h from the start of fluid infusion. After that protocol, the scenario for our simulation was a 70 kg man administered isotonic crystalloid solution at 40 or 5 mL/kg over 3 h. During the subsequent 24 h-period, isotonic crystalloid solution was administered at a rate of 1 mL 䡠 kg⫺1 䡠 h⫺1 to replace 1700 mL drinking water that would normally be consumed by healthy volunteers.12 The time course of ⌬VPL and ⌬VIT in the whole body was calculated during the 24 h period using our mathematical model describing fluid distribution in a healthy subject (i.e., without surgery).10 Net fluid balances (i.e., infused fluid volume, urine volume, insensible water losses) were also calculated from simulation data, and compared with weight changes measured in healthy volunteers, assuming that net fluid balance is equal to weight change. The marginal values of ⌬VPL and ⌬VIT in the whole body, which were obtained from simulation at which no adverse effects were observed in healthy volunteers, were determined as critical values for ⌬VPL and ⌬VIT.

surgery between uninjured and injured tissues for different duration of surgery measured by segmental bioimpedance for abdominal surgery patients in previously published study.11 ⌬VIT ⫽ changes to interstitial volume expressed as percent changes relative to preoperative volume; VINF ⫽ total fluid volume infused during surgery expressed in terms of mL/kg. The data were obtained from patients undergoing abdominal surgeries of duration ⬍3 h (n ⫽ 5) and ⱖ3 h (n ⫽ 25) in previously published study.11 Significant differences of ⌬VIT/VINF ratios at the end of surgery were found between uninjured and injured tissues for surgery of duration ⱖ3 h (P ⬍ 0.0001) and between surgeries of duration ⬍3 h and ⱖ3 h for injured tissue (P ⫽ 0.003).

Fluid Volume Simulation During Abdominal Surgery

RESULTS

The mathematical model used for fluid volume simulation during abdominal surgery was the microvascular exchange model modified for abdominal surgery, which predicts fluid and protein distribution and transport in the vascular and interstitial compartments and lymphatics.10 The interstitial compartment was subdivided into uninjured and injured tissue compartments. Urinary dynamics were included in the model formulation. Table 1 provides normal steady-state values for fluid and protein in the compartments and parameters related to capillary exchange, lymphatics, and kidney in a 70 kg man. The interstitial compliance for a severely over-hydrated segment (i.e., ⌬VIT ⬎100%) was obtained from the interstitial fluid pressure-volume curve measured in the isolated hindlimb of dogs.15 Details on the formulation, including mass balance equations and model parameter values used for the simulation, are given in the Appendix available online only at www.anesthesia-analgesia.org. The scenario for a simulation was a 70 kg man who underwent elective abdominal surgery lasting between 1 and 8 h. The patient received isotonic crystalloid solution at constant infusion rates of 2–30 mL 䡠 kg⫺1 䡠 h⫺1 throughout surgery. Hemorrhage was not included in the model. On the basis of mass balance equations for fluid and protein in body fluid

Interstitial Volume Changes at Different Surgical Duration

Vol. 109, No. 1, July 2009

Figure 1. Comparison of ⌬VIT/VINF ratios at the end of

compartments, we calculated the time course of ⌬VPL and ⌬VIT in uninjured and injured tissues during surgery using the mathematical model. Fluid infusion rates that maintain ⌬VPL and ⌬VIT at the end of surgery within critical values previously calculated were determined for different durations of surgery with the mathematical model.

No differences were found for ⌬VIT/VINF ratios at the end of surgery between uninjured and injured tissues grouped in the surgery of duration ⬍3 h (0.40 ⫾ 0.25% 䡠 kg/mL vs 0.45 ⫾ 0.35% 䡠 kg/mL, Fig. 1). However, significant differences in ⌬VIT/VINF ratios at the end of surgery between uninjured and injured tissues were found for patients undergoing surgery of ⱖ3 h duration (0.30 ⫾ 0.17% 䡠 kg/mL vs 1.55 ⫾ 0.73% 䡠 kg/mL, P ⬍ 0.0001, Fig. 1). The ⌬VIT/VINF ratios in uninjured tissue were not different between surgeries of duration ⬍3 and ⱖ3 h with mean values of ⌬VIT/VINF ratios of 0.40% 䡠 and 0.30% 䡠 kg/mL, respectively. In contrast, the injured tissue showed large values of ⌬VIT/VINF ratios at the end of surgery for the surgery of duration ⱖ3 h (mean value: 1.55% 䡠 kg/mL), compared with those for the surgery of duration ⬍3 h (mean value: 0.45% 䡠 kg/mL, P ⫽ 0.003).

Determination of Critical Values for Changes in Plasma and Interstitial Volume Healthy volunteers who were infused with crystalloid solution of 5 mL/kg did not show significant adverse effects at the end of the assessment period (i.e., 27 h from the start of fluid infusion). Given that our model predicted that the decrease of plasma © 2009 International Anesthesia Research Society

213

Figure 2. Time course of changes to plasma volume, interstitial volume in the whole body and net fluid balance when a 70 kg healthy man is infused IV with crystalloid solution of 40 or 5 mL/kg during the first 3 h-period and at a rate of 1 mL 䡠 kg⫺1 䡠 h⫺1 during the subsequent 24 h-period. ⌬VPL ⫽ changes to plasma volume expressed as percent changes relative to baseline; ⌬VIT ⫽ changes to interstitial volume in the whole body expressed as percent changes relative to baseline. Solid and unfilled arrows denote the times at which there was impaired and normal pulmonary function, respectively, in healthy volunteers who received IV crystalloid solution of 40 mL/kg over 3 h.12 Solid and unfilled circles denote weight changes in healthy volunteers who received IV crystalloid solution of 40 and 5 mL/kg over 3 h, respectively.12 Gray bars denote critical values for ⌬VPL and ⌬VIT. volume at that time was 15.8% compared with baseline (Fig. 2, top panel), we set the critical value for ⌬VPL to ⫺15%. In healthy volunteers who were infused with crystalloid solution of 40 mL/kg, pulmonary function was impaired at 3, 5, 7, and 11 h from the start of infusion, and normal at 27 h. Given that corresponding ⌬VIT values were 22.9%, 23.2%, 22.3%, 21.2%, and 18.1%, respectively (Fig. 2, middle panel), we set the critical value for ⌬VIT to 20%. Net fluid balances predicted by the mathematical model were almost identical with weight changes measured in healthy volunteers (Fig. 2, bottom panel).

Fluid Volume Simulation During Abdominal Surgery Figure 3 shows volume changes at the end of surgery in plasma and interstitium (uninjured and injured tissues) relative to preoperative volume as a function of fluid infusion rate for a simulated abdominal surgery of duration 2, 4, or 8 h in a 70 kg man. Values of ⌬VPL increased almost linearly with fluid infusion rates for 2 and 4 h-surgeries. For a 2 h-surgery, fluid infusion at a rate of 10 mL 䡠 kg⫺1 䡠 h⫺1 maintained normal plasma volume, whereas a 4 h-surgery required infusion rate as high as 17.5 mL 䡠 kg⫺1 䡠 h⫺1 to maintain normal plasma volume. 214


Figure 3. Volume changes at the end of surgery of plasma (top panel), interstitium in uninjured tissue (middle panel), and injured tissue (bottom panel) relative to preoperative volume at different fluid infusion rates for a simulated abdominal surgery of duration 2, 4, or 8 h in a 70 kg man. ⌬VPL ⫽ changes to plasma volume expressed as percent changes relative to preoperative volume; ⌬VIT ⫽ changes to interstitial volume expressed as percent changes relative to preoperative volume.

The 8 h-surgery caused severe hypovolemia compared with the 2 and 4 h-surgeries, for which fluid infusion at a rate higher than 20 mL 䡠 kg⫺1 䡠 h⫺1 maintained normal plasma volume, but with an abrupt increase of ⌬VPL (19% at 25 mL 䡠 kg⫺1 䡠 h⫺1). As fluid infusion rates were increased, the values of ⌬VIT increased monotonically both for uninjured and injured tissues. However, for a 4 h-surgery, injured tissue exhibited a significantly larger ⌬VIT than uninjured tissue (59% vs 18% at 10 mL 䡠 kg⫺1 䡠 h⫺1), with this difference being even greater for an 8 h-surgery (118% vs 38% at 10 mL 䡠 kg⫺1 䡠 h⫺1). In contrast, the 2 h-surgery did not show significant differences of ⌬VIT between injured and uninjured tissues (23% vs 9% at 10 mL 䡠 kg⫺1 䡠 h⫺1) compared with the 4 and 8 h-surgeries. As shown in Figure 4, the range of fluid infusion rates required to maintain ⌬VPL at a level above ⫺15%, and ⌬VIT in uninjured tissue at ⬍20% at the end of surgery, was wide for short-duration surgery (2–18.5 mL 䡠 kg⫺1 䡠 h⫺1 for a 2 h-surgery). As the duration of surgery increased, this range of fluid infusion rate became narrower, with an infusion rate of 5– 8 mL 䡠 kg⫺1 䡠 h⫺1 for a 6 h-surgery.

DISCUSSION Using a mathematical model, we assessed the effects of surgical duration and crystalloid infusion rate on the time course of plasma volume and interstitial ANESTHESIA & ANALGESIA

Figure 4. Optimal fluid infusion rates as a function of duration of surgery for a simulated abdominal surgery in a 70 kg man. Unfilled circles and triangles denote marginal infusion rates which maintain plasma volume at above the critical value and interstitial volume in uninjured tissue at less than the critical value at the end of surgery, respectively. Shaded zone shows safety range of fluid infusion rates.

volume during abdominal surgery. The absolute values of fluid volume changes calculated for the simulated patient may not be valid for the individual patient,10 but a relative comparison of fluid volume changes for different infusion rates of crystalloid between different durations of surgery demonstrated that fluid distribution between the intravascular space and the interstitium is affected both by infusion rate and also by the duration of surgery. Consistent with our previous study,10 despite aggressive fluid infusion, surgical injury caused progressive hypovolemia. Hypovolemia was significant for surgeries of 4 and 8 h duration, which required fluid infusion rates higher than 15 mL 䡠 kg⫺1 䡠 h⫺1 to maintain normal plasma volume at the end of surgery (Fig. 3). Injured tissue showed a greater tendency for interstitial swelling (“third spacing”) compared with that in uninjured tissue for 4 and 8 h surgeries, whereas for a 2 h-surgery the difference of relative interstitial volume change between uninjured and injured tissues was small (Fig. 3). This finding was confirmed by bioimpedance analysis data from patients who underwent elective abdominal surgery, where the difference in the ⌬VIT/VINF ratio at the end of surgery between uninjured and injured tissues was significant for surgery of duration ⱖ3 h (Fig. 1). The primary goal of fluid management during the perioperative period is to maintain organ perfusion by maintaining normal blood flow. However, our results demonstrated that, for surgery duration of 8 h, such a goal may be achieved by fluid infusion at a rate as high as 18 mL 䡠 kg⫺1 䡠 h⫺1. However, this infusion rate increased interstitial volume in uninjured and injured tissues by 93% and 133% of preoperative values, respectively (Fig. 3). One possible solution is to maintain plasma volume at less than the normal value and thereby limit interstitial fluid accumulation in uninjured tissue. On this basis, fluid infusion rates required to maintain plasma volume and interstitial volume in uninjured tissue (i.e., lung) at the end of Vol. 109, No. 1, July 2009

surgery within critical values were computergenerated according to parameters used in the mathematical model. We set the critical value for ⌬VPL to ⫺15%. According to our model, plasma volume at this value produces urine at a rate of 0.5 mL 䡠 kg⫺1 䡠 h⫺1, which is a minimum requirement for perioperative fluid management.4,5 Pulmonary function in healthy volunteers was impaired at ⌬VIT larger than 20%. In the previous study that measured extravascular lung water in patients who underwent esophagectomies, oxygenation of the lung was most impaired 24 h after surgery, with a resultant 140 mL increase in extravascular lung water relative to that before surgery.16 Given that this increase corresponds to a 22% increase of interstitial volume, the finding lends support to the critical value we arrived at for ⌬VIT. Our study demonstrated that fluid infusion rates required to maintain plasma volume and interstitial volume in uninjured tissue at the end of surgery within critical levels depend on the duration of surgery. The safety range of fluid infusion rates was wide for short-duration surgery (2–18.5 mL 䡠 kg⫺1 䡠 h⫺1 for a 2 h-surgery), whereas it became narrow for long-duration surgery (5– 8 mL 䡠 kg⫺1 䡠 h⫺1 for a 6 h-surgery). In a recent study of physiological recovery after fast-track colonic surgery (duration of surgery: 2 h), liberal fluid administration (18 mL 䡠 kg⫺1 䡠 h⫺1 of crystalloid plus 7 mL/kg of colloid) significantly impaired pulmonary function at 6 h after surgery compared with restrictive fluid administration (6 mL 䡠 kg⫺1 䡠 h⫺1 of crystalloid plus 7 mL/kg of colloid).7 This finding is consistent with our model prediction for the safety range of fluid infusion rates. Our study does not provide direct evidence regarding the fluid regimen that provides the best postoperative outcome. Furthermore, surgery of long duration is likely to involve a complicated surgical procedure, and thereby cause severe and extensive tissue damage compared with that of short duration. However, our findings suggest that the duration of surgery may be responsible, at least in part, for discrepancies reported in published studies. For surgery duration of ⬍3 h, it is possible to increase the fluid infusion rate without causing significant interstitial edema and thus reduce adverse effects by improving tissue circulation. This is not the case for surgery duration of ⬎6 h, which requires restrictive fluid management to prevent excessive interstitial edema. Additionally, different end points for outcome assessment may result in a different optimal fluid infusion rate for abdominal surgery. Given that pulmonary function is a primary assessment of outcome, the interstitial fluid accumulation in uninjured tissue may be responsible for postoperative outcome of abdominal surgery. If the restoration of bowel function after abdominal surgery is considered, the prevention of interstitial fluid accumulation in injured tissue may lead to an improved outcome. According to our findings, the assessment of interstitial edema in © 2009 International Anesthesia Research Society

215

injured tissue may require more restrictive fluid management compared with that in uninjured tissue. For long procedures, highly restrictive fluid management may cause hypovolemia. We have not assessed whether other options, such as the use of vasopressors or substitution of colloid solutions for crystalloid solutions, would help maintain adequate cardiac filling pressure and cardiac output during prolonged surgery. REFERENCES 1. Bellamy MC. Wet, dry or something else? Br J Anaesth 2006;97:755–7 2. Hahn RG. Fluid therapy might be more difficult than you think. Anesth Analg 2007;105:304 –5 3. Neal JM, Wilcox RT, Allen HW, Low DE. Near-total esophagectomy: the influence of standardized multimodal management and intraoperative fluid restriction. Reg Anesth Pain Med 2003;28:328 –34 4. Joshi GP. Intraoperative fluid restriction improves outcome after major elective gastrointestinal surgery. Anesth Analg 2005;101:601–5 5. Nisanevich V, Felsenstein I, Almogy G, Weissman C, Einav S, Matot I. Effect of intraoperative fluid management on outcome after intraabdominal surgery. Anesthesiology 2005;103:25–32 6. Holte K, Klarskov B, Christensen DS, Lund C, Nielsen KG, Bie P, Kehlet H. Liberal versus restrictive fluid administration to improve recovery after laparoscopic cholecystectomy. A randomized, double-blind study. Ann Surg 2004;240:892–9 7. Holte K, Foss NB, Andersen J, Valentiner L, Lund C, Bie P, Kehlet H. Liberal or restrictive fluid administration in fast-track colonic surgery: a randomized, double-blind study. Br J Anaesth 2007;99:500 – 8

216


8. Holte K, Sharrock NE, Kehlet H. Pathophysiology and clinical implications of perioperative fluid excess. Br J Anaesth 2002;89:622–32 9. Fantone JC, Ward PA. Inflammation. In: Rubin E, Farber JL, eds. Pathology. 3rd ed. Philadelphia: Lippincott-Raven Publishers, 1999:37–75 10. Tatara T, Tashiro C. Quantitative analysis of fluid balance during abdominal surgery. Anesth Analg 2007;104:347–54 11. Tatara T, Tsuzaki K. Segmental bioelectrical impedance analysis improves the prediction for extracellular water volume changes during abdominal surgery. Crit Care Med 1998;26:470 – 6 12. Holte K, Jensen P, Kehlet H. Physiological effects of intravenous fluid administration in healthy volunteers. Anesth Analg 2003;96:1504 –9 13. Tedner BT, Jacobson HS, Linnarsson D, Lins LE. Impedance fluid volume monitoring during intravenous infusion in healthy subjects. Acute Care 1983–1984;10:200 – 6 14. Thomas BJ, Cornish BH, Ward LC. Bioelectrical impedance analysis for measurement of body fluid volumes: a review. J Clin Eng 1992;17:505–10 15. Guyton AC. Interstitial fluid pressure. II. Pressure-volume curves of interstitial space. Circ Res 1965;16:452– 60 16. Sato Y, Motoyama S, Maruyama K, Okuyama M, Hayashi K, Nakae H, Tajimi K, Ogawa J. Extravascular lung water measured using single transpulmonary thermodilution reflects perioperative pulmonary edema induced by esophagectomy. Eur Surg Res 2007;39:7–13 17. Gyenge CC, Bowen BD, Reed RK, Bert JL. Transport of fluid and solutes in the body I. Formulation of a mathematical model. Am J Physiol 1999;277:H1215–H1227 18. Gyenge CC, Bowen BD, Reed RK, Bert JL. Preliminary model of fluid and solute distribution and transport during hemorrhage. Ann Biomed Eng 2003;31:823–39 19. Tatara T, Tsunetoh T, Tashiro C. Crystalloid infusion rate during fluid resuscitation from acute haemorrhage. Br J Anaesth 2007;99:212–17


Special Article

2009 Anesthesia & Analgesia Guide for Authors 2008 –2009 Editorial Board, Anesthesia & Analgesia*

Summary of Changes to Guide for Authors • We have removed the requirement for Implication Statements. • We have added several article types. • We have clarified requirements for supplemental digital content (formerly “data supplements”). • We have amended requirements for author responsibility. • We have amended the requirements for referencing websites. • We have added the requirement for patient consent for Letters to the Editor, Case Reports, and Echo Rounds. • We have amended our conflict of interest disclosure requirements. • We have updated the requirements for clinical trial registration. • We have created a title page generator to accommodate several of these requirements for manuscript submission. Introduction Anesthesia & Analgesia publishes articles that are novel or definitive and improve clinical care or guide future research. This Guide for Authors was written for authors preparing manuscripts for submission to Anesthesia & Analgesia. It explains the Editorial Board’s *Sorin J. Brull, Xavier Capdevila, Vincent W. S. Chan, Neal Cohen, Marie E. Csete, Peter J. Davis, Franklin Dexter, François Donati, Marcel E. Durieux, Kazuhiko Fukuda, Thomas J. Gal, Tong J. Gan, Adrian W. Gelb, Tony Gin, Peter S. A. Glass, Keith M. Gregg, Jeffrey B. Gross, George M. Hall, Quinn H. Hogan, Charles W. Hogue, Jr., Terese T. Horlocker, Jonas S. Johansson, Igor Kissin, Jerrold H. Levy, Spencer S. Liu, Martin J. London, Edward C. Nemergut, Nancy A. Nussmeier, Paul S. Pagel, Richard C. Prielipp, Carl E. Rosow, Lawrence J. Saidman, John Sear, Steven L. Shafer, Edward R. Sherwood, Peter Douglas Slinger, Gary R. Strichartz, Jukka Takala, Dwayne R. Westenskow, Paul F. White, Cynthia A. Wong, Tony L. Yaksh, and Mark H. Zornow. Reprints will not be available. Address correspondence to Steven L. Shafer, MD, Editor-in-Chief, Anesthesia & Analgesia, 100 Pine Street, Suite 230, San Francisco, CA 94111. Address e-mail to [email protected]. Copyright © 2009 International Anesthesia Research Society DOI: 10.1213/ane.0b013e3181a977c0

Vol. 109, No. 1, July 2009

expectations for submitted manuscripts and Journal policies on manuscript handling. The Guide contains extensive information on manuscript preparation that may be helpful regardless of where the authors submit their manuscript. The Guide for Authors is available at http://www. anesthesia-analgesia.org/misc/GuideForAuthors.pdf. When preparing a manuscript, please be certain to download the most recent version of the guide. Please contact our Editorial Office for any questions: Anesthesia & Analgesia 100 Pine Street Suite 230 San Francisco, CA 94111 e-mail: [email protected] Phone: (415) 777-2750 Fax: (415) 777-2803

About Anesthesia & Analgesia Anesthesia & Analgesia, the oldest publication for the specialty of anesthesiology, is the official scientific journal of the following societies and foundations: • • • • • • • • •

International Anesthesia Research Society (IARS) Society of Cardiovascular Anesthesiologists (SCA) Society for Pediatric Anesthesia (SPA) Society for Ambulatory Anesthesia (SAMBA) International Society for Anaesthetic Pharmacology (ISAP) Society for Technology in Anesthesia (STA) Anesthesia Patient Safety Foundation (APSF) American Society of Critical Care Anesthesiologists (ASCCA) Society of Obstetric Anesthesia and Perinatology (SOAP)

Anesthesia & Analgesia is divided into the following sections: • Editorials • Cardiovascular Anesthesiology, which includes Hemostasis and Transfusion Medicine • Pediatric Anesthesiology • Ambulatory Anesthesiology • Anesthetic Pharmacology, which includes PreClinical Pharmacology and Clinical Pharmacology 217

• • • • • • • •

Technology, Computing, and Simulation Patient Safety Economics, Education, and Policy Critical Care and Trauma Neurosurgical Anesthesiology and Neuroscience Obstetric Anesthesiology General (otherwise unspecified) Analgesia, consisting of 䡩 Pain Mechanisms, 䡩 Pain Medicine, 䡩 Regional Anesthesia • Cochrane Corner • Correspondence • Book, Multimedia, and Meeting Reviews We assign all manuscripts to one of these sections. Each section other than Editorials has one or more designated Section Editors, responsible for shepherding manuscripts through the peer review process. Authors may request a section at the time of the submission. Requests are considered by the Editor-in-Chief when the manuscript is assigned to a Section Editor.

Responsible Conduct of Research The following pages describe the standards set by the Editorial Board of Anesthesia & Analgesia for responsible conduct of research. The Editorial Board does not permit publication of any manuscripts not adhering to these rules. The name of the research ethical review committee varies with country and local custom. In the United States the committee is called the Institutional Review Board. Other countries call their research ethical review committee a “Research Ethics Committee.” Some institutions refer to the board that reviews animal studies as the “Animal Care and Use Committee.” In this document, “Institutional Review Board” is used generically to refer to the local board that reviews the ethical treatment of human or animal experimental subjects and grants institutional approval for the study. Human Subjects Regardless of the country of origin, all clinical investigators describing human research must abide by the Ethical Principles for Medical Research Involving Human Subjects outlined in the Declaration of Helsinki, and adopted in October 2000 by the World Medical Association. This document can be found at http://ohsr.od.nih.gov/guidelines/helsinki.html. Investigators are encouraged to read and follow the Declaration of Helsinki. Clinical studies not meeting the Declaration of Helsinki criteria will be denied peer review. If published research is subsequently found to be noncompliant, it will be withdrawn or retracted. On the basis of the Declaration of Helsinki, Anesthesia & Analgesia requires that all manuscripts reporting clinical research state in the first paragraph of the Methods section that: 1. The study was approved by the appropriate Institutional Review Board, and 218

Guide for Authors

2. Written informed consent was obtained from all subjects, a legal surrogate, or the parents or legal guardians for minor subjects, OR the requirement for written informed consent was waived by the Institutional Review Board. Human subjects should not be identifiable. Do not disclose patients’ names, initials, hospital numbers, dates of birth, or other protected health care information. Keep copies of your Institutional Review Board approval and written informed consents. The editor or reviewers may request copies of these documents to address questions about Institutional Review Board approval and study conduct. Investigational Drugs The Editorial Board of Anesthesia & Analgesia may exercise judgment about the ethics of a clinical trial involving investigational drugs that is more stringent than the investigator’s Institutional Review Board. Compliance with the Journal’s guidelines should be specified in the Methods section, when appropriate. Neuraxial or Perineural Drug Administration Studies using drugs injected into the neuraxial (caudal, intrathecal, or epidural) or perineural space must meet one of three criteria: 1. The drug is approved for neuraxial or perineural administration by the United States Food and Drug Administration (FDA) or the equivalent regulatory agency for the country in which the study took place. 2. The drug is not approved for neuraxial or perineural use, but it is widely used and accepted for neuraxial (e.g., fentanyl) or perineural administration. The publication of dosing guidelines in multiple textbooks represents a reasonable demonstration that a drug is widely used and accepted for neuraxial or perineural administration. 3. The study is performed under an Investigational New Drug (IND) application approved by the FDA or the equivalent agency in the investigator’s country. Investigators in the United States are directed to http://www.fda.gov/cder/forms/ 1571-1572-help.html for further information on obtaining an investigator IND. To obtain an investigator IND, the investigator must complete forms 1571 and 1572, which are mailed to the FDA along with the investigator’s curriculum vitae. Should the investigator’s country not have an equivalent process, the investigator must submit a statement from the Institutional Review Board that the preclinical toxicity data were reviewed for safety by a qualified expert prior to approval of the human trial. Anesthesia & Analgesia will not publish a retrospective paper involving neuraxial or perineural drug administration if the treatment would be considered inappropriate or unethical in a prospective trial. ANESTHESIA & ANALGESIA

Drug Studies in Children Anesthesia & Analgesia is committed to expanding knowledge of the clinical pharmacology of drugs in children. However, studying drugs in children when there is no pediatric indication poses ethical concerns.1 Therefore, studies of drugs in children must meet one of three criteria: 1. The drug is approved for pediatric administration by the FDA or an equivalent regulatory agency. 2. The drug is not approved for use in children, but it is widely used and accepted for pediatric administration. A reasonable demonstration that the drug is clinically accepted for use in children is when the administration in the study is consistent with the route, dose, and indication reported in multiple textbooks. 3. The study is done under an IND application approved by the FDA or the equivalent agency in the investigator’s country, as described by Schultheis et al.2 Investigators in the United States are directed to http://www.fda.gov/ cder/forms/1571-1572-help.html for further information on obtaining an investigator IND. Anesthesia & Analgesia will not publish a retrospective paper involving pediatric drug administration if the treatment would be considered inappropriate or unethical in a prospective trial. Nonconformity in Dose, Route, or Indication (“OffLabel” Use) In the United States, FDA regulations state that drug use conforms to the package insert (“on-label”) when the dose, route of administration, and indication match the guidelines in the package insert. If the dose, route, or indication does not match the package insert, then the drug use is “off-label.” Drugs are commonly used off-label in clinical trials, and the practice is generally acceptable. However, the Editorial Board of Anesthesia & Analgesia reserves the right not to review a manuscript describing off-label administration of a drug if the Editorial Board believes the study has posed unacceptable risk to the study subjects. To preclude such a determination, investigators are encouraged to obtain an Investigator IND from the FDA, as described in http://www.fda.gov/cder/forms/ 1571–1572-help.html, or an equivalent agency in their country before initiating studies involving off-label drug administration. Animal Subjects Manuscripts that describe investigations performed in vertebrate animals must explicitly state that the study was approved by the authors’ Institutional Review Board for animal research (e.g., the Institutional Animal Care and Use Committee). The Journal expects humane and ethical treatment of all experimental animals, as outlined by the United States Public Health Service Policy on Humane Care and Use of Laboratory Animals and the Guide for the Care and Use of Laboratory Animals (1996), prepared by the Vol. 109, No. 1, July 2009

National Academy of Sciences’ Institute for Laboratory Animal Research. This statement should appear at the beginning of the Methods section. Multiple Publications of Human or Animal Trials In the interest of minimizing risk to human and animal subjects, as well as promoting efficient use of scarce research funds, investigators may pose several questions and make multiple measurements in a single study, with the intent of publishing multiple manuscripts. This may be a laudable practice, or it may be an inappropriate attempt to slice a single study into “minimum publishable units.” Division of data from a single research study into multiple manuscripts is acceptable, provided the following three requirements are met: 1. The cover letter for every paper derived from the study explains the need for dividing the study into multiple manuscripts. This requirement applies even if only one of the submissions is to Anesthesia & Analgesia. The Journal will consider the appropriateness of the division as part of the review process. 2. In all manuscripts after the first, the investigator must disclose any data that have been previously reported, with appropriate citation to the first manuscript. This practice is essential for scientific continuity. For example, should a question arise about the conduct of the study in one manuscript, readers should be able to identify all manuscripts based on the same experimental data. 3. Measurements must not interfere with each other. Such interference may happen in ways not evident at the time of the study. For example, measurements of pain thresholds may make it impossible to measure sedative effects. The potential for interfering measurements may not be evident if the pain thresholds and sedation effects are reported in separate manuscripts that are not appropriately cross-referenced. Registration of Clinical Trials All clinical trials involving investigational drugs supported by a pharmaceutical firm or investigational devices supported by a device manufacturer must be registered at the time that a manuscript is submitted to Anesthesia & Analgesia for publication. The registry, registration number, and date of registration must be stated in the first paragraph of the Methods section of the manuscript. All clinical trials involving investigational drugs or devices supported by a pharmaceutical firm or device manufacturer that began after January 1, 2008 must be registered prior to patient enrollment. A number of registries have been approved by the International Committee of Medical Journal Editors, including www.clinicaltrials.gov (the most commonly used registry in the United States), isrctn.org, www. umin.ac.jp/ctr/index/htm, www.actr.org.au, and © 2009 International Anesthesia Research Society

219

www.trialregister.nl. Submissions that have registered with the European Clinical Trials Database, EudraCT (https://eudract.emea.europa.eu/eudract/index.do), meet this requirement. Conflict of Interest A conflict of interest exists when an author’s judgment about a manuscript may be influenced by secondary gain. Secondary gain typically involves personal, financial, academic, or political advancement. Examples of financial gain are easiest to identify and include direct monetary benefits, such as investments, stocks, honoraria, etc. When study results (as differentiated from publication per se) may affect an author’s bonus, incentive payment (e.g., from likely changes in clinical workload), or salary (e.g., research about academic appointments and salary), this is also considered a conflict of interest. Academic recognition and advancement resulting from publishing high quality papers are the appropriate reward for good work and do not represent a conflict of interest. Potential conflicts of interest occur frequently. In some disciplines, they may be unavoidable. Authors of scientific studies sponsored by industry necessarily possess a conflict of interest. Although this conflict is understood and accepted, it must be disclosed. Investigators frequently also have consulting or lecturing relationships with companies sponsoring their research. These relationships may be entirely appropriate, but must be disclosed. Conflict of interest disclosure should be made at the time of manuscript submission for every author so that a decision can be made on whether the competing interests may have influenced the manuscript in any manner. Conflicts of interest must be disclosed in every submission, including Editorials and Letters to the Editor. You must disclose your conflicts of interest when you create your title page using the form at http://www.aaeditor.org/title.html. A complete title page will be generated that includes all conflict of interest disclosures. This title page must then be pasted at the beginning of your manuscript. A manuscript will not be rejected solely because of conflict of interest. However, appearance of a potential conflict of interest could result in a request that the conflict of interest be stated in the published manuscript. Anesthesia & Analgesia does not have a threshold monetary value to determine “relevant” or “significant” conflicts of interest. Similarly, the Journal believes that there is no specific time at which a potential conflict of interest ceases to exist. All relevant potential conflicts of interest should be declared, regardless of monetary value or the date of the relationship. Conversely, we recognize that extensive disclosures of trivial or ancient relationships may unintentionally obfuscate relevant conflicts. Authors are encouraged to err on the side of full disclosure. Full disclosure at the time of submission 220

Guide for Authors

has fewer repercussions than subsequent exposure. This does not mean that everything disclosed will appear in the manuscript. Only what is deemed relevant will be included in the published manuscript, based on discussion between the editors and the author(s).

Preparing Your Manuscript The following pages describe the types of manuscripts published by Anesthesia & Analgesia. The guidelines offer general rules on length, format, and content. These guidelines are intended to help authors write manuscripts meeting the expectations of reviewers and editors, improving chances that a manuscript will be accepted for publication. If a manuscript must deviate from these rules in any significant manner, please contact the Editorial Office in advance of submitting the manuscript to be certain that the Journal will consider publication. Additionally, please explain any significant deviations from the expected format in the “Enter Comments” section when submitting the manuscript using Editorial Manager. Submissions to Anesthesia & Analgesia should use grammatically accurate English and American spellings. All submissions will be edited for syntax, grammar, and spelling. Manuscript Types and Word Count Please review the following descriptions of manuscript types and recommended word counts. Word counts are for guidance and are not absolute limits. Manuscripts should be as succinct as possible. All submissions must include a title page. Research Reports describe original clinical or laboratory investigations. A meta-analysis of a series of research papers is also a Research Report. Research Reports include a structured Abstract (maximum 400 words), Introduction, Methods, Results, and Discussion. Research Reports should not exceed 3000 words. Anesthesia & Analgesia currently accepts about 25% of submitted Research Reports. Case Reports describe “truly exceptional” cases making an important teaching point or scientific observation. They may describe unusual and instructive cases, novel anesthetic techniques, novel use of equipment, or new information on diseases of importance to anesthesiology. Case Reports are frequently suitable for documenting unusual cases of toxicity or equipment failure. They are almost never appropriate for describing efficacy of a drug or a treatment, which should be demonstrated by an adequately powered and wellcontrolled clinical trial. The only exception is a demonstration of efficacy in a population, or a clinical scenario, so uncommon that a clinical trial cannot be performed. Case reports describing successful management of complex cases will only be considered if they make a truly exceptional observation. Case Reports include an unstructured Abstract (maximum of 100 words), Introduction, Case Description, and a ANESTHESIA & ANALGESIA

Discussion. The Case Description and Discussion should not exceed 1500 words. Case reports about one or more patients must include a statement that the authors sought, and received, written permission from the patient or patients to report the case. If such permission has not been obtained, this must be disclosed in the case report, as well as the reason for not obtaining patient permission. A case report becomes a research study if the authors intend to publish the outcome at the time they are providing treatment for the patient. The authors should obtain Institutional Review Board approval and written informed consent prior to treating the patient. If that is not possible, then the author should obtain Institutional Review Board approval and patient consent to pursue publication shortly after providing treatment, and in advance of submission to Anesthesia & Analgesia. Interesting but not truly exceptional cases should be submitted as Letters to the Editor. Anesthesia & Analgesia accepts approximately 10% of submitted Case Reports. Echo Rounds are short reports providing a focused discussion of one or more unique or interesting perioperative echocardiographic images (transesophageal, precordial, epicardial, or epiaortic) from a clinical situation in which echocardiography was central to clinical management. Submissions must provide succinct teaching points on echocardiographic views, techniques, or calculations. Teaching points must be supported by the current literature or standard reference texts of echocardiography, preferably those most accessible to the general reader. Echo Rounds should not be construed as “mini-Case Reports” and as such only the most relevant clinical details should be succinctly presented. The suggested format is to present clinical details and specific echo findings in the first third of the report and didactic discussion of the echo topic(s) in the subsequent two-thirds, followed by 2–5 references. The report should be accompanied by 1–2 echocardiographic still images and video clip(s), with legends, which will be available online. The still images should usually, but not always, correspond to the respective video clip(s). Authors should provide appropriate labeling (e.g., arrows, abbreviations of anatomic structures, etc.) of figures and clips (if possible) and may elect to consolidate consecutive time segments into one clip (although adequate viewing time for each segment must be provided to clearly illustrate the primary findings being discussed in the text). Selected reports may benefit from addition of a short table or schematic figure. Authors are advised to examine previously published Echo Rounds (either via the Table of Contents or via the online Echo Rounds database at www.scahq.org or via www. anesthesia-analgesia.org) to avoid submission of topics previously published in this series. See page 227 for video formatting details. Echo Rounds do not include an Abstract and should be 500 –750 words in length. Echo Rounds must include a statement that the authors sought, and received, written permission from Vol. 109, No. 1, July 2009

the patient or patients to report the case. If such permission has not been obtained, this must be disclosed in the text, as well as the reason for not obtaining patient permission. An echo round becomes a research study if the authors intended to publish the outcome at the time they recorded the echocardiographic images. The authors should obtain Institutional Review Board approval and written informed consent prior to producing the images. If that is not possible, then the author should obtain Institutional Review Board approval and patient consent to pursue publication shortly afterwards, and in advance of submission to Anesthesia & Analgesia. A detailed checklist for submitting Echo Rounds is available at http://www.anesthesiaanalgesia.org/misc/EchoRoundsCheckList.doc. Brief Reports describe clinical or laboratory investigations not requiring the breadth of experimentation or documentation expected of a Research Report. Brief Reports require an Abstract, which may be structured or unstructured, depending on the topic (maximum of 100 words). Brief Reports contain an Introduction, Methods, Results, and a very brief (1 paragraph) Discussion. The Introduction, Methods, Results, and Discussion together should not exceed 1000 words. Technical Communications describe instrumentation and analytic techniques. Technical Communications include an unstructured Abstract (maximum of 400 words), and the text of the communication, not exceeding 1500 words. Review Articles synthesize previously published material into an integrated presentation of our current understanding of a topic. Review Articles should describe aspects of a topic in which scientific consensus exists, as well as aspects that remain controversial and are the subject of ongoing scientific disagreement and research. Review Articles are expected to be comprehensive in scope. If the author used a formal strategy to search the medical literature, this strategy should be described. Review Articles should include an unstructured Abstract of ⬍400 words. Review Articles should not exceed 5000 words. A meta-analysis is a formal statistical analysis of an existing body of literature, with the intention of producing new knowledge. A meta-analysis should be written and submitted as a Research Report, not as a Review Article. Medical Intelligence Articles collate and evaluate previously published material to aid in evaluating new concepts or updating old concepts germane to anesthesiology. Medical Intelligence Articles are expected to be highly focused in scope. They should include an unstructured Abstract (maximum of 100 words), and the text of the review, which should not exceed 2000 words. Special Articles are manuscripts not fitting any of the above categories. They are typically invited by the Editorial Board to examine a particular topic. Occasionally, authors produce publishable scholarly texts not fitting the above models. These may be submitted as Special Articles. There are no word limits or rules © 2009 International Anesthesia Research Society

221

for the structure of Special Articles, although they must have an Abstract (maximum of 400 words). Editorials provide editorial perspective on articles published in the Journal or express the general policies or opinions of the Editorial Board. They are solicited by the Editorial Board. Editorials do not have Abstracts and should not exceed 1500 words. Pro/Con Editorials are scholarly discussions of clinically relevant topics providing opposing, wellfounded viewpoints. They are solicited by the Editorial Board. Pro/Con Editorials do not have Abstracts and should not exceed 1500 words. Pro/Con/Core Reviews present a focused review accompanied by expert commentary for and against a specific clinical topic or technique. The Core Review includes an Abstract (unstructured, maximum of 100 words), and the text of the review, which should not exceed 2500 words. It may be accompanied by figures or a video supplement. Pro/Con/Core Reviews are solicited by the Editorial Board. Book and Multimedia Reviews report current literature in perioperative medicine, critical care, and pain management. Publishers interested in having their book or multimedia material reviewed by the Journal should first contact Dr. Paul White, Section Editor for Book, Multimedia, and Meeting Reviews at paul. [email protected] prior to sending the material. All books and multimedia material for review should be sent to Paul F. White, Ph.D., M.D., FANZCA, Department of Anesthesiology and Pain Management, University of Texas Southwestern Medical Center, 5323 Harry Hines Blvd., RM F2.208, Dallas, TX 75390-9068. Book Reviews should not exceed 750 words. Meeting Reports are scholarly outlines of the program and content of a scientific meeting. They may be organized temporally (day by day) or thematically (topic by topic). Authors interested in submitting meeting reports should first contact Dr. Paul White, Section Editor for Book, Multimedia, and Meeting Reviews at [email protected] to confirm that the meeting is of general interest to the readership. Meeting reports do not have Abstracts and should not exceed 1500 words. Focused Reviews summarize recent advances in a particular field with direct application to clinical practice. They are intended to efficiently communicate new knowledge to make clinical practice safer, more efficient, and up-to-date. They are solicited by the Editorial Board. Focused Reviews contain an unstructured abstract, text, and references. They should not exceed 1500 words. Commentaries provide expert perspective on articles or topics published in the Journal. They are typically solicited from reviewers who provide unusually thoughtful insight during the peer review process that should be shared with the Anesthesia & Analgesia readership. They are solicited by the Editorial Board. Commentaries contain only a title page, text and 222

Guide for Authors

references and do not have an Abstract. They should not exceed 1500 words. The Open Mind is a forum for thoughtful, scholarly, and well-referenced reader perspectives. The Open Mind is intended to stimulate discussion. Submissions to The Open Mind must pass the scrutiny of peer review. The Open Mind is not a forum for rants, tirades, or complaints about being overworked and underpaid.8 Submissions to The Open Mind do not have an Abstract and should not exceed 1500 words. Letters to the Editor Submit Letters to the Editor using Editorial Manager (http://aa.edmgr.com). Consider the following points when composing a Letter to the Editor.9 Consent Letters to the Editor about one or more patients must include a statement that the authors sought, and received, written permission from the patient or patients to report the case. If such permission has not been obtained, this must be disclosed in the case report, as well as the reason for not obtaining patient permission. A Letter to the Editor becomes a research study if the authors intend to publish the outcome at the time they are providing treatment for the patient. The authors should obtain Institutional Review Board approval and written informed consent prior to treating the patient. If that is not possible, then the author should obtain Institutional Review Board approval and patient consent to pursue publication shortly after providing treatment, and in advance of submission to Anesthesia & Analgesia. Brevity Letters that respond to a published paper should not exceed 300 words. Long critiques are difficult to follow and will likely generate a response that is also too lengthy. Letters describing an interesting or uncommon clinical experience should be limited to relevant clinical details. Unlike Case Reports, these letters should not delve into the background of diseases or therapeutic interventions. A letter describing a new gadget or technique should not exceed several paragraphs. References should be limited to a few key articles. Focus A letter should address a single issue. It should not discuss an entire subject, but should briefly identify the reason for submission (e.g., a flaw in methodology, relevant observations, or alternative explanation). A letter should be of interest to more than the correspondent and the author of the article in question. Quibbles involving a complex and sophisticated subject or methodology should be settled privately rather than in the Correspondence Section of the Journal. Scientific Accuracy Letters do not necessarily have the imprimatur of external peer review. Nevertheless, scientific accuracy is crucial. If letters deal with complex or arcane issues, they will be reviewed by members of our Editorial ANESTHESIA & ANALGESIA

Board or, occasionally, outside reviewers, especially when letters propose a new idea or methodology. Tone Letters must be respectful. Letters that attack authors, the Journal, or our readership will not be published. Letters that are self-promoting will also not be published. Just as we discourage authors of peerreviewed articles from claiming to be first to make an observation, we similarly are not interested in letters claiming prior publication of an observation, although we will publish letters to correct the record if we believe that the claim is meritorious. Timeliness A letter written in response to a published paper should be submitted no later than 4 months after the paper has been published. A longer interval detracts from the interest, relevance, and “punch.” Writing All letters are edited and occasionally completely rewritten, to be highly focused, readable, and succinct. Accepted letters may or may not be forwarded to the author to approve the edited text. Conflict of Interest Conflict of interest disclosure is required for all submissions to the Journal, including letters.

includes general principles applicable to many types of investigations. • Follow these rules when composing your manuscript: 䡩 Create your manuscript using Microsoft Word or a compatible program. 䡩 Please use “Standard US Paper” or “Letter” page format (width of 8.5 inches or 21.4 centimeters, length of 11 inches or 27.7 centimeters) for your manuscript before uploading the document to Editorial Manager. 䡩 Double-space all text, including references and table and figure legends. 䡩 Begin each part (title page, abstract, text, acknowledgments, references, tables, and legends) on a new page by inserting a page break before each part. 䡩 Number pages consecutively, beginning with the title page. • A Transfer of Copyright form is required only for manuscripts accepted for publication. Please do not send a Transfer of Copyright until you receive a letter of acceptance from the Editorial Office. The form can be downloaded from http://www. anesthesia-analgesia.org/misc/copyrighttransfer.pdf.

General Guidelines and Set-up Instructions Authors are encouraged to follow these guidelines carefully, which will improve the timeliness and quality of the review process. The Editors of Anesthesia & Analgesia may return manuscripts to authors without peer review if the manuscripts do not conform to the Journal guidelines.

The Editorial Office has prepared a series of templates in Microsoft Word format that can be downloaded and used for manuscript preparation. Each template includes the appropriate formatting defaults, instructions for the type of manuscript being submitted and a checklist for manuscript submission. The instructions and checklist should be deleted before submitting the manuscript electronically. Templates exist for

• Follow the specifications in Uniform Requirements for Manuscripts Submitted to Biomedical Journals, as updated in October 2008, available at http://www.icmje.org. • Carefully think through the overall organization of your manuscript. Make sure you prepare all parts. Follow the guidance given in the subsections below to prepare each part. • Write clearly. Be straightforward, unambiguous, and succinct, as described in Strunk and White’s The Elements of Style3 (see http://www.bartleby. com/141/). • Follow the technical styles found in these texts: 䡩 Scientific Style and Format: The CSE manual for Authors, Editors, and Publishers. 7th ed.4 See http://www.councilscienceeditors.org/ publications/style.cfm 䡩 American Medical Association. Manual of Style. 10th ed.5 See https://catalog.ama-assn.org/ Catalog/product/product_detail.jsp?productId⫽ prod1020022?checkXwho⫽done • Clinical trials should be presented in accordance with the CONSORT statement (http://www. consort-statement.org). The CONSORT statement Vol. 109, No. 1, July 2009

• Research Reports (http://www.anesthesia-analgesia. org/templates/general.doc) • Case Reports (http://www.anesthesia-analgesia. org/templates/case.doc) • Echo Rounds (http://www.anesthesia-analgesia. org/templates/echo.doc) • Brief Reports (http://www.anesthesia-analgesia. org/templates/brief.doc) • Technical Communications (http://www.anesthesiaanalgesia.org/templates/technical.doc) • Letters to the Editor (http://www.anesthesiaanalgesia.org/templates/letter.doc) Title Page All submissions require a title page. Please create the title page of your manuscript by going to http://www.aaeditor.org/Authors/home.html and completing the form. A complete title page will be generated, which you must paste into your manuscript file. The title page generated by the web site will contain the following elements: Title of the article: Be concise but informative. Include species when appropriate. © 2009 International Anesthesia Research Society

223

Short Title: An abbreviated title of no more than 60 characters, including letters and spaces. The short title is used on the Journal’s cover, and also appears in the footer of the published article. List of Authors: First name, middle initial, and last name of each author, with highest academic degree(s) and e-mail address, for each author. Each author must: • Indicate his or her Institution at the time the work was performed. If the author has moved since the work was performed, the current institution may appear in parentheses, e.g., (current affiliation: Columbia University). • Disclose his or her role in the manuscript. Identified roles include study design, conduct of the study, data collection, data analysis, and manuscript preparation. • Attest to having approved the final manuscript. Additionally, for research reports, brief reports, and technical communications involving more than one author, at least two authors must attest to having reviewed the original study data and data analysis. One author must also be designated as the archival author, who is responsible for maintaining the study records. • Disclose all conflicts of interest, or indicate that no conflict of interest exists. All relationships between authors and any company or organization with a vested interest in the outcome of the study should be disclosed, including current or previous relationships with a potential interest in the outcome of the research project. More information on conflict of interest can be found on page 220. Authors must contribute intellectually to the work, as described in the Uniform Requirements for Manuscripts Submitted to Biomedical Journals: Writing and Editing for Biomedical Publication, updated February 2006 (see http://www.icmje.org). Attributing authorship to those who have not contributed intellectually is not acceptable. For example, it is unacceptable to include senior members of a research group, chairs of academic departments, or representatives of the commercial sponsor, if they do not meet the requirements for authorship. It is similarly unacceptable to exclude individuals who meet the requirements for authorship. For example, scientists from a sponsoring company who are involved in the study design, execution of the study, data analysis, and preparation of the manuscript should be coauthors of the paper, with appropriate disclosure. If two authors are to be considered “co-first authors,” this should be identified as a footnote to each co-first author. The footnote will appear in the published paper, but does not appear in PubMed. Name of Department(s) and Institution(s) to which the work should be attributed. Multiple institutions may be listed if appropriate. The National Library of Medicine (PubMed) determines institutional affiliation from the affiliation of the first author. Anesthesia & Analgesia has no control over this process. 224

Guide for Authors

Corresponding Author: Name, address, telephone number, FAX number, and e-mail address of author responsible for manuscript correspondence. Reprints: Name and address of author to whom requests for reprints should be addressed or a statement that reprints will not be available from the author. Funding Statement: The source(s) of funding, including foundations, institutions, pharmaceutical and device manufacturers, private companies, or intramural departmental sources. Abstract Location: The Abstract should appear after the title page(s). Content: Structured Abstracts include Background, Methods, Results, and Conclusions. Structured abstracts should provide enough detail to permit the reader to quickly understand the study and findings. • Background: State the context and purpose of the research, and the hypothesis being tested. • Methods: Define the study subjects or experimental animals, study groups, controls, data collected, primary endpoints, and analytic and statistical methods. • Results: State the number of subjects studied, key findings, and statistical significance. • Conclusions: State whether the hypothesis was proven or not, and the scientific and clinical conclusions drawn from the study. Unstructured Abstracts summarize the article, including salient observations and conclusions. Word Count: Please state the number of words in the Abstract after the Abstract text. Abstract Requirements: Manuscript Type Research Reports Case Reports Echo Rounds Brief Reports Technical Communications Review Articles Medical Intelligence Articles Special Articles Editorials Pro/Con Editorials Pro/Con/Core Reviews Book and Multimedia Reviews Letters to the Editor Focused Reviews

Abstract Type

Word Limit

Structured Unstructured None Structured or unstructured Unstructured

400 100 — 100

Unstructured Unstructured

400 100

Structured or unstructured None None Unstructured None

400

None Unstructured

— 100

400

— — 100 —

Text The text of Research Reports is usually, but not necessarily, divided into the following sections: Introduction, Methods, Results, and Discussion, as described below. ANESTHESIA & ANALGESIA

Introduction • Summarize the background in one or two sentences. • Offer only fundamental background information for the work. • Succinctly state the purpose of the study. • If the study tests a specific hypothesis, state the hypothesis. • Do not review the topic. • The introduction typically should be fewer than 500 words. Methods • State the study’s conformance with the Journal’s requirements for human and animal trials, as described in Responsible Conduct of Research, page 218. • If the study involves neuraxial or perineural drug administration, drug administration in children, or “off-label” use of drugs, please state how the study conforms to the Investigational Drugs guidelines on page 218. If the drug is used “off-label” and an investigator IND was not obtained, this should be stated. If an investigator IND was obtained, please include the IND number. • State the clinical trial registry, registration number, and date of registration if the trial is registered. • Inclusion and exclusion criteria: describe how observational or experimental subjects (patients or experimental animals, including controls) were selected. • Describe methods, materials, devices (manufacturer’s name and city, state, country in parentheses), computer software (including revision numbers), and procedures in sufficient detail so that the experiment can be reproduced by other investigators. If the text and the references provide inadequate detail, include an Appendix or other material that would be published as Supplemental Digital Content. • Disclose molecular structures when describing novel compounds. Structural disclosure may be waived, at the discretion of the Editorial Board, when there is a compelling reason to publish a manuscript before the sponsor is ready to disclose the molecular structure. • Provide references to established methods. • Provide references and brief descriptions for published methods that are not well known. The methods section should be interpretable on its own to a knowledgeable reader, who should not need access to another manuscript to understand yours. • Describe new or substantially modified methods, give reasons for using them, and define their limitations. Vol. 109, No. 1, July 2009

• Identify all drugs and chemicals, including generic name(s), dosage(s), and route(s) of administration. Refer to the drugs throughout the text by their generic names, unless the subject of the research is a comparison of branded formulations, in which case the use of the brand name is more precise. • Describe all data handling and statistical methods. • If you use a methodology that you previously reported, it is acceptable to use wording identical to your previous wording. If you are not the author of the previous description of the methodology, then the methodology must be rewritten with reference to the original description of the methodology. • Present methodologies in the same order in which the results are presented. Results • The results are the most important part of the manuscript. • Present results in logical sequence in the text, tables, and illustrations. • Account for all subjects, e.g., number enrolled but not randomized, number withdrawn and for what reasons, etc. • Do not repeat large amounts of material in the text that are also presented in tables or figures. However, commenting on key data from tables or figures is necessary to highlight the main findings. • Emphasize important observations. • In text, tables, and illustrations, present P values as the actual value, rounded to the nearest onehundredth if greater than 0.01 (e.g., P ⫽ 0.04) rather than as an inequality (e.g., P ⬍ 0.05). Inequality may be used in footnotes describing symbols that designate statistical significance in tables and figures (e.g., *P ⬍ 0.05) and when statistical software uses an inequality to report very small P values (e.g., P ⬍ 0.001). • Use consistent rules for presenting numerical results. For example, if a numeric result appears in the abstract, the results, and a table, be certain that it is reported with the same precision in each instance. • In general, determining that the difference between two groups is greater than 0 at P ⬍ 0.05 is not an interesting result. Even the most trivial difference might be statistically significant if enough subjects were studied. The important questions are 1) what are the confidence bounds for the difference between groups, and 2) is the difference large enough to matter scientifically or clinically? Discussion • Discussions should be focused and succinct. • The discussion should not be a comprehensive review of the literature. © 2009 International Anesthesia Research Society

225

• • • • • •

•

• The discussion need not cite every previous study in the field. • The discussion should not exceed 1000 words. • The discussion should not contain inflated claims or product advertisements, e.g., “this new product is conveniently packaged and may transform anesthesia and perioperative medicine.” Claims of being the first to publish a finding are best made in retrospect. Do not claim to be the first to report something. It only invites angry Letters to the Editor. Where possible, structure your discussion in the same order that the results were presented in the Results section. Emphasize new and important aspects of the study and the conclusions that follow. State the limitations of the study, including the limitations of the materials and methods. State how the limitations temper the conclusions. Succinctly relate the observations to other relevant studies. Do not repeat data presented in the Results section, except as required for clarity. In the last paragraph, link the conclusions with goals of the study. If the study tested a hypothesis, state whether the hypothesis was proven, not proven, or the study was inconclusive. Avoid statements and conclusions not completely supported by the data.

• •

• •

•

Tables • Use a separate page for each table. • Double-space each table’s entries. • Do not submit tables as photographs or pasted images. • Tables must be submitted as text in the same manuscript document. • Number the tables consecutively. Each table should have a brief title. Each column in a table should have a brief name. • Use footnotes (not table titles or column headings) for explanatory matter and definitions of abbreviations. Abbreviations must be described with footnotes, even if they are defined in the text or in other tables. • For footnotes, use lower-case italicized letters in alphabetical order. • Cite each table in the text in consecutive order. • If you include a block of data, a table, or a figure from another source, whether published or unpublished, acknowledge the original source. Figures and Illustrations • For useful information on preparing digital art, please review the detailed instructions at http://art.cadmus.com/da/index.jsp. • You are encouraged to read The Visual Display of Quantitative Information by Edward Tufte6 (http:// 226

Guide for Authors

•

www.edwardtufte.com/tufte/books_vdqi), a superb treatise on statistical graphics, charts, and tables. Design figures and illustrations with their published size in mind, i.e., 1 or 2 columns wide. Large figures will be reduced. If Microsoft Excel is used to create figures, use fonts that are clear and appropriately sized for all axis names and labels. In general san-serif fonts (e.g., Arial or Helvetica) are better for tables than serif fonts. Avoid unnecessary use of colors. The default formatting provided with Microsoft Excel is not acceptable for scientific graphics. Number figures consecutively. Supply a brief title for each. Cite figures in the text in consecutive, numerical order. Color figures may be published at no charge at the discretion of the Editor-in-Chief. Authors willing to bear the additional expense of color figures should indicate this in their cover letter. Regardless of whether or not color figures are reproduced in color in the printed edition of Anesthesia & Analgesia, color figures are reproduced in full color in the PDF files that can be downloaded from HighWire Press (www.anesthesia-analgesia. org) at no cost to authors. If a figure has already been published, acknowledge the original source. You must obtain and submit written permission from the copyright holder to reproduce the material when you submit the manuscript for review. Unpublished figures require permission of the author. Permission is required to reproduce any previously published material except for documents or figures in the public domain. Define all abbreviations used in each figure. Repeat definitions of any abbreviations used in subsequent legends.

References • All references must be generally available to readers. Cite references to articles only if they are published in peer-reviewed journals included in the Index Medicus. Unacceptable references include abstracts appearing only in meeting programs or abstracts more than 3 years old. These should be listed as footnotes. Number references consecutively in the order in which they are first mentioned in the text. Double-space between all lines of each reference and between references. • Cite references in text, tables, and legends using superscripted numbers after the punctuation in the order in which the citations appear in the text, tables and figure legends (e.g., Wong et al.1 described . . .). • The titles of journals must be abbreviated according to the style used in Index Medicus. • Verify all references against the original documents or Medline. (http://www.pubmed.gov) ANESTHESIA & ANALGESIA

Table 1. Reference Formats Document type

Example format

Standard journal article (list all authors, do not use “et al”)

Dalal PG, Murray D, Cox T, McAllister J, Snider R. Sedation and anesthesia protocols used for magnetic resonance imaging studies in infants: provider and pharmacologic considerations. Anesth Analg 2006;103:863–8 Zar JH. Biostatistical Analysis. 3rd ed. Upper Saddle River, NJ: Prentice-Hall, 1996 Eger EI II. Uptake and distribution. In: Miller RD, ed. Miller’s Anesthesia. 6th ed. Philadelphia: Elsevier Churchill Livingstone, 2005:131–53 DuPont B. Bone marrow transplantation in severe combined immunodeficiency with a paper unrelated MLC compatible donor. In: White HJ, Smith R, eds. Proceedings of the third annual meeting of the International Society for Experimental Hematology. Houston: International Society for Experimental Hematology, 1974:44–6 Do not use as a reference. May be used as a footnote (3 or fewer) listing the URL and the date it was last accessed by the author, e.g., NIH Request for Applications. Available at: http://grants.nih.gov/grants/guide/ rfa-files/RFA-HL-08-005.html. Accessed May 15, 2008 For 4 or more websites, please create a table and number each listing. In the text, site the table and number, e.g., “Table 1, Ref. 3.”

Books/monographs Book chapter Published proceedings

Website

• Submit copies of “in press” references to Editorial Manager when the manuscript is submitted. • Check the citation list for duplicate entries. • Use the formats of the example references shown in Table 1 above as guides for formatting references. Supplemental Digital Content Supplemental Digital Content (SDC) provides additional material too detailed for inclusion in the manuscript or material not readily presented in printed form. For example, SDCs include audio and video files, spreadsheets, additional figures and tables, appendices, data files, and statistical analysis programming code. If SDC is submitted, be sure to remove all patient identifiers from the material. SDCs should be labeled as to whether they are to be published in the print journal or as an online supplement, or not published and are for reviewers only. Please also cite SDCs in the text along with a very brief description, for example, “Supplemental Video 1, dilated right coronary artery . . .” Because SDC is part of the overall submitted manuscript, make every effort to have the supplement clearly formatted and organized. More detailed instructions can be found online at http://links.lww.com/A142. Authors are urged to share raw data whenever possible. Raw data are invaluable to the community of investigators working to move a discipline forward. Excel spreadsheets are commonly used to share raw data. If authors are not comfortable sharing data online as Supplemental Digital Content, they should indicate as a footnote on the title page which author (if any) can be contacted via e-mail for the raw data. All data shared as a web supplement should be appropriately de-identified to protect patient privacy. Please follow the guidelines below for submitting supplemental video and audio files: Video, including video for Echo Rounds • The preferred video file formats are MPEG-4 (MP4), QuickTime (MOV), and Windows Media Vol. 109, No. 1, July 2009

• • •

•

• •

Video (WMV). Please preview video clips on both Windows and Macintosh platforms to be certain they play correctly. The review process will be delayed if the Editorial Office cannot play the video clip. Deliver still images from video clips in highresolution JPEG or TIFF formats. Individual video clips should not exceed 10 MB. Use video-compression software to reduce video size if necessary. Optimal video frame dimensions are 480 ⫻ 360 pixels and 640 ⫻ 480 pixels. Videos of 320 ⫻ 240 pixels have inadequate resolution for teaching. Video clips are typically 15–25 seconds. Combinations of clips: If several video clips are combined, for example, several TEE echocardiographic loops, please provide adequate time for each segment and leave a suitable gap between the videos. Use appropriate labeling to ensure that the viewer can understand the timing of the pathology and events. Labeling can be added with video editing programs, such as Adobe Premiere or iMovie. Patient Identifiers: Remove all patient identifiers from video clips and still images. For echocardiographic video, please consult Rokey and Vick. Masking Personal Health Information on Real-time Echocardiographic Images.7

Audio • Submit audio files in WAV or MP3 formats. Units of Measurement Anesthesia & Analgesia serves an international audience. For this reason, Syste`me International (SI) units are preferred. We recognize that authors and readers unfamiliar with SI units have difficulty interpreting them. Authors may make undetected errors if they convert their measurements to SI units. To minimize the chance of © 2009 International Anesthesia Research Society

227

conversion errors, authors should submit manuscripts using the units of measurement used in the study, or the units that are used clinically at the author’s institution. These are the units that will appear in the published manuscript. Readers may readily convert published units to units of their choice using commonly available conversion tables. Anesthesia & Analgesia provides a spreadsheet for unit conversion, available at http:// www.anesthesia-analgesia.org/misc/units.xls. A more complete conversion table can be found in the American Medical Association Manual of Style, A Guide for Authors and Editors, Chapter 15: Units of Measure, Table 4: Conversions from Conventional Units to Syste`me International (SI) Units, pp 486 –503. (10) Abbreviations AAWP. Avoid Abbreviations Whenever Possible. The added length of spelling out words is more than compensated for by the increased readability when words are spelled out. Idiosyncratic abbreviations make text particularly difficult to read. Abbreviations widely used within a narrow discipline can make a manuscript uninterpretable to the interested reader from outside that discipline. Do not create new or unusual abbreviations. For example, if the paper refers to the paw pressure test, just call it the paw pressure test throughout the paper, not the PPT. At first mention of an abbreviated term in the abstract, text, each figure legend, and each table, spell out in full and follow immediately with the abbreviation (enclosed within parentheses). For subsequent uses of the term in the same section, use only the abbreviation, without parentheses. Write as you speak. An electrocardiogram might be called an ECG, or EKG, so it is acceptable to abbreviate it as ECG or EKG (after it is spelled out on first use). However, spell out words if there is any possible ambiguity. This will help clarify the manuscript on morphine sulfate kinetics in multiple sclerosis patients with severe mitral stenosis. Consult the following sources for abbreviations: • Scientific Style and Format: The CSE manual for Authors, Editors, and Publishers. 7th ed.4 See http:// www.councilscienceeditors.org/publications/style. cfm • American Medical Association. Manual of Style. 10th ed.5 See https://catalog.ama-assn.org/ Catalog/product/product detail.jsp?productId⫽ prod1020022?checkXwho⫽done

Submitting The Manuscript 1. Go to Editorial Manager at http://aa.edmgr. com, or access the same site from either www. iars.org or www.anesthesia-analgesia.org, where the link can be found under “Authors.” 228

Guide for Authors

2. If you have not previously submitted a manuscript to Anesthesia & Analgesia or reviewed for the Journal, then you must click “Register” to create a username and password, and create an account. Otherwise, just log in with your username and password. What Happens After Submission? Manuscripts are reviewed by the Editorial Office to make certain that the submission contains all parts and is properly formatted. The Editorial Office will not forward manuscripts to the Editor-in-Chief if the manuscript is not complete. An e-mail will be sent to the author describing how to finalize the manuscript. Manuscripts are then forwarded to the Editor-inChief, who makes an initial assessment of the manuscript. If the manuscript does not appear meritorious, or is not appropriate for Anesthesia & Analgesia, the manuscript will be returned with an explanation that it has not been forwarded for external peer review. If the manuscript appears meritorious and appropriate for the Journal, the Editor-in-Chief assigns the manuscript to the appropriate Section Editor, or serves as Section Editor if the manuscript falls within the “General” section of the Journal. Authors are encouraged to suggest the section of the Journal they believe is best suited for their manuscript. The Editorin-Chief considers authors’ suggestions when assigning a manuscript to a section within Anesthesia & Analgesia. The Section Editor makes an initial assessment of the manuscript. The Section Editor determines whether the manuscript is meritorious and verifies that the assignment of the manuscript to his or her section is appropriate. If the manuscript meets these criteria, it is sent for peer review. Acceptance of manuscripts is based on the importance of the findings, originality, and scientific rigor. Reviewers submit their critiques of the manuscript to the Section Editor, using Editorial Manager. The Section Editor drafts an initial decision letter, weighing the views of the reviewers and his or her own impressions of the manuscript. This decision letter is forwarded to the Editor-in-Chief, who may modify the decision. The decision letter is then forwarded to the Editorial Office for final editing and then sent to the author by e-mail. Anesthesia & Analgesia currently accepts about onequarter of the manuscripts submitted, making it among the most selective journals in the specialty. If a manuscript is rejected and the author believes that the reviewers’ critiques can be addressed, the author may appeal the rejection by sending a letter to editor@ anesthesia-analgesia.org. If the author chooses to submit the paper elsewhere, we strongly encourage the author to first revise the manuscript, based on the reviews received from Anesthesia & Analgesia, in order to have the best chance of publication elsewhere. Rejected manuscripts re-submitted without permission will be rejected without further review. ANESTHESIA & ANALGESIA

Sometimes we recommend that a rejected manuscript or Case Report be resubmitted to Anesthesia & Analgesia as a Letter to the Editor. This occurs when the manuscript is not exceptional enough for publication, but contains an interesting observation that our readers would value. If the manuscript or Case Report is rejected, with a recommendation to resubmit as a Letter to the Editor, the manuscript or Case Report must be revised to meet the guidelines for Letters to the Editor. Additionally, the Correspondence Editor is under no obligation to accept a Letter to the Editor submitted at the suggestion of a Section Editor. The letter will be evaluated on its own merits. Authors can expect an initial decision on submitted manuscripts or letters within 6 weeks. Nearly all accepted manuscripts undergo several rounds of revision and copy-editing to produce the best possible published paper. Once the Section Editor decides that a manuscript is ready for publication, a provisional acceptance letter is forwarded to the Editor-in-Chief. If the Editor-inChief accepts the recommendation, then the manuscript is reviewed by the Editorial Office for proper English and clear writing. The Editorial Office may return the manuscript to the Section Editor or Editorin-Chief if portions are incomprehensible. The Editorial Office may contact authors for clarifications about unclear text or references. If you receive an e-mail from the Editorial Office, please respond immediately, as your manuscript will not be accepted until it has been approved by the Editorial Office. Once the Editorial Office approves the final text, the author will receive a final “acceptance” letter from the Journal, and the manuscript will be forwarded to the publisher. The publisher further edits the manuscript and prepares a “galley proof” of the typeset manuscript. You will receive this from the publisher by e-mail link. It is essential that you carefully review your galley proof, including the spelling of the names and affiliations of every author. Manuscripts are in the publication queue when the galley proof is created, so please review your galley proof within two working days to ensure that any corrections are caught before the Journal is printed. Errors in printed manuscripts are almost always present in the galley proof, and would have been caught before publication if the author had carefully reviewed the galley proof. The galley proof will also include “author queries,” which must be answered. The usual time from acceptance to publication is 4 – 6 months. This can be as short as 2 months for very brief communications, and as long as 1 year for manuscripts included in a collection of related papers.

Academic Misconduct The Editorial Board of Anesthesia & Analgesia adheres to the Committee on Publications Ethics (COPE) Code of Conduct for Editors of Biomedical Journals and its Guidelines on Good Publication Practice. See Vol. 109, No. 1, July 2009

http://publicationethics.org/ for details. The US Public Health Service’s Office of Research Integrity has devoted a considerable amount of effort to help institutions and authors understand responsible conduct of research. We strongly recommend that authors utilize this excellent resource, available at http:// ori.dhhs.gov/. We will limit our discussion to just three areas of academic misconduct: plagiarism, duplicate publication, and data falsification. Plagiarism is the use of previously published material without attribution. There is an excellent summary of what constitutes plagiarism at http://www.indiana. edu/⬃wts/pamphlets/plagiarism.shtml. The ability to rapidly search text fragments using Google and similar search tools on the Internet makes plagiarism very easy to identify. Self-plagiarism is the use of your own previously published material without attribution. This is a common practice when a laboratory frequently uses the same methodology. In this setting, the description of an assay or an analysis technique may be identical in multiple papers. This is acceptable. However, with this sole exception, manuscripts that plagiarize previously published material, even if it is the author’s own work, will be rejected if identified during peer review and will be retracted if the plagiarism is discovered after publication. Authors uncomfortable with writing in English occasionally use sentences from a published manuscript simply to obtain grammatically correct text. This is still plagiarism. This is generally discovered during the review process and will result in rejection of the submission and possible sanction. Authors uncomfortable with writing in English are strongly encouraged to ask for editing help from colleagues proficient with scientific English. Duplicate publication is prior publication of a manuscript with considerable content overlap by the same author or co-authors. Prior publication may be in the same language, or it may be a translation (usually to English from the author’s native language). If a manuscript has been published previously, the submission to Anesthesia & Analgesia will be rejected, unless it has already been published, in which case it will be retracted. The editors of Anesthesia & Analgesia regularly check the “Deja Vu” database for duplicate publications. Authors can access this database at http://spore.swmed.edu/dejavu/. Prior publication of an abstract does not count as duplicate publication. We request that authors inform the Journal when parts of a manuscript have previously been published as an abstract. There is sometimes value in publishing in English an important manuscript previously published in another language. Anesthesia & Analgesia will consider such submissions. However, they must be accompanied by a letter from the copyright holder of the © 2009 International Anesthesia Research Society

229

original publication granting Anesthesia & Analgesia permission to publish the work. Duplicate submission is concurrent submission of a nearly identical manuscript to two journals. Duplicate submissions identified during peer review will be immediately rejected. Duplicate submissions that are discovered after publication will be retracted. Data falsification is any manipulation of data that is not disclosed in the publication. This can include editing data (removing outliers, altering values), creating false data, or misrepresenting data analysis (e.g., describing an intent-to-treat analysis but actually performing a per-protocol analysis). Any manuscript in which the data have been falsified will be rejected if the falsification is discovered during peer review. If the manuscript has been published, it will be retracted. Anesthesia & Analgesia reviews all allegations of academic misconduct. This review includes inquiries to the author for clarification. If inquiries to the author do not generate satisfactory replies, we will request that the author’s institution assess the facts. Anesthesia & Analgesia will cooperate fully with any institutional inquiry into allegations of academic misconduct and will provide the inquiry board with copies of all submissions and correspondence. The conclusion of the author’s institution is not binding on Anesthesia & Analgesia. Sanctions against authors range from requesting a Letter to the Editor acknowledging the error and voluntarily withdrawing a manuscript, to a lifetime ban on publication in Anesthesia & Analgesia. Expectations of Authors The Journal has an overriding interest in research integrity. To fulfill this obligation, all authors must fulfill the following expectations: 1. All authors must attest to having reviewed and approved the final manuscript. 2. For research reports, brief reports, and technical communications a single author must be designated as the archival author. The archival author is responsible for maintaining the study records, and will be contacted should questions about the data arise following publication. 3. For research reports, brief reports, and technical communications with more than one authors, at least two authors must attest to a. Having seen the original study data. b. Having reviewed the data analysis. 4. Authors are expected to participate in any external review that assesses the integrity of their research. Failure to fully cooperate with any inquiry into the integrity of their research may result in retraction of any papers related to the research by Anesthesia & Analgesia. 5. Authors must be prepared to make original research records, informed consents, IRB approvals, spreadsheets, and analyses available for 230

Guide for Authors

review by the Journal should questions arise during or after the peer review process about the integrity of the data or the ethical conduct of the research. The Journal acknowledges that authors may need to first redact protected healthcare and proprietary business information from research records to protect patient and organizational confidentiality. 6. Authors understand that the Journal may make available to an authors’ institution any communications between the author and the Journal if requested to do so by a review panel investigating possible academic misconduct under the auspices of the institution. 7. Authors understand that journal editors may share information about academic misconduct with the editors of other journals, as misconduct frequently involves more than a single journal.

Conclusion Anesthesia & Analgesia exists for the benefit of current and future patients under the care of health care professionals engaged in the disciplines broadly related to anesthesiology: perioperative medicine, critical care, and pain management. The Journal furthers the care of these patients by reporting the fundamental advances in the sciences of these clinical disciplines and by documenting the clinical, administrative, and educational advances that guide therapy. The Journal thus seeks a balance between outstanding basic scientific reports and definitive clinical and management investigations. The Journal welcomes original manuscripts reflecting rigorous analysis, even if unusual in style and focus. Anesthesia & Analgesia accepts a limited number of the manuscripts submitted for publication. However, the Journal is genuinely honored by every submission. In exchange for authors’ following the Guide for Authors, the Journal promises to consider every manuscript thoughtfully. In addition, the Journal promises to treat all authors with the respect and dignity they have so thoroughly earned by their dedication to improving the health and well-being of patients. Addendum: Many members of the Editorial Board of Anesthesia & Analgesia also serve on the editorial boards of other journals. Anesthesia & Analgesia acknowledges the contribution of these editorial boards to these guidelines through our overlapping editors. Neither Anesthesia & Analgesia nor the International Anesthesia Research Society (IARS) wishes to claim ownership of the principles espoused in these guidelines. The IARS hereby grants societies, journals, and individuals the right to paraphrase or quote verbatim sections of any length from these guidelines without attribution. REFERENCES 1. Tobin JR, Shafer SL, Davis PJ. Pediatric research and scholarship: another Gordian knot? Anesth Analg 2006;103:43– 8


2. Schultheis LW, Mathis LL, Roca RA, Simone AF, Hertz SH, Rappaport BA. Pediatric drug development in anesthesiology: an FDA perspective. Anesth Analg 2006;103:49 –51 3. Strunk W Jr, White EB. The Elements of Style. 4th ed. New York: Allyn & Bacon, 2000 4. Council of Science Editors, Style Manual Committee. Scientific Style and Format: the CSE manual for authors, editors, and publishers. 7th ed. Reston (VA): The council; 2006 5. JAMA and Archives Journals, Annette Flanagin (Editor), Stacy Christiansen (Editor) American Medical Association manual of style: a guide for authors and editors. 10th ed. Oxford University Press USA, 2007

Vol. 109, No. 1, July 2009

6. Tufte ER. The Visual Display of Quantitative Information. 2nd ed. Cheshire: Graphics Press, 2001 7. Rokey R, Vick GW 3rd. Masking personal health information on real-time echocardiographic images. J Am Soc Echocardiogr 2005;18:970 – 8 8. Shafer SL. Changes for 2009. Anesth Analg 2009;108:1–2. 9. Saidman LJ. Let us hear from you. Anesth Analg 2006; 103:1347– 8 10. Manual of Style, A Guide for Authors and Editors. 9th ed. Baltimore: Lippincott Williams & Wilkins, 1998


231

Pain Medicine Section Editor: Spencer S. Liu

Intravascular Flow Patterns in Transforaminal Epidural Injections: A Comparative Study of the Cervical and Lumbar Vertebral Segments Do Wan Kim, MD Kyung Ream Han, MD Chan Kim, MD Yun Jeong Chae, MD

BACKGROUND: Transforaminal epidural injection (TEI) is commonly used in the treatment of radicular pain. However, there have been many published cases of serious complications after a TEI, occurring most often in cervical levels. One of the presumptive reasons for this complication is inadvertent intravascular injection. We sought to identify the incidence of intravascular injections in cervical and lumbar spinal segments during TEI. METHODS: All patients with radicular symptoms or herpes zoster-associated pain underwent cervical and lumbar TEIs (LTEIs) prospectively by one of the authors. After an ideal needle position was confirmed by biplanar fluoroscopy, 3 mL of a mixture containing nonionic contrast and normal saline was continuously injected at the rate of 0.3– 0.5 mL/s with real-time fluoroscopic visualization. RESULTS: One hundred eighty-two TEIs were performed. Fifty-six cases (30.8%) showed intravascular spreading patterns, 45 cases occurring during a cervical TEI (CTEI) and 11 during a LTEI. The incidences of simultaneous perineural and vascular injection in cervical and LTEIs were 52.1% and 9%, respectively, and pure vascular flow pattern rates in cervical and LTEIs were 11.3% and 0.9%, respectively. CONCLUSION: The incidence of vascular injection in CTEIs is significantly higher than in LTEIs, suggesting that CTEIs should be performed more cautiously. Furthermore, the vascular injection rate of CTEIs is much higher than that previously reported. This finding suggests the need for a proper volume of contrast injection (3 mL) to detect vascular flow, especially in simultaneous perineural and vascular injections. (Anesth Analg 2009;109:233–239)

E

pidural injections with or without corticosteroid have been used worldwide for the management of acute or chronic pain since the first introduction of steroids into the epidural space in 1952.1 Epidural injection may be classified as being interlaminar, caudal, or transforaminal, depending upon the approach taken to the epidural space. The transforaminal approach is target specific, using the smallest volume of injectate closer to the dorsal root ganglion and facilitating better ventral epidural flow to the involved nerve root complex, compared with other methods.2,3 However, there are many published reports of serious complications after transforaminal epidural injections (TEIs) occurring more frequently at the cervical level than others.2,4 –9 The most serious complications caused by TEI include paraplegia or quadriplegia from spinal From the Department of Anesthesiology and Pain Medicine, Pain Clinic, College of Medicine, Ajou University, Suwon, Korea. Accepted for publication March 20, 2009. Address correspondence and reprint requests to Kyung Ream Han, MD, Department of Anesthesiology and Pain Medicine, Ajou University Hospital, San 5 Won-Cheon Dong Pal-Dal Gu, Suwon 442-721, Korea. Address e-mail to [email protected]. Copyright © 2009 International Anesthesia Research Society DOI: 10.1213/ane.0b013e3181a826db

Vol. 109, No. 1, July 2009

cord infarction and coma, death or disability from infarction in the distribution of the posterior cerebral circulation, including brainstem and/or cerebella infarction. Causes of these complications have been explained by vasospasm, direct vascular trauma, and an embolus made of particulate steroids, which are related to intravascular injections.2,4 –9 The purpose of this study was to identify the incidence of intravascular injection in cervical and lumbar spinal segments during a TEI, using a mixture of contrast media and saline in the same volume of injectate as treatment drug with real-time fluoroscopic guidance.

METHODS The study was approved by the IRB at Ajou University Hospital. From November 2007 to March 2008, 182 consecutive patients who were scheduled to undergo TEI were enrolled. The inclusion criteria of this study were patients with radicular pain caused by herniated nucleus pulposus, herpes zoster-associated pain, spinal stenosis, postlaminectomy syndrome, and other conditions, including complex regional pain syndrome, central pain, and sprain. Patients who had nonradicular pain, known allergies to contrast media, or coagulopathy were excluded. All patients 233

Figure 1. Right anterior oblique radiography demonstrates a needle in a position to touch the upper part of the C7 superior articular process at the initial step of the procedure (A). Anterior posterior (AP) radiography shows the needle placed at the lateral margin of the C6 –7 articular pillar (B). The needle is placed in the middle of the pillar and inferolateral to the C6 pedicle on the AP view and contrast mixture spreads into epidural space and C7 nerve root (C). Contrast mixture spreads up to C4 vertebral segment on the oblique view (D). were provided with an explanation of the purpose of this study, and an informed consent was obtained. All procedures were performed by one of the authors under fluoroscopic guidance with contrast enhancement. Patients were prepared and draped in a sterile fashion on a fluoroscopic table. For cervical TEI (CTEI), we preferred patients to be in a lateral oblique position, and the fluoroscope (OEC® series 9800, G.E., Farifield, CT) was adjusted to get the proper oblique view for showing the biggest neural foramen. The patients were placed in a lateral oblique position with the injected side of the body tilted about 30° anteriorly. To obtain the best oblique view of the cervical intervertebral foramen, a true lateral view was performed first, and then the fluoroscope was rotated to a 35°– 45° anterior oblique position on the side to be injected, in which the pedicle was in the upper half of the height and anterior half of the width of the vertebral body, and the superior articular process was placed in the lower half of the posterior intervertebral foramen (Fig. 1A). With an optimal oblique view, a 234

Vascular Uptake in Transforaminal Epidural Block

needle (8 cm, 21 gauge, Neurotic Nerve Block Needle®, Hakko, Japan) was inserted by targeting the posterior and upper one third of the superior articular process. Once a needle touched the superior articular process (Fig. 1B), the needle was advanced 1–2 mm slightly anteriorly and medially. At this point, an anterior posterior (AP) view of the fluoroscopic image was needed to evaluate the depth of the needle. While frequently checking the AP view under live fluoroscopic visualization, the needle was slowly advanced until it reached the lateral portion of the pedicle from the AP view. The destination of the needle on the AP view was the point that was not beyond the midline of the articular pillar (Fig. 1C). For the lumbar TEI (LTEI), the fluoroscope was adjusted to get a proper oblique view while the patient was in a prone position. To obtain the best oblique view of the lumbar intervertebral foramen, a true AP view was made first and the fluoroscope was rotated toward the side to be injected approximately 30°. A 21-G needle was advanced to the pars interarticularis and into the neural foramen. ANESTHESIA & ANALGESIA

Figure 2. Anteriorposterior (AP) radiographies demonstrate the three patterns of cervical transforaminal epidural injections (TEIs). (A) Left C6 nerve root and epidural spreading of contrast agent (perineural pattern). (B) Right C7 nerve root and epidural spreading of contrast agent and vascular flow away from the spinal canal (arrows) (simultaneous perineural and peripheral vascular pattern). (C) Right C5 nerve root and epidural spreading of contrast agent with vascular flow toward the midline of vertebral column (arrow) (simultaneous perineural and central vascular pattern). (D) Vascular spreading away from the spinal canal without flow to perineural structure in a moment (arrow) and then the vascular flow washes out completely immediately after contrast medium injection (purely vascular pattern). After the ideal needle position at all levels was confirmed by biplanar fluoroscopy, 1 min was allowed to pass to detect spontaneous bleeding through the needle hub. Subsequently, gentle aspiration from the needle was done, and 3 mL of a mixture (2 mL of 370 mgI/mL of nonionic contrast media, iopamidol [IOPAMIRO®, Bracco s.p.a., Millano, Italy] and 1 mL of normal saline) was continuously injected at the rate of 0.3– 0.5 mL/s under a real-time fluoroscopic visualization to detect any intravascular injection. After fluoroscopically confirming a nonvascular injection, 3 mL of 0.3% mepivacaine with 10 mg of triamcinolone was injected. We defined contrast patterns as perineural, simultaneous perineural and peripheral vascular, simultaneous perineural, and central vascular and purely vascular flow (Fig. 2). Only perineural pattern was Vol. 109, No. 1, July 2009

defined as contrast agent spread along the neural sheath without any flow disappearance (Fig. 2A). Simultaneous perineural and vascular pattern was defined as contrast agent spreading along the neural sheath with partial contrast flow disappearance in any direction (Figs. 2B and C). If vascular flow occurred along with perineural contrast spread, we identified the flow direction as going to the periphery (Fig. 2B) or toward the midline of vertebral column (Fig. 2C). Only vascular pattern was defined as contrast agent spreading out through the vascular channel first (Fig. 2D), and then it was washed out completely and immediately after showing vascular uptake. Data were collected, including patient’s age, gender, diagnosis, and duration of symptoms. We classified the patients into two age groups of younger than and older than 65 yr. © 2009 International Anesthesia Research Society

235

Table 1. Characteristics and Overall Radiologic Results of the Study Patients Age (yr) Sex (M/F) Height (cm) Weight (kg) Pain site (Rt/Lt) Disease HNP HZAP SS PLPS Others Duration of Sx (mo) Radiologic pattern Perineural Vascular Perineural ⫹ pVascular Perineuroal ⫹ cVascular Vascular only

Cervical

Lumbar

P

52.5 ⫾ 13.7 37/34 162.8 ⫾ 9.4 63.2 ⫾ 11.6 29/42

55.2 ⫾ 14.5 41/70 160.0 ⫾ 8.9 59.4 ⫾ 10.6 52/59

0.061 0.413 0.135 0.043* 0.943 0.093

36 (50.7) 17 (24.0) 5 (7.0) 4 (5.6) 9 (12.7) 5.6 ⫾ 9.8

58 (52.3) 27 (24.3) 11 (9.9) 7 (6.4) 8 (7.2) 4.7 ⫾ 7.2

26 (36.6) 45 (63.4) 26 (36.6)

100 (89.1) 11 (9.9) 9 (8.1)

0.000†

11 (15.5)

1 (0.9)

0.001‡

8 (11.3)

1 (0.9)

0.006‡

0.174 0.000†

Rt ⫽ right; Lt ⫽ left; HNP ⫽ Herniated nucleus pulposus; HZAP ⫽ herpes zoster-associated pain; PLPS ⫽ postlaminectomy pain syndrome; perinerual ⫹ pVascular ⫽ perineural ⫹ peripheral vascular; perineural ⫹ cVascular ⫽ peirneural ⫹ central vascular. * Independent sample t-test. † ␹2 test. ‡ Fisher’s exact test.

Before performing the study, we hypothesized that the estimated probability of vascular injection in LTEI and CTEI was 10% and 30%, respectively, after reviewing related articles and our experience. Based on a Type 1 error level of 0.05, Type 2 error level of 0.2, and a two-sided test, we needed 59 patients for each treatment group. Seventy-one CTEI and 110 LTEI patients were consecutively allocated. All statistical analyses were performed with SPSS version 13 (SPSS, Chicago, IL). Student’s t-test was used to calculate statistical differences in continuous variables, and ␹2 test and Fisher’s exact test were used for comparing categorical variables. A logistic regression was performed to identify factors associated with the vascular injection.

RESULTS This study included 182 TEIs performed in 71 cervical and 111 lumbar segments without any complications. The characteristics of study patients are presented in Table 1. There was no significant difference in any characteristic between the cervical and lumbar patients. Approximately half of the patients suffered from a herniated nucleus pulposus and one quarter of them had herpes zoster-associated pain (Table 1). Overall vascular injections, including simultaneous vascular and perineural injections and only vascular injections, were observed in 45 of the 71 CTEIs (63.4%) and 11 of the 111 LTEIs (9.9%). The total intravascular injection rate, including purely vascular and simultaneous perineural and vascular uptake, in the cervical 236


segment was significantly higher (P ⬍ 0.001) compared with the lumbar injections. Fifty-two percentage of the cervical injections resulted in simultaneous perineural and vascular spread, whereas 9% of the lumbar injections showed a dual contrast spreading pattern (P ⬍ 0.001). The purely vascular injection rate in the cervical region was 11.3% (8/71) compared with 0.9% (1/111) in the lumbar region (P ⬍ 0.006). Eleven cases of 71 CTEIs (15%) showed a vascular flow pattern on AP view running to the vertebral column. However, there was a significantly different result in LTEI; there was one case of intravascular uptake (0.9%) running to the midline of the vertebral column (P ⬍ 0.001) (Table 1). Tables 2 and 3 present the incidence of intravascular injections at each of the cervical and lumbar vertebral segments. The size of samples was too small in the individual vertebral levels to make statistically significant conclusions about the incidence of vascular injection at each level. Nevertheless, all C5 TEIs produced both vascular and neural patterns, and 75% of the vascular injections associated with C3 and C6 TEIs resulted in both vascular and neural patterns. A C8 TEI (33.3%) was less likely to result in vascular injections than other CTEIs. In L3 and L4 TEIs, no vascular injections occurred. Flashback of blood upon preinjection aspiration was not observed in 80% (45/56) of the intravascular uptake cases. There was no statistically significant correlation between the observed contrast pattern and subject’s age, gender, diagnosis, and pain site (Table 4).

DISCUSSION The overall incidence of vascular injections during CTEIs and LTEIs was 30.7%, with an incidence of 9.9% at lumbar levels and a much higher incidence of 63.4% at cervical levels. In the lumbar segments, our results are similar to those previously described.10 –12 On the other hand, the incidence of vascular injections associated with CTEIs was much higher than the 19.4% reported by Furman et al.,13 which is the only other report of this kind. One possible reason for the discrepancy could be the volume of contrast solution. Furman et al. used 0.5–2 mL of nonionic contrast for detecting intravascular injection instead of a fixed volume of contrast agent. For the CTEI technique, the use of a small volume of contrast agent (0.5–1 mL) was recommended to visualize the intravascular injection under real-time fluoroscopic guidance,14,15 and Furman et al. did not mention the point at which they stopped contrast injection. Nevertheless, they might have stopped injecting a small volume of contrast agent once they confirmed the contrast flow to spread into the epidural space through the spinal nerve root or vascular fleet. In this study, we adjusted contrast mixture volume, injection speed, and viscosity of pure contrast media to be similar to the real treatment solution. We used 3 ANESTHESIA & ANALGESIA

Table 2. Contrast Patterns in Cervical Vertebral Levels No. of injections Neural only Vascular ⫹ neural Vascular only Total vascular (%)

C3

C4

C5

C6

C7

C8

Total (%)

4 1 1 2 3 (75)

7 4 3 0 3 (42.9)

6 0 6 0 6 (100)

22 5 13 4 17 (77.3)

20 8 10 2 12 (60)

12 8 4 0 4 (33.3)

71 (100) 26 (36.6) 37 (52.1) 8 (11.3) 45 (63.4)

Table 3. Contrast Patterns in Lumbar Vertebral Levels No. of injections Neural only Vascular ⫹ neural Vascular only Total vascular (%)

L1

L2

L3

L4

L5

Total (%)

22 19 2 1 3 (13.6)

5 3 2 0 2 (40)

4 4 0 0 0 (0)

8 8 0 0 0 (0)

72 65 8 0 8 (11.1)

111 (100) 99 (89.1) 10 (9.0) 1 (0.9) 11 (9.9)

Table 4. Factors Contributing to the Vascular Flow Age (⬎65 yr) Sex (male) Height Weight Site (right) Disease HZAP Spinal stenosis PLPS Others Vertebral segment (cervical)

B

SE

OR

0.019 ⫺1.136 ⫺0.017 ⫺0.032 ⫺0.677

0.028 0.979 0.065 0.047 0.548

1.019 0.321 0.984 0.968 0.508

⫺0.296 0.052 1.267 ⫺0.799 3.249

0.782 1.019 1.042 0.753 0.639

0.744 1.053 3.552 0.450 25.760

95% CI 0.965–1.077 0.047–2.189 0.866–1.118 0.883–1.062 0.174–1.486 0.161–3.441 0.143–7.765 0.461–27.51 0.103–1.968 7.363–90.120*

HNP ⫽ herniated nucleus pulposus; HZAP ⫽ herpes zoster-associated pain; PLPS ⫽ postlaminectomy pain syndrome. * P ⬍ 0.001.

mL of contrast mixture in the same volume as the local anesthetics with steroid for the treatment. The reason for using the mixture of contrast media and normal saline instead of pure contrast agent was to reduce its viscosity. In our experience, a small volume of contrast agent is not sufficient to detect the incidence of a TEI-associated intravascular injection, because we observed a simultaneous perineural and vascular flow pattern in some cases only after injecting more than 0.5 mL of contrast mixture. Even though the use of real-time fluoroscopy could detect the incidence of vascular injection more than twice than intermittent fluoroscopy, it could miss the spread of the contrast agent through the tiny vessels.16 In some cases of a simultaneous perineural and vascular spreading, we could detect the vascular flow at the end of injecting the contrast mixture. Therefore, if most of the contrast spread to the perineural structure and little to the vessels, we would not be able to detect the vascular flow. In this study, we used same sized needles and same rate of 0.3– 0.5 mL/s contrast injection, and one of the authors, who has 12 yr of experience in interventional pain management, performed the study procedures. There are several factors that could influence the occurrence of intravascular injection during a TEI: age, Vol. 109, No. 1, July 2009

spinal level of injection, practitioner’s experience, injection technique, injection speed, needle size, etc. Sullivan et al.12 reported that spinal level of injection and needle size did not have any relationship with the incidence of TEI-associated intravascular injections, which physicians with ⬍2 yr of experience had a higher incidence of TEI-associated intravascular injection of TEIs than those with more than 2 yr of experience, and there was a strong trend toward an increase of incidence of intravascular uptake as a patient’s age increased.12 In this study, however, there was no significant difference in intravascular uptake rates between age groups, similar to the studies by Furman et al.10,13 Similar to previous results on CTEIs and LTEIs, in this study, we observed a significant difference in the intravascular injection rates between the cervical and lumbar spine. One of the reasons for the difference is that a different anatomy is involved in the arterial supply of the intervertebral foramina between the cervical and lumbar levels. In general, spinal arterial branches arise from the aorta and iliac vessels at the lumbar and thoracic levels. In the cervical spine, however, spinal arterial branches arise from the vertebral, ascending cervical, superior intercostals, and deep cervical arteries.17 In the CTEIs, the technique used in this study appears to be proper, avoiding vertebral artery puncture by keeping the needletip placed immediately anterior to the portion of the superior articular process, where the posterior portion of the dorsal root ganglion is located.14,15 On the other hand, cervical radicular arteries arise from ascending cervical or deep cervical arteries, which are located in the posterior aspect of the intervertebral foramen. Huntoon18 demonstrated that ascending and deep cervical arterial branches enter the external opening of the posterior intervertebral foramen near the classic target area for TEIs, and that those arteries could possibly be cannulated during the CTEI procedure because of their large size at the external foraminal openings. They concluded that there are variable © 2009 International Anesthesia Research Society

237

anastomoses between the vertebral and cervical arteries, and that the posterior, caudal aspect of the foramen is the most vulnerable area, among the arteries supplying the spinal cord, to be punctured. According to Huntoon,18 there is no safer area of the cervical intervertebral foramen to perform TEIs. In this study, two thirds of the CTEIs resulted in vascular flows, regardless of the sizes of the vessels, arteries, or veins. In humans, one or two of the cervical segmental medullary vessels may feed into the anterior spinal artery.18 Among our vascular flows, we could not unravel exactly which vessels were critical for reinforcing the spinal cord. However, we could postulate that 15.5% of cases of CTEIs running to the midline of the vertebral column would be feeding the artery to the spinal cord because of the vascular direction. Among 15.5% of the vascular flow cases in CTEIs, flow patterns supplying the superficial area of the spine might be included because a fluoroscopic view shows only two-dimensional images. On the other hand, we could estimate that 26 cases (36.6%) of CTEIs and nine (8.1%) of the LTEIs showing the perineural and peripheral vascular pattern would be venous flows. The most serious concern about vascular injection in TEIs is that administration of drug into arteries feeding the spinal cord possibly leads to disastrous neurologic complications. An inadvertent venous injection may also result in deceasing the treatment efficacy and central nervous system toxicity. Most intravascular uptake patterns from our data were postulated venous flows, which would be free from catastrophic complications. This study has some limitations. First, we did not differentiate between venous or arterial flow. Although intraarterial injection is manifested by a more rapid clearance of injected contrast agent than venous flow and vascular flow direction to the spinal cord, an exact differentiation might be difficult during contrast injection in TEIs. Using AP view, we evaluated whether vascular flow direction was running to the midline of the vertebral body or toward the periphery. Although vascular flow toward the midline implies arterial flow, this is not certain. Fluoroscopy cannot show threedimensional radiographic images. Recently, threedimensional digital subtraction angiography clearly provided clinicians with better information about the relationships between the feeding arteries and draining veins in the intervetral foramen and helped to discriminate between intramedullary or perimedullary flow.19 However, it has started to be used in diagnosis and performance of the endovascular procedures in spinal vascular malformation but has not yet been used in pain practice. Second, we could not reach statistically significant differences in the intravascular incidences according to the individual spinal segment because of the small sample size. When we conducted statistics, however, our present data showed that TEIs in cervical segment C8 had the smallest number of intravascular injections. 238


Third, we did not analyze the number of times that the needle was repositioned to avoid a vascular injection. In some cases of CTEIs, an intravascular injection could not be avoided despite repeated needle repositioning, as Furman et al.13 have demonstrated. However, we were able to obtain a nonvascular pattern with just one needle repositioning in almost all lumbar vascular injection cases. We have not yet found the definite needle position in CTEIs to avoid a vascular injection. More research and a future study with larger sample size should be performed to analyze the incidence of intravascular injection according to the individual spinal segment and to find the best needle position to avoid a vascular injection. In conclusion, the overall incidence of vascular injection in CTEIs is significantly higher than in LTEIs and also much higher than previously reported. Majority of the intravascular uptake in TEIs are postulated to be venous; however, 15.5% of the CTEIs should be considered to be possible dangerous arterial injection. This finding might be explained by the fact that an appropriate volume of contrast injection (3 mL) needs to be used to detect vascular flow, especially in simultaneous perineural and vascular injections. Our present observation also suggests that CTEIs should be performed based on careful detection of vascular flow patterns by an experienced practitioner, and further research is needed to find a technique to reduce intravascular injection. REFERENCES 1. Robechhi A, Capra R. L’idrocortisone (composto F); prime esperienze cliniche in campo reumatologico. Minerva Med 1952;98:1259 – 63 2. Baker R, Dreyfuss P, Mercer S, Bogduk N. Cervical transforaminal injection of corticosteroids into a radicular artery: a possible mechanism for spinal cord injury. Pain 2003;103:211–5 3. Batson OV. The function of the vertebral veins and their role in the spread of metastases. Ann Surg 1940;112:138 – 49 4. Scanlon GC, Moeller-Bertram T, Romanowsky SM, Wallace MS. Cervical transforaminal epidural steroid injections. Spine 2007;32:1249 –56 5. Muro K, O’Shaughnessy B, Ganju A. Infarction of the cervical spinal cord following multilevel transforaminal epidural steroid injection: case report and review of the literature. J Spinal Cord Med 2007;30:385– 8 6. Tiso RL, Cutler T, Catania JA, Whalen K. Adverse central nervous system sequelae after selective transforaminal block: the role of corticosteroids. Spine J 2004;4:468 –74 7. Brouwers PJ, Kottink EJ, Simon MA, Prevo RL. A cervical anterior spinal artery syndrome after diagnostic blockade of the right C6-nerve root. Pain 2001;91:397–9 8. Rozin L, Rozin R, Koehler SA, Shakir A, Ladham S, Barmada M, Dominick J, Wecht CH. Death during transforaminal epidural steroid nerve root block (C7) due to perforation of the left vertebral artery. Am J Forensic Med Pathol 2003;24:351–5 9. Houten JK, Errico TJ. Paraplegia after lumbosacral nerve root block: report of three cases. Spine J 2002;2:70 –5 10. Furman MB, O’Brien EM, Zgleszewski TM. Incidence of intravascular penetration in transforaminal lumbosacral epidural steorid injections. Spine 2000;25:2628 –32 11. Smuch M, Fuller BJ, Yoder B, Huerta J. Incidence of simultaneous epidural and vascular injection during lumbosacral transforaminal epidural injections. Spine J 2007;7:79 – 82 12. Sullivan WJ, Willick SE, Chira-Adisai W, Zuhosky J, Tyburski M, Dreyfuss P, Prather H, Press JM. Incidence of intravascular uptake in lumbar spinal injection procedures. Spine 2000;25:481– 6


13. Furman MB, Giovanniello MT, O’Brien EM. Incidence of intravascular penetration in transforaminal cervical epidural steroid injections. Spine 2003;28:21–5 14. Windsor RE, Storm S, Sugar R, Nagula D. Cervical transforaminal injection: review of the literature, complications, and a suggested technique. Pain Physician 2003;6:457– 65 15. Rathmell JP, Aprill C, Bogduk N. Cervical transforaminal injection of steroids. Anesthesiology 2004;100:1595– 600 16. Smuch M, Fuller BJ, Ahiodo A, Beny B, Singaracharlu B, Tong H, Ho S. Accuracy of intermittent fluoroscopy to detect intravascular injection during transforaminal epidural injections. Spine 2008;33:E205–10

Vol. 109, No. 1, July 2009

17. Hoeft MA, Rathmell JP, Monsey RD, Fonda BJ. Cervical transforaminal injection and the radicular artery: variation in anatomical location within the cervical intervertebral foramina. Reg Anesth Pain Med 2006;31:270 – 4 18. Huntoon MA. Anatomy of the cervical interertebral foramina: vulnerable arteries and ischemic neurologic injuries after transforaminal epidural injections. Pain 2005;117:104 –11 19. Matsubara N, Miyachi S, Izumi T, Ohshima T, Tsurumi A, Hososhima O, Kinkori T, Yohida J. Usefulness of threedimensional digital subtraction angiography in endovascular treatment of a spinal dural arteriovenous fistula. J Neurosurg Spine 2008;8:462–7


239

The Analgesic Properties of Scalp Infiltrations with Ropivacaine After Intracranial Tumoral Resection He´le`ne Batoz, MD Olivier Verdonck, MD Christelle Pellerin, MD Gae¨lle Roux, MD Pierre Maurette, PhD

BACKGROUND: The issue of postoperative pain after neurosurgery is controversial. It has been reported as mild to moderate and its treatment may be inadequate. Infiltration of the surgical site with local anesthetics has provided transient benefit after craniotomy, but its effect on chronic pain has not been evaluated. Accordingly, we designed the present study to test the hypothesis that ropivacaine infiltration of the scalp reduces acute and persistent postoperative pain after intracranial tumor resection. METHODS: This was a prospective, single-blinded study. Inclusion criteria were intracranial tumor resection, age ⱖ18 or ⱕ80 yr, and ability to understand and use a visual analog scale (VAS). Exclusion criteria were history of craniotomy, chronic drug abuse, and neurologic disorders. All eligible patients were randomly included in Group I (infiltration) or C (control). Postoperative analgesia was IV acetaminophen combined with nalbuphine. At the end of the surgery, Group I received an infiltration of the surgical site with 20 mL of ropivacaine 0.75%. Acute pain was evaluated hourly by VAS during the first 24 h. The analgesic effect of ropivacaine was evaluated based on total consumption of nalbuphine and VAS scores. The incidence of persistent pain and neuropathic pain was assessed at the 2-mo postoperative evaluation. We used the Student’s t-test to compare total nalbuphine consumption, repeated measures analysis of variance with post hoc Bonferroni t-test for VAS score and the Fisher’s exact test for chronic and neuropathic pain. RESULTS: Fifty-two patients were enrolled, 25 in Group I and 27 in Group C. Demographic and intraoperative data were similar between groups. Group I showed a nonsignificant trend toward reduced nalbuphine consumption during the first postoperative day, 11.2 ⫾ 9.2 mg vs 16.6 ⫾ 11.0 mg for Group C (mean ⫾ sd, P ⫽ 0.054). VAS scores were significantly higher in Group C. Two months after surgery, persistent pain was significantly lower in Group I, 2/24 (8%) vs 14/25 (56%), P ⫽ 0.0003. One patient (4.1%) in Group I versus six (25%) patients in Group C (P ⫽ 0.04) experienced neuropathic pain. CONCLUSIONS: Because pain is moderate after intracranial tumor resection, there is limited interest in scalp infiltrations with ropivacaine in the acute postoperative period. Nevertheless, these infiltrations may be relevant for the rehabilitation of neurosurgical patients and their quality of life by limiting the development of persistent pain and particularly neuropathic pain. (Anesth Analg 2009;109:240 –4)

N

eurosurgery often causes minimal pain after surgery.1,2 However, it has been reported that up to 60%– 80% of patients may experience moderate to severe pain after craniotomy,3 mainly because this postoperative pain is underestimated and inadequately treated.4 Perfect anesthetic management after neurosurgery includes not only a fast recovery so that early neurological examination and complications can

From the Department of Anesthesiology and Intensive Care Unit III, Pellegrin Hospital, Bordeaux, France. Accepted for publication January 19, 2009. No financial support was needed to achieve this work. Address correspondence and reprint requests to Olivier Verdonck, MD, Department of Anesthesiology III, Pellegrin Hospital, Place Amelie Raba León, 33076 Bordeaux, France. Address e-mail to [email protected]. Copyright © 2009 International Anesthesia Research Society DOI: 10.1213/ane.0b013e3181a4928d

240

be detected, but also optimal postoperative comfort. The use of an ultrashort context-sensitive half-life opiate, such as remifentanil, has become the standard to achieve rapid recovery. Nevertheless, its use can lead to excessive postoperative pain and might induce pain sensitization.5 Acute postoperative pain management should, therefore, be based on the pharmacokinetics of the opiate used during the intraoperative period and should be started before the end of anesthesia.6 Although inadequate pain control is associated with many negative physiologic and psychological consequences, morphine is often avoided because of its sedative and respiratory depressing effects. However, acetaminophen alone is not powerful enough to control pain after craniotomy, and other analgesic drugs are usually required.7 Some authors have studied the efficacy of scalp infiltrations with local anesthetics after craniotomy and found beneficial effects Vol. 109, No. 1, July 2009

on acute postoperative pain.8 Neurosurgery can also lead to the development of chronic pain. Up to 34% of patients report chronic pain after craniotomy.9 Several predictive factors have been described by de Gray and Matta10; one factor is pain intensity during the acute postoperative period. Given these two points, ideal analgesic management after neurosurgery should, in all cases, be efficient for controlling both acute and chronic postoperative pain. The aim of this study was to assess the efficacy of scalp infiltration with ropivacaine for reducing the intensity of acute pain, and the incidence of persistent pain, after craniotomy.

METHODS This was a prospective, randomized, single-blinded, controlled study performed between June 2006 and April 2007. Our study protocol was approved by the IRB of our institution. After written informed consent, all patients scheduled for elective craniotomy for excision of an intracranial tumor were enrolled. Eligible patients were between 18 and 80 yr old and ASA physical status I-III. We excluded patients presenting one of the following: pre- or postoperative aphasia, neurological disorders preventing a good understanding of the protocol and the visual analog scale (VAS), use of narcotic analgesics, history of chronic pain, alcohol or drug abuse, stroke or neurosurgery, suspicion of high intracranial pressure, Glasgow coma scale (GCS) ⬍15, or pregnancy. Our primary end point was nalbuphine consumption. Based on a study conducted by Verchere et al.,7 we found that two groups of 25 patients would be required to assess a difference of 6 mg of nalbuphine during the first 24 h after surgery with an 80% power and an ␣ risk of 0.05. The day before surgery, the study protocol and the VAS were explained to all patients. All patients were given 1.5 mg/kg of hydroxyzine orally the night before and 2 h before surgery. Other medications, such as corticosteroids, were prescribed at the discretion of the neurosurgeons. All patients fasted for at least 6 h before surgery. Patient randomization was conducted before induction of anesthesia, according to a computer-generated table of random number assignments; each patient was blindly assigned to one of the two following groups: patients in Group I (Infiltration) received a scalp infiltration with ropivacaine at the end of surgery, whereas patients in Group C (Control) did not receive any infiltration. Only the surgeon and the anesthesiologist in charge of the intraoperative period were not blinded, whereas the patient, nurses, and anesthesiologist in charge of the postoperative period and pain evaluation were. After 0.05 mg/kg of midazolam given IV, anesthesia was induced and maintained with target-controlled infusions (TCI Orchestra Primea威, Fresenius Kabi, Paris, Vol. 109, No. 1, July 2009

France) of propofol and remifentanil. During surgery, patients received standard monitoring that included invasive and noninvasive arterial blood pressure, continuous electrocardiogram, pulse oximetry, urine output, intravesical temperature, and end-tidal CO2. Depth of anesthesia was controlled using bispectral index monitoring (BIS XP™ Aspect Medical Systems, Newton, MA) installed before induction. The patient’s temperature was kept between 35°C and 37°C with warming blankets. Just before scalp suturing, neurosurgeons infiltrated the wound margins (1/3 in muscle and 2/3 in the subcutaneous tissue) with 20 mL of ropivacaine 0.75% for Group I patients. Group C was not infiltrated. Patients recovered from anesthesia in the intensive care unit. Tracheal extubation was performed when hemodynamic, respiratory, and neurologic evaluations were satisfactory. Postoperative pain was then assessed using a VAS. Analgesia was managed as follows: all patients received acetaminophen 1 g IV 1 h before the end of surgery, then every 6 h. If the VAS score was more than 30/100, patients were given nalbuphine 10 mg IV every 4 h. Patients experiencing nausea or vomiting received ondansetron 4 mg IV, three times a day maximum. Recorded data were demographic criteria (age, height, weight), type of surgery (supra or infratentorial), and incision (temporal, frontoparietal, or occipital), length of surgery and anesthesia, total doses of propofol and remifentanil and time of tracheal extubation (H0). The following measurements and observations were made in the intensive care unit at H0, H1, H2, H4, H6, H8, H12, and H24: GCS score, VAS score, cumulative doses of nalbuphine and ondansetron required, pulse oximetry, mean arterial blood pressure, heart rate, and respiratory rate. At H24, patients were asked about their satisfaction with pain management. A patient satisfaction score was used (0: unsatisfied, 1: moderately satisfied, 2: satisfied, 3: very satisfied). The surgical wound was examined on the second postoperative day to detect infection or hematoma. Two months after surgery, patients were contacted to detect the presence of persistent pain and to characterize its type. Chronic pain was defined as headaches occurring daily10 or abnormal sensations or pain during everyday activities. The DN4 score was used to classify pain as neuropathic.11 DN4 is a French questionnaire based on 10 items grouped in four questions. A positive answer to four or more items is a sign of neuropathic pain. Seven of these items are sensory descriptors, such as the presence of burning or painful cold, and the three others are sensory examination items (e.g., hypoesthesia). If patients had one or more positive answers on the first seven sensory descriptors, they were asked to undergo a medical examination to complete the DN4. Statistical analyses were made using StatView威 5.0 program (SAS Institute, Cary, NC) and EpiInfo威 6.04d © 2009 International Anesthesia Research Society

241

(CDC, Atlanta, USA). All data are expressed and displayed as mean ⫾ sd or frequency distribution tables unless otherwise stated. Continuous data were analyzed using the t-test and the Mann–Whitney U-test as appropriate. VAS scores were compared with repeated measures of analysis of variance with Bonferroni correction as post hoc test. Categorical variables were examined by a ␹2 test or by a Fisher’s exact test in case of small size samples. A probability value of ⬍0.05 was considered to be statistically significant.

Nonetheless, this narcotic can lead to hyperalgesia when inadequate or nonpreemptive analgesia is administered.5 There is controversy regarding the management of postoperative pain after craniotomy.3,12 Inadequate analgesia must be avoided because of the agitation, hypertension, shivering, or vomiting it can cause, whereas the use of morphine can lead to sedation or myosis which can delay the diagnosis of a surgical complication. A study conducted by Verchere

RESULTS

Table 1. Demographic and Intraoperative Data

Fifty-three patients were enrolled between June 2006 and April 2007. One patient was excluded because data were lost. Fifty-two patients completed the study. Of these, 25 received scalp infiltration (Group I) and 27 did not (Group C). Only 48 patients were successfully contacted at 2-mo postoperatively as three patients died and one moved. Demographic data, intraoperative management, and surgical procedures were similar in the two groups (Table 1). The total amount of Nalbuphine administered during the first postoperative day was lower for Group I, but it was not statistically significant (11.2 ⫾ 9.2 mg vs 16.6 ⫾ 11 mg in Group C, P ⫽ 0.054). However, Group I patients had significantly reduced VAS scores during the first 24 postoperative hours (P ⫽ 0.046; Fig. 1). The incidence of postoperative nausea and vomiting was low and similar between groups: 4 patients (16%) in Group I versus 5 (18%) in Group C, NS. The GCS at recovery was 14 or 15 for every patient in both groups. Forty-nine of our patients (94%) were satisfied with pain management when they were asked on the first postoperative day (satisfaction score ⱖ2). There were no cases of local or general complications during the study period. On the second month after surgery, the number of patients suffering from persistent pain was significantly lower in the infiltrated group: 2 (8%) in Group I versus 14 (56%) in Group C, P ⬍ 0.001. The global incidence of neuropathic pain was also significantly reduced from six patients (25%) in Group C to 1 patient (4%) in Group I (P ⫽ 0.04) (Table 2).

DISCUSSION The goal of this study was to evaluate the effects of scalp infiltration with ropivacaine performed at the end of surgery on acute and chronic pain after intracranial tumoral resection. During the first 24 h, acute postoperative pain was assessed with nalbuphine consumption and VAS. Chronic pain was detected and characterized on the second month after surgery. The anesthetic protocol we used included remifentanil. This narcotic is widely used by neuroanesthesiologists for its particular pharmacokinetic properties that allow a rapid recovery from anesthesia, which is essential for detecting postoperative complications. 242

Scalp Infiltrations After Craniotomy

Age (yr) Height (cm) Sex (M) Weight (kg) Duration of anesthesia (min) Duration of surgery (min) Preoperative corticosteroids Gabapentin Remifentanil total (mg) Types of craniotomy (supratentorial)

Group C (n ⫽ 27)

Group I (n ⫽ 25)

55 ⫾ 16 166 ⫾ 8 12 (44%) 67 ⫾ 14 301 ⫾ 76

50 ⫾ 10 168 ⫾ 9 12 (48%) 77 ⫾ 28 335 ⫾ 69

NS NS NS NS NS

134 ⫾ 41

158 ⫾ 55

NS

P

19 (70%)

17 (68%)

NS

10 (38%) 4.32 ⫾ 2.32

7 (29%) 5.70 ⫾ 3.12

NS NS

23 (85%)

20 (80%)

NS

Group C ⫽ Control; Group I ⫽ Infiltration; NS ⫽ not significant.

Figure 1. Representation of means ⫾ se. Visual analog scale scores between groups. Black bars represents the control (C) group and gray bars the infiltration (I) group. Repeated measures analysis of variance shows a statistical difference (P ⫽ 0.046 with Bonferroni correction). H0 ⫽ extubation time.

Table 2. Two-month Postoperative Persistent and Neuropathic Pain (Patients With a DN4 Score Greater than 4/10)

Two-month postoperative persistent pain, n (%) Two-month postoperative neuropathic pain, n (%)

Group C (n ⫽ 25)

Group I (n ⫽ 24)

P

14 (56%)

2 (8%)

0.0003

6 (25%)

1 (4%)

0.04

Group C ⫽ Control; Group I ⫽ Infiltration.


et al.7 showed that postoperative pain cannot be managed with acetaminophen alone and that the addition of nalbuphine was efficient for maintaining a VAS score ⬍30. In this study, scalp infiltration with ropivacaine did not appear to be effective for acute postoperative pain, as we did not reach our primary end point though VAS scores were significantly different between groups. This lack of effectiveness may appear surprising considering the long effect duration of this local anesthetic. Most authors who have assessed the benefit of scalp infiltration on acute postoperative pain have found a transient effect with ropivacaine or bupivacaine lasting 1 or 2 h.8,13 Few studies have evaluated scalp infiltration as transition analgesia when a short-acting narcotic, such as remifentanil, was used. However, our study might have been underpowered because of the small study group size, which could explain the absence of difference. This result could also have been because of the difficulty of obtaining a reliable pain evaluation. The VAS score might not be the most accurate scale for pain evaluation after neurosurgery as some patients can suffer transient confusion. Some studies have found major differences in analgesic consumption with similar VAS scores.8,14 The numerical rating scale could have been a better alternative as it has been successfully used in other studies.15 We did not evaluate the evolution of acute postoperative pain for longer than 24 h, as previous studies showed a significant decrease in pain intensity after this period.4,14 Despite this lack of power concerning the early postoperative period, our results regarding persistent pain are interesting. As anesthesiologists, we rarely meet our patients after the acute postoperative period. We were not aware that chronic pain was such a frequent issue for neurosurgical patients (56% of incidence in our control group). Kaur et al.16 reported a lower rate (17.6%) of persistent headaches after an anterior temporal lobectomy at the 2-mo follow-up visit. However, we asked patients not only whether they were suffering from headaches, but also whether they were experiencing abnormal scalp sensitivity or pain during daily activities, such as hair-brushing or hair-washing. Moreover, we included patients who underwent supra- or infratentorial craniotomy. Base of skull surgery provokes more severe postoperative headaches as reported by Schaller et al.,9 who found 34% of their patients suffering from severe headaches 3 mo after a retrosigmoid craniotomy. Both these differences can explain our much higher rate of 2 mo of persistent pain in our control group. This rate, however, dramatically decreased in the infiltrated group with the incidence decreasing from 56% to 8%. This technique has proven effective during the first 24 h postoperatively.6,8 Nevertheless, our study is the first, to our knowledge, to have assessed the long-term effect of scalp infiltrations on persistent pain. Vol. 109, No. 1, July 2009

A longer follow-up would be required to assess the exact duration of these differences on persistent and neuropathic pain. Postsurgical nerve injury is one of the frequent causes of neuropathic pain.17 The DN4 score is based on a simple questionnaire that has proven its efficacy in the detection of neuropathic pain.11 We used it to characterize the 2-mo postoperative persistent pain. Even if the overall incidence of neuropathic pain was lower in the infiltrated group (4% vs 25%), our small sample size prevents us from discussing a specific effect of ropivacaine on neuropathic pain. This result shows only that both nociceptive/inflammatory pain and neuropathic pain were present at the 2-mo follow-up visit and was reduced with scalp infiltration. Gabapentin is recommended to treat or prevent the development of neuropathic pain and has proven to be effective for acute and chronic postoperative pain, regardless of its neuropathic or inflammatory components.18 –21 It was initially introduced as an antiepileptic drug, especially for partial seizures, which are frequently associated with intracranial tumors. Some of our patients were thus treated with gabapentin before and after surgery. However, we observed the same rate of gabapentin-treated patients in each group (38% in Group C, 29% in Group I, NS) with an equal median dose of 1800 mg a day. Therefore, we did not consider gabapentin a bias in our results. In a review, Kehlet et al.22 explained that prolonged inflammatory pain is more likely to lead to the development of a chronic pain state via central sensitization, and that surgical nerve injury can produce persistent maladaptive neuronal plasticity leading to neuropathic pain. This is why aggressive early pain relief can prevent central sensitization and neuropathic pain after a surgical tissue injury.23 Although we could not demonstrate an effect of ropivacaine infiltration on early postoperative pain in our study, we observed a trend toward less intense pain and smaller analgesic consumption. Other studies have assessed its beneficial effects on the early postoperative period that lasted up to 24 h.6,8 Ropivacaine infiltrations might thus have played a crucial role in early central sensitization. Peripheral sensitization results from the local action of inflammatory mediators, such as prostanoids, that reduce the pain threshold and increase excitability of nociceptor neurons.24 Many studies have assessed the antiinflammatory properties of local anesthetics.25–27 As a consequence, local infiltration of the surgical site with ropivacaine can reduce the inflammatory response and the accompanying peripheral sensitization. Even though the analgesic effects of scalp infiltration are not obvious in our study of the early postoperative period, they might have played a crucial role in the inhibition of peripheral sensitization and, later, of chronic pain. © 2009 International Anesthesia Research Society

243

CONCLUSION Scalp infiltrations with ropivacaine revealed a limited efficacy in decreasing acute postoperative pain after intracranial tumor resection. Nevertheless, our results suggest that its effects are much more pronounced in limiting the development of the chronic pain state, regardless of its inflammatory or neuropathic component, which seems crucial in improving the quality of life and the rehabilitation of neurosurgical patients. REFERENCES 1. Dunbar PJ, Visco E, Lam AM. Craniotomy procedures are associated with less analgesic requirements than other surgical procedures. Anesth Analg 1999;88:335– 40 2. Quiney N, Cooper R, Stoneham M, Walters F. Pain after craniotomy. A time for reappraisal? Br J Neurosurg 1996;10:295–9 3. De Benedittis G, Lorenzetti A, Migliore M, Spagnoli D, Tiberio F, Villani RM. Postoperative pain in neurosurgery: a pilot study in brain surgery. Neurosurgery 1996;38:466 –9; discussion 469 –70 4. Gottschalk A, Berkow LC, Stevens RD, Mirski M, Thompson RE, White ED, Weingart JD, Long DM, Yaster M. Prospective evaluation of pain and analgesic use following major elective intracranial surgery. J Neurosurg 2007;106:206 –16 5. Angst MS, Koppert W, Pahl I, Clark DJ, Schmelz M. Short-term infusion of the mu-opioid agonist remifentanil in humans causes hyperalgesia during withdrawal. Pain 2003;106:49 –57 6. Nguyen A, Girard F, Boudreault D, Fugere F, Ruel M, Moumdjian R, Bouthilier A, Caron JL, Bojanowski MW, Girard DC. Scalp nerve blocks decrease the severity of pain after craniotomy. Anesth Analg 2001;93:1272– 6 7. Verchere E, Grenier B, Mesli A, Siao D, Sesay M, Maurette P. Postoperative pain management after supratentorial craniotomy. J Neurosurg Anesthesiol 2002;14:96 –101 8. Law-Koune JD, Szekely B, Fermanian C, Peuch C, Liu N, Fischler M. Scalp infiltration with bupivacaine plus epinephrine or plain ropivacaine reduces postoperative pain after supratentorial craniotomy. J Neurosurg Anesthesiol 2005;17:139 – 43 9. Schaller B, Baumann A. Headache after removal of vestibular schwannoma via the retrosigmoid approach: a long-term follow-up-study. Otolaryngol Head Neck Surg 2003;128:387–95 10. de Gray LC, Matta BF. Acute and chronic pain following craniotomy: a review. Anaesthesia 2005;60:693–704 11. Bouhassira D, Attal N, Alchaar H, Boureau F, Brochet B, Bruxelle J, Cunin G, Fermanian J, Ginies P, Grun-Overdyking A, JafariSchluep H, Lanteri-Minet M, Laurent B, Mick G, Serrie A, Valade D, Vicaut E. Comparison of pain syndromes associated with nervous or somatic lesions and development of a new neuropathic pain diagnostic questionnaire (DN4). Pain 2005;114:29 –36

244

Scalp Infiltrations After Craniotomy

12. Stoneham MD, Walters FJ. Post-operative analgesia for craniotomy patients: current attitudes among neuroanaesthetists. Eur J Anaesthesiol 1995;12:571–5 13. Bloomfield EL, Schubert A, Secic M, Barnett G, Shutway F, Ebrahim ZY. The influence of scalp infiltration with bupivacaine on hemodynamics and postoperative pain in adult patients undergoing craniotomy. Anesth Analg 1998;87:579 – 82 14. Jeffrey HM, Charlton P, Mellor DJ, Moss E, Vucevic M. Analgesia after intracranial surgery: a double-blind, prospective comparison of codeine and tramadol. Br J Anaesth 1999;83:245–9 15. Ayoub C, Girard F, Boudreault D, Chouinard P, Ruel M, Moumdjian R. A comparison between scalp nerve block and morphine for transitional analgesia after remifentanil-based anesthesia in neurosurgery. Anesth Analg 2006;103:1237– 40 16. Kaur A, Selwa L, Fromes G, Ross DA. Persistent headache after supratentorial craniotomy. Neurosurgery 2000;47:633– 6 17. Bouhassira D, Lanteri-Minet M, Attal N, Laurent B, Touboul C. Prevalence of chronic pain with neuropathic characteristics in the general population. Pain 2008;136:380 –7 18. Attal N, Cruccu G, Haanpaa M, Hansson P, Jensen TS, Nurmikko T, Sampaio C, Sindrup S, Wiffen P. EFNS guidelines on pharmacological treatment of neuropathic pain. Eur J Neurol 2006;13: 1153– 69 19. Mao J, Chen LL. Gabapentin in pain management. Anesth Analg 2000;91:680 –7 20. Patel S, Naeem S, Kesingland A, Froestl W, Capogna M, Urban L, Fox A. The effects of GABA(B) agonists and gabapentin on mechanical hyperalgesia in models of neuropathic and inflammatory pain in the rat. Pain 2001;90:217–26 21. Turan A, Karamanlioglu B, Memis D, Hamamcioglu MK, Tukenmez B, Pamukcu Z, Kurt I. Analgesic effects of gabapentin after spinal surgery. Anesthesiology 2004;100:935– 8 22. Kehlet H, Jensen TS, Woolf CJ. Persistent postsurgical pain: risk factors and prevention. Lancet 2006;367:1618 –25 23. Dirks J, Moiniche S, Hilsted KL, Dahl JB. Mechanisms of postoperative pain: clinical indications for a contribution of central neuronal sensitization. Anesthesiology 2002;97:1591– 6 24. Julius D, Basbaum AI. Molecular mechanisms of nociception. Nature 2001;413:203–10 25. Blumenthal S, Borgeat A, Pasch T, Reyes L, Booy C, Lambert M, Schimmer RC, Beck-Schimmer B. Ropivacaine decreases inflammation in experimental endotoxin-induced lung injury. Anesthesiology 2006;104:961–9 26. Cassuto J, Sinclair R, Bonderovic M. Anti-inflammatory properties of local anesthetics and their present and potential clinical implications. Acta Anaesthesiol Scand 2006;50:265– 82 27. Hollmann MW, Durieux ME. Local anesthetics and the inflammatory response: a new therapeutic indication? Anesthesiology 2000;93:858 –75


Case Report

Neuromodulation in Patients Deployed to War Zones Anthony Dragovich, MD* Thomas Weber, DO* Daniel Wenzell, MD† Michael H. Verdolin, MD‡

Four active duty military personnel and two retired soldiers/military contractors were treated with spinal or peripheral nerve stimulators. All six personnel were able to deploy after the stimulators were placed. Five patients had no incidents during their deployments. One patient completed four deployments but had mechanical complications that necessitated eventual revisions. Considering the risks and limitations of reoperation, nerve blocks, and pharmacotherapy in a forward-deployed area, spinal cord stimulation provides an appealing alternative in soldiers who desire to remain deployable on active duty. (Anesth Analg 2009;109:245–8)

Steven P. Cohen, MD§储

S

pinal cord stimulation (SCS) has been used to treat chronic pain syndromes for more than 40 yr. Since the first description by Shealy et al.1 in 1967, numerous clinical and technologic improvements have been made. Most studies published in the 1970s and 1980s reported success rates of approximately 40%.2 These high failure rates could generally be attributed to a combination of factors, including mechanical failures, surgical complications, and poor patient selection. However, improvement in all of these areas has dramatically increased success rates. In a systematic review of patients with complex regional pain syndrome and failed back surgery syndrome, Taylor et al.3 reported that 62% of patients experienced a significant reduction in pain, 40% returned to work, 53% discontinued analgesics, and 70% were satisfied with treatment. Notwithstanding these advances, SCS is still considered by many to be a last resort for patients who have failed all other treatment modalities. There are no reports of any patients returning to physically strenuous occupations after SCS implantation. In fact, many physicians consider SCS a relative contraindication to vigorous physical activity. This is one reason many patients opt for repeat spinal surgery rather than SCS despite evidence demonstrating SCS may be superior to reoperation.4 The reluctance to implant From the *Department of Surgery, Womack Army Medical Center Ft. Bragg, North Carolina; †Department of Surgery, Madigan Army Medical Center, Tacoma, Washington; ‡Department of Anesthesiology, Naval Medical Center San Diego, San Diego, California; §Departments of Anesthesiology and Critical Care Medicine, Johns Hopkins School of Medicine, Baltimore, Maryland; and 储Department of Surgery, Walter Reed Army Medical Center, Washington, DC. Accepted for publication January 7, 2009. Address correspondence and reprint requests to Anthony Dragovich, MD, 25 Bay Point, Sanford, NC 27332. Address e-mail to [email protected]. Anthony Dragovich is currently at TF 10, Ibn Sina Hospital, Baghdad, Iraq, APO AE 09348, until June 2009. Copyright © 2009 International Anesthesia Research Society DOI: 10.1213/ane.0b013e3181a3368e

Vol. 109, No. 1, July 2009

physically active patients is not unreasonable considering that the complication rate from technical problems, such as lead failure, migration, or fracture, ranges between 8.3% to 42.8%, with many attributed to falls or excessive activity.5,6 The current operations in Iraq and Afghanistan have produced unique medical challenges in pain medicine. Soldiers are surviving more severe injuries that leave them with residual pain and disability. Due to advances in medical science and the physical and mental fortitude of our soldiers, many previously career-ending and life-altering disabilities have been overcome. Many soldiers desire to remain in military service, but unremitting pain is often the last standing hurdle. There are several treatment options in patients with chronic nonmalignant pain, but few are conducive to the physically and mentally demanding lifestyle of service personnel. In this article, we report five cases from our treatment records in which service personnel or contractors deployed or were preparing to deploy to a combat zone with a SCS, and one case whereby a marine was treated with occipital nerve stimulation for headaches sustained after an improvised explosive device injury and subsequently redeployed to Afghanistan. These are the only known cases of neuromodulation use in deployed personnel. This case series has been approved for publication by the Womack Army Medical Center IRB (Table 1).

CASE HISTORIES Case 1 A 37-yr-old male soldier left active duty to join the army reserve in 2002. In the 18 mo before exiting active duty, he underwent an L5-S1 discectomy for radiculopathy and subsequent spinal fusion after his pain failed to resolve. After his surgeries, he was maintained on gabapentin and oxycodone as needed, with moderate relief. While in the reserves, he continued to work full-time as a supervisor in a supermarket. In 2003, his pain began to worsen. His pain was predominantly in his left leg, with associated tingling but no weakness. He sought help from a civilian pain 245

Table 1. Demographic and Clinical Variables of Six Personnel Who Deployed into Combat Areas With Spinal Cord or Peripheral Nerve Stimulators Symptom Age Sex Pain site Diagnosis durationa Follow-up 37 26 32 33 48 29

M Axial and FBSS Bil LE F Forearm CRPS II M Headaches Occipital neuralgia M Axial and FBSS Bil LE M Axial and FBSS Bil LE M Right LE CRPS I

6 yr

4 yr

2 mo 6 mo

12 mo 8 mo

2 yr

4 yr

7 yr

8 mo

14 mo

20 mo

Bil ⫽ bilateral; LE ⫽ lower extremities; FBSS ⫽ failed back surgery syndrome; CRPS ⫽ complex regional pain syndrome. a Pretreatment.

specialist who, in 2004, placed a single lead SCS after multiple failed trials with adjuvants and nerve blocks. In 2006, the soldier’s military police unit deployed to Afghanistan. At the time, he was taking venlafaxine and hydrocodone once or twice a month as needed, and using the SCS for up to 2 h several times per week. After consultation with his civilian and military physicians, he elected to accompany his unit. He completed a 12-mo deployment with only one visit to a battalion surgeon for back pain, which was treated with a nonsteroidal antiinflammatory drug.

Case 2 A 26-yr-old woman sailor sustained a sports injury to her left ulnar nerve at the elbow. The patient was treated with an ulnar nerve repair and transposition that was complicated by the development of complex regional pain syndrome II. Subsequent treatment with stellate ganglion blocks, multiple neuropathic and opioid pain medications, and physical therapy, failed to provide her with significant relief. Two months after her surgery, a cervical SCS was implanted, which reduced her pain from 8/10 to 2/10 and enabled her to discontinue opioid therapy. Eight weeks later she was sent on an overseas assignment to Japan. One-year postimplant she was pain-free and in the process of applying to dive school.

Case 3 A 32-yr-old Marine sustained a fragmentation injury secondary to an improvised explosive device. The fragments traveled through his neck to the base of his brain, lodging next to the right vertebral artery. The patient suffered from severe headaches, but was deemed to be a poor surgical candidate because of the location of fragments in the circle of Willis. Medical management with opioids, nonsteroidal antiinflammatory drugs, and adjuvants was unsuccessful. An atlanto-axial joint injection resulted in 2 wk of pain relief, with contrast injection outlining the fragment and vertebral artery. The injection was not repeated due to distorted anatomy and transient benefit. Occipital nerve blocks and subsequent occipital nerve radiofrequency ablation also provided only short-term relief. An occipital stimulator trial conducted with ultrasound guidance to map the occipital nerve and vessels resulted in excellent pain relief. He was implanted with two octopolar leads. Activation of the stimulator alleviated the headaches. He redeployed to Iraq with his unit 8 mo postimplant, fulfilling his tour of duty without incident.

Case 4 A 33-yr-old man with a long history of radicular back pain was status post failed L5-S1 discectomy. Subsequent 246

Case Report

treatment with epidural steroids, lumbar facet radiofrequency denervation, and opioid and non-opioid pharmacotherapy failed to alleviate his symptoms. In 2004, he was boarded out of the army. Shortly after discharge from the army, he underwent SCS implantation and was able to be weaned from all opioid and non-opioid medications. Within a year, he deployed to Iraq as a civilian contractor. In Iraq, he fell out of a transport truck and experienced a recurrence of his axial and lower extremity symptoms requiring evacuation to the United States. During a pain clinic consultation, his SCS was reprogrammed, and he was restarted on low-dose opioids. He subsequently deployed for 4 – 6 mo on 3 more occasions. After his fourth deployment he was again seen in the pain clinic with increased pain. The generator was found to be nonfunctional upon interrogation. He underwent two surgical revisions but eventually had the entire system replaced with a dual octrode arrangement and a rechargeable generator. At his last interrogation, he was found to be using the stimulator 87% of the time. He continues to have severe pain despite high dose-opioid use. He was subsequently diagnosed with posttraumatic stress disorder and spent 9 wk in an inpatient facility last year.

Case 5 A 48-yr-old male military contactor with low back pain, status post laminectomy had spinal fusion 7 yr earlier that was complicated by cauda equina syndrome. His initial symptoms included a marked decrease in sensation in the perineum and bilateral lower extremities, along with urinary hesitancy. Over time he regained some sensation but continued to experience a constant dull, throbbing pain bilaterally in his low lumbar region, perineum and lower legs, rated between 5 and 8 of 10. The urinary hesitancy did not improve. Gabapentin and hydrocodone provided him with some relief, but he had difficulty performing his job secondary to medication-induced cognitive dysfunction. Tramadol, duloxetine, nortriptyline, and nonsteroidal antiinflammatory drugs provided no relief. In an attempt to improve his pain control and decrease his medication usage, he underwent a SCS trial that successfully reduced his back pain by 50% and his leg pain by 90%. A permanent SCS was implanted in October 2007. By December 2007, he was able to be weaned from all pain medications and reported a 60% reduction in pain. He deployed to Iraq in February of 2008 and successfully completed an uneventful 6-mo tour of duty that included “multiple combat related missions and sleeping on rocks.” He currently takes tramadol as needed for groin pain but his back and leg symptoms are well-controlled with the SCS.

Case 6 A 29-yr-old active duty man in a special forces unit suffered a severely sprained right ankle while training for a marathon. He was treated with standard conservative therapy but failed to improve. Four months postinjury he developed progressively increasing allodynia enveloping his right foot and lower leg without sudomotor or vasomotor changes. Further studies, including a triple-phase bone scan, computed tomography scan, and eletromyogram, failed to reveal an etiology. A diagnostic lumbar sympathetic block relieved his symptoms for 2 wk. A second lumbar sympathetic block was performed but provided only 1 wk of relief. Multiple medication trials either failed to relieve his pain or were discontinued secondary to side effects. After a successful trial, a SCS was implanted in December 2006. This relieved his pain by 50%. The soldier deployed to Iraq in early 2007 with a selfmodified uniform that allowed him to easily change stimulator programs (Fig. 1). He completed a physically demanding 6-mo deployment without complications or need of medical ANESTHESIA & ANALGESIA

Table 2. Selection Criteria for Neuromodulation in Motivated Soldiers Inclusion criteria Desire to remain deployable overseas Absence of a pending medical board (equivalent of pending litigation or worker’s compensation claim) Currently employed in their preinjury occupation or aggressively engaged in rehabilitation with the stated goal of returning to their prior occupation Pain condition amenable to neuromodulation Presence of social support Realistic expectations (patient defined success as a 50% reduction in pain) Ability to understand and manipulate the device Exclusion criteria Absence of a history of drug or alcohol abuse Absence of drug seeking behavior Absence of major psychiatric disorders (active psychosis, severe depression, hypochondriasis) Absence of coagulation disorders Absence of current infection Absence of an additional source of pain not amenable to neuromodulation

Figure 1. Antenna pouch mounted to trousers with antenna in (top) and out (bottom) to demonstrate position. The antenna is held inside the pouch by hook and loop fasteners on the inner sides of the pouch. resources. Since his return, he continues to perform all his duties, which include long road marches with heavy ruck sacks, running, scuba diving, and rock climbing.

DISCUSSION A major dilemma faced by medical officers is how to control chronic pain in a motivated soldier who might otherwise be an asset to his unit. SCS has been successfully used to treat war injuries,7 but in the military community it is widely acknowledged that neuromodulation renders a serviceman nondeployable. Yet, the convergence of several factors, such as technological advances in equipment, more refined selection criteria (Table 2), forward-deployed medical assets, and the unique value of veteran soldiers with combat experience, have led some specialists to reexamine this premise. In this series, we present six cases that demonstrate SCS can be a viable option for motivated patients in a physically and mentally challenging environment. The use of SCS becomes even more conceptually appealing when one considers the limitations of alternative therapies. The use of nerve blocks in theaters of Vol. 109, No. 1, July 2009

operation is limited by the lack of trained personnel outside of combat support hospitals and the need for repeated treatments. Pharmacotherapy has similar downsides in that all neuropathic pain medications have the potential to cause lethargy and cognitive dysfunction, which can endanger lives in combat situations. The use of opioids can pose particularly daunting challenges in austere environments. By their depressant effects on the central nervous system and hypothalamicpituitary-adrenal axis, opioids can cause depression, lethargy, cognitive impairment, impaired wound healing, immunosuppression, and loss of muscle mass, all of which may place soldiers at increased risk of injury.8,9 If opioids are lost, stolen or unable to be refilled, withdrawal can occur in a hostile environment. Even under the best of circumstances, less than half of patients with noncancer pain treated with opioids will obtain long-term benefit, and 25% will experience significant adverse effects.9 All of these factors make opioids and high doses of adjuvants poor choices in soldiers preparing for deployment. In view of the risks and limitations associated with more conventional pain treatment, SCS seems to be an attractive choice for motivated soldiers who wish to remain deployable. In a 22-yr retrospective review in 410 patients with chronic pain treated with SCS, Kumar et al.10 found significant improvements in mood, sleep, and energy levels sustained throughout the mean 97-mo follow-up period. After weighing the risk: benefit ratio of SCS to other therapies in soldiers with chronic pain wishing to stay on active duty, we believe the scales tilt in favor of SCS in carefully selected patients. Although the average follow-up was only 24 mo in our series, the potential for SCS to enable soldiers and other patients in physically demanding occupations to © 2009 International Anesthesia Research Society

247

meet their occupational goals should be strongly considered in patients who fail other treatments. Five of the six patients did not have any mechanical complications during the deployment. None of these patients could have deployed before placement of the SCS, and all five successfully completed their tour of duty with minimal need for medical resources. Case 4 illustrates that this treatment is not without potential complications. Patient selection is of paramount importance in selecting SCS candidates, and a 6-mo trial period whereby a soldier with a stimulator engages in training that simulates his deployment duties might reduce the risk of pain recurrence or mechanical malfunction necessitating medical evacuation. The failed deployment of patient 4 underscores the inherent risk of reinjury any time a person is engaged in a physically demanding environment. In conclusion, this small series suggests that SCS should be considered in motivated patients who suffer from chronic pain and wish to return to physically strenuous occupations. Both the patient and treating physicians should be aware of the limitations and potential complications of dorsal column and peripheral nerve stimulators. However, with appropriate selection criteria, SCS can be a life-altering therapy that enables soldiers and other patients to achieve their professional goals.

248

Case Report

REFERENCES 1. Shealy C, Mortimer J, Reswick J. Electrical inhibition of pain by stimulation of the dorsal columns: preliminary clinical report. Anesth Analg 1967;46:489 –91 2. Bedder M. Spinal cord stimulation and intractable pain: patient selection. In: Waldman, Winne, eds. Interventional pain management. 1st ed. Philadelphia: WB Saunders, 1996:412–18 3. Taylor RS, Van Buyten JP, Buchser E. Spinal cord stimulation for chronic back and leg pain and failed back surgery syndrome: a systematic review and analysis of prognostic factors. Spine 2005;30:152– 60 4. North RB, Kidd DH, Farrokhi F, Piantadosi SA. Spinal cord stimulation versus repeated lumbosacral spine surgery for chronic pain: a randomized, controlled trial. Neurosurgery 2005;56:98 –107 5. Kumar K, Hunter G, Demeria D. Spinal cord stimulation in the treatment of chronic benign pain: challenges in treatment planning and present status, a 22-year experience. Neurosurgery 2006;58:481–96 6. Grabow T, Raja S. Spinal cord stimulation for complex regional pain syndrome: an evidenced-based medicine review of the literature. Clin J Pain 2003;19:371– 83 7. Verdolin M, Stedje-Larsen E, Hickey A. Ten consecutive cases of complex regional pain syndrome of less than 12 months duration in active duty United States military personnel treated with spinal cord stimulation. Anesth Analg 2007;104:1557– 60 8. Daniell H. Hypogonadism in men consuming sustained-action oral opioids. J Pain 2002;3:377– 84 9. Kalso E, Edwards J, Moore R, McQuay H. Opioids in chronic non-cancer pain: systematic review of efficacy and safety. Pain 2004;112:372– 8 10. Kumar K, Hunter G, Demeria D. Spinal cord stimulation in treatment of chronic benign pain: challenges in treatment planning and present status, a 22-year experience. Neurosurgery 2006;58:481–96


Pain Mechanisms Section Editors: Tony L. Yaksh/Quinn H. Hogan

Instilled or Injected Purified Natural Capsaicin Has No Adverse Effects on Rat Hindlimb Sensory-Motor Behavior or Osteotomy Repair Susan M. Kramer, DrPH* Jonelle R. May, MS† Daniel J. Patrick, DVM, DACVP† Luc Chouinard, DVM, DES‡ Marilyne Boyer, BSc‡ Nancy Doyle, BSc‡ Aurore Varela, DVM, IPSAV, MSc‡ Susan Y. Smith, MSc‡ Eric Longstaff, PhD§

BACKGROUND: A novel formulation of ⱖ98% pure capsaicin (4975) is currently undergoing clinical investigation using novel routes of delivery to provide selective analgesia lasting weeks to months with a single dose. We conducted this study to assess the safety and effects of instilled and injected 4975 in rat models of wound healing osteotomy repair and sensory-motor nerve function. METHODS: Adult male and female Sprague-Dawley rats were used. To assess the effects of 4975 on nerve or muscle, 0.0083 or 0.025 mg 4975 or vehicle (25% polyethylene glycol-300) was applied to exposed sciatic nerve, or 0.1 mg 4975 or vehicle was injected into the surrounding muscle (Group 1). To assess the effect of 4975 on bone healing, an osteotomy was made in one femur and 0.5 mg of 4975 or vehicle was instilled into the site (Group 2). Behavioral testing was performed on both groups of rats and histological evaluation of the sciatic nerve, and surrounding soft tissue and bone was done at days 3, 14, and 28 after surgery. Femurs from osteotomy rats were assessed using peripheral quantitative computed tomography and biomechanical testing. Standard statistical tests were used to compare groups. RESULTS: Rats with direct application of 4975 to the sciatic nerve and surrounding muscle were no different from the controls in nociceptive sensory responses (F ⫽ 0.910, P ⫽ 0.454), grip strength (F ⫽ 0.550, P ⫽ 0.654), or histology of the muscle or sciatic nerve. In osteotomy rats, there were no statistical differences between 4975 and vehicle-treated rats for bone area (H ⫽ 2.858, P ⫽ 0.414), bone mineral content (F ⫽ 0.945, P ⫽ 0.425), or bone mineral density (F ⫽ 0.87, P ⫽ 0.462) and no difference in soft tissue healing. There were neither differences in bone stiffness (F ⫽ 1.369, P ⫽ 0.268) nor were there noticeable differences in the macro- or microscopic appearance of the right femur osteotomy healing site and surrounding soft tissues between the control group and the 4975-treated animals. CONCLUSION: A single, clinically relevant application of instilled or injected 4975 has no observable adverse effect on wound and bone healing after osteotomy or on the structural integrity of exposed muscle and nerve. (Anesth Analg 2009;109:249 –57)

T

he management of moderate to severe postoperative pain continues to be an unmet medical need and a significant impediment to the timely recovery and resumption of routine activities.1,2 Despite multimodal postoperative treatments, including anesthetics, systemic nonsteroidal antiinflammatory drugs or

From the *Anesiva, Inc., South San Francisco, California; †MPI Research, Inc., Mattawan, Michigan; ‡Charles River Laboratories Preclinical Services Montreal Inc., Senneville, Quebec, Canada; and §Preclinical Development Services Ltd., Bramhall, Stockport, Cheshire, UK. Accepted for publication January 29, 2009. The studies reported in this manuscript were conducted as contract research funded by Anesiva, Inc. and AlgoRx Pharmaceuticals, Inc. Susan M. Kramer is an employee of Anesiva, Inc. and Eric Longstaff is a former consultant for AlgoRx Pharmaceuticals, Inc. Address correspondence and reprint requests to Susan M. Kramer, DrPH, Anesiva, Inc., 400 Oyster Point Blvd, Suite 502, South San Francisco, CA 94080. Address e-mail to [email protected]. Copyright © 2009 International Anesthesia Research Society DOI: 10.1213/ane.0b013e3181a7f589

Vol. 109, No. 1, July 2009

cyclooxygenase-2 inhibitors, acetaminophen, and opioids, it is often difficult to achieve acceptable levels of analgesia.2 Topically applied capsaicin, the active ingredient of chili Capsicum fruits, has long been used as a local analgesic but the recent finding that locally delivered (instilled or injected), highly purified capsaicin (4975) may provide substantial long-term relief for postoperative pain has made this compound a focus for clinical investigation.3,4 Capsaicin selectively activates C- and thinly myelinated A␦-fibers that transmit signals from noxious stimuli to the central nervous system. Capsaicin exposure initially excites nociceptors resulting in a profound algesia, including dose-related burning and stinging pain, thermal and mechanical hyperalgesia, generalized axon reflex vasodilation, and erythema. These algesic effects subsequently give way to analgesia as a result of an increase in nociceptive thresholds that lasts from hours5,6 to weeks4 for topically applied capsaicin-containing creams, and weeks to months for injected or instilled 4975.7,8 This pattern of sensory 249

change is generally considered to result from the activation of C- and A␦-nociceptors, followed by the desensitization of the same fibers as a consequence of prolonged activation.9 –11 Although there is a large literature on the application of topical capsaicin,12–15 there are little data on the effects of capsaicin delivered by nontopical routes. Recently reported clinical trials using nontopical application of capsaicin3,16 –19 have increased the interest in what effects capsaicin has on tissues that are not normally reached by topical application. We have therefore examined the histological and behavioral consequences of 4975 applied directly to nerve, muscle, and bone at concentrations suggested for use in clinical settings to support the safety of these novel routes of delivery.

METHODS One hundred twenty-eight male and 20 female CD (Crl:CD[sd]) Charles River rats were used. Rats were housed individually, provided water and food ad libitum and were maintained on a 12 h light/dark cycle. This study was conducted in compliance with Good Laboratory Practice regulations of the United States Food and Drug Administration (21 CFR Part 581) and approved by the Institutional Animal Care and Use Committees of MPI Research and Charles River Laboratories.

Extrafascicular (EF) and IM (IM) Application of 4975 Surgery Anesthesia was induced directly with isoflurane and oxygen (0.5%–5%) and the rats remained under anesthesia for the entire procedure. Ophthalmic ointment was administered to each eye and a subcutaneous injection of buprenorphine (0.4 mg/kg) was given. Skin over the surgery site was shaved and washed with antiseptic solution. EF Application of 4975 onto the Sciatic Nerve Rats were placed in lateral recumbency and a 3- to 4-cm incision was made in the skin overlying the sciatic nerve. The nerve was exposed by blunt dissection and the peroneal branch was identified. One hundred micro liter purified ⱖ98% capsaicin (4975, Anesiva, South San Francisco, CA) in 25% polyethylene glycol (PEG) 300 in sterile water was instilled onto the sciatic nerve approximately 0.5 cm proximal to the peroneal nerve branch. Treatment groups consisted of 0.083 mg/mL (0.0083%) or 0.25 mg/mL (0.025%) 4975 or vehicle alone (25% PEG 300 in sterile water). Capsaicin concentrations were based on previous clinical studies.7,8,20 A sham group underwent the same surgical procedure but did not receive 4975 or vehicle. Subcutaneous and muscle tissue were closed with 4-0 Vicryl suture, and the skin was closed with skin staples. IM Application of Capsaicin A 3– 4-cm incision was made in the skin overlying the sciatic nerve and two 200 ␮L IM injections of 0.1 mg 4975 in 25% PEG 300 or vehicle alone (25% PEG 250

4975 Has No Effect on Surgical Healing

300) were made in the muscle surrounding the nerve, one on either side, at approximately 0.5 cm proximal to the peroneal nerve branch. Subcutaneous and muscle tissue were closed with 4-0 Vicryl suture, and the skin was closed with skin staples. Behavioral Testing The following behavioral tests were performed the day before surgery (day-1) then on days 3, 14, and 28 after surgery. Clinical Evaluation All rats were examined for 42 different variables, including condition of skin and fur, muscle wasting, and swelling at defined body locations every day beginning the day of surgery (immediately before the procedure) and continuing until the end of the experimental observations. Open Field Testing For all tests, experimenters were blinded to the treatment groups. Individual rats were placed in a black plexiglass observation box measuring 20 ⫻ 20 ⫻ 8 in and observed for a minimum of 3 min. The variables evaluated were based on those outlined in Moser et al.21,22 The observations included posture, rearing, bizarre behavior, gait, mobility, and stereotypy (repetitive movement or postures). Testing was performed the day before surgery (day-1) then on days 3, 14, and 28 after surgery. Forelimb and Hindlimb Grip Strength and Hindlimb Splay Forelimb and hindlimb grip strength was measured using the procedure described by Meyer et al.23 The rat was held around the dorsal thorax with one hand and placed in the testing apparatus and allowed to grasp a ring attached to a strain gauge. The rat was then smoothly and steadily pulled backward by the base of the tail until it released its grip on the forelimb ring. The strain gauge recorded the maximum force in kilograms. The hindpaws of the rat then came into contact, and gripped a bar attached to a second strain gauge. The backward pull of the rat was continued until it released its grip on the hindlimb bar and the strain gauge recorded the maximum force in kilograms. This procedure was repeated three times and the force measurement for each limb was averaged. Testing was performed the day before surgery (day-1) then on days 3, 14, and 28 after surgery. Hindlimb splay was measured as described by Edwards and Parker.24 For this test, the plantar surface of each hindpaw was inked and the rat held horizontally 32 cm above a sheet of white paper. The rat was then dropped onto the paper and the distance between the center of the right and left heels was measured from the ink mark. The mean of three measurements was recorded for each rat. Testing was performed the day before surgery (day-1) then on days 3, 14, and 28 after surgery. ANESTHESIA & ANALGESIA

Hot Plate Test Hindpaw withdrawal latency was measured using the hotplate test as described by Ankier.25 The hotplate was set at 52°C and the latency to paw withdrawal measured. The test was administered three times and mean values were recorded. Histology Rats were killed by intraperitoneal injection of sodium pentobarbital (90 mg/kg) followed by transcardiac perfusion with 0.9% saline, followed by 3% paraformaldehyde, and 3% glutaraldehyde in 0.1 M phosphate buffer. The entire length of the sciatic nerve (with peroneal branch) was excised and postfixed in the same fixative, then embedded in glycol methacrylate. Representative sections were taken from the injection site and proximal and distal sites and stained with hematoxylin and eosin (H&E). The sections were examined microscopically and any lesions in or around the nerve (epineurium, perineurium, endoneurium, Schwann cells, myelin sheath, and axons) were noted and scored for severity. Rats receiving IM injection were killed and perfused on day 28, and tissue from the injection site was collected and postfixed in 10% neutral buffered formalin. Paraffin sections were H&E stained and examined microscopically for lesions.

Osteotomy Surgery: Osteotomy with 4975 Irrigation A 2.0-cm skin incision was made along the lateral aspect of the right femur. A parapatellar incision was then made into the joint capsule and the patellar ligament laterally retracted. An intercondylar entry point was made and muscles retracted to expose the femur. The periosteum was cut longitudinally at the midshaft (approximately 0.5 cm) and retracted from the bone along the incision using a scalpel blade and retractor. Care was taken to minimize the periosteum incision. The femur was cut at the midshaft using a micro air drill with an oscillating saw blade with care taken to avoid muscle damage. The osteotomy site was thoroughly irrigated with sterile fluid during the osteotomy to prevent heat-mediated microdamage. Once the osteotomy was completed, an intramedullary Kirshner wire (1.6 mm diameter for males, 1.25 mm for females) was inserted into the right femur for fracture reduction so that it did not interfere with knee movement. One hundred microliters of 4975 (0.5 mg in 20% PEG 300) or vehicle control (20% PEG 300, 10 mM histidine and 5% [w/v] sucrose in sterile water) was then instilled into the osteotomy site. The soft tissues were carefully closed in layers and the skin was closed using surgical skin staples which were removed 5–12 days after surgery. Radiographs of the right femur were taken of all animals after surgery to confirm the extent of the osteotomy and the placement of the intramedullary pin. During the insertion of the pin, slight fissures in Vol. 109, No. 1, July 2009

Figure 1. Clinical variables for extrafascicular (EF) and IM (IM) 4975-treated rats. EF instillation with 0.025 or 0.0083 mg 4975 or vehicle resulted in reduced weight gain (A) over the first 10 postoperative days compared with sham-operated rats. There was no difference (P ⬎ 0.05) among sham, vehicle, or 4975 EF-instilled rats in body temperature (C) or activity/arousal (D). There was no difference (P ⬎ 0.05) between the vehicle and 4975 IM-injected rats for weight (B) body temperature (D) or activity/arousal (F). *P ⬍ 0.05; **P ⬍ 0.01. Day-1, n ⫽ 18; Day 3, n ⫽ 18; Day 14, n ⫽ 12; Day 28, n ⫽ 6. the medial trochlea were noted for a few animals from all groups. These fissures were undetectable on the radiographs and therefore considered to have no impact on the outcome of the study. Rats received two doses of analgesic (buprenorphine 0.4 mg/kg) after surgery and a second dose (0.2 mg/kg) 9 –12 h later. Clinical Observations All rats were observed twice daily to assess possible clinical signs and symptoms related to the treatment. A detailed physical examination was performed on all animals weekly, starting during the acclimation period. Observed clinical signs were individually recorded and particular attention was paid to limb use. Forty-two different variables were recorded, including condition of skin and fur, muscle wasting, and swelling at defined body locations. Gross Pathology and Histology Rats were killed by exsanguination from the abdominal aorta although under deep isoflurane anesthesia. The hindlimbs were removed and immediately © 2009 International Anesthesia Research Society

251

Figure 2. Measurements of hindpaw function in extrafascicular (EF) and IM (IM) 4975-treated rats. There was a significant decrease in hindpaw grip strength on postoperative days 3 and 14 for vehicle and 4975 EF-instilled (A) compared with shamoperated rats, but there was no significant difference (P ⬍ 0.05) between the vehicle and 4975-injected rats. There were no consistent statistical differences among sham, vehicle, or 4975 EF-injected rats for hindpaw splay (C) or thermal withdrawal (E) or for hindpaw grip (B), hindpaw splay (D) or thermal withdrawal (F) for the IM injected group. *P ⬍ 0.05; **P ⬍ 0.01 vs sham; #P ⬍ 0.05 vs vehicle. Day-1, n ⫽ 18; Day 3, n ⫽ 18; Day 14, n ⫽ 12; Day 28, n ⫽ 6.

assessed for callus rigidity by manual palpation. The callus was described as rigid when no motion of the distal segments was detected by a gentle manual palpation and as flexible when the manual palpation induced motion of the distal segment. On completion of the gross examination, the soft tissues were fixed in 10% neutral buffered formalin and both femurs were placed in 70% ethanol. After biomechanical testing, femora were decalcified and processed for histopathological evaluation. Femora and soft tissue were embedded in paraffin, cut at 4 ␮m and stained with H&E. Biomechanical Testing Both osteotomized and contralateral whole femurs were tested in 4-point bending with a servohydrolic material test system (858 Mini Bionix) to determine the material properties of the cortical bone. The femurs were cleaned of soft tissue and the midspan was marked between the two loading points and served as the upper loading point for each femur. The actuator was set at a rate of 1 mm/s until failure occurred. Load and displacement data were collected using TestWorks (version 3.8) for TestStar II software (version 4.0c). Peak load, energy to break (area under the curve), and stiffness were derived. Peak load was measured as the maximum height of the loaddisplacement curve, and stiffness was measured as the 252


slope of the linear portion of the load-displacement curve. Statistics Levene’s test was used to assess homogeneity of group variances for each specified end point and for each data collection interval. If Levene’s test was not significant (P ⬎ 0.01), a pooled estimate of the variance (Means Square Error) was computed from a one-way analysis of variance and post hoc analysis done using Dunnett’s comparison of each treatment group with the control group. If Levene’s test was significant (P ⬍ 0.01), comparisons with the control group were made using Welch’s t-test with a Bonferroni correction.

RESULTS EF and IM Injection Clinical Signs Neither EF nor IM application of 4975 had any measurable effect on the overall health of the rats. There was no long-term change in weight gain between vehicle and 4975-injected rats (Figs. 1A and B), although all EF-instilled groups showed an initial reduction in weight gain (Fig. 1A) that normalized about 10 days after surgery. For both the EF and IM experiments, there was no difference among the sham, ANESTHESIA & ANALGESIA

Figure 3. Hematoxylin and eosinstained sections of sciatic nerve and muscle from 4975 extrafascicular (EF) and IM (IM) 4975-injected rats. At 3 days (B-E) and 28 days post-EF application, there was no difference in appearance between the sciatic nerve from a sham-operated rat (A) and 0.025 mg 4975-treated nerve (C and G are high magnifications of B and F, respectively). At 28 days postinjection, there was no difference between muscle from vehicle (J) or 0.1 mg 4975 (K) IM injected rats. Bar ⫽ 10 ␮m. Inj site ⫽ injection site; Prox ⫽ proximal to the injection site; Dist ⫽ distal to the injection site.

Table 1. Summary of Microscopic Observations; Intrafascicular Injections Number of animals examined

Severity Day 3 Sciatic nerve—distal to injection site Hemorrhage Inflammation, subacute Within normal limits Sciatic nerve—injection site Hemorrhage Inflammation, subacute Within normal limits Sciatic nerve—proximal to injection site Hemorrhage Inflammation, subacute Within normal limits Day 14 Sciatic nerve—distal to injection site Inflammation, subacute Within normal limits Sciatic nerve—injection site Inflammation, subacute Within normal limits Sciatic nerve—proximal to injection site Inflammation, subacute Within normal limits

Minimal Minimal Minimal Minimal Minimal Minimal

Minimal Minimal Minimal

4975 and vehicle-treated groups for any examined physical sign, including core temperature (Figs. 1C and D), fur quality, skin quality, discharges, missing nails, or material in bedding (data not shown). There were also no differences among groups in activity and arousal (Figs. 1E and F), posture, vocalization, bizarre behavior, gait, mobility, or stereotypy (data not shown). Behavior Because the sciatic nerve carries sensory and motor information for the hindpaw, we tested whether the Vol. 109, No. 1, July 2009

Sham

Vehicle

0.0083 mg

0.025 mg

6

6

6

6

6 1 1 5 6 2 5 1 6 0 4 2

6 0 1 5 6 1 5 1 6 0 5 1

6 0 1 5 6 1 6 0 6 1 5 1

6 0 2 4 6 0 6 0 6 0 5 1

6 3 3 6 6 0 6 3 3

6 1 5 6 4 2 6 5 1

6 0 6 6 3 3 6 6 0

6 0 6 6 4 2 6 4 2

4975 had any effect on motor or sensory-related behavior. The only consistent difference seen was in the EF instillation experiments where there was an initial decrease in hindpaw grip strength for all injected groups (4975 and vehicle) compared with shamoperated rats (Fig. 2A). It is not clear why the injected rats differed from the sham rats other than expected intergroup variability. It could be that injection of any liquid has a transitory effect. However, the most meaningful comparison in these experiments was the comparison between 4975 and vehicle-injected groups © 2009 International Anesthesia Research Society

253

and notably there was no difference between the 4975 and vehicle groups at any tested time point. For all other variables, for both EF and IM treatments (Figs. 2B–F) there were no consistent differences among sham, 4975 and vehicle-treated rats. Histology The sciatic nerve and surrounding muscle were examined at the injection site and proximal and distal to the injection site (Fig. 3). In both the EF and IM groups, the subacute inflammation (consisting of a mixture of neutrophils, lymphocytes, macrophages, and plasma cells) and hemorrhage present in the 4975-treated animals was no different than shamoperated or vehicle-treated rats. By day 14 (Tables 1 and 2), there were no signs of hemorrhage in any group, the subacute inflammation was reduced in all groups and by day 28 there was no evidence of hemorrhage or inflammation in any group (Fig. 3).

Osteotomy Clinical Signs and Histology During the 13-wk observation period, there was no difference between the 4975 and vehicle-treated rats in any of the recorded clinical observations, such as weight gain (Fig. 4), fur or skin condition, swelling, or muscle wasting. At the end of the 13-wk observation period, there was no difference in the macroscopic appearance of the osteotomy site (callus) or in the callus rigidity (as assessed by manual palpation) from the right femur of vehicle and 4975-treated rats (Table 3). The morphology of the soft and hard callus did not differ between the animals treated with vehicle or 4975 (Fig. 5). There were no differences between 4975 and vehicle-treated rats for bone area (H ⫽ 2.858, P ⫽ 0.414, DF ⫽ 3), bone mineral content (F ⫽ 0.945, P ⫽ 0.425), or bone mineral density (F ⫽ 0.87, P ⫽ 0.462). Biomechanical Testing Figure 6 presents the results of the biomechanical testing of the 4975 and vehicle-injected (right) femurs and the unoperated contralateral femur. The only significant difference in the biomechanical measurements was in the work-to-failure values (Fig. 6A) and peak load (Fig. 6C) on the operated (right) femur of the male rats However, in both cases the vehicletreated animals had lower (i.e., less bone integrity) values than the 4975 injected counterparts indicating reduced bone integrity.

DISCUSSION The overall conclusion of these studies is that highly purified (ⱖ98%) capsaicin (4975) does not seem to induce morphological changes in nerves or muscle or interfere with bone or tissue healing when applied in clinically relevant concentrations. Furthermore, 4975 applied in this manner does not result in quantifiable changes in the overall health of the experimental animals or in measured motor and sensory behavior. 254


Table 2. Histological Observation. Intramuscular Injection Severity Day 3 Sciatic nerve—distal to injection site Inflammation, subacute Within normal limits Sciatic nerve—injection site Inflammation, subacute Within normal limits Sciatic nerve—proximal to injection site Inflammation, subacute Within normal limits Day 14 Sciatic nerve—distal to injection site Inflammation, subacute Within normal limits Sciatic nerve—injection site Inflammation, subacute Sciatic nerve—proximal to injection site Inflammation, subacute Within normal limits

Minimal Minimal

Minimal

Minimal Minimal Minimal

Vehicle

0.1 mg

6

5

3 3 6 6 0 6

4 1 5 4 1 4

3 3

3 1

6

6

3 3 6 6 6

3 3 6 6 6

4 2

5 1

Figure 4. Weight gain for male and female osteotomy rats. There was no statistical difference (P ⬎ 0.05) in weight gain between osteotomy rats treated with vehicle compared with 0.5 mg 4975-treated rats.

In tested variables in which there were differences between 4975- and sham-treated animals, the same differences were seen in vehicle-treated rats showing that any changes were the result of the surgery, or less likely the vehicle, and not related to the 4975 application. Caterina et al.26,27 cloned an ion channel present on C- and thinly myelinated A␦ fibers that is activated by capsaicin which has been named the transient receptor potential vanilloid Type 1 (TRPV1) receptor.28 Because the actions of capsaicin are mediated by TRPV1 receptors, it would be predicted that capsaicin-induced changes are most likely to occur in tissues that express this receptor and are therefore susceptible to capsaicin activation. Because small diameter primary sensory ANESTHESIA & ANALGESIA

Table 3. Summary of Grading by Organ/Group and Sex Male

Female

Dose group destination

Vehicle

0.5 mg 4975

Number of animals examined Bone-femur, right: number examined Soft callus Grade 2 Grade 3 Grade 4 Grade 5 Total number of tissues affected: Average grade of tissues affected Hard callus Grade 1 Grade 2 Grade 3 Total number of tissues affected Average grade of tissues affected Necrosis: cortex Grade 2 Grade 3 Grade 4 Total number of tissues affected Average Grade of tissues affected Surgical site number examined Inflammation: granulomatous Grade 4 Total number of tissues affected Average Grade of tissues affected

10 10 1 2 7 — 10 3.6 — 7 3 10 2.3 — 9 1 10 3.1 10 1 1 4.0

10 10 — 4 4 1 9 3.7 1 7 2 10 2.1 2 6 2 10 3.0 10 — — —

Vehicle

0.5 mg 4975

10 10

10 10

8 2 — 10 3.2 2 7 1 10 1.9 4 6 — 10 2.6

3 6 1 10 3.8 3 6 1 10 1.8 6 3 1 10 2.5

Figure 5. Hematoxylin and eosin stained section of the osteotomy site from vehicle-treated (A, low magnification; B, high magnification) and 0.5 mg 4975-treated (C, low magnification; D, high magnification) femurs. The soft callus was mainly composed of fibrocartilaginous tissue containing hypertrophic chondrocytes mixed with variable amounts of fibrous connective tissue (fibroblast and collagen). The hard callus consisted of intramembranous bone formed under the periosteum mixed with bone trabeculae formed at the periphery of the hypertrophic cartilaginous matrix (endochondral bone formation). There was no complete bone union because complete bone healing in this model requires a longer time than the study period.

neurons have the highest TRPV1 expression,29 it could be expected that EF application of 4975 to the sciatic nerve would affect thermal sensation mediated by small diameter fibers but neither the histology nor behavior indicates that is the case. A recent study by Neubert et al.30 found an increase in heat withdrawal threshold 1 day after application to the sciatic nerve but thresholds had returned to baseline 1 wk after application. They did find a long lasting effect of the resiniferatoxin treatment on heat withdrawal after caregeenan-induced inflammation. These results are not dissimilar to ours as we found no change in heat withdrawal thresholds 3 days after capsaicin application. It should be denoted that, because of the study Vol. 109, No. 1, July 2009

design (anesthesia, surgery, postoperative analgesics), histological and behavioral assessment was only begun on day 3 and, therefore, if there were acute, transient changes noted by Neubert et al.30 and others,4,5 these would not have been seen. Our results, and those of Neubert et al.,30contrast with some previous studies that found capsaicin application directly to the sciatic nerve resulted in analgesia for at least 7 days.31 This difference in the results is probably because the this study31 used a much higher concentration of capsaicin (1%) than our maximum concentration (0.025%). The lack of damage to small diameter fibers is supported by a previous electron microscope study © 2009 International Anesthesia Research Society

255

Figure 6. Bone variables after right femur osteotomy irrigated with vehicle or 0.5 mg 4975 (n ⫽ 10 for each group). The osteotomy reduced all measured variables when compared with the unoperated left femur, but for stiffness there was no difference between measurements from the vehicle and 4975-instilled right femurs. Although there was a significant difference for area under the curve (AUC) and peak load, the 4975-instilled side was closer to control values than the vehicle-treated side. Peak load is the force required to break the bone and reflects the general integrity of the bone structure, whereas the AUC is a measure of the work to failure. Stiffness is closely related to the mineralization of the bone. that found no changes in small diameter fibers after direct capsaicin application.32 In contrast, the same group33 reported an effect of capsaicin on small diameter fibers examined by electrophysiology. However, the latter study33 removed the connective tissue surrounding the nerve and dissolved the capsaicin in a detergent (Tween 20), and there was no behavioral assessment. This suggests that in the present experiments the intact fascia surrounding the nerve and the fact that the small diameter fibers are protected by Schwann cells in Remak bundles prevented the 4975 gaining access to the small diameter fibers. The histological analysis and normal hindlimb motor performance suggests there were no changes in large diameter myelinated fibers as might be expected as these fibers do not express the TRPV1 receptor. The TRPV1 receptor is expressed by other neuronal cell types and in nonneuronal tissue.34 –38 Most relevant to the present study is the finding that TRPV1 is present in human and rat skeletal muscle.39,40 In muscle, the receptor is most abundant in the sarcoplasmic reticulum and seems to function as a leakage channel for Ca2⫹ when muscle is in the resting state, although a 10-fold higher concentration of capsaicin is required to activate these internal TRPV1 channels as opposed to membrane channels.41 If the IM injections of 4975 did activate the sacroplasmic reticulum TRPV1 channels, there was no observable consequence in muscle integrity or motor-related behavior suggesting that there were no effects on muscle TRPV1 channels. Recent studies showing the presence of TRPV1 receptors in adipose tissue have implicated these receptors in metabolic regulation and obesity42 and TRPV1 receptors on bone marrow-derived dendritic cells43 have suggested an immune role for the receptor. It is unlikely that the routes of administration of the 4975 used in this study could produce systemic concentrations high enough to affect energy metabolism or the immune system, and there was no such indication in the present study. For all the capsaicin-treated rats, weight gain, skin and fur condition, immune response at surgery sites, and overall health of the experimental animals were no different than sham or vehicletreated rats. 256


As clinical trials using injection or instillation progress, there will be a need to study the effects of 4975 on different tissues by using analysis and tissue collection methods not available in the clinical setting. The results of the present study, that the instillation or injection of 4975 in clinically relevant concentrations had no adverse effects on nerve, muscle, or bone integrity or on bone healing, provides basic biological support for the safety of 4975 in clinical settings of local delivery. REFERENCES 1. Apfelbaum JL, Chen C, Mehta SS, Gan TJ. Postoperative pain experience: results from a national survey suggest postoperative pain continues to be undermanaged. Anesth Analg 2003;97: 534 – 40 2. Callesen T, Kehlet H. Postherniorrhaphy pain. Anesthesiology 1997;87:1219 –30 3. Aasvang EK, Hansen JB, Malmstrom J, Asmussen T, Gennevois D, Struys MM, Kehlet H. Effect of wound instillation of a novel purified capsaicin formulation on postherniotomy pain: a double blind, randomized, placebo controlled study. Anesth Analg 2008;107:282–91 4. Knotkova H, Pappagallo M, Szallasi A. Capsaicin (TRPV1 Agonist) therapy for pain relief: farewell or revival? Clin J Pain 2008;24:142–54 5. Lynn B. Capsaicin: actions on nociceptive C-fibres and therapeutic potential. Pain 1990;41:61–9 6. Watson CP, Evans RJ, Watt VR. Post-herpetic neuralgia and topical capsaicin. Pain 1988;33:333– 40 7. Richards PT, Vasko G, Stasko I, Lacko M, Hewson G. ALGRX 4975 reduces pain of acute lateral epicondylitis: preliminary results from a randomized, double-blind, placebo-controlled, phase II multicenter clinical trial. J Pain 2006;7:S3 8. Diamond E, Richards PT, Miller T. ALGRX 4975 reduces pain of intermetatarsal neuroma: preliminary results from a randomized, double-blind, placebo-controlled, phase II multicenter clinical trial. J Pain 2006;7:S41 9. Jancso G, Lawson SN. Transganglionic degeneration of capsaicin-sensitive C-fiber primary afferent terminals. Neuroscience 1990;39:501–11 10. Jancso N, Jancso-Gabor A, Szolcsanyi J. Direct evidence for neurogenic inflammation and its prevention by denervation and by pretreatment with capsaicin. Br J Pharmacol Chemother 1967;31:138 –51 11. Bley KR. Recent developments in transient receptor potential vanilloid receptor 1 agonist-based therapies. Expert Opin Investig Drugs 2004;13:1445–56 12. Zhang WY, Li Wan Po A. The effectiveness of topically applied capsaicin. A meta-analysis. Eur J Clin Pharmacol 1994;46:517–22 13. Ellison N, Loprinzi CL, Kugler J, Hatfield AK, Miser A, Sloan JA, Wender DB, Rowland KM, Molina R, Cascino TL, Vukov AM, Dhaliwal HS, Ghosh C. Phase III placebo-controlled trial of capsaicin cream in the management of surgical neuropathic pain in cancer patients. J Clin Oncol 1997;15:2974 – 80


14. Bernstein JE, Korman NJ, Bickers DR, Dahl MV, Millikan LE. Topical capsaicin treatment of chronic postherpetic neuralgia. J Am Acad Dermatol 1989;21:265–70 15. Watson CP, Tyler KL, Bickers DR, Millikan LE, Smith S, Coleman E. A randomized vehicle-controlled trial of topical capsaicin in the treatment of postherpetic neuralgia. Clin Ther 1993;15:510 –26 16. Rapoport AM, Bigal ME, Tepper SJ, Sheftell FD. Intranasal medications for the treatment of migraine and cluster headache. CNS Drugs 2004;18:671– 85 17. Davis J, Williams H, Bramlett K, Powell T, Schuster A, Yu KP, Gennevois D. A single intra-operative administration of 4975 provides well-tolerated, long-term analgesia for postsurgical pain after total knee arthroplasty. Pain Med 2005;10:1 18. Cantillon M, Vause E, Sykes D, Tagoe E. Safety, tolerability and efficacy of intraoperative ALGRX 4975 in a randomized, doubleblind, placebo-controlled study of subjects undergoing bunionectomy. J Pain 2005;6:S48 19. Cantillon M, Vause E, Sykes D, Russell R, Moon A, Hughes S. Preliminary safety, tolerability and efficacy of ALGRX 4975 in osteoarthritis (OA) of the knee. J Pain 2005;6:S39 20. Stoker DG, Gotleib IJ, Comfort S. A single instillation of a highly purified capsaicin formulation decreases postoperative pain and analgesic use after bunionectomy surgery: a randomized, double-blind, placebo-controlled study. Pain Med. In press 21. Moser VC, McCormick JP, Creason JP, MacPhail RC. Comparison of chlordimeform and carbaryl using a functional observational battery. Fundam Appl Toxicol 1988;11:189 –206 22. Moser VC, Ross JF. Training video and reference manual for a functional observation battery. Washington DC: US Environmental Protection Agency, 1996 23. Meyer OA, Tilson HA, Byrd WC, Riley MT. A method for the routine assessment of fore- and hindlimb grip strength of rats and mice. Neurobehav Toxicol 1979;1:233– 6 24. Edwards PM, Parker VH. A simple, sensitive, and objective method for early assessment of acrylamide neuropathy in rats. Toxicol Appl Pharmacol 1977;40:589 –91 25. Ankier SI. New hot plate tests to quantify antinociceptive and narcotic antagonist activities. Eur J Pharmacol 1974;27:1– 4 26. Caterina MJ, Schumacher MA, Tominaga M, Rosen TA, Levine JD, Julius D. The capsaicin receptor: a heat-activated ion channel in the pain pathway. Nature 1997;389:816 –24 27. Caterina MJ, Julius D. The vanilloid receptor: a molecular gateway to the pain pathway. Annu Rev Neurosci 2001;24: 487–517 28. Gunthorpe MJ, Benham CD, Randall A, Davis JB. The diversity in the vanilloid (TRPV) receptor family of ion channels. Trends Pharmacol Sci 2002;23:183–91 29. Sanchez JF, Krause JE, Cortright DN. The distribution and regulation of vanilloid receptor VR1 and VR1 5⬘ splice variant RNA expression in rat. Neuroscience 2001;107:373– 81

Vol. 109, No. 1, July 2009

30. Neubert JK, Mannes AJ, Karai LJ, Jenkins AC, Zawatski L, Abu-Asab M, Iadarola MJ. Perineural resiniferatoxin selectively inhibits inflammatory hyperalgesia. Mol Pain 2008;4:3 31. Szolcsanyi J. Capsaicin and nociception. Acta Physiol Hung 1987;69:323–32 32. Ainsworth A, Hall P, Wall PD, Allt G, MacKenzie ML, Gibson S, Polak JM. Effects of capsaicin applied locally to adult peripheral nerve. II. Anatomy and enzyme and peptide chemistry of peripheral nerve and spinal cord. Pain 1981;11:379 – 88 33. Wall PD, Fitzgerald M. Effects of capsaicin applied locally to adult peripheral nerve. I. Physiology of peripheral nerve and spinal cord. Pain 1981;11:363–77 34. Birder LA, Kanai AJ, de Groat WC, Kiss S, Nealen ML, Burke NE, Dineley KE, Watkins S, Reynolds IJ, Caterina MJ. Vanilloid receptor expression suggests a sensory role for urinary bladder epithelial cells. Proc Natl Acad Sci U S A 2001;98:13396 – 401 35. Biro T, Maurer M, Modarres S, Lewin NE, Brodie C, Acs G, Acs P, Paus R, Blumberg PM. Characterization of functional vanilloid receptors expressed by mast cells. Blood 1998;91:1332– 40 36. Inoue K, Koizumi S, Fuziwara S, Denda S, Inoue K, Denda M. Functional vanilloid receptors in cultured normal human epidermal keratinocytes. Biochem Biophys Res Commun 2002;291: 124 –9 37. Sasamura T, Sasaki M, Tohda C, Kuraishi Y. Existence of capsaicin-sensitive glutamatergic terminals in rat hypothalamus. Neuroreport 1998;9:2045– 8 38. Cortright DN, Krause JE, Broom DC. TRP channels and pain. Biochim Biophys Acta 2007;1772:978 – 88 39. Xin H, Tanaka H, Yamaguchi M, Takemori S, Nakamura A, Kohama K. Vanilloid receptor expressed in the sarcoplasmic reticulum of rat skeletal muscle. Biochem Biophys Res Commun 2005;332:756 – 62 40. Cavuoto P, McAinch AJ, Hatzinikolas G, Janovska A, Game P, Wittert GA. The expression of receptors for endocannabinoids in human and rodent skeletal muscle. Biochem Biophys Res Commun 2007;364:105–10 41. Karai LJ, Russell JT, Iadarola MJ, Olah Z. Vanilloid receptor 1 regulates multiple calcium compartments and contributes to Ca2⫹-induced Ca2⫹ release in sensory neurons. J Biol Chem 2004;279:16377– 87 42. Zhang LL, Yan Liu D, Ma LQ, Luo ZD, Cao TB, Zhong J, Yan ZC, Wang LJ, Zhao ZG, Zhu SJ, Schrader M, Thilo F, Zhu ZM, Tepel M. Activation of transient receptor potential vanilloid type-1 channel prevents adipogenesis and obesity. Circ Res 2007;100:1063–70 43. Lu T, Newton C, Perkins I, Friedman H, Klein TW. Role of cannabinoid receptors in Delta-9-tetrahydrocannabinol suppression of IL-12p40 in mouse bone marrow-derived dendritic cells infected with Legionella pneumophila. Eur J Pharmacol 2006;532:170 –7


257

Mexiletine and Lidocaine Suppress the Excitability of Dorsal Horn Neurons Andrea Olschewski, MD, PhD*† Rose Schnoebel-Ehehalt, MD* Yingji Li, MD, PhD‡§ Bi Tang, MD‡§ Michael E. Braü, MD, PhD* Matthias Wolff, MD, PhD*

BACKGROUND: Spinal sensitization and facilitatory processes in dorsal horn neurons after nerve injury alter spinal outflow leading to enhanced pain perception and chronic pain syndromes. Clinically used Na⫹ channel blockers at doses which do not block conduction can relieve such chronic pain. Although much attention has been paid to their effect upon afferents, less work has been done with their effect on the excitability of central sensory neurons. Thus, we investigated the effects of the Na⫹ channel blockers mexiletine and lidocaine on sensory spinal dorsal horn neurons. METHODS: Patch-clamp recordings were directly performed in visualized neurons of the substantia gelatinosa in the spinal cord of young rats to investigate the effect of mexiletine and lidocaine in different types of dorsal horn neurons (tonically firing, adapting-firing, and single spike neurons). RESULTS: All three different types of neurons responded dose-dependently to mexiletine and lidocaine. Both local anesthetics reversibly inhibited Na⫹ and K⫹ currents. The half-maximal inhibitory concentration for Na⫹ conductance block was 89 ⫾ 2 or 54 ⫾ 6 ␮M and for delayed-rectifier K⫹ conductance block was 582 ⫾ 36 or 398 ⫾ 14 ␮M for lidocaine and mexiletine, respectively. The inhibition of Na⫹ and K⫹ currents consecutively altered the properties of single action potentials and reduced the firing rate of tonically firing and adapting-firing neurons. CONCLUSIONS: In clinically relevant concentrations, lidocaine and mexiletine reduced the excitability of sensory dorsal horn neurons via a blockade of Na⫹ and K⫹ channels. Our work confirms that, in addition to the peripheral effects of lidocaine and mexiletine, modulation of voltage-gated ion channels in the central nervous system contributes to the antinociceptive effects of these drugs used in pain therapy. (Anesth Analg 2009;109:258 –64)

T

he mechanisms of pain secondary to peripheral nerve injury are complex. Experimental evidence collected thus far indicates that peripheral nerve injury causes signal generation in the damaged nerves and in their sensory neurons.1 The initiated action potentials This article has supplementary material on the Web site: www.anesthesia-analgesia.org.

From the *Departments of Anaesthesiology, Intensive Care Medicine, Pain Therapy, University Clinic Giessen and Marburg GmbH, Giessen, Germany; †Experimental Anaesthesiology, University Clinic of Anaesthesia and Intensive Care Medicine, Medical University of Graz, Austria; ‡Department of Pulmonology, University Clinic of Internal Medicine, Medical University of Graz, Austria; and §Department of Physiology, Justus-Liebig-University, Giessen, Germany. Accepted for publication January 11, 2009. MW was supported by the Justus Liebig University Giessen, AO was supported by the Justus Liebig University Giessen and by the DFG (SFB 547 Kardiopulmonales Gefa¨ßsystem). Address correspondence and reprint requests to Andrea Olschewski, MD, PhD, Department of Anesthesiology, University Clinic of Anaesthesiology and Intensive Care Medicine, Medical University Graz, Auenbruggerplatz 29, A-8036 Graz, Austria. Address e-mail to [email protected]. Copyright © 2009 International Anesthesia Research Society DOI: 10.1213/ane.0b013e3181a3d5d8

258

(AP) lead to spinal activation sensed as painful by humans and animals.2 In addition to the altered afferent traffic, spinal sensitization and facilitatory processes in dorsal horn neurons may contribute to the aberrant encoding of the afferent traffic leading to spinal outflow interpreted by supraspinal centers as noxious.3 Voltage-gated sodium channel blockers may play crucial roles in nociception.4 Clinically used Na⫹ channel blockers, such as lidocaine and mexiletine, show beneficial effects in treating neuropathic pain when added systemically as a sole drug or as adjuvants.5–7 Importantly, these effects may be observed at plasma concentrations which do not alter the conducted potential. Because of the high first-pass effect, lidocaine can only be administered IV. The drawbacks of continuous IV therapy, cost, and invasiveness of the treatment usually preclude its use in long-term treatment. Mexiletine, the oral congener of lidocaine, extends the use of IV lidocaine therapy and provides an alternative long-term approach in neuropathic pain treatment. The analgetic actions of lidocaine and mexiletine might be generated through central, peripheral, or mixed mechanisms.8 Although some evidence suggests the central nervous system as the main target of Vol. 109, No. 1, July 2009

these two drugs,9,10 the related mechanisms must still be described in detail. Based on behavioral, anatomical, and electrophysiological data, substantia gelatinosa neurons of the spinal cord form the first relay for a variety of different fiber types, particularly for those conveying nociceptive information via small diameter afferent fibers.11 Thus, they play a pivotal role in the maintenance of aberrant somatosensory transmissions associated with nerve injury.12,13 Regarding their firing patterns as a response to a long depolarizing pulse, we classified spinal dorsal horn neurons into three major physiological groups: tonically firing neurons (TFNs) or category 1 neurons, adapting-firing (AFNs) or category 2 neurons, and single spike neurons (SSNs) or category 3 neurons, supported by investigations showing that the substantia gelatinosa is formed by neurons with diverse intrinsic firing properties.11,14 –17 TFNs are characterized by maintained firing to intracellular depolarizing current pulses and by little spike frequency adaptation during sustained depolarization. They respond to cutaneous and visceral nociceptive stimulation and to nociceptive thermal stimuli.18 AFNs generate adapted series of APs dependent on the received information from the same nocireceptors.18 Stimulation of the excitatory field results in depolarization of TNFs and AFNs and increased AP firing. SSNs generate only up to two APs and can act as coincidence detectors encoding information by detecting the occurrence of simultaneous yet separate input signals.19 Thus, TFNs and AFNs neurons represent an important pharmacological site for the antinociceptive action of different drugs in the central nervous system. Thus, this study focuses on the underlying mechanisms of how lidocaine and mexiletine affect the excitability of different types of dorsal horn neurons involved in pain transmission.

METHODS Preparation of Dorsal Horn Neurons Experiments were performed by means of the patch-clamp technique20 on 200 ␮m slices, prepared from the lumbar spinal cord (L3-6) of young rats (2–7 wk old) of both sexes.21 All animals were killed by concussion and rapid decapitation according to the standards of the German guidelines. The procedure was approved by the Institutional Animal Care Committee and reported to the Local Veterinarian Authority in Giessen (Regierungspra¨sidium Giessen, Deutschland). The spinal cord was carefully removed and put into ice-cold preparation solution enriched with O2–CO2 (95%–5%). The pial membrane of the spinal cord was removed and the spinal cord was embedded in 2% agar. The spinal cord was sliced and then incubated for 1 h at 32°C. The standard procedure of cell cleaning by repetitive blowing and suction of the bath solution via a broken Vol. 109, No. 1, July 2009

patch pipette was not applied because each slice contained numerous dorsal horn neurons with clean surfaces.

Chemicals and Solutions Detailed description is given in Online Supplement (available at www.anesthesia-analgesia.org).

Electrophysiology Identification of Dorsal Horn Neurons In spinal cord slices, dorsal horn neurons were identified as multipolar cells with a soma (8 –12 ␮m diameter) located in the substantia gelatinosa.22 Details are given in the Online Supplement (available at www.anesthesia-analgesia.org). Resting potentials in intact neurons were measured between ⫺78 and ⫺50 mV and the input resistance was 1.2 ⫾ 0.4 G⍀. Entire Soma Isolation (ESI) Method Experiments in voltage clamp mode were performed using the method of ESI to reduce series resistance. Identification of a neuron in the spinal cord slice was followed by the isolation procedure monitored under infrared optics (Hamamatsu Photonics, Japan). A detailed description of the ESI method has been given in the Online Supplement (available at www.anesthesia-analgesia.org). Current Recording Whole-cell recordings were performed as previously described.21,23 For a detailed description see the Online Supplement (available at www.anesthesia-analgesia.org).

Statistical Analysis and Fitting Numerical values are expressed as mean ⫾ se of the mean. The normalized current amplitudes in the concentration-effect curves were fitted using a nonlinear least-squares method with the equation: f(C) ⫽ 1 ⫻ (1 ⫹ C(IC50)⫺1)n)⫺1. C is the blocker concentration, half-maximal inhibitory concentration (IC50) the half-maximal inhibiting concentration, and n the Hill coefficient.21 For each individual recording, the firing frequency was determined as f ⫽ (N ⫺ 1) ⫻ (⌬T)⫺1, whereas N is the number of spikes and ⌬T the time interval between the first and the last spike. Intergroup differences were assessed by a factorial analysis of variance with post hoc analysis using Fisher’s least significant difference test. Student’s paired t-test was used to compare the frequency of five repetitive current pulses before and after mexiletine or lidocaine application. Significance is assumed at the value P ⬍ 0.05.

RESULTS The effects of the Na⫹ channel blockers mexiletine and lidocaine were evaluated in voltage-clamp and current-clamp experiments. Na⫹ currents were recorded in external TEA-solution. Pipettes were filled © 2009 International Anesthesia Research Society

259

Figure 1. Concentration dependence of current suppression by mexiletine and lidocaine. (A) Concentration–inhibition curves for tonic (䡺 lidocaine; E mexiletine) and use-dependent (f lidocaine; ● mexiletine) block. Current amplitudes (I) were normalized by the amplitude of the corresponding current recorded in control solution (I0). Data points were fitted using the Hill equation. The IC50 value was 89 ⫾ 2 ␮M (n ⫽ 5) for lidocaine and 54 ⫾ 6 ␮M (n ⫽ 5) for mexiletine for tonic inhibition and 32 ⫾ 3 ␮M (n ⫽ 5) for lidocaine and 18 ⫾ 2 ␮M (n ⫽ 5) for mexiletine for use-dependent block. The Hill coefficient was 0.9. (B) Concentration dependence of KDR current suppression by lidocaine and mexiletine (n ⫽ 5 for each concentration). The IC50 value was 582 ⫾ 36 ␮M (n ⫽ 5) for lidocaine and 398 ⫾ 14 ␮M (n ⫽ 5) for mexiletine. The Hill coefficients were 0.9 and 0.8.

with high-Csin solution. At a holding potential (E) of ⫺80 mV, lidocaine and mexiletine reversibly inhibited tetrodotoxin-sensitive (TTXs) Na⫹ currents (Fig. 1A). The current inhibition induced by these two drugs was concentration-dependent and complete at high concentrations. The concentration-effect curves of lidocaine or mexiletine revealed tonic inhibition IC50 of 89 ⫾ 2 ␮M (n ⫽ 5) or 54 ⫾ 6 ␮M (n ⫽ 5), respectively (Fig. 1A). Lidocaine and mexiletine produced a usedependent block of the currents at 10-Hz stimulation. The half-maximum inhibiting concentration for the use-dependent inhibition was 18 ⫾ 3 ␮M (n ⫽ 5) for lidocaine and 15 ⫾ 2 ␮M (n ⫽ 5) for mexiletine. Delayed-rectifier K⫹ (KDR) currents were recorded in external choline-Cl solution supplemented with TTX using pipettes filled with high-Kin solution. Figure 1B demonstrates the inhibitory effect of lidocaine and 260

Excitability of Dorsal Horn Neurons

mexiletine showing that mexiletine and lidocaine reduced the KDR current. The concentration-effect curves of lidocaine or mexiletine revealed tonic inhibition IC50 of 582 ⫾ 36 ␮M (n ⫽ 5) or 398 ⫾ 14 ␮M (n ⫽ 5), respectively (Fig. 1B). The KDR channels inhibition was reversible. To investigate the impact of current inhibition on the excitability of the neurons involved in pain transmission, intact TFNs, AFNs, and SSNs dorsal horn neurons were examined by current clamp. The SSNs, AFNs, and TNFs had resting membrane potentials of ⫺55 ⫾ 2 mV (n ⫽ 50), ⫺54 ⫾ 2 mV (n ⫽ 31), or ⫺58 ⫾ 2 mV (n ⫽ 27), respectively. Application of lidocaine or mexiletine had no effect on resting membrane potential or input resistance of the neurons at clinically relevant concentrations. First, we compared the single APs after application of lidocaine or mexiletine using 1 ms depolarizing current pulses. All three types of neurons responded to mexiletine and lidocaine in a dose-dependent manner. The peak amplitude of the single AP was decreased, the duration of the APs (measured at half-maximal amplitude) was increased, and the maximum positive slope and negative slope decreased, indicating a Na⫹ or KDR current blockade, respectively (Tables 1 and 2). The changes in repetitive firing behavior after application of mexiletine and lidocaine were investigated on TFNs and AFNs dorsal horn neurons. Figure 2 illustrates the effects of mexiletine and lidocaine in clinically relevant concentrations on a series of APs in TFNs. The lowest concentration of mexiletine (0.5 ␮M) reduced the number of APs from 9.6 ⫾ 1.9 to 4.2 ⫾ 1 (n ⫽ 5). It is noteworthy that 60% of the neurons were still able to generate a series of APs. The maximum firing frequency was reduced to 44%. In contrast, 5 and 50 ␮M mexiletine virtually abolished the repetitive firing neuronal activity. Mexiletine (5 ␮M) decreased the number of APs from 9.2 ⫾ 1.6 to 1.8 ⫾ 0.3 (n ⫽ 6; Fig. 2) and 50 ␮M decreased from 7.2 ⫾ 1.6 to 1.0 ⫾ 0.6 (n ⫽ 5; not shown). The effects of lidocaine in clinically relevant concentrations corresponded to that observed with mexiletine in low concentrations. Lidocaine (5 ␮M) significantly reduced the generation of APs (from 7.2 ⫾ 1.3 to 1.6 ⫾ 0.2; n ⫽ 5). After lidocaine application, a series of APs was detected only in one neuron. Lidocaine applied in a higher concentration (30 ␮M) completely abolished the AP series generated during the long depolarization pulses (number of APs from 8.2 ⫾ 0.9 to 1 ⫾ 0; n ⫽ 6; not shown). Mexiletine and lidocaine were also tested in AFNs (not shown). Mexiletine (0.5, 5, and 50 ␮M) reduced the maximum number of APs from 3 ⫾ 0 to 1.2 ⫾ 0.2 (n ⫽ 5; P ⬍ 0.05), 5.3 ⫾ 1.9 to 1.3 ⫾ 0.3 (n ⫽ 4; P ⬍ 0.05), or 4 ⫾ 0.4 to 1 ⫾ 0 (n ⫽ 7; P ⬍ 0.05), respectively. Similar results were obtained with lidocaine. The decrease in the maximum number of APs after application of 5 ␮M lidocaine was 63% (from 5.3 ⫾ 0.6 to 2 ⫾ 0.3; n ⫽ 4). ANESTHESIA & ANALGESIA

Table 1. The Effects of Increasing Mexiletine Concentrations on Single Action Potentials Overshoot (mV) Duration (ms) Max. positive slope (Vsⴚ1) Max. negative slope (Vsⴚ1) SSNs Control (␮M) 0.5 5 50 TFNs Control (␮M) 0.5 5 50 AFNs Control (␮M) 0.5 5 50

n

25.2 ⫾ 1.8 23.5 ⫾ 3.2 9.1 ⫾ 1.9* 7.2 ⫾ 3.5*

2.8 ⫾ 0.1 3.1 ⫾ 0.3 3.9 ⫾ 0.3* 5.4 ⫾ 0.6*

81.1 ⫾ 5.7 67.8 ⫾ 8.4 32.6 ⫾ 5.6* 18.8 ⫾ 7.6*

⫺37.2 ⫾ 1.6 ⫺34.6 ⫾ 2.3 ⫺19.3 ⫾ 2.5* ⫺13.3 ⫾ 3.6*

9 9 11 6

39.3 ⫾ 0.4 38.3 ⫾ 2.3 28.3 ⫾ 3.1† 11.5 ⫾ 7.1*

2.2 ⫾ 0.1 2.2 ⫾ 0.1 3.0 ⫾ 0.1 5.7 ⫾ 1.3*

131.6 ⫾ 1.5 119.7 ⫾ 6.6 74.5 ⫾ 8.2* 41.3 ⫾ 23.6*

⫺57.8 ⫾ 2.8 ⫺54.7 ⫾ 5.2 ⫺40.3 ⫾ 4.1† ⫺16.8 ⫾ 6.7*

7 7 6 5

34.5 ⫾ 1.7 30.1 ⫾ 5.8 21.7 ⫾ 4.8* 14.9 ⫾ 4*

2.9 ⫾ 0.2 3.3 ⫾ 0.5 4.1 ⫾ 0.7 5.2 ⫾ 0.9*

106.7 ⫾ 6.5 85.0 ⫾ 18.5 63.4 ⫾ 17.7† 42.1 ⫾ 11.2*

⫺40.6 ⫾ 3.2 ⫺36.6 ⫾ 7.9 ⫺28.4 ⫾ 7.9 ⫺17.3 ⫾ 3.8*

5 5 5 5

The duration of the action potentials is measured at the half-maximum potential. Positive and negative slope are the maximum values measured at the rising and the falling phase. SSNs ⫽ single spike neurons; TFNs ⫽ tonically firing neurons; AFNs ⫽ adapting-firing neurons. Significance levels are expressed as * P ⬍ 0.001, † P ⬍ 0.01, and ‡ P ⬍ 0.05 compared with controls.

Table 2. The Effect of Increasing Lidocaine Concentrations on Single Action Potentials Overshoot (mV) Duration (ms) Max. positive slope (Vs⫺1) Max. negative slope (Vs⫺1) n SSNs Control (␮M) 5 30 TFNs Control (␮M) 5 30 AFNs Control (␮M) 5 30

24.2 ⫾ 1.8 13.5 ⫾ 1.8* 5.0 ⫾ 0.9*

3.0 ⫾ 0.2 4.4 ⫾ 0.5† 5.4 ⫾ 0.7*

81.0 ⫾ 7.6 44.3 ⫾ 8.9* 20.2 ⫾ 2.5*

⫺37.5 ⫾ 3.0 ⫺23.4 ⫾ 4.0* ⫺12.2 ⫾ 1.3*

7 7 6

38.0 ⫾ 3.1 28.0 ⫾ 5.9 21.9 ⫾ 3.1*

2.5 ⫾ 0.2 2.8 ⫾ 0.3 3.9 ⫾ 0.3*

122 ⫾ 9.6 93.2 ⫾ 10.7 63.7 ⫾ 10.4*

⫺52.4 ⫾ 3.3 ⫺39.6 ⫾ 3.8† ⫺27.3 ⫾ 3.4*

5 5 7

28.3 ⫾ 1.8 18.9 ⫾ 2.4* 6.4 ⫾ 2.3*

2.6 ⫾ 0.3 3.6 ⫾ 0.5 4.8 ⫾ 0.5*

96.5 ⫾ 7.2 68.3 ⫾ 6.1† 23.4 ⫾ 6.8*

⫺46.2 ⫾ 3.1 ⫺32.6 ⫾ 4.2† ⫺14.2 ⫾ 3.0*

7 7 8

The duration of the action potentials is measured at the half-maximum potential. Positive and negative slope are the maximum values measured at the rising and the falling phase. SSNs ⫽ single spike neurons; TFNs ⫽ tonically firing neurons; AFNs ⫽ adapting-firing neurons. Significance levels are expressed as * P ⬍ 0.001, † P ⬍ 0.05, and ‡ P ⬍ 0.01 compared with controls.

Figure 2. Decrease of firing frequency by mexiletine and lidocaine. (A) Series of action potentials are shown under control conditions (left) and after application of 0.5 (middle) and 5 ␮M (right) mexiletine. Dotted lines present holding potential of ⫺80 mV. (B) The maximum firing frequency in control solutions was compared with the frequency after application of 5 ␮M lidocaine. Action potentials were evoked by 500 ms current pulses. Dotted lines indicate holding potential of ⫺80 mV.

After administration of 30 ␮M lidocaine, the maximum number of APs was reduced to 20% (from 4.9 ⫾ 0.6 to 1.3 ⫾ 0.4; n ⫽ 8). Vol. 109, No. 1, July 2009

Voltage-dependent K⫹ channels and Na⫹ channels have been shown to influence interspike intervals and firing rates in neurons of the substantia gelatinosa.24,25 © 2009 International Anesthesia Research Society

261

Figure 3. Mexiletine reduces the maximum action potential frequency after repetitive pulses with different interspike intervals. Figures show the effects under control conditions (left), after application of 0.5 (middle, above) and 5 ␮M (middle, below) mexiletine and superimposed images (right). Impulse protocol is presented below the registration.

Figure 4. Lidocaine reduces the maximum action potential frequency after repetitive pulses with different interspike intervals. The maximum action potential frequency was decreased with increasing concentrations of lidocaine. Figures show currents in control solutions (left) and after applications of 5 (middle, above) and 30 ␮M (middle, below) lidocaine and superimposed images (right). Membrane potential was adjusted to ⫺80 mV. Impulse protocol is shown below the registration. Therefore, we examined the effects of mexiletine and lidocaine on interspike intervals in response to five repetitive pulses in dorsal horn neurons. The mean frequency of the dorsal horn neurons was 231 ⫾ 9 Hz (n ⫽ 18), 207 ⫾ 10 Hz (n ⫽ 29), 204 ⫾ 10 Hz (n ⫽ 19), in SSNs, AFNs, or TFNs, respectively. Analysis of the action of mexiletine and lidocaine on all types of dorsal horn neurons showed a dose-dependent increase of the interspike intervals and a dose-dependent decrease of the maximum possible frequency (Figs. 3 and 4 and Table 3).

DISCUSSION Despite the increasing number of drugs tested for the treatment of neuropathic pain conditions during the last decade, Na⫹ channel blockers, such as local 262


anesthetics, still remain an important and effective therapeutic tool in modern pain therapy.5,26,27 In chronic pain therapy, lidocaine infusions are frequently used to help to identify those patients most likely to benefit from an oral treatment of Na⫹ channel blockers. Patients, who have once achieved therapeutic lidocaine levels, usually show a good response to orally administered mexiletine.28,29 Although the clinical responses to both local anesthetics are relatively well described, it is not characterized in detail how lidocaine and mexiletine affect the excitability of central sensory neurons. Therefore, we have compared the effects of lidocaine and mexiletine on the repetitive firing behavior of substantia gelatinosa neurons via direct patch-clamp recording from visualized intact neurons of the spinal cord and found that lidocaine and mexiletine exert similar effects on the neuron excitability by blocking both, the TTXs Na⫹ (only TTXs Na channels are expressed in dorsal horn neurons)30 and the delayed-rectifier K⫹ conductances, at clinically relevant concentrations. Substantia gelatinosa neurons of the spinal cord receive primary afferent input that encodes nociceptive information conducted by C- and/or A␦-fibers.31–33 In the chronic pain state, abnormal activities are produced in the periphery. The dorsal horn neurons become sensitized and respond more vigorously to peripheral input, indicating that they represent a key element in the transmission of pain-related information and thus are of particular interest for pain therapy. During the last decades, different cell types, like TFNs, AFNs, and SSNs were described in the substantia gelatinosa of the spinal cord. Our investigations show that lidocaine and mexiletine both administered in clinically relevant concentrations significantly affect the single action potential properties in all three types of neurons. Furthermore, lidocaine and mexiletine decreased the frequency and stability of neuron firing involved in nociceptive pain processing. The lidocaine and mexiletine concentrations that evoke these cellular effects correspond to the concentrations necessary for antinociceptive action during systemic administration in clinical studies. Plasma concentrations of lidocaine after IV administration of 2–2.5 mg/kg generally reach 10 – 60 ␮M.5,34 The steady state of the plasma mexiletine level after an oral dose of 450 mg/d ranged from 4 to 7 ␮M.35 In this study, both local anesthetics decreased the Na⫹ current, reduced the amplitude, decreased the maximum rate of rise, and increased the width of a single AP in a concentrationdependent manner. Inhibition of Na⫹ channels by lidocaine or mexiletine have been also shown in peripheral nerve and dorsal ganglion neurons36 –38 and in neurons of the central nervous system.22,39 Our results support the assumption that reduced numbers of voltage-gated Na⫹ channels affect single impulse generation, which is also in good agreement with previous findings from electrophysiological studies.16,38 ANESTHESIA & ANALGESIA

Table 3. Mexiletine and Lidocaine Decreases the theoretical Maximum Possible Frequency in Dorsal Horn Neurones Group

Drug

Concentration (␮M)

Reduction in maximum firing frequency

n

SSNs

Mexiletine

0.5 5 50 5 30 0.5 5 50 5 30 0.5 5 50 5 30

10 ⫾ 4% 21 ⫾ 7%* 41 ⫾ 6%† 23 ⫾ 11% 27 ⫾ 4%* 3 ⫾ 2% 23 ⫾ 8%* 44 ⫾ 9%* 8 ⫾ 4% 20 ⫾ 2%* 9 ⫾ 4% 16 ⫾ 4%* 40 ⫾ 9%‡ 10 ⫾ 4% 22 ⫾ 6%*

n⫽7 n⫽9 n⫽6 n⫽6 n⫽6 n⫽7 n⫽6 n⫽5 n⫽5 n⫽5 n⫽5 n⫽5 n⫽5 n⫽8 n⫽5

Lidocaine TFNs

Mexiletine Lidocaine

AFNs

Mexiletine Lidocaine

For a detailed description of the frequency calculation after application of repetitive short current pulses see Methods. The theoretical maximum firing frequency is the maximum frequency the neuron is able to follow stimulation. SSNs ⫽ single spike neurons; TFNs ⫽ tonically firing neurons; AFNs ⫽ adapting-firing neurons. Significance levels are expressed as * P ⬍ 0.05, † P ⬍ 0.001, and ‡ P ⬍ 0.01 compared with control.

The delayed-rectifier K⫹ current underlies the major K⫹ conductance in AFNs and TFNs.16,17 In clinically relevant concentrations both lidocaine and mexiletine caused a significant decrease in the delayed-rectifier K⫹ current. Until recently, only a few studies could show that both drugs are able to block neuronal voltagegated K⫹ channels,40 – 42 whereas two-pore domain K⫹ channels43,44 or ATP-dependent K⫹ channels in other cell types45 are sensitive to local anesthetics. The reduction of delayed-rectifier K⫹ channels in dorsal horn neurons induced by lidocaine and mexiletine could account for the significant reduction of the firing frequency as shown in studies by Melnik et al.16,17 Although the inhibitory effects of lidocaine and mexiletine on the delayed-rectifier K⫹ current was less pronounced than on the Na⫹ current, the blockade of the delayed-rectifier K⫹ current could become important if these neurons had small resources of K⫹currents, because the safety factor for K⫹-channels seems to be lower than two.24 Progress in understanding the role of ion channels in repetitive firing behavior of subtantia gelatinosa neurons has demonstrated a complex interaction between Na⫹ and delayed-rectifier K⫹ conductances.17,24 In TFNs the voltage-gated Na⫹ and delayed-rectifier K⫹ channels were shown to generate the basic pattern of tonic firing, whereas Ca2⫹-dependent conductances stabilized firing and regulated discharge frequency.17 In AFNs Ca2⫹-dependent conductances do not contribute to adapting firing but Na⫹ channels seem to be critical for determining the appearance of spike frequency adaptation.16 Because of the finding that AFNs and TFNs are key elements for nociception in the central nervous system, the modulation of their ion channels generating APs provides an important therapeutic approach for the treatment of neuropathic pain. In this study, lidocaine and mexiletine changed the single AP properties through a blockade of the Na⫹ and delayed-rectifier K⫹ channels. Consecutively, the Vol. 109, No. 1, July 2009

frequency of APs is decreased. Furthermore, both drugs show similar effectiveness for reducing the excitability of all types of dorsal horn neurons at lower concentrations. Thus, our work clearly confirms that, in addition to their peripheral effects, the modulation of voltage-gated ion channels in the central nervous system contributes to the antinociceptive effects of lidocaine and mexiletine used in clinical pain therapy. ACKNOWLEDGMENTS The authors thank Brigitte Agari and Otto Becker for their excellent technical assistance. REFERENCES 1. Study RE, Kral MG. Spontaneous action potential activity in isolated dorsal root ganglion neurons from rats with a painful neuropathy. Pain 1996;65:235– 42 2. Matzner O, Devor M. Hyperexcitability at sites of nerve injury depends on voltage-sensitive Na⫹ channels. J Neurophysiol 1994;72:349 –59 3. Woolf CJ. Physiological, inflammatory and neuropathic pain. Adv Tech Stand Neurosurg 1987;15:39 – 62 4. Cummins TR, Sheets PL, Waxman SG. The roles of sodium channels in nociception: implications for mechanisms of pain. Pain 2007;131:243–57 5. Tanelian DL, Brose WG. Neuropathic pain can be relieved by drugs that are use-dependent sodium channel blockers: lidocaine, carbamazepine, and mexiletine. Anesthesiology 1991;74: 949 –51 6. Stracke H, Meyer UE, Schumacher HE, Federlin K. Mexiletine in the treatment of diabetic neuropathy. Diabetes Care 1992;15: 1550 –5 7. Jarvis B, Coukell AJ. Mexiletine. A review of its therapeutic use in painful diabetic neuropathy. Drugs 1998;56:691–707 8. Ossipov MH, Lai J, Malan TP Jr, Porreca F. Spinal and supraspinal mechanisms of neuropathic pain. Ann N Y Acad Sci 2000; 909:12–24 9. Woolf CJ, Wiesenfeld-Hallin Z. The systemic administration of local anaesthetics produces a selective depression of C-afferent fibre evoked activity in the spinal cord. Pain 1985;23:361–74 10. Bach FW, Jensen TS, Kastrup J, Stigsby B, Dejgard A. The effect of intravenous lidocaine on nociceptive processing in diabetic neuropathy. Pain 1990;40:29 –34 11. Thomson AM, West DC, Headley PM. Membrane Characteristics and synaptic responsiveness of superficial dorsal horn neurons in a slice preparation of adult rat spinal cord. Eur J Neurosci 1989;1:479 – 88 © 2009 International Anesthesia Research Society

263

12. Chapman V, Suzuki R, Dickenson AH. Electrophysiological characterization of spinal neuronal response properties in anaesthetized rats after ligation of spinal nerves L5–L6. J Physiol 1998;507:881–94 13. Waxman SG, Hains BC. Fire and phantoms after spinal cord injury: Na⫹ channels and central pain. Trends Neurosci 2006; 29:207–15 14. Grudt TJ, Perl ER. Correlations between neuronal morphology and electrophysiological features in the rodent superficial dorsal horn. J Physiol 2002;540:189 –207 15. Lu Y, Perl ER. A specific inhibitory pathway between substantia gelatinosa neurons receiving direct C-fiber input. J Neurosci 2003;23:8752– 8 16. Melnick IV, Santos SF, Safronov BV. Mechanism of spike frequency adaptation in substantia gelatinosa neurons of rat. J Physiol 2004;559:383–95 17. Melnick IV, Santos SF, Szokol K, Szucs P, Safronov BV. Ionic basis of tonic firing in spinal substantia gelatinosa neurons of rat. J Neurophysiol 2004;91:646 –55 18. Lopez-Garcia JA, King AE. Membrane properties of physiologically classified rat dorsal horn neurons in vitro: correlation with cutaneous sensory afferent input. Eur J Neurosci 1994;6:998 –1007 19. Prescott SA, De KY. Four cell types with distinctive membrane properties and morphologies in lamina I of the spinal dorsal horn of the adult rat. J Physiol 2002;539:817–36 20. Hamill OP, Marty A, Neher E, Sakmann B, Sigworth FJ. Improved patch-clamp techniques for high-resolution current recording from cells and cell-free membrane patches. Pflugers Arch 1981;391:85–100 21. Schnoebel R, Wolff M, Peters SC, Braeu ME, Scholz A, Hempelmann G, Olschewski H, Olschewski A. Ketamine impairs excitability in superficial dorsal horn neurones by blocking sodium and voltage-gated potassium currents. Br J Pharmacol 2005;146:826 –33 22. Olschewski A, Hempelmann G, Vogel W, Safronov BV. Blockade of Na⫹ and K⫹ currents by local anesthetics in the dorsal horn neurons of the spinal cord. Anesthesiology 1998;88:172–9 23. Wolff M, Olschewski A, Vogel W, Hempelmann G. Meperidine suppresses the excitability of spinal dorsal horn neurons. Anesthesiology 2004;100:947–55 24. Olschewski A, Hempelmann G, Vogel W, Safronov BV. Suppression of potassium conductance by droperidol has influence on excitability of spinal sensory neurons. Anesthesiology 2001; 94:280 –9 25. Macica CM, von Hehn CA, Wang LY, Ho CS, Yokoyama S, Joho RH, Kaczmarek LK. Modulation of the kv3.1b potassium channel isoform adjusts the fidelity of the firing pattern of auditory neurons. J Neurosci 2003;23:1133– 41 26. Tremont-Lukats IW, Challapalli V, McNicol ED, Lau J, Carr DB. Systemic administration of local anesthetics to relieve neuropathic pain: a systematic review and meta-analysis. Anesth Analg 2005;101:1738 – 49 27. Tremont-Lukats IW, Hutson PR, Backonja MM. A randomized, double-masked, placebo-controlled pilot trial of extended IV lidocaine infusion for relief of ongoing neuropathic pain. Clin J Pain 2006;22:266 –71 28. Galer BS, Harle J, Rowbotham MC. Response to intravenous lidocaine infusion predicts subsequent response to oral mexiletine: a prospective study. J Pain Symptom Manage 1996;12: 161–7

264


29. Nathan A, Rose JB, Guite JW, Hehir D, Milovcich K. Primary erythromelalgia in a child responding to intravenous lidocaine and oral mexiletine treatment. Pediatrics 2005;115:e504 – e07 30. Olschewski A, Brau ME, Hempelmann G, Vogel W, Safronov BV. Differential block of fast and slow inactivating tetrodotoxinsensitive sodium channels by droperidol in spinal dorsal horn neurons. Anesthesiology 2000;92:1667–76 31. LaMotte C. Distribution of the tract of Lissauer and the dorsal root fibers in the primate spinal cord. J Comp Neurol 1977;172: 529 – 61 32. Rethelyi M. Preterminal and terminal axon arborizations in the substantia gelatinosa of cat’s spinal cord. J Comp Neurol 1977;172:511–21 33. Light AR, Trevino DL, Perl ER. Morphological features of functionally defined neurons in the marginal zone and substantia gelatinosa of the spinal dorsal horn. J Comp Neurol 1979;186: 151–71 34. Wallace MS, Dyck JB, Rossi SS, Yaksh TL. Computer-controlled lidocaine infusion for the evaluation of neuropathic pain after peripheral nerve injury. Pain 1996;66:69 –77 35. Ohashi K, Ebihara A, Hashimoto T, Hosoda S, Kondo K, Oka T. Pharmacokinetics and the antiarrhythmic effect of mexiletine in patients with chronic ventricular arrhythmias. Arzneimittelforschung 1984;34:503–7 36. Brau ME, Vogel W, Hempelmann G. Fundamental properties of local anesthetics: half-maximal blocking concentrations for tonic block of Na⫹ and K⫹ channels in peripheral nerve. Anesth Analg 1998;87:885–9 37. Brau ME, Dreimann M, Olschewski A, Vogel W, Hempelmann G. Effect of drugs used for neuropathic pain management on tetrodotoxin-resistant Na⫹ currents in rat sensory neurons. Anesthesiology 2001;94:137– 44 38. Scholz A, Vogel W. Tetrodotoxin-resistant action potentials in dorsal root ganglion neurons are blocked by local anesthetics. Pain 2000;89:47–52 39. Akada Y, Ogawa S, Amano K, Fukudome Y, Yamasaki F, Itoh M, Yamamoto I. Potent analgesic effects of a putative sodium channel blocker M58373 on formalin-induced and neuropathic pain in rats. Eur J Pharmacol 2006;536:248 –55 40. Mitcheson JS, Hancox JC. Modulation by mexiletine of action potentials, L-type Ca current and delayed rectifier K current recorded from isolated rabbit atrioventricular nodal myocytes. Pflugers Arch 1997;434:855– 8 41. Trellakis S, Benzenberg D, Urban BW, Friederich P. Differential lidocaine sensitivity of human voltage-gated potassium channels relevant to the auditory system. Otol Neurotol 2006;27: 117–23 42. Bischoff U, Brau ME, Vogel W, Hempelmann G, Olschewski A. Local anaesthetics block hyperpolarization-activated inward current in rat small dorsal root ganglion neurones. Br J Pharmacol 2003;139:1273– 80 43. La JH, Kang D, Park JY, Hong SG, Han J. A novel acid-sensitive K⫹ channel in rat dorsal root ganglia neurons. Neurosci Lett 2006;406:244 –9 44. Kindler CH, Yost CS, Gray AT. Local anesthetic inhibition of baseline potassium channels with two pore domains in tandem. Anesthesiology 1999;90:1092–102 45. Olschewski A, Brau ME, Olschewski H, Hempelmann G, Vogel W. ATP-dependent potassium channel in rat cardiomyocytes is blocked by lidocaine. Possible impact on the antiarrhythmic action of lidocaine. Circulation 1996;93:656 –9


Regional Anesthesia Section Editor: Terese T. Horlocker

A Prospective, Randomized, Controlled Trial Comparing Ultrasound Versus Nerve Stimulator Guidance for Interscalene Block for Ambulatory Shoulder Surgery for Postoperative Neurological Symptoms Spencer S. Liu, MD* Victor M. Zayas, MD* Michael A. Gordon, MD* Jonathan C. Beathe, MD* Daniel B. Maalouf, MD* Leonardo Paroli, MD* Gregory A. Liguori, MD* Jaime Ortiz, MD* Valeria Buschiazzo* Justin Ngeow, BA* Teena Shetty, MD† Jacques T. Ya Deau, MD, PhD*

BACKGROUND: Visualization with ultrasound during regional anesthesia may reduce the risk of intraneural injection and subsequent neurological symptoms but has not been formally assessed. Thus, we performed this randomized clinical trial comparing ultrasound versus nerve stimulator-guided interscalene blocks for shoulder arthroscopy to determine whether ultrasound could reduce the incidence of postoperative neurological symptoms. METHODS: Two hundred thirty patients were randomized to a standardized interscalene block with either ultrasound or nerve stimulator with a 5 cm, 22 g Stimuplex® insulated needle with 1.5% mepivacaine with 1:300,000 epinephrine and NaCO3 (1 meq/10 mL). A standardized neurological assessment tool (questionnaire and physical examination) designed by a neurologist was administered before surgery (both components), at approximately 1 wk after surgery (questionnaire), and at approximately 4 – 6 weeks after surgery (both components). Diagnosis of postoperative neurological symptoms was determined by a neurologist blinded to block technique. RESULTS: Two hundred nineteen patients were evaluated. Use of ultrasound decreased the number of needle passes for block performance (1 vs 3, median, P ⬍ 0.001), enhanced motor block at the 5-min assessment (P ⫽ 0.04) but did not decrease block performance time (5 min for both). No patient required conversion to general anesthesia for failed block, and patient satisfaction was similar in both groups (96% nerve stimulator and 92% ultrasound). The incidence of postoperative neurological symptoms was similar at 1 wk follow-up with 11% (95% CI of 5%–17%) for nerve stimulator and 8% (95% CI of 3%–13%) for ultrasound and was similar at late follow-up with 7% (95% CI of 3%–12%) for nerve stimulator and 6% (95% CI of 2%–11%) for ultrasound. The severity of postoperative neurological symptoms was similar between groups with a median patient rating of moderate. Symptoms were primarily sensory and consisted of pain, tingling, or paresthesias. CONCLUSIONS: Ultrasound reduced the number of needle passes needed to perform interscalene block and enhanced motor block at the 5 min assessment; however, we did not observe significant differences in block failures, patient satisfaction or incidence, and severity of postoperative neurological symptoms. (Anesth Analg 2009;109:265–71)

U

ltrasound guidance for regional anesthesia has increased in popularity. A recent systematic review of randomized controlled trials (RCTs) comparing ultrasound guidance with conventional techniques1 noted This article has supplementary material on the Web site: www.anesthesia-analgesia.org.

From the Departments of *Anesthesiology, and †Neurology, Hospital for Special Surgery, Weill College of Medicine of Cornell University, New York. Accepted for publication January 8, 2009. Supported by the Department of Anesthesiology, Hospital for Special Surgery. Spencer S. Liu is the Section Editor of Pain Medicine for the Journal. This manuscript was handled by Terese T. Horlocer, Section Editor of Regional Anesthesia, and Dr. Liu was not involved in any way with the editorial process or decision. Vol. 109, No. 1, July 2009

similar efficacy between ultrasound guidance versus nerve stimulator when regional anesthesia was performed by experts. However, the ability of ultrasound to dynamically visualize needle placement to avoid intraneural contact and injection of local anesthetic may offer a greater safety margin to avoid neurological injury after peripheral nerve blocks.2 Although nerve stimulator guidance is the current technique of Presented in part at the annual meeting of the American Society of Regional Anesthesia, Cancun, Mexico, May 2, 2008. Reprints will not be available from the author. Address correspondence to Dr. Spenser S. Liu, Department of Anesthesiology, Hospital for Special Surgery, 535 East 70th St., New York City, NY 10021. Address e-mail to [email protected]. Copyright © 2009 International Anesthesia Research Society DOI: 10.1213/ane.0b013e3181a3272c

265

choice, it may have suboptimal ability to detect intraneural needle placement. Both laboratory and clinical studies suggest that use of a minimal nerve stimulator current threshold (e.g., mA ⫽ 0.3– 0.5) as a cutoff to detect intraneural needle placement and subsequent increased risk for neurological injury may be neither sensitive nor specific.3–5 Interscalene blocks are commonly performed to provide anesthesia and analgesia for shoulder surgery but have a comparatively frequent incidence of postoperative neurological symptoms.6 Incidences of postoperative neurological symptoms within the first week after interscalene block for shoulder surgery typically range from 4% to 16%.7–9 The etiology of these postoperative neurological symptoms is unclear, and permanent nerve injury after shoulder arthroscopy appears to be rare (approximately 0.1% to 0.2% without specifications on type of anesthesia).7,8 Nonetheless, the frequent initial incidence of this distressing complication led us to compare ultrasound versus nerve stimulator-guided interscalene blocks for shoulder arthroscopy to determine whether ultrasound could reduce the incidence of postoperative neurological symptoms at 1 wk after surgery.

METHODS After obtaining approval by our IRB to conduct this prospective, randomized clinical trial, 230 patients scheduled to undergo an outpatient shoulder arthroscopy under interscalene block and sedation gave written, informed consent. Exclusion criteria were age younger than 18 or older than 75 yr and typical contraindication to interscalene block including patient refusal, pregnancy, dementia, severe pulmonary disease, and known preexisting neurological disorders involving the operative limb. Each subject underwent a standardized sensory and motor neurological evaluation and physical examination to determine baseline neurological function (Appendix 1, available at: www.anesthesia-analgesia. org). This tool was prospectively designed by our neurologist coinvestigator (TS), who subsequently trained two coinvestigators to administer this tool (VB and JN). Patients were then randomized to either ultrasound guidance or nerve stimulator guidance for interscalene block with a computer-generated random number table, using a sealed envelope sequence, and with the sequence concealed until after the enrollment of the subject. Each subject was placed supine with the usual American Society of Anesthesiologist monitors. Midazolam up to 5 mg was used for sedation at the discretion of the anesthesiologist to allow comfort and cooperation from the patient. The “interscalene area” was prepared with an antiseptic solution. The time to complete each block (from needle insertion to final needle withdrawal) and number of needle movements (each forward movement of the needle after halting or 266

Postoperative Neurological Symptoms

retracting until acquisition of end point) was recorded by an investigator not performing the block.

Nerve Stimulator Group A 5 cm, 22 g Stimuplex® insulated needle (B Braun Medical, Bethlehem, PA) was placed into the interscalene groove with the bevel oriented parallel to the groove. The initial settings for the nerve stimulating unit were a current of 0.6 –1.5 mA at 2 Hz. A motor response in the distribution of the axillary, musculocutaneous, ulnar, radial, or median nerve was accepted as evidence of correct needle placement. The current was decreased to a range between 0.2 mA and 0.5 mA while maintaining a motor response. If a motor response was still evident at a current ⬍0.2 mA or more than 0.5 mA, the needle placement was adjusted accordingly. This range of stimulation end points was meant to reflect common practice and has been used in several previous large surveys and RCTs of nerve stimulator-guided interscalene blocks for shoulder surgery.9 –12 After a 1 mL test dose to exclude severe pain or resistance on injection, local anesthetic was injected in divided doses with frequent aspiration. If pain or resistance with injection was evident, the needle placement was adjusted accordingly. For patients below 50 kg, a total dose of 45–55 mL was used. For patients ⱖ50 kg, a total dose of 55– 65 mL was used. The local anesthetic consisted of mepivacaine 1.5% with 1:300,000 epinephrine and NaCO3 (1 meq/10 mL).

Ultrasound-Guided Group A linear 10 –13 MHz ultrasound probe was used to visualize the brachial plexus. Initial ultrasound visualization was at the interscalene area. If the brachial plexus was not well visualized, then the probe was repositioned at the supraclavicular fossa, the brachial plexus visualized, and the probe tracked cephalad to follow the brachial plexus to the interscalene area. A 5 cm, 22 g Stimuplex® insulated needle (B Braun Medical) was placed through the middle scalene muscle, into the interscalene groove, and adjacent to the brachial plexus via in-plane ultrasound guidance to visualize the entire needle with the bevel oriented parallel to the interscalene groove. After a 1 mL test dose to exclude obvious intraneural injection, local anesthetic was injected in divided doses with frequent aspiration under ultrasound visualization. If intraneural injection or resistance to injection was observed at any time, then the needle was repositioned, and this observation was recorded. Type and dose of local anesthetic was identical to the nerve stimulator group. After block placement, the patient was positioned in the beach chair position for surgery. Sensory and motor block were evaluated by an investigator (VB or JN) who was aware of type of block. Motor block was evaluated by testing deltoid motor function and biceps motor function on a 0 (no movement), 1 (weak), and 2 (normal) scale every 5 min until a score of 0 was ANESTHESIA & ANALGESIA

reached or surgery commenced. Sensory function was evaluated at the same time in the distribution of the median nerve (“money sign”)13 using a 0 (numb), 1 (dysesthesia), and 2 (normal) scale. Upon commencement of surgery, the anesthesiologist who performed the block rated the effectiveness for surgical anesthesia on a scale of 2 (complete), 1 (adequate), and 0 (inadequate). A block was considered successful if rescue general anesthesia was not required. After surgery, patients were discharged from the postanesthesia care unit using our standard discharge criteria. The patients were contacted by telephone at 1 wk after surgery and the same neurological questionnaire (Appendix 1) was administered by VB or JB. Upon surgical follow-up (usually at 4 – 6 wk postoperatively), the patient was again evaluated with the same neurological questionnaire and physical examination (Appendix 1) administered by VB or JB. Results from neurological testing were recorded and compared with baseline by our neurologist coinvestigator (TS), who was blinded to interscalene block technique. Postoperative neurological symptoms were defined as neurological symptoms within the operative site brachial plexus that were related to brachial plexus irritation but were unrelated to the surgical procedure as determined by our neurologist TS. Symptoms involving the axillary or suprascapular nerves were considered to be potentially related to the surgical procedure8,14 and were not considered to be postoperative neurological symptoms. Patients were asked to rate overall severity of postoperative neurological symptoms as mild ⫽ barely noticeable, moderate ⫽ definitely noticeable, or severe ⫽ very preoccupied. All patients with postoperative neurological symptoms were offered a complete neurological evaluation and standard diagnostic testing (e.g., nerve conduction velocities, electromyography) by TS to define the cause and determine prognosis of postoperative neurological symptoms. Any patient with postoperative neurological symptoms at each follow-up was followed monthly by phone until resolution of symptoms, until lost to follow-up or submission of this manuscript.

Statistics Power Analysis The largest series from Borgeat et al.11 found an approximately 16% incidence of postoperative neurological symptoms at 1 wk after the nerve stimulator technique for interscalene block. No studies have evaluated the incidence of postoperative neurological symptoms with the use of ultrasound for interscalene block. Direct ultrasound visualization of the brachial plexus and block needle throughout the block would theoretically prevent any intraneural injection and should potentially decrease the risk of postoperative neurological symptoms. Thus, we assumed that the risk of postoperative neurological symptoms with Vol. 109, No. 1, July 2009

Figure 1. CONSORT diagram of patient flow through study protocol.

ultrasound guidance was similar to the lowest previous rate of postoperative neurological symptoms of approximately 4% at 1 wk follow-up.15 Power analysis indicated a sample size of 109 patients in each group was needed to detect a difference (4% vs 16%) between techniques (␣ ⫽ 0.05, ␤ ⫽ 0.8). To compensate for expected dropouts, we planned to enroll 230 patients. Analyses were performed using SAS version 9.1 (SAS Institute, Cary, NC). ␹2 was used to compare rates of postoperative neurological symptoms between groups and for other incident data. Continuous variables were compared with t test or Wilcoxon’s rank sum for nonparametric data. Because each group had a large sample size (⬎100), according to the central limit theorem, the sample proportion was approximately normally distributed. Thus, confidence intervals for the incidences and for the relative risks were calculated based on the normal theory tests.16

RESULTS Figure 1 displays the CONSORT flow of patients through the study protocol. Demographics were similar between groups (Table 1). Two hundred nineteen patients were available for early follow-up (nerve stimulator ⫽ 108 and ultrasound ⫽ 111), and all analyses were based on intent to treat for these patients. Perioperative characteristics are displayed in Table 2. Use of ultrasound significantly decreased the number of needle passes (1 vs 3, median, P ⬍ 0.001) © 2009 International Anesthesia Research Society

267

Table 1. Patient Demographics Nerve stimulator

Ultrasound

49 ⫾ 14 174 ⫾ 10 85 ⫾ 24 28 ⫾ 7

48 ⫾ 16 173 ⫾ 9 86 ⫾ 34 29 ⫾ 11

4 40 8 4

1 41 9 3

7 19 23 3

9 23 23 2

Nerve stimulator

Ultrasound

Age Height Weight (kg) Body mass index Shoulder arthroscopy procedure types Diagnostic Rotator cuff repair Stabilization Acromioclavicular joint resection Debridement Labral repair All decompressions Other

Table 2. Perioperative Characteristics

Needle passes* (median/ mode) Time to perform block (min) Accentuation on injection History of diabetes Attending/trainee Postoperative pain at needle site Satisfaction Definite PONS assessment at 1 wk Pons severity at 1 wk (median/mode); 1 ⫽ mild; 2 ⫽ moderate 3 ⫽ severe Definite PONS assessment at late follow-up Pons Severity at 4 wk (median/mode); 1 ⫽ mild; 2 ⫽ moderate 3 ⫽ severe

3 (1)

1 (1)

5⫾3

5⫾3

22 (20%) 4 (4%) 26/82 23 (21%)

26 (23%) 6 (5%) 40/71 16 (14%)

96% 12 (11%)

92% 9 (8%)

2 (2)

2 (1)

8 (7%)

7 (6%)

1.5 (1)

1 (1)

PONS ⫽ postoperative neurological symptoms. * Different between groups, P ⬍ 0.05.

compared with the nerve stimulator, but time to perform blocks was similar (5 min for both groups). Motor block at the biceps was enhanced (P ⫽ 0.04) at the 5-min assessment for ultrasound (Fig. 2). No patient required conversion to general anesthesia for failed block. The number of patients who were satisfied with anesthesia was similar between groups (96% nerve stimulator and 92% ultrasound, Table 2). The incidence of postoperative neurological symptoms was similar at 1 wk follow-up with 11% (95% CI of 6%–18%) for nerve stimulator and 8% (95% CI of 4%–15%) for ultrasound and was similar at late follow-up with 7% (95% CI of 3%–13%) for nerve stimulator and 6% (95% CI of 3%–12%) for ultrasound. Tables 3 and 4 display individual characteristics of 268


patients with postoperative neurological symptoms. The relative risk for postoperative neurological symptoms at 1-wk follow-up was not statistically significant at 1.37 (nerve stimulator versus ultrasound) with 95% CI of 0.6 –3.1. The relative risk for postoperative neurological symptoms at late follow-up was not statistically significant at 1.2 (nerve stimulator versus ultrasound) with 95% CI of 0.4 –3.1. The severity of postoperative neurological symptoms was similar in both groups with a median self-report of moderate (Tables 2– 4). All patients declined to return to the hospital for formal diagnostic evaluation by our neurologist coinvestigator.

DISCUSSION Our primary finding was that the use of ultrasound guidance did not significantly reduce the incidence or severity of postoperative neurological symptoms after interscalene block for outpatient shoulder arthroscopy when compared with a nerve stimulator technique. Real-time visualization with ultrasound has been proposed to improve the safety of peripheral nerve blocks due to the ability to avoid intraneural needle placement,5,17 whereas other techniques may often result in unintentional intraneural placement.18 Despite this theoretical advantage, ultrasound was not associated with a significant reduction in postoperative neurological symptoms within the framework of our study. A potential reason is that we used a fixed twodimensional cross-sectional image plane on the ultrasound, thus a similar rate of neural contact may have occurred due to the inability to fully visualize all three planes in real time. In addition, common clinical steps, such as monitoring for difficult injection or complaints of pain upon injection,19 were included for both techniques and may have narrowed a potential difference between groups. Our study examined the efficacy of anesthesia as secondary outcomes. Use of ultrasound guidance reduced the number of needle passes and provided more complete motor block at the 5 min assessment. This is in agreement with most previous RCTs comparing the two techniques. However, block success was not improved with the use of ultrasound, as no patient required conversion to general anesthesia due to failed block. This is likely explained by the already high-success rate of interscalene block with a nerve stimulator (97%–99%) in experienced hands,10 –12 thus there may be little room for improvement with the use of ultrasound. One other RCT recently compared nerve stimulator versus ultrasound specifically for interscalene block.20 This study reported greater block success for surgical anesthesia with ultrasound (99% vs 91%). However, clinical applicability of the study technique may be limited, as multiple needle passes under real-time ultrasound visualization were performed to ensure complete spread of local anesthetic around the nerve roots of C5-T1. Such a technique may require more skill than most clinicians possess, ANESTHESIA & ANALGESIA

Figure 2. Evidence of sensory or

motor block at 5 min. NS ⫽ nerve stimulator group. US ⫽ ultrasound group. * ⫽ different between groups, P ⫽ 0.04.

Table 3. Nerve Stimulator Patients with Postoperative Neurological Symptoms (PONS) Patient no. 1

Location of twitch

Location of accentuation with injection

Location of PONS

Description of PONS

Deltoid

Shoulder

Arm and hand

Tingling

2 11

Fingers-radial Deltoid

None None

Tips of fingers Wrist and fingers

30

Trapezius, phrenic, triceps Trapezius, deltoid Deltoid

Neck

Elbow and shoulder

Parasthesias Pain and weakness Pain

Neck

Hand, upper arm Biceps and shoulder

55 97

107

Biceps, deltoid

Neck, shoulder, and upper arm None

150

Triceps

None

173

Triceps

None

179

Trapezius, biceps Deltoid Trapezius, triceps

Biceps

193 210

None Neck

Mid-arm to shoulder and neck Fingers and hand, biceps Triceps, deltoid, scapula Ring, pinky fingers Fingertips Fingertips, neck, shoulder

PONS severity (early/late)

Duration of PONS

Rotator cuff Repair Diagnostic Decompression

Severe/–

3 days

Moderate/moderate Mild/mild

Continues (17 mo) Continues (17 mo)

Decompression

Moderate/moderate

3 mo

Parasthesia and pain Tightness

Decompression

Moderate/mild

4 mo

Stabilization

Moderate/mild

Continues (11 mo)

Pain

Rotator cuff

Severe/–

1 wk

Tingling fingers, pain in biceps Pain

Rotator cuff

Mild/–

2 wk

Decompression

Moderate/moderate

Continues (7 mo)

Parasthesia

Rotator cuff

Mild/–

3 days

Tingling Tingling fingers; pain elsewhere

Labral repair Debridement

Mild/mild Severe/severe

Continues (6 mo) Continues (5 mo)

which was noted in an accompanying editorial.21 In contrast, our methodology probably more closely resembles common clinical practice. Similar to previous RCTs,1 we noted no difference inpatient satisfaction between the two techniques despite the reduced needle passes with ultrasound guidance. This somewhat surprising finding may be related to the lack of validated tools to assess perioperative satisfaction,22 use of amnestics before block performance or the Vol. 109, No. 1, July 2009

Procedure type

relatively high-skill levels of operators for both techniques in the published RCTs. There are several limitations to our study. The incidence of postoperative neurological symptoms in our control group was within the range of those previously reported by Borgeat et al.11,12 (8%–14% at 10 days postoperative), which adds external validity to our findings. However, there is no standard definition for postoperative neurological symptoms and © 2009 International Anesthesia Research Society

269

Table 4. Ultrasound Patients with Postoperative Neurological Symptoms (PONS) Location of Patient accentuation no. with injection 5 17

None

66

Neck and shoulder None

82

None

83

None

Location of PONS

Description of PONS

Up and down arm Arm and hand

Pain and weakness

Labral repair

Pins and needles

Rotator cuff repair Mild/mild

Continues (17 mo) 3 mo

Scapula, biceps

Scapula pain, weak Bicep Pain

Rotator cuff repair Moderate/mild

3 mo

Labral repair

Tingling several minutes at a time Pain

Rotator cuff repair Mild/mild

Continues (12 mo) 12 mo

Labral repair

Moderate/–

1 wk

Pain

Labral repair

Mild/mild

3 mo

Biceps

155

Armpit, upper bicep None Shoulder to elbow Between elbow Wrist to elbow and shoulder None Fingers, arm

168

None

118 135

Procedure type

Tingling and pain Labral repair postoperative diagnosis of CRPS Wrist, fingertips, Tingling Labral repair thumb

PONS severity (early/late) Severe/severe

Moderate/mild

Duration of PONS

Moderate/moderate Continues (7 mo) Mild/–

3 days

CRPS ⫽ Complex regional pain syndrome.

incidences vary depending on measurement tool and time of assessment (incidences decrease with time). Our incidences of postoperative neurological symptoms were similar, but not identical, for the nerve stimulator and ultrasound. We note that our 95% confidence intervals do not suggest a significant clinical difference between groups and that a power analysis based on our results indicates that a follow-up RCT would require approximately 3000 subjects.* Finally, we used postoperative neurological symptoms as a surrogate measure for neurological injury after interscalene block for shoulder arthroscopy. We considered postoperative neurological symptoms to be a reasonable primary end point, as symptoms are clinically neurological in nature, relatively frequent, distressing to the patient, and have been well established as an outcome for clinical studies.9 –11 However, the exact pathophysiology of postoperative neurological symptoms is unclear. We cannot exclude a nonregional anesthesia etiology for postoperative neurological symptoms, as we did not include a general anesthesia only group in our study. Other proposed etiologies for postoperative nerve injury after shoulder arthroscopy include patient position,23 compression due to fluid extravasation,8 amount of traction,7 selection of arthroscopy ports,14 or toxic effects of local anesthetics.24 In conclusion, we observed that ultrasound guidance for interscalene block does not appear to offer major advantages over nerve stimulator guidance. The use of ultrasound reduced the number of needle *The power analysis was conducted with Power Analysis and Sample Size (PASS) software (NCSS, Kayesville, UT) with a target alpha 0.05. With 80% power, 1499 subjects per group are needed to detect a 3% difference. The sample size calculation is based on the normal theory tests.19

270


passes needed to perform the block; however, we did not observe significant differences in block failures, patient satisfaction or incidence, and severity of postoperative neurological symptoms. REFERENCES 1. Liu SS, Ngeow J, Yadeau JT. Ultrasound guided regional anesthesia and analgesia: a qualitative systematic review. Reg Anesth Pain Med 2009;34:47–59 2. Marhofer P, Chan VW. Ultrasound-guided regional anesthesia: current concepts and future trends. Anesth Analg 2007;104: 1265–9 3. Tsai TP, Vuckovic I, Dilberovic F, Obhodzas M, Kapur E, Divanovic KA, Hadzic A. Intensity of the stimulating current may not be a reliable indicator of intraneural needle placement. Reg Anesth Pain Med 2008;33:207–10 4. Sinha SK, Abrams JH, Weller RS. Ultrasound-guided interscalene needle placement produces successful anesthesia regardless of motor stimulation above or below 0.5 mA. Anesth Analg 2007;105:848 –52 5. Chan VW, Brull R, McCartney CJ, Xu D, Abbas S, Shannon P. An ultrasonographic and histological study of intraneural injection and electrical stimulation in pigs. Anesth Analg 2007;104: 1281– 4 6. Brull R, McCartney CJ, Chan VW, El-Beheiry H. Neurological complications after regional anesthesia: contemporary estimates of risk. Anesth Analg 2007;104:965–74 7. Rodeo SA, Forster RA, Weiland AJ. Neurological complications due to arthroscopy. J Bone Joint Surg Am 1993;75:917–26 8. Weber SC, Abrams JS, Nottage WM. Complications associated with arthroscopic shoulder surgery. Arthroscopy 2002;18:88 –95 9. Liguori GA, Zayas VM, YaDeau JT, Kahn RL, Paroli L, Buschiazzo V, Wu A. Nerve localization techniques for interscalene brachial plexus blockade: a prospective, randomized comparison of mechanical paresthesia versus electrical stimulation. Anesth Analg 2006;103:761–7 10. Candido KD, Sukhani R, Doty R Jr, Nader A, Kendall MC, Yaghmour E, Kataria TC, McCarthy R. Neurologic sequelae after interscalene brachial plexus block for shoulder/upper arm surgery: the association of patient, anesthetic, and surgical factors to the incidence and clinical course. Anesth Analg 2005;100:1489 –95 11. Borgeat A, Ekatodramis G, Kalberer F, Benz C. Acute and nonacute complications associated with interscalene block and shoulder surgery: a prospective study. Anesthesiology 2001;95: 875– 80


12. Borgeat A, Dullenkopf A, Ekatodramis G, Nagy L. Evaluation of the lateral modified approach for continuous interscalene block after shoulder surgery. Anesthesiology 2003;99:436 – 42 13. Brown AR, Ragukonis TP. Early sign of successful bupivacaine interscalene block: the “money sign.” Reg Anesth 1996;21:166 –7 14. Meyer M, Graveleau N, Hardy P, Landreau P. Anatomic risks of shoulder arthroscopy portals: anatomic cadaveric study of 12 portals. Arthroscopy 2007;23:529 –36 15. Urban MK, Urquhart B. Evaluation of brachial plexus anesthesia for upper extremity surgery. Reg Anesth 1994;19:175– 82 16. Fleiss JL, Levin B, Paik MC. Statistical methods for rates and proportions. 2nd ed. Somerset: Wiley, 2003 17. Schafhalter-Zoppoth I, Zeitz ID, Gray AT. Inadvertent femoral nerve impalement and intraneural injection visualized by ultrasound. Anesth Analg 2004;99:627– 8 18. Bigeleisen PE. Nerve puncture and apparent intraneural injection during ultrasound-guided axillary block does not invariably result in neurologic injury. Anesthesiology 2006;105: 779 – 83

Vol. 109, No. 1, July 2009

19. Hadzic A, Dilberovic F, Shah S, Kulenovic A, Kapur E, Zaciragic A, Cosovic E, Vuckovic I, Divanovic KA, Mornjakovic Z, Thys DM, Santos AC. Combination of intraneural injection and high injection pressure leads to fascicular injury and neurologic deficits in dogs. Reg Anesth Pain Med 2004;29:417–23 20. Kapral S, Greher M, Huber G, Willschke H, Kettner S, Kdolsky R, Marhofer P. Ultrasonographic guidance improves the success rate of interscalene brachial plexus blockade. Reg Anesth Pain Med 2008;33:253– 8 21. Salinas FV, Neal JM. A tale of two needle passes. Reg Anesth Pain Med 2008;33:195– 8 22. Liu SS, Wu CL. The effect of analgesic technique on postoperative patient-reported outcomes including analgesia: a systematic review. Anesth Analg 2007;105:789 – 808 23. Coppieters MW, Van de Velde M, Stappaerts KH. Positioning in anesthesiology: toward a better understanding of stretchinduced perioperative neuropathies. Anesthesiology 2002;97: 75– 81 24. Hogan QH. Pathophysiology of peripheral nerve injury during regional anesthesia. Reg Anesth Pain Med 2008;33:435– 41


271

Regional Blockade in Patients with a History of a Seizure Disorder Sandra L. Kopp, MD Kimberly P. Wynd, MBBCh Terese T. Horlocker, MD James R. Hebl, MD Jack L. Wilson, MD

BACKGROUND: Systemic local anesthetic toxicity is a potential complication in patients undergoing regional anesthesia, particularly during procedures requiring large doses of local anesthetic, such as epidurals, caudals, and peripheral nerve blocks. It is unknown whether patients with a history of a seizure disorder are at an increased risk of central nervous system toxicity (seizures) after local anesthetic administration. METHODS: We retrospectively reviewed the medical records of all patients with documented history of a seizure disorder who underwent epidural, caudal, or peripheral nerve block from January 1, 1988 to December 31, 2001. Patient demographics, character of the seizure disorder, details of the regional procedure, and seizure activity in the perioperative period were recorded. The rate of seizure due to local anesthetic toxicity per 10,000 anesthetics was estimated using a point estimate and corresponding 95% confidence interval (CI). RESULTS: During the 14-yr study period, 411 procedures in 335 patients with a seizure disorder were identified. Twenty-four patients experienced postoperative seizure activity. The timing of the most recent (preoperative) seizure was found to be significantly related to the likelihood of experiencing a postoperative seizure (P ⬍ 0.001). Based on the extended time interval between local anesthetic injection and/or termination of the infusion and the event, it was determined that the regional anesthetic was neither the primary etiology nor a contributing factor for the seizure in 19 of the 24 patients. In the remaining five patients, perioperative seizure activity was characteristic of their usual seizures. Although unlikely to be the cause of the seizure, local anesthetic toxicity could not be absolutely excluded as a contributing factor to the event in these five patients. Assuming that none of the seizures was related to local anesthetic toxicity the estimated incidence is 0 per 10,000 (95% CI 0 – 89 per 10,000). Conversely, if the seizures were related to local anesthetic toxicity in the five cases, the incidence is increased to 120 per 10,000 (95% CI 40 –280 per 10,000). CONCLUSIONS: We conclude that majority of seizures occurring in the perioperative period in patients with a preexisting seizure disorder are likely related to the patient’s underlying condition and that regional anesthesia in these patients is not contraindicated. Furthermore, because the likelihood of a postoperative seizure is increased in patients with a recent seizure, it is essential to be prepared to treat seizure activity, regardless of the anesthetic and analgesic technique. (Anesth Analg 2009;109:272–8)

L

ocal anesthetic toxicity is a potential risk for all patients undergoing regional anesthesia, particularly during procedures that require a large dose of local anesthetic, such as epidural, caudal, or peripheral nerve block. Systemic local anesthetic toxicity

From the Department of Anesthesiology, Mayo Clinic College of Medicine, Rochester, Minnesota. Accepted for publication January 21, 2009. Terese T. Horlocker is the Section Editor of Regional Anesthesia. This manuscript was handled by Steven L. Shafer, Editor-in-Chief, and Dr. Horlocker was not involved in any way with the editorial process or decision. Address correspondence and reprint requests to Sandra L. Kopp, MD, Department of Anesthesiology, Mayo Clinic College of Medicine, 200 First St. SW, Rochester, MN 55905. Address e-mail to [email protected]. Copyright © 2009 International Anesthesia Research Society DOI: 10.1213/ane.0b013e3181a832da

272

presents as a spectrum of neurologic signs and symptoms that worsen as plasma-drug levels continue to rise. Symptoms of central nervous system (CNS) toxicity generally follow a progression from lightheadedness, dizziness, and perioral numbness to visual or auditory disturbances and tinnitus. These are usually followed by peripheral muscle twitching and ultimately generalized tonic-clonic convulsions. Previous studies have determined the time-to-peak plasma levels after a single bolus of local anesthetic to range from 15 to 120 min depending on anatomic location.1,2 The limited data available regarding continuous infusions of local anesthetics indicate that plasma levels continue to rise throughout the duration of the infusion.3,4 The published incidence of CNS local anesthetic toxicity (seizure) in the general population undergoing regional anesthesia ranges from 1.2 Vol. 109, No. 1, July 2009

to 11 per 10,000 epidural anesthetics,5–9 1.3– 69 per 10,000 caudal anesthetics,5,10 and 0 –25.4 per 10,000 peripheral nerve blocks.5–7,9 However, these series did not report the number of patients with a preexisting seizure disorder. Although literature to support withholding a regional anesthetic from patients with a history of a seizure disorder is lacking, modifications to standard practice are recommended, such as selection of less potent local anesthetics, minimizing local anesthetic dose, and opting for a postoperative epidural infusion of opioid rather than local anesthetic.11 Historically, local anesthetics have been used to treat status epilepticus because of their concentrationdependent effect on seizures. At low blood levels, local anesthetics decrease cerebral blood flow, metabolism, and electrical activity, and are potent anticonvulsants. Conversely, at higher levels, they act as proconvulsants by lowering the seizure threshold within the cerebral cortex, amygdala, and hippocampus, usually leading to a generalized convulsion.12 As the serum drug level continues to rise, both the inhibitory and excitatory neural pathways are blocked, resulting in generalized CNS depression. Currently, it is unknown whether patients with a history of a seizure disorder are at an increased risk of CNS toxicity after the administration of large doses of local anesthetics. This retrospective study evaluated the frequency of seizure activity in patients with a history of a seizure disorder who subsequently underwent regional blockade.

METHODS After IRB approval, all patients with a documented history of a seizure disorder (seizure disorder listed as a medical diagnosis in the medical record) who subsequently underwent an epidural, caudal, or peripheral nerve block from January 1, 1988 through December 31, 2001 were identified. Patient demographics including age, gender, ASA physical status classification, and the urgency (elective or emergent) of the surgery were recorded. We recorded the characteristics and clinical course of the seizure disorder, including type of seizure disorder (absence, simple partial, complex partial, or convulsive), seizure frequency, most recent seizure, antiepileptic drug use, medication administration that may interact with antiepileptic drugs (e.g., antibiotics, anesthetics, antihistamines), and blood levels of antiepileptic drugs within 2 wk of the procedure. Details of the regional anesthetic technique were also documented, including procedure, type and dose of local anesthetic bolus, subsequent local anesthetic administration or infusions, adjuvants used, time of last sedative administration, and the presence of any CNS activity (tinnitus, perioral numbness, myoclonic jerks, tonic/clonic seizure, or partial seizure) during the block placement. For all patients experiencing seizure activity during their hospitalization, the circumstances surrounding the event were documented, Vol. 109, No. 1, July 2009

Table 1. Patient Characteristics Characteristics

%

n

Overall Procedures Patients Gender Male Female ASA statusa I II III IV Emergency Yes No Type of regional Epidural Caudal Axillary Other upper extremityb Lower extremityc Purpose of block Surgical Obstetric Postoperative pain control Predominate type of seizure disorder Absence Complex partial Convulsive Simple partial Other Seizure during hospitalization Yes No Unknown

411 335 191 220

46 54

30 202 169 10

7 49 41 2

18 393

4 96

204 45 89 23 50

50 11 22 5 12

171 77 163

41 19 40

18 56 292 19 26

4 14 71 5 6

24 384 3

6 93 1

a

American Society of Anesthesiologists physical status. Interscalene, infraclavicular, and supraclavicular. c Ankle, femoral, psoas, and fascia iliaca. b

Table 2. Timing of Most Recent Seizure n

Seizure

Pa

30 41 43 21 271

8 (27%) 4 (10%) 3 (7%) 0 (0%) 8 (3%)

⬍0.001

b

Time since last seizure Within 1 wk 1 wk to 1 mo 1 wk to 6 mo 6 mo to 1 yr Greater than 1 yr a

Fisher’s exact test. Data were missing for five patients overall, including one patient who experienced a perioperative seizure and four patients who did not experience a perioperative seizure. b

including time of seizure, type of seizure, recent local anesthetic administration, antiepileptic drugs, medications administered at the time of seizure, and recent antiepileptic drug blood levels. Information pertaining to the seizure activity was derived from anesthetic records, postanesthesia care unit notes, and daily progress notes of the primary service, medical consultation team(s), and the anesthesia pain service. The cause of the seizure was determined using published data on the timing of peak blood levels, frequency of preoperative seizures, antiepileptic drug levels, results © 2009 International Anesthesia Research Society

273

Table 3. Patients Experiencing Seizures Unlikely Related to Local Anesthetic Toxicity

Gender (age)

ASA statusa

Seizure disorder

Continuous infusion of local anesthetic Male (17 mo) III Intractable mixed seizure disorder

Male (66 yr)

Female (74 yr)

Male (7 yr)

III

III

II

Simple partial

Generalized convulsive

Generalized convulsive and complex partial

Single injection of local anesthetic Female (30 yr) III Generalized convulsive or pseudoseizures

Preoperative antiseizure medications

Date of last seizure ⬍1 wk

⬎1 mo ⬍6 mo

⬎1 wk ⬍1 mo

Felbamate, phenobarbital, and levetiracetam

Phenytoin

Phenytoin

Type of regional block Caudal

Local anesthetic administration time course 08:42: Bupivacaine 0.25% (9 mL) 09:50: Bupivacaine 0.25% (7 mL)

Epidural

11:10: Infusion–bupivacaine 0.1% at 3 mL/h POD 1: Infusion stopped at 09:00 (23 h total) 09:00: Bupivacaine 0.25% (10 mL)

Epidural

09:30: Infusion–bupivacaine 0.075% at 12 mL/h POD 3: Infusion stopped (72.5 h total) 12:00: Bupivacaine 0.5% (10 mL)

⬎1 wk ⬍1 mo

Phenobarbital

Caudal

14:50: Bupivacaine 0.5% (10 mL) 20:00: Infusion–bupivacaine 0.125% ⫹ fentanyl 5 mcg/mL at 8 mL/h POD 1: Infusion stopped (10 h total) 08:50: Bupivacaine 0.25%, 40 mL given incrementally over a 7-h period 16:40: Infusion–bupivacaine 0.1% at 10 mL/h POD 3: Infusion stopped (65.5 h total)

⬎1 yr

None

Axillary

08:45: Mepivacaine 1.5% (50 mL)

09:40: Mepivacaine 1.5% (10 mL) ⫹ 2-chloroprocaine 3% (10 mL) 10:10: Bupivacaine 0.25% (10 mL)

Male (9 yr)

IV


⬍1 wk

Phenytoin and phenobarbital

Femoral

Male (79 yr)

IV

Complex partial

⬎1 yr

Carbamazepine

Inguinal block

Male (43 yr)

II


⬍1 wk

Phenytoin and phenobarbital

Epidural

09:35: Lidocaine 2% (20 mL) 12:55: Bupivacaine 0.5% (22 mL)

Female (52 yr)

III


⬎1 mo ⬍6 mo

Phenobarbital

Interscalene

07:45: Bupivacaine 0.5% with epinephrine (20 mL)

Male (72 yr)

IV


⬎1 yr

Phenytoin

Epidural

11:17: Lidocaine 1.5% (11 mL)

Male (4 yr)

III


⬍1 wk

Caudal

14:38: Bupivacaine 0.25% (15 mL)

Male (53 yr)

II

Generalized convulsive and complex partial

⬎1 yr

Valproic acid and felbamate Phenytoin

Supraclavicular

14:14: Bupivacaine 0.5% (30 mL) ⫹ clonidine 100 mcg

Female (3 yr)

II


⬎1 yr

Phenobarbital

Caudal

8:10: Bupivacaine 0.25% (11 mL) ⫹ clonidine 15 mcg

Comments PODb 1, 14:00: Seizure activity Seizure 6 h after infusion was discontinued

Seizure occurred 6 d after infusion was discontinued

1st seizure on POD 9, multiple seizures afterwards

Seizure activity noted 17 d after infusion was discontinued

09:35: Post anesthesia care unit–jerking of head, stuttering, oriented Long history of psychiatric disorders Anesthesiologist documented it was not a seizure or preseizure, rather a pseudoseizure No medications administered 16:15: Seizure activity Multiple seizures during hospitalization 18:00: Seizure activity Neurology consult: seizure likely due to hypoglycemia and missed dose of carbamazepine 20:45: Seizure activity Multiple seizures while hospitalized Facial twitching noted approximately 12 h after block placement POD 1, 05:00: Seizure activity Subtherapeutic phenytoin level at time of seizure Seizure activity 20 h postoperatively 1st seizure 23.5 h after block Several seizures in the postoperative period Neurology consult: seizures likely due to decreased dose of phenytoin Seizure 24 h after bolus dose (Continued)

274

Epilepsy and Local Anesthetic Administration


Table 3. Continued Preoperative antiseizure medications

ASA statusa

Seizure disorder

Date of last seizure

Male (76 yr)

III

⬎1 mo ⬍6 mo

Carbamazepine and phenobarbital

Lumbar plexus (psoas)


Seizure activity 24 h later in the setting of fever

Female (78 yr)

III

Unknown

None

Axillary


POD 1, 13:01: Seizure activity

Female (66 yr)

III

Generalized convulsive and complex partial Generalized convulsive and simple partial Generalized convulsive and generalized absence Generalized absence Simple partial Generalized convulsive

⬍1 wk

Carbamazepine, phenytoin, and phenobarbital

Epidural


POD 1, 17:00: Seizure activity


Multiple absence seizures during hospitalization

⬍1 wk

Valproic acid

Caudal

⬎1 yr ⬎1 wk ⬍1 mo

None None

Ankle Caudal

11:25: Bupivacaine 0.125% (15 mL) 10:45: Lidocaine 1.5% (17 mL) 12:50: Bupivacaine 0.125% (3 mL)

Seizure activity noted on POD 2 Seizure activity on POD 4 Seizure occurred 19 d postoperatively in the setting of multiple medical problems

Gender (age)

Male (2 yr)

III

Male (78 yr) Female (19 d)

IV II

a b

Type of regional block

Local anesthetic administration time course

Comments

American Society of Anesthesiologists physical status. POD ⫽ postoperative day (defined as the first full day after surgery).

from neurology consultation, and documentation detailing the seizure activity. Patient and procedure characteristics were summarized for the entire cohort using mean ⫾ SD for continuous variables and percentages for categorical variables. The rate of seizure due to local anesthetic toxicity per 10,000 anesthetics was estimated using a point estimate and corresponding 95% confidence interval (CI).

RESULTS During the 14-yr study period, 411 caudal, epidural, or peripheral nerve blocks were performed in 335 patients with a documented history of a seizure disorder. Twenty-four patients had at least one episode of CNS (seizure) activity in the perioperative period. Sixteen patients had a single injection of local anesthetic and eight patients had a continuous infusion of local anesthetic. No patient experienced a seizure during or immediately after (⬍50 min) administration of the local anesthetic initial loading dose. The mean patient age was 42.8 ⫾ 24.5 yr (range, 1 mo–94 yr). Patient characteristics including gender, ASA physical status, type of regional block, reason for surgery, seizure type, and seizure rates during hospitalization are provided in Table 1. Patients with a recent preoperative seizure were significantly more likely to experience a seizure during the perioperative period (P ⬍ 0.001) (Table 2). Of the 24 patients who had a perioperative seizure, the regional anesthetic was determined to be coincidental and not causal in 19 patients due to the extended time interval between injection of local anesthetic and the seizure. Of these 19 patients, 15 received only a bolus dose of local anesthetic and four received a bolus of local anesthetic followed by a continuous infusion (Table 3). One patient experiVol. 109, No. 1, July 2009

enced pseudoseizure activity 50 min after the local anesthetic bolus, four patients within 6 to 12 h, five patients within 12 to 24 h, and five patients more than 24 h after the local anesthetic bolus. In all four patients with a continuous infusion of local anesthetic, the seizure did not occur until more than 5 h after the infusion was discontinued. In five of the 24 patients, local anesthetic toxicity could not be definitively excluded as a cause or contributing factor of the seizure activity (Table 4). Two patients experienced frequent (almost daily) seizures before hospital admission. Although the seizures noted postoperatively were similar to their “normal seizure activity” they occurred during the local anesthetic infusion and therefore CNS toxicity cannot be excluded. One patient who preoperatively experienced at least monthly seizures had bradycardia and facial twitching 90 min after a bolus of local anesthetic for an axillary block. Although the neurology consult documented this was likely not a seizure, preseizure activity facilitated by local anesthetic absorption cannot be absolutely excluded as a contributory factor. Two patients receiving antiepileptic medication who had not experienced a seizure for at least 1 yr preoperatively had an episode of seizure activity during an epidural infusion of local anesthetic. One patient with a subtherapeutic antiepileptic drug blood level had an epidural placed for pain control after a pneumonectomy and experienced a seizure 15 min after the infusion (bupivacaine 0.075%, 8 mL/h) had been discontinued. The second patient taking phenobarbital (unknown blood level) had been seizure-free for more than 6 yr and had an epidural placed for labor and delivery. Fifty minutes after the local anesthetic initial loading dose (bupivacaine 0.125%, 15 mL) and 40 min into the infusion (bupivacaine 0.125%, 10 © 2009 International Anesthesia Research Society

275

Table 4. Patients in Whom Local Anesthetics Potentially Contributed to Seizure Activity

Gender (age) Male (61 yr)

ASA statusa III

Seizure disorder Complex partial

Preoperative antiseizure medication

Date of last seizure Greater than 1 yr prior

Phenytoin

Type of regional block Epidural

Local anesthetic administration time course 11:53: Lidocaine 1% with 1:200,000 epinephrine (5 mL) 13:58: Bolus bupivacaine 0.25% (19 mL)

16:08: Infusion–bupivacaine 0.075% at 8 mL/h PODb 3: Infusion discontinued (65 h total)

Female (29 yr)

II


Greater than 1 yr prior

Phenobarbital

Epidural

15:50: Bolus bupivacaine 0.125% (15 mL) 16:00: Infusion–bupivacaine 0.125% at 10 mL/h 16:40: Infusion stopped 17:30: Restarted bupivacaine 0.125% at 10 mL/h

19:20: Bolus bupivacaine 0.125% (5 mL)

Female (8 mo)

III


⬍1 wk

Phenobarbital and clonazepam

Caudal

19:45: Epidural infusion discontinued following delivery 15:39: Bupivacaine 0.125% (5 mL) 19:00: Infusion–bupivacaine 0.1% ⫹ hydromorphone 2 mcg/mL at 3.5 mL/h POD 2: Infusion stopped (52 h total)

Female (6 yr)

III

Complex partial

⬍1 wk

Phenobarbital

Epidural

11:07: Bupivacaine 0.25% (10 mL) 13:15: Bupivacaine 0.25% (6 mL) 15:30: Infusion–bupivacaine 0.1% at 4 mL/h POD 1: Infusion stopped (20 h total)

Female (83 yr)

a b

III


⬎1 wk ⬍1 mo

Carbamazepine and phenobarbital

Axillary


Comments Epidural for postoperative pain control following pneumonectomy Seizure activity 15 min after infusion discontinued on POD 3 Resolved spontaneously within minutes Phenytoin level subtherapeutic No further seizures during hospitalization Epidural for labor Seizure-free for more than 6 yr 16:40: Patient states “I felt like I had a seizure” No witnessed tonic clonic activity, no loss of consciousness, no postictal symptoms Denied symptoms of local anesthetic toxicity No further seizures during hospitalization PODb 1, 12:45: Seizure activity during infusion Multiple postoperative seizures felt to be similar to preoperative seizures Seizure activity during infusion, therefore could not rule out that local anesthetic contributed POD 1, 07:10: Seizure activity during infusion Identical to typical seizure activity Usual seizures occur at least once per day Had daily seizures while hospitalized Seizure activity during infusion, therefore could not rule out that local anesthetic contributed 12:15 Patient became bradycardic and was given atropine, comment in record about facial twitching Neurology consult: unlikely this was a seizure Preseizure activity could not be ruled out given time of local anesthetic bolus

American Society of Anesthesiologists physical status. Postoperative day (defined as the first full day after surgery).

mL/h), the patient felt that she had a seizure. However, there was no witnessed tonic-clonic activity, loss of consciousness, or postictal symptoms. The epidural infusion was stopped for 50 min and subsequently restarted with no further episodes of seizure activity. Assuming that local anesthetic toxicity did not play a role in any of the seizures (0 of 24 events), the seizure incidence during the perioperative period would be 0 per 10,000 (95% CI 0 – 89 per 10,000) anesthetics. 276


Although unlikely, if local anesthetic toxicity contributed to the five seizures described above, the incidence would be 120 per 10,000 (95% CI 40 –280 per 10,000) anesthetics.

DISCUSSION Seizure is defined as the clinical manifestation of abnormally hyperexcitable cortical neurons. Although ANESTHESIA & ANALGESIA

Table 5. Time to Peak Plasma Level After a Single Injection Author year Misra et al. 199119 Robison et al. 19912 Farny et al. 199420 Atanassoff et al. 19961 Downing et al. 199721 Vanterpool et al. 199622

Total patients

Block type

22 22 45 20 12 17 10 10 10 10

Femoral 3-in-1 ⫹ sciatic Femoral 3-in-1 ⫹ sciatic Posterior lumbar plexus ⫹ sciatic Direct obturator Bilateral obturator 3-in-1 obturator Epidural Lumbar plexus Lumbar plexus ⫹ sciatic

Local anesthetic dose

Time-to-peak level (min)

Bupivacaine 3 mg/kg Bupivacaine 168.75 mg Lidocaine 680 mg Lidocaine 225 mg Lidocaine 450 mg Lidocaine 600 mg Lidocaine 6.1 mg/kg Bupivacaine 1.5 mg/kg Ropivacaine 175 mg Ropivacaine 300 mg

60 ⫾ 7 60 ⫺ 120 61.7 ⫾ 66.2 45 15 60 31 41 80 ⫾ 49 38 ⫾ 22

Table 6. Time to Peak Plasma Levels During Continuous Infusion

Author year

Total patients

Block type

Denson et al. 198423 Tuominen et al. 198724

12 40

Epidural Interscalene

Dahl et al. 198825

10

Femoral 3-in-1

Tuominen et al. 198926

24

Interscalene

Tuominen et al. 19914

20

Interscalene

Ekatodramis et al. 200327

12 12 11 11

Interscalene Interscalene Femoral 3-in-1 Psoas

Kaloul et al. 20043

all patients with epilepsy have seizures, many patients have a single seizure during their lives and are not considered to have epilepsy. A population-based epidemiologic study from Rochester, MN, found that the cumulative incidence of epilepsy through 74 yr of age is 3.0%, with an incidence of any convulsive disorder approaching 10%.13 Therefore, it is not uncommon for a patient with the diagnosis of a seizure disorder to present for a surgical procedure in which regional anesthesia or analgesia may be considered. Furthermore, there are no data regarding the overall incidence of perioperative seizures in patients with a history of a seizure disorder and the effect of anesthetic and analgesic management on seizure risk. Our investigation identified 24 patients who experienced perioperative seizure activity. It is difficult to determine whether a perioperative seizure is due to the seizure disorder itself, or any of the numerous factors that influence the frequency of seizures in this population. Common factors that may provoke seizure activity include fluctuations in antiepileptic drug blood levels, sleep deprivation, fatigue, stress, excessive alcohol intake, and menstruation.14,15 The reduction in antiepileptic drug serum levels is a well known cause of seizure provocation during the perioperative period.16 Factors that have been shown to contribute to fluctuations in Vol. 109, No. 1, July 2009

Local anesthetic infusion rate Bupivacaine 0.125 mg/h Bupivacaine 0.25 mg · kg⫺1 · h⫺1 Bupivacaine 0.35 mg · kg⫺1 · h⫺1 Bupivacaine 0.25 mg · kg⫺1 · h⫺1 Bupivacaine 0.25 mg · kg⫺1 · h⫺1 Bupivacaine 12 mg/h Bupivacaine 18 mg/h Ropivacaine 24 mg/h Ropivacaine 24 mg/h

Infusion length (h)

Time-to-peak level (h)

3.25 24

0.58 24

16

1

24

24

48 48 48 48 48

Rising from 24 to 48 54 (30–56) 54 (12–54) 48 48

antiepileptic drug levels during the perioperative period include preoperative antiepileptic drug noncompliance, changes in the antiepileptic dosing schedule, anesthetic drugs, perioperative medications, and alteration of gastric motility leading to delayed absorption and reduced bioavailability.15,17 Conversely, results from animal studies suggest that specific anticonvulsants may provide a degree of protection against lidocaine-induced seizures. For example, benzodiazepines (diazepam and clonazepam) seem to confer the greatest protection against local anesthetic-induced neurotoxicity within animal models. Phenobarbital and carbamazepine are also protective; however, phenytoin and primidone may enhance local anesthetic convulsions.18 Four patients experienced a postoperative seizure during or immediately after an epidural/caudal infusion of local anesthetic. One patient experienced bradycardia and facial twitching 90 min after an axillary block. Although the cause of the seizures was likely multifactorial, local anesthetic toxicity could not be absolutely excluded in these patients. In the remaining 19 patients who experienced a postoperative seizure, local anesthetic toxicity was determined to be unlikely based upon published data estimating time to peak blood levels after local anesthetic administration or a combination of variables, such as altered administration of © 2009 International Anesthesia Research Society

277

antiepileptic drugs and overall preoperative seizure frequency. Previous studies have determined that the time to peak plasma levels after a bolus of local anesthetic varies from 15 to 120 min (Table 5). A limited number of studies evaluating infusions through continuous catheters indicate that local anesthetic plasma levels continue to rise throughout the duration of the infusion (Table 6). The peak blood level measured in patients with a continuous psoas or femoral 3-in-1 block occurred at 48 h, which also corresponded to the time that the infusion was discontinued.3 As the use of prolonged perineural infusions increases, further studies are needed to determine the risk of local anesthetic toxicity in this patient population. Patients who had frequent seizures near the time of admission were significantly more likely to experience a seizure during their hospital course when compared to patients with less frequent seizures. A previous study at our institution, during the same time period, reported the incidence of CNS toxicity among all patients undergoing regional anesthesia (epidurals, caudals, and peripheral nerve blocks) to be 10 per 10,000 regional anesthetics.5 It is difficult to interpret whether our results represent an increased risk, because determining the exact etiology of the perioperative seizure in this population is challenging. The difficulty arises due to incomplete documentation regarding the preoperative and perioperative seizure, as well as limited information on local anesthetic and antiepileptic blood levels. In summary, we conclude that regional anesthesia in patients with a history of a seizure disorder is not contraindicated. However, the relative risk of inducing seizure activity with a regional anesthetic technique in patients with a diagnosis of seizure disorder, compared with patients without such a diagnosis, is unknown. Furthermore, because the likelihood of a postoperative seizure is increased in patients with a recent seizure, it is essential to be prepared to treat seizure activity, regardless of the anesthetic or analgesic technique. REFERENCES 1. Atanassoff PG, Weiss BM, Brull SJ. Lidocaine plasma levels following two techniques of obturator nerve block. J Clin Anesth 1996;8:535–9 2. Robison C, Ray DC, McKeown DW, Buchan AS. Effect of adrenaline on plasma concentrations of bupivacaine following lower limb nerve block. Br J Anaesth 1991;66:228 –31 3. Kaloul I, Guay J, Cote C, Halwagi A, Varin F. Ropivacaine plasma concentrations are similar during continuous lumbar plexus blockade using the anterior three-in-one and the posterior psoas compartment techniques. Can J Anaesth 2004;51:52– 6 4. Tuominen MK, Pere P, Rosenberg PH. Unintentional arterial catheterization and bupivacaine toxicity associated with continuous interscalene brachial plexus block. Anesthesiology 1991;75:356 – 8 5. Brown DL, Ransom DM, Hall JA, Leicht CH, Schroeder DR, Offord KP. Regional anesthesia and local anesthetic-induced systemic toxicity: seizure frequency and accompanying cardiovascular changes. Anesth Analg 1995;81:321– 8

278


6. Borgeat A, Ekatodramis G, Kalberer F, Benz C. Acute and nonacute complications associated with interscalene block and shoulder surgery: a prospective study. Anesthesiology 2001;95:875– 80 7. Auroy Y, Ecoffey C, Messiah A, Rouvier B. Relationship between complications of pediatric anesthesia and volume of pediatric anesthetics. Anesth Analg 1997;84:234 –5 8. Tanaka K, Watanabe R, Harada T, Dan K. Extensive application of epidural anesthesia and analgesia in a university hospital: incidence of complications related to technique. Reg Anesth 1993;18:34 – 8 9. Auroy Y, Benhamou D, Bargues L, Ecoffey C, Falissard B, Mercier FJ, Bouaziz H, Samii K. Major complications of regional anesthesia in France: the SOS Regional Anesthesia Hotline Service. Anesthesiology 2002;97:1274 – 80 10. Giaufre E, Dalens B, Gombert A. Epidemiology and morbidity of regional anesthesia in children: a one-year prospective survey of the French-Language Society of Pediatric Anesthesiologists. Anesth Analg 1996;83:904 –12 11. Faccenda KA, Finucane BT. Complications of regional anaesthesia. Incidence and prevention. Drug Saf 2001;24:413– 42 12. DeToledo JC, Minagar A, Lowe MR. Lidocaine-induced seizures in patients with history of epilepsy: effect of antiepileptic drugs. Anesthesiology 2002;97:737–9 13. Hauser WA, Annegers JF, Rocca WA. Descriptive epidemiology of epilepsy: contributions of population-based studies from Rochester, Minnesota. Mayo Clin Proc 1996;71:576 – 86 14. Sokic D, Ristic AJ, Vojvodic N, Jankovic S, Sindjelic AR. Frequency, causes and phenomenology of late seizure recurrence in patients with juvenile myoclonic epilepsy after a long period of remission. Seizure 2007;16:533–7 15. Paul F, Veauthier C, Fritz G, Lehmann TN, Aktas O, Zipp F, Meencke HJ. Perioperative fluctuations of lamotrigine serum levels in patients undergoing epilepsy surgery. Seizure 2007;16:479 – 84 16. Specht U, Elsner H, May TW, Schimichowski B, Thorbecke R. Postictal serum levels of antiepileptic drugs for detection of noncompliance. Epilepsy Behav 2003;4:487–95 17. Tan JH, Wilder-Smith E, Lim EC, Ong BK. Frequency of provocative factors in epileptic patients admitted for seizures: a prospective study in Singapore. Seizure 2005;14:464 –9 18. Sawaki K, Ohno K, Miyamoto K, Hirai S, Yazaki K, Kawaguchi M. Effects of anticonvulsants on local anaesthetic-induced neurotoxicity in rats. Pharmacol Toxicol 2000;86:59 – 62 19. Misra U, Pridie AK, McClymont C, Bower S. Plasma concentrations of bupivacaine following combined sciatic and femoral 3 in 1 nerve blocks in open knee surgery. Br J Anaesth 1991;66:310 –3 20. Farny J, Girard M, Drolet P. Posterior approach to the lumbar plexus combined with a sciatic nerve block using lidocaine. Can J Anaesth 1994;41:486 –91 21. Downing JW, Johnson HV, Gonzalez HF, Arney TL, Herman NL, Johnson RF. The pharmacokinetics of epidural lidocaine and bupivacaine during cesarean section. Anesth Analg 1997;84:527–32 22. Vanterpool S, Steele SM, Nielsen KC, Tucker M, Klein SM. Combined lumbar-plexus and sciatic-nerve blocks: an analysis of plasma ropivacaine concentrations. Reg Anesth Pain Med 2006;31:417–21 23. Denson DD, Knapp RM, Turner P, Thompson GA. Serum bupivacaine concentrations in term parturients following continuous epidural analgesia for labor and delivery. Ther Drug Monit 1984;6:393– 8 24. Tuominen M, Pitkanen M, Rosenberg PH. Postoperative pain relief and bupivacaine plasma levels during continuous interscalene brachial plexus block. Acta Anaesthesiol Scand 1987;31:276 – 8 25. Dahl JB, Christiansen CL, Daugaard JJ, Schultz P, Carlsson P. Continuous blockade of the lumbar plexus after knee surgery—postoperative analgesia and bupivacaine plasma concentrations. A controlled clinical trial. Anaesthesia 1988;43:1015–18 26. Tuominen M, Haasio J, Hekali R, Rosenberg PH. Continuous interscalene brachial plexus block: clinical efficacy, technical problems and bupivacaine plasma concentrations. Acta Anaesthesiol Scand 1989;33:84 – 8 27. Ekatodramis G, Borgeat A, Huledal G, Jeppsson L, Westman L, Sjovall J. Continuous interscalene analgesia with ropivacaine 2 mg/ml after major shoulder surgery. Anesthesiology 2003;98:143–50


Brief Report

Ultrasound-Guided Axillary Brachial Plexus Block with 20 Milliliters Local Anesthetic Mixture Versus General Anesthesia for Upper Limb Trauma Surgery: An Observer-Blinded, Prospective, Randomized, Controlled Trial Brian D. O’Donnell, MB, FCARCSI, MSc Helen Ryan, MB, BCh, BAO Owen O’Sullivan, MB, FCARCSI Gabrielle Iohom, FCARCSI, MD, PhD

OBJECTIVE: We performed a randomized, controlled trial comparing low-dose ultrasound-guided axillary block with general anesthesia evaluating anesthetic and perioperative analgesic outcomes. METHODS: Patients were randomized to either ultrasound-guided axillary block or general anesthesia. Ultrasound-guided axillary block was performed using a needle-out-of-plane approach. Up to 5 mL of local anesthetic injectate (equal parts 2% lidocaine with 1:200,000 epinephrine and 0.5% bupivacaine with 7.5 mg/mL clonidine) was injected after identifying the median, ulnar, radial, and musculocutaneous nerves. A maximum of 20 mL local anesthetic injectate was used. General anesthesia was standardized to include induction with fentanyl and propofol, maintenance with sevoflurane in an oxygen/nitrous oxide mixture. Pain scores were measured in the recovery room and at 2, 6, 24, 48 h, and 7 days. Ability to bypass the recovery room and time to achieve hospital discharge criteria were also assessed. RESULTS: All ultrasound-guided axillary block patients achieved satisfactory anesthesia. The ultrasound-guided axillary block group had lower visual analog scale pain scores in the recovery room (0.3 [1.3] vs 55.8 [36.5], P ⬍ 0.001), and visual rating scale pain scores at 2 h (0.3 [1.3] vs 45 [29.6], P ⬍ 0.001), and at 6 h (1.1 [2.7] vs 4 [2.8], P ⬍ 0.01). All ultrasound-guided axillary block patients bypassed the 30 recovery room and attained earlier hospital discharge criteria (30 min vs 120 min 240 P ⬍ 0.0001 median [range]). CONCLUSIONS: Ultrasound-guided axillary brachial plexus block with 20 mL local anesthetic mixture provided satisfactory anesthesia and superior analgesia after upper limb trauma surgery when compared with general anesthesia. (Anesth Analg 2009;109:279 –83)

R

egional anesthesia has several reported advantages when compared with general anesthesia for patients undergoing upper limb trauma surgery, including improved perioperative analgesia,1 reduced opiate consumption,2 reduced postoperative nausea and vomiting (PONV),3 shorter postanesthesia care unit stay,3 and earlier hospital discharge.4,5 Nerve localization using ultrasound-guidance has been shown to improve success rates,6,7 decrease onset times,6 –9 and reduce local anesthetic dose.10 –12 Traditionally, local anesthetic volumes of up to and more than From the Department of Anesthesia, Cork University Hospital, Wilton, Cork, Ireland. Accepted for publication January 21, 2009. This study was financed entirely from internal departmental resources. Reprints will not be available from the author. Address correspondence to Brian O’Donnell, MB, FCARCSI, MSc, Department of Anesthesia, Cork University Hospital, Wilton Rd., Cork, Ireland. Address e-mail to [email protected]. Copyright © 2009 International Anesthesia Research Society DOI: 10.1213/ane.0b013e3181a3e721

Vol. 109, No. 1, July 2009

40 mL have been used to perform axillary brachial plexus block.4,13,14 It is not known whether lower dose ultrasound-guided axillary block confers similarly reported advantages when compared with general anesthesia. We hypothesized that low-dose ultrasound-guided axillary block using 20 mL of solution provides excellent anesthesia for and superior analgesia after upper limb trauma surgery when compared with general anesthesia. We conducted a prospective, randomized, observer-blinded study comparing ultrasound-guided axillary block to general anesthesia evaluating anesthetic and perioperative analgesic outcomes.

METHODS After ethical approval and written informed consent, patients scheduled for upper limb trauma surgery to the forearm or hand of ⬍90 min expected duration were enrolled in the study. Patients were 18 yr or older and ASA physical status I, II, or III. Exclusion criteria included language barrier, contraindication to regional anesthesia, multiple injuries, intolerance to nonsteroidal antiinflammatory drugs, 279

nonsteroidal antiinflammatory drug-sensitive asthma, chronic pain history, and pregnancy. Patients were randomized using a computergenerated random number table. The general anesthesia group received standardized anesthesia, which included fentanyl 1 mg/kg and propofol 1–2 mg/kg. Anesthesia was maintained with oxygen/nitrous oxide and sevoflurane titrated to effect. Morphine was administered in 0.05 mg/kg aliquots at the attending anesthesiologists discretion to treat intraoperative pain. A SonoSite Titan unit (SonoSite®, Bothwell, WA) with a 38 mm linear array 5–10 MHz transducer (L38) was used to identify the axillary brachial plexus in the ultrasound-guided axillary block group. With the operative arm abducted and externally rotated and the elbow flexed to 90°, the axilla was scanned to reveal the axillary vessels. Each nerve was identified individually. Nerve stimulation was not used to confirm needletip placement. The block solution consisted of 9.5 mL 2% lidocaine with 1:200,000 epinephrine and 9.5 mL 0.5% bupivacaine with clonidine 150 mg (1 mL).15,16 A single operator performed all ultrasound-guided axillary blocks. A Stimuplex A 50 mm needle (Stimuplex®, BBraun, Melsungen, Germany) was directed under ultrasound guidance toward each of the neural structures. The needle was moved to several points around each nerve to facilitate circumferential perineural local anesthetic spread. A total of 3–5 mL of local anesthetic solution was injected and observed to surround each nerve. A maximum of 20 mL of local anesthetic mixture was injected. Motor function was assessed using a modified Bromage Scale and sensory function was assessed using pinprick (25G hypodermic needle), soft touch (gauze), and cold (ethyl chloride spray). Motor and sensory modalities were assessed every 5 up to 30 min postblock. Surgical anesthesia was defined as a motor score of 2 or lower, with absent appreciation of cold and pinprick sensation in the territory of all four terminal branches. Failure of surgical anesthesia at 30 min was managed with either an ultrasound-guided “rescue” block at a site distal to the axilla or a general anesthesia at the patient’s discretion. Both groups received 1 g acetaminophen and 75 mg diclofenac sodium IV during surgery. Intraoperative local anesthetic wound infiltration by the surgeon was permitted if it was their usual practice. Cyclizine 50 mg and ondansetron 4 mg were administered if patients received opiate analgesia during surgery. Anxiolysis was provided at the patient’s request with midazolam 2– 4 mg, titrated to effect. In the recovery room, morphine was administered IV in 2 mg aliquots until satisfactory analgesia was achieved. On discharge, patients were prescribed diclofenac sodium 75 mg for 12 h and acetaminophen 1 g for 6 h to be taken regularly for 48 h after surgery and as required thereafter. 280

Brief Report

Primary Outcome Measure The primary outcome measure was pain assessed with a 100 mm visual analog scale (0 –100) immediately after surgery in either the recovery room or the operating room in the case of recovery room bypass.

Secondary Outcome Measures The effectiveness of ultrasound-guided axillary block anesthesia was recorded, as was the need for supplemental block or conversion to general anesthesia. Postoperative pain was measured at 2, 6, 24, 48 h, and at 7 days after surgery using a Verbal Rating Scale (0 –10). The anesthesia induction time was defined as the interval in seconds from the patient lying on the operating table (all anesthesia was induced in the operating room) to the time surgical anesthesia was achieved (including performing and assessing ultrasound-guided axillary block or acquiring adequate depth of general anesthesia). Other variables recorded included, the duration of surgery (minutes); morphine consumption (mg); PONV on a four-point scale (0 ⫽ No PONV, 1 ⫽ Mild; 2 ⫽ Moderate; 3 ⫽ Severe); the time to first request for supplemental analgesia (minutes); the ability to bypass the recovery room17; the duration of recovery room stay (min); the total operating room transit time (minutes from start of anesthesia to discharge from the operating room complex); and the time to suitability for hospital discharge (min).18

Statistics Data were analyzed on an intention-to-treat basis using EpiInfo™ 2002 (Centers for Disease Control and Prevention) statistics software. Normally distributed data were analyzed using the unpaired Student t or analysis of variance as appropriate. Nonnormally distributed data were analyzed using the Mann–Whitney/ Wilcoxon two-sample test. Differences in proportions were compared by Yates ␹2 test. Statistical significance was considered at P ⬍ 0.05. To detect a 50% difference in visual analog scale pain scores with a Type I error rate of 0.01 and a power of 0.80, it was estimated that 60 patients would be required. We planned an interim analysis at 50% recruitment (30 patients).

RESULTS The study was terminated at interim analysis (n ⫽ 30). No patients meeting entry criteria were excluded from data analysis and none was lost to follow-up. The groups’ baseline characteristics were similar (Table 1). One patient required radial nerve block at the elbow because of incomplete block at 20 min. There were no conversions to general anesthesia. Three ultrasound-guided axillary block patients requested sedation, each received 4 mg midazolam. Pain scores were significantly lower in the ultrasound-guided axillary block group in the recovery room, at 2 and 6 h postoperatively. There were no significant differences in pain scores at 24, 48 h, or at ANESTHESIA & ANALGESIA

Table 1. Baseline Group Characteristics and Intraoperative Data Variable

General anesthesia (n ⫽ 15)

Ultrasound-guided axillary block (n ⫽ 15)

Significance

Age (yr) Sex M/F Weight (kg) Duration of surgery (min) Intraoperative morphine (n) Intraoperative morphine dose (mg) Anesthesia time (s) Fracture surgery Y/N Subcutaneous wound infiltration

37.7 (11.6) 4/11 74.1 (11.3) 46.5 (24.8) 11/4 3.7 (2.9) 560 (109) 7/8 4/15

40.6 (18.7) 9/6 70.9 (9.8) 54.7 (31.5) 0/15 0 (0) 783(314) 9/6 0/15

P ⫽ 0.87* P ⫽ 0.69† P ⫽ 0.41‡ P ⫽ 0.38* P ⫽ 0.00015† P ⫽ 0.0001‡ P ⫽ 0.001* P ⫽ 0.71† P ⫽ 0.1†

Data expressed as mean ⫾ SD . * Mann–Whitney/Wilcoxon two-sample test (nonnormally distributed data). † ␹2 Yates correction (proportions). ‡ ANOVA One Way (normally distributed data). § Independent t-test for samples of unequal variance (normally distributed data).

Figure 1. (a) and (b) Points on graph represent sample means with bars representing standard deviation. RR ⫽ recovery room; 2 h ⫽ 2 h postoperatively; 6 h ⫽ 6 h postoperatively; GA ⫽ general anesthesia group; RA ⫽ regional anesthesia group. *Statistically significant result. Significance testing at RR and 2 h P ⬍ 0.001 at rest and movement using Mann–Whitney/Wilcoxon TwoSample test for nonnormally distributed data; at 6 h rest P ⫽ 0.026 and 6 h movement P ⫽ 0.01 using analysis of variance One Way test for normally distributed data.

7 days (Figs. 1a and b). The ultrasound-guided axillary block group had significantly higher recovery room bypass scores, longer time to first analgesia, shorter recovery room stay, shorter total operating room Vol. 109, No. 1, July 2009

transit time, and achieved hospital discharge criteria sooner (Table 2). Morphine consumption was higher in the general anesthesia group at all time points (Fig. 2). There were no differences between PONV scores in © 2009 International Anesthesia Research Society

281

Table 2. Efficiency Measures Variable

General anesthesia (n ⫽ 15)

Ultrasound-guided axillary block (n ⫽ 15)

Significance

Recovery room bypass (Y/N) Recovery room bypass score Recovery room duration (min) Time to request supplemental analgesia (min) Total theatre transit time Time to hospital discharge criteria

0 9 (1–11) 65 (30–135) 15 (5–240) 129 (61–193) 120 (30–240)

15 (100%) 14 (12–14) 0 ⬎240 54 (30–142) 30

P ⬍ 0.0001* P ⬍ 0.0001† P ⬍ 0.0001† P ⬍ 0.0001† P ⫽ 0.002† P ⬍ 0.0001†

Data expressed as median (range). * ␹2 Yates correction (proportions). † Mann–Whitney/Wilcoxon two-sample test (nonnormally distributed data).

Figure 2. Points on graph represent sample means with bars representing standard deviation. * ⫽ statistically significant result; RR ⫽ recovery room; 2 h ⫽ 2 h postoperatively; 6 h ⫽ 6 h postoperatively; 24 h ⫽ 24 h postoperatively; GA ⫽ general anesthesia group; RA ⫽ regional anesthesia group. Significance testing at all time points P ⬍ 0.002 using Mann–Whitney/Wilcoxon TwoSample test for nonnormally distributed data.

either group (general anesthesia 0 20 and ultrasoundguided axillary block 0 median [range] at all time points P ⫽ 0.06). Only one patient in the general anesthesia group required antiemetic rescue in the recovery room. Four patients (two in either group) violated the postoperative analgesia protocol. Three patients took analgesics as required rather than on a regular basis during the first 48 h after surgery. One patient added tramadol to her analgesia regimen. Only two patients (one in either group) required supplementation of analgesia with codeine phosphate 16 mg for 6 h on the first postoperative day. There were no adverse clinical incidents relating to the conduct of general or regional anesthesia.

DISCUSSION Our data demonstrated improved perioperative analgesia, reduced opiate consumption, recovery room bypass, and earlier achievement of hospital discharge criteria with the use of regional anesthesia. This compares favorably with previous reports.1– 4 Low injectate dose did not alter the anesthetic or analgesic efficacy of axillary brachial plexus block. The potential for dose reduction with ultrasound guidance is promising. Systemic local anesthetic toxicity is still frequently reported in the anesthesia literature.19,20 Lipid rescue therapy has been associated with improved outcomes after local anesthetic toxicity.21,22 The systemic toxicity of local anesthetics is dose dependent. Reducing 282

Brief Report

the dose of drug administered may contribute to improved safety of regional anesthesia. We acknowledge that the study has limitations which may limit the generalizability of the data, including small patient numbers (n ⫽ 30) and that a single anesthesiologist performed all ultrasoundguided axillary blocks. The study may be criticized as measuring the obvious. Patients with an insensate limb have no pain sensation for the duration of the block. It is therefore not surprising to report superior analgesia when compared with patients who received general anesthesia and parenteral analgesia. Importantly, however, our study demonstrated that previously reported benefits of regional anesthesia are maintained despite the use of low-dose local anesthetic mixture. In conclusion, when compared with general anesthesia, low-dose ultrasound-guided axillary block provided excellent anesthesia, superior analgesia, reduced opiate consumption, facilitated recovery room bypass, and earlier hospital discharge readiness in patients undergoing upper limb surgery. REFERENCES 1. Richman JM, Liu SS, Courpas G, Wong R, Rowlingson AJ, McGready J, Cohen SR, Wu CL. Does continuous peripheral nerve block provide superior pain control to opioids? A metaanalysis. Anesth Analg 2006;102:248 –57 2. Liu SS, Strodbeck WM, Richman JM, Wu CL. A comparison of regional versus general anesthesia for ambulatory anesthesia: a meta-analysis of randomized controlled trials. Anesth Analg 2005;101:1634 – 42


3. Hadzic A, Williams BA, Karaca PE, Hobeika P, Unis G, Dermksian J, Yufa M, Thys DM, Santos AC. For outpatient rotator cuff surgery, nerve block anesthesia provides superior same-day recovery over general anesthesia. Anesthesiology 2005;102: 1001–7 4. McCartney CJ, Brull R, Chan VW, Katz J, Abbas S, Graham B, Nova H, Rawson R, Anastakis DJ, von Schroeder H. Early but no long-term benefit of regional compared with general anesthesia for ambulatory hand surgery. Anesthesiology 2004; 101:461–7 5. Liu SS, Wu CL. Effect of postoperative analgesia on major postoperative complications: a systematic update of the evidence. Anesth Analg 2007;104:689 –702 6. Kapral S, Greher M, Huber G, Willschke H, Kettner S, Kdolsky R, Marhofer P. Ultrasonographic guidance improves the success rate of interscalene brachial plexus blockade. Reg Anesth Pain Med 2008;33:253– 8 7. Perlas A, Brull R, Chan VW, McCartney CJ, Nuica A, Abbas S. Ultrasound guidance improves the success of sciatic nerve block at the popliteal fossa. Reg Anesth Pain Med 2008;33:259 – 65 8. Marhofer P, Schrogendorfer K, Koinig H, Kapral S, Weinstabl C, Mayer N. Ultrasonographic guidance improves sensory block and onset time of three-inone blocks. Anesth Analg 1997; 85:854 –7 9. Williams SR, Chouinard P, Arcand G, Harris P, Ruel M, Boudreault D, Girard F. Ultrasound guidance speeds execution and improves the quality of supraclavicular block. Anesth Analg 2003;97:1518 –23 10. Marhofer P, Schrogendorfer K, Wallner T, Koinig H, Mayer N, Kapral S. Ultrasonographic guidance reduces the amount of local anesthetic for 3-in-1 blocks. Reg Anesth Pain Med 1998;23:584 – 8 11. Casati A, Baciarello M, Di Cianni S, Danelli G, De Marco G, Leone S, Rossi M, Fanelli G. Effects of ultrasound guidance on the minimum effective anaesthetic volume required to block the femoral nerve. Br J Anaesth 2007;98:823–7 12. Willschke H, Bo¨senberg A, Marhofer P, Johnston S, Kettner S, Eichenberger U, Wanzel O, Kapral S. Ultrasonographic-guided ilioinguinal/iliohypogastric nerve block in pediatric anesthesia: what is the optimal volume? Anesth Analg 2006;102:1680 – 4

Vol. 109, No. 1, July 2009

13. Chan VW, Perlas A, McCartney CJ, Brull R, Xu D, Abbas S. Ultrasound guidance improves success rate of axillary brachial plexus block. Can J Anaesth 2007;54:176 – 82 14. Casati A, Danelli G, Baciarello M, Corradi M, Leone S, Di Cianni S, Fanelli G. A prospective, randomized comparison between ultrasound and nerve stimulation guidance for multiple injection axillary brachial plexus block. Anesthesiology 2007; 106:992– 6 15. Iohom G, Machmachi A, Diarra DP, Khatouf M, Boileau S, Dap F, Boini S, Mertes PM, Bouaziz H. The effects of clonidine added to mepivacaine for paronychia surgery under axillary brachial plexus block. Anesth Analg 2005;100:1179 – 83 16. Hutschala D, Mascher H, Schmetterer L, Klimscha W, Fleck T, Eichler HG, Tschernko EM. Clonidine added to bupivacaine enhances and prolongs analgesia after brachial plexus block via a local mechanism in healthy volunteers. Eur J Anaesthesiol 2004;21:198 –204 17. White PF, Song D. New criteria for fast-tracking after outpatient anesthesia: a comparison with the modified Aldrete’s scoring system. Anesth Analg 1999;88:1069 –72 18. Chung F, Chan VW, Ong D. A post-anesthetic discharge scoring system for home readiness after ambulatory surgery. J Clin Anesth 1995;7:500 – 6 19. McCutchen T, Gerancher JC. Early intralipid therapy may have prevented bupivacaine-associated cardiac arrest. Reg Anesth Pain Med 2008;33:178 – 80 20. Warren JA, Thoma RB, Georgescu A, Shah SJ. Intravenous lipid infusion in the successful resuscitation of local anestheticinduced cardiovascular collapse after supraclavicular brachial plexus block. Anesth Analg 2008;106:1578 – 80 21. Weinberg GL, Di Gregorio G, Ripper R, Kelly K, Massad M, Edelman L, Schwartz D, Shah N, Zheng S, Feinstein DL. Resuscitation with lipid versus epinephrine in a rat model of bupivacaine overdose. Anesthesiology 2008;108:907–13 22. Felice K, Schumann H. Intravenous lipid emulsion for local anesthetic toxicity: a review of the literature. J Med Toxicol 2008;4:184 –91


283

Section Editor: Lawrence Saidman

Letters to the Editor Patient-Controlled Epidural Analgesia Regimens for Labor Analgesia: Background Infusion or Demand-Only?

To the Editor: In a recent article on patientcontrolled epidural analgesia (PCEA) for labor pain, Lim et al.1 conclude that demand-only PCEA resulted in an increased incidence of breakthrough pain, higher pain scores, and lower maternal satisfaction, when compared with PCEA with background infusion (ropivacaine 0.1% with fentanyl 2 ␮g/mL; 5-mL bolus, 10 –15 min lockout interval, and 510 mL/h infusion). These conclusions raise several issues. First, we acknowledge that performing this study as a randomized, controlled, double-blind trial is a challenge, but we are concerned that several aspects of the study design make it susceptible to expectation bias. Although breakthrough pain occurred more in the demand-only group, the hourly consumption of ropivacaine remained a low mean of 7.3 mg, corresponding with approximately 1.5 boluses. This may be the result of several hours of effective analgesia, but might also be due to inadequate use of the PCEA-system or failure to administer top-up boluses in a timely manner. Moreover, breakthrough pain may have been influenced by other factors not studied, e.g., fetal position. Parturients were withdrawn from the study when visual analog scale remained greater than 30 mm, despite a maximum of two supplemental topup boluses of 5 mL ropivacaine 0.2%. Surprisingly, Table 2 shows a median number of zero administered supplemental boluses with a maximum of three in the demand-only group. The three different PCEA regimens compared do not only differ in background infusion rates, but also have different lockout times. The authors 284

explain this to be standard PCEA settings used in their institution, as recommended in a review by D’Angelo.2 Not only were these settings designed for PCEA with bupivacaine 0.125%, but they were also suggested for meeting different specific goals. We, therefore, find comparing different PCEA regimens with different lock-out intervals to be inappropriate. Finally, Lim et al. state in the conclusions section of their abstract that demand-only PCEA resulted in lower maternal satisfaction. These conclusions are not supported by their data. Niels Koopmans, MD Go¨tz J. K. Wietasch, MD, PhD Michel M. R. F. Struys, MD, PhD Department of Anesthesiology University Medical Center Groningen University of Groningen Groningen, The Netherlands [email protected]

REFERENCES 1. Lim Y, Ocampo CE, Supandji M, Teoh WH, Sia AT. A randomized controlled trial of three patient-controlled epidural analgesia regimens for labor. Anesth Analg 2008;107:1968 –72 2. D’Angelo R. New techniques for labor analgesia: PCEA and CSE. Clin Obstet Gynecol 2003;46:623–32 DOI: 10.1213/ane.0b013e3181a67096

In Response: We thank Koopmans et al.1 for their comments and observations regarding this study.2 Our results concurred with several previous studies, using both bupivacaine3,4 and ropivacaine5,6 epidural solutions. We did not analyze the trend of hourly demand in the three groups of patients, but the low consumption of ropivacaine in patients using the demand-only patient-controlled epidural analgesia (PCEA) regimen

may well be because of the low requirements of local anesthetic required at early stage of labor. Our patients are educated on the use of the PCEA upon recruitment and instructed to self-administer an epidural bolus dose when they experience mild to moderate pain, and before the pain intensity becomes severe. It is unlikely that inadequate use of the PCEA system is the cause of increased breakthrough pain, specifically for patients in the demand-only group. Our institution has an anesthesiologist assigned to the delivery suite 24 h a day, specifically attending to patients in labor. It is our institutional policy to respond to a patient’s complaint of breakthrough pain within 30 min of being informed by the nurse. A delay in the administering a top-up bolus would therefore be limited. We agree that many obstetric and patient factors other than the differing anesthetic regimens may have influenced the incidence of breakthrough pain. However, this being a randomized trial involving 300 patients, these factors would not be significantly different across the three groups recruited. We withdrew patients from the study when pain intensity on the visual analog scale remained more than 30 mm, despite a maximum of two supplemental top-up boluses of 5 mL ropivacaine 0.2%. This was intended to prevent catheter migration and/or improper placement of epidural catheter from skewing the results. The supplemental boluses in each group were not normally distributed and, as such, we analyzed

Table 1. Number of patients receiving supplemental boluses No of supplemental boluses

Group 0 n ⫽ 100

Group 5 n ⫽ 100

Group 10 n ⫽ 100

0 1 2 3

57 37 4 2

83 14 3 0

89 8 3 0 Vol. 109, No. 1, July 2009

Letter to the Editor

the data using nonparametric test and the data were presented as median (min-max). Our center’s initial PCEA regimens were based on a review by D’Angelo.7 We agree with Koopmans et al. that having differing lockout times may affect the outcomes. In our future studies, we have chosen to fix the lock-out time for all groups. Patients’ satisfaction with their pain relief was one of our secondary outcomes. Although the clinical difference was small (Group 0: 87 ⫾ 12, Group 5: 91 ⫾ 10, P ⫽ 0.028), it was statistically significant. Yvonne Lim, MMED Cecilia E. Ocampo, MD Mia Supandji, MD Wendy H. L. Teoh, FANZCA Alex T. Sia, MMED Department of Women’s Anesthesia KK Women’s and Children’s Hospital Singapore, 100 Bukit Timah Road Singapore [email protected]

REFERENCES 1. Koopmans N, Wietasch G, Struys M. PCEA regiments for labor analgesia, background infusion or demand-only? Anesth Analg 2009;109:284 2. Lim Y, Ocampo CE, Supandji M, Teoh WH, Sia AT. A randomized controlled trial of three patient-controlled epidural analgesia regimens for labor. Anesth Analg 2008;107:1968 –72 3. Ferrante FM, Rosinia FA, Gordon C, Datta S. The role of continuous background infusion in parturient-controlled epidural analgesia for labor and delivery. Anesth Analg 1994;79:80 – 4 4. Petry J, Vercauteren M, Van Mol I, Van Houwe P, Adriaensen HA. Epidural PCA with bupivacaine 0.125%, sufentanil 0.75 microgram and epinephrine 1/800.000 for labor analgesia: is a background infusion beneficial? Acta Anaesthesiol Belg 2000;51: 163– 6 5. Bremerich DH, Waibel HJ, Mierdl S, Meininger D, Byhahn C, Zwissler BC, Ackermann HH. Comparison of continuous background infusion plus demand dose and demand-only parturientcontrolled epidural analgesia (PCEA) using ropivacaine combined with sufentanil for labor and delivery. Int J Obstet Anesth 2005;14:114 –17 6. Boselli E, Debon R, Cimino Y, Rimmele´, Allaouchiche B, Chassard D. Background infusion is not beneficial during labor patient-controlled analgesia with ropivacaine 0.1% plus 0.5 ␮g/ml sufentanil. Anesthesiology 2004;100:968 –72 Vol. 109, No. 1, July 2009

7. D’Angelo R. New techniques for labor analgesia: PCEA and CSE. Clin Obstet Gynecol 2003;46:621–32 DOI: 10.1213/ane.0b013e3181a670ab

Stepwise Logistic Regression To the Editor: Through simulations, Pace1 demonstrates in an editorial the difficulties of stepwise automatic variable selection as applied to logistic regression. I agree with Dr. Pace that one needs to exercise caution with any kind of model selection technique and that prior knowledge in the area of study is extremely important in covariate selection. In his editorial, Pace refers to three variants of automatic variable selection: forward selection, backward elimination, and stepwise regression and the simulations were presented for model selection using stepwise regression. From his supplementary data analysis codes, the model selection technique employed in the simulation was backward elimination. (For his specification of the model the software chooses backward elimination as the default method). Using the backward elimination method, there were 825 instances in the 1000 simulations with at least one significant covariate at P ⬍ 0.05. When the same simulations were repeated with either forward selection or stepwise regression, no covariate was found significant at P ⬍ 0.05. Pace also uses the Akaike Information Criterion2 to choose the model in backward elimination. The Bayesian Information Criterion,2 provides a greater penalty for the addition of an extra covariate. Simulation using the Bayesian Information Criterion failed to choose a single covariate at P ⬍ 0.05 in each of the 1000 simulations. Furthermore, with Pace’s simulation (i.e., backward elimination using Akaike Information Criterion for model selection), if one used multiple hypothesis tests3 This article has supplementary material on the Web site: www.anesthesia-analgesia.org.

and computed adjusted P values, then the number of instances where a significant covariate was selected would decrease to 414 instances (332 with 1, 73 with 2, and 9 with 3 covariates) (see our simulations in the online appendix, available at www.anesthesia-analgesia.org). These results demonstrate the effects the various model selection techniques and selection criteria have on the selection of the final model. Therefore, to interpret the data properly, it is always advisable to consider various model building and selection strategies in statistical analysis, as suggested by Dr. Pace. Srikesh G. Arunajadai, PhD Department of Anesthesiology and Biostatistics Columbia University New York City, New York [email protected]

REFERENCES 1. Pace NL. Independent predictors from stepwise logistic regression may be nothing more than publishable P values. Anesth Analg 2008;107:1775– 8 2. Claeskens G, Hjort NL. Model selection and model averaging. Cambridge University Press, 2008 3. Hothorn T, Bretz F, Westfall P. Simultaneous inference in general parametric models. Biometrical J 2008;50:346 – 63 DOI: 10.1213/ane.0b013e3181a7b51a

In Response: We thank Dr. Arunajadai for his comments about the statistical simulations in our editorial (text NLP, algorithm WMB) demonstrating the perils of stepwise logistic regression.1 This allows us to clarify an ambiguity in the nomenclature of the stepwise automatic variable selection algorithm. Correctly specified, the algorithm should be described as either stepwise forward selection, stepwise backward elimination, or stepwise with forward selection and/or backward elimination; however, the word stepwise itself is also commonly used to refer to any of the three variants or This article has supplementary material on the Web site: www.anesthesia-analgesia.org.


285

Letters to the Editor

to just the third variant. Arunajadai2 has correctly stated that our particular simulations used the stepwise backward elimination variant. Our simulations used randomly created covariates to demonstrate how commonly there was the creation of spurious associations by stepwise modeling (backward elimination variant). Dr. Arunajadai has also provided R software code to perform the other two variants; he reports that there were no spurious associations with no covariate significant at P ⬍ 0.05 using either the forward selection or the forward selection/ backward elimination variants. In his code, Arunajadai estimates a mean intercept model object, i.e., “fit ⬍glm(y ⬃ 1, data ⫽ w, family ⫽ binomial),” for submission to the stepwise function. The submission of a mean intercept model to the stepwise process cannot identify any association, true or spurious. When a full (all covariates) model, i.e., “fit ⬍- glm(y ⬃., data ⫽ w, family ⫽ binomial)” is used, all three variants have qualitatively the same results of numerous spurious associations (appendix available at www.anesthesia-analgesia.org). The inclusion of noise variables during stepwise modeling regardless of the variant has been demonstrated elsewhere.3–5 Dr. Arunajadai also raised the very interesting question of which information criterion should be used at each step for adding or removing a covariate; he advocates the Bayesian Information Criterion (BIC) in contrast to the Akaike Information Criterion (AIC) used in our simulation. Both the AIC and the BIC are indexes in which twice the negative maximized log likelihood of the model fit is penalized by subtracting either twice the number of model parameters (AIC) or the number of model parameters multiplied by the log of the sample size (BIC). Of the candidate models possible, the model with the higher AIC or higher BIC is favored. As Arunajadai noted, the BIC is more heavily penalized and will produce more parsimonious models 286


(fewer significant covariates). However, there is a competition in choosing between AIC and BIC; the AIC will yield optimal regression estimation while the BIC represents consistent model identification rules. It is not possible to create models with the properties favored by both the AIC and the BIC.6 Using the BIC index in our simulation still produces spurious associations. Automatic variable selection via a stepwise process is a hazardous undertaking. As J. B. Copas3 humorously noted, “If you torture the data for long enough, in the end they will confess . . . . What more brutal torture can there be than subset selection? The data will always confess, and the confession will usually be wrong.” Nathan L. Pace, MD, MStat Department of Anesthesiology University of Utah Salt Lake City, Utah

William M. Briggs, PhD Department of Emergency Medicine New York Methodist Hospital Brooklyn, New York [email protected]

REFERENCES 1. Pace NL. Independent predictors from stepwise logistic regression may be nothing more than publishable P values. Anesth Analg 2008;107:1775– 8 2. Arunajadai S. Stepwise logistic regression. Anesth Analg 2009;109:285 3. Copas JB. quoted by: Miller AJ. Selection of subsets of regression variables. J R Stat Soc [Ser A] 1984;147:412. Cited by Derksen S, Keselman HJ. Backward, forward and stepwise automated subset selection algorithms: frequency of obtaining authentic and noise variables. Br J Math Stat Psychol 1992;45:265– 82 4. Flack VF, Chang PC. Frequency of selecting noise variables in subset regression analysis: a simulation study. Am Stat 1987;41:84 – 6 5. Mundry R, Nunn CL. Stepwise model fitting and statistical inference: turning noise into signal pollution. Am Nat 2009;173:119 –23 6. Yang Y. Can the strengths of AIC and BIC be shared? A conflict between model identification and regression estimation. Biometrika 2005;92:937–50 DOI: 10.1213/ane0b013e3181a7b52d

Elastin Arteriopathy and William Syndrome: Do You Feel Lucky? To the Editor: Over the past 2 yr, we have anesthetized 10 patients with Williams

syndrome and concur with the recommendations by Burch et al.1 in their recent article. In addition, we have additional suggestions based on the successful resuscitation of two infants who suffered a cardiac arrest under anesthesia. In both infants, induction of anesthesia was with incremental doses of sevoflurane (2% and 3%), a remifentanil infusion and rocuronium. Cardiac arrest occurred within 5 min of induction and before intubation. The electrocardiograms demonstrated marked ST segment depression that was followed soon after by bradycardia. Neither patient had discernible systemic output (no palpable or Doppler evidence of pulses). Resuscitation with phenylephrine, epinephrine, and small fluid boluses occurred during the institution of cardiopulmonary resuscitation. Within 5 min from the start of resuscitation, the transthoracic echocardiography (TTE) demonstrated biventricular akinesis in the first and severe biventricular hypokinesis in the second patient. We used TTE to guide our therapy, allowing us to monitor wall motion abnormalities, guide fluid administration, and carefully titrate vasopressors without worsening tachycardia or the dynamic component of subarterial outflow tract obstruction. In addition, we avoided administration of epinephrine once the rhythm returned, but pulses were still absent (pulseless electrical activity) because the TTE demonstrated severely hypertrophied, underfilled ventricles. In this population, the inappropriate administration of epinephrine as per the pediatric advanced life support protocol for pulseless electrical activity could have compromised recovery. The second patient required urgent extracorporeal membrane oxygenation cannulation after sustaining another cardiac arrest several hours after arrival in the pediatric intensive care unit within 3 min after receiving a 1 ␮g/kg dose of fentanyl. Both the patients underwent successful surgical repair. These patients ANESTHESIA & ANALGESIA


made in our article.2 First, patients with elastin arteriopathy resulting in bilateral ventricular outflow tract obstruction seem to be at greatest risk of anesthesia and sedation-related cardiac arrest and should be cared for in institutions with the personnel and infrastructure necessary to rapidly provide the type of response described by Joffe et al.1 Second, patients with isolated supravalvular aortic stenosis, regardless of its apparent severity, are also at risk. Because risk stratification of these patients is imperfect, institutions caring for any patient with elastin arteriopathy should anticipate the need for rapid, aggressive resuscitation. James A. DiNardo, MD Francis X. McGowan, MD

Figure 1. The PA hypoplasia.

Barry D. Kussman, MD Andrew J. Powell, MD

had severe bilateral outflow track obstruction with what Stamm et al.2 have termed “macaroni pulmonary arteries” (Fig. 1). It is noteworthy that the presence of an intracardiac shunt (patent foramen ovale and small restrictive ventricular septal defect) did not decrease the risk of arrest. Despite an induction technique designed to minimize changes in coronary perfusion, our experience demonstrates that these patients continue to have a very highanesthetic risk of cardiac arrest. We postulate that, during induction, a decrease in coronary perfusion pressure occurred secondary to vasodilation and in combination with relative hypovolemia, and severe biventricular hypertrophy precipitated ischemia leading to cardiac arrest. In the second patient, a cardiac arrest occurred within minutes after a small dose of fentanyl, highlighting the precarious hemodynamic balance of these children. We recommend preinduction hydration, ideally with an IV, but at minimum liberal PO fluids until 2 h preinduction and the early use of TTE to help treat hemodynamic abnormalities and guide therapy. Vol. 109, No. 1, July 2009

Denise C. Joffe, MD Michael Richards, BM Michael Eisses, MD Brian Emerson, MD Jeremy Geiduschek, MD Department of Pediatric Anesthesiology University of Washington Seattle, Washington [email protected]

REFERENCES 1. Burch TM, McGowan FX Jr, Kussman BD, Powell AJ, DiNardo JA. Congenital supravalvular aortic stenosis and sudden death associated with anesthesia: what’s the mystery. Anesth Analg 2008;107: 1848 –54 2. Stamm C, Friehs I, Moran AM, Zurakowski D, Bacha E, Mayer JE, Jonas RA, Del Nido PJ. Surgery for bilateral outflow tract obstruction in elastin arteriopathy. J Thorac Cardiovasc Surg 2000;120:755– 63 DOI: 10.1213/ane.0b013e3181a7fdfc

In Response: We agree with Joffe et al.1 that resuscitation of patients with congenital supravalvular aortic stenosis, particularly those with concomitant right ventricular outflow tract obstruction, is notoriously difficult and that prompt use of echocardiography seems to be a useful adjuvant in successful resuscitation. The experience of Joffe et al. also serves to emphasize two points

Children’s Hospital Boston Boston, Massachusetts [email protected]

Thomas M. Burch, MD Medical University of South Carolina Charleston, South Carolina

REFERENCES 1. Joffe D, Joffe DC, Richards M, Eisses M, Emerson B, Geiduschek J, Richards M, Eisses M. Elastin arteriopathy and William’s syndrome: do you feel lucky? Anesth Analg 2009;109:286 –7 2. Burch TM, McGowan FX, Kussman BD, Powell JA, DiNardo JA. Congenital supravalvular aortic stenosis and sudden death associated with anesthesia: what’s the mystery? Anesth Analg 2008;107:1848 –54 DOI: 10.1213/ane.0b013e3181a7ff03

Patients Undergoing Transurethral Resection of the Prostate Should Receive the Same Amount of Intravenous Fluids To the Editor: In the recent article by Riesmeier et al.1 studying the effect of different types of fluid on hemodynamic in elderly patients before instituting spinal anesthesia, an important defect in the study design is that the subjects received different amounts of fluid (500 mL saline vs 500 mL saline plus 500 mL hydroxyethyl starch). Although the authors mention this as a limitation that may in and of itself


287


influence the primary end point (hemodynamic changes), the practical effect of this limitation is that no firm conclusions can be made in terms of type of fluid and its effects on hemodynamics. In other words, to study the efficacy of preload volume and compare the effects of crystalloid/ colloid versus crystalloid administration before spinal anesthesia on cardiac output and blood pressure, both groups should receive the same total amount of fluids.

hetastarch plus LR was more effective in stabilizing hemodynamic status than LR alone. However, this article is also flawed in that different volumes were given in the two groups, albeit less in the colloid group than in the crystalloid group.2 Further studies with many patients should be undertaken to fully elucidate the value of crystalloid and colloid preload in the prevention of spinal anesthesiainduced hypotension. Stefan Suttner, MD, PhD

Ates Duman, MD Department of Anesthesiology Selcuk University Hospital Konya, Turkey [email protected]

Seza Apiliogullari, MD Department of Anesthesiology Faruk Sukan Obstetrics and Childrens Hospital Konya, Turkey

REFERENCE 1. Riesmeier A, Schellhaass A, Boldt J, Suttner S. Crystalloid/colloid versus crystalloid intravascular volume administration before spinal anesthesia in elderly patients: the influence on cardiac output and stroke volume. Anesth Analg 2009;108:650 – 4 DOI: 10.1213/ane.0b013e3181a7fedc

In Response: We do not fully agree with the interpretation of our data by Duman and Apiliogullari1 that no firm conclusions can be made regarding the type of fluid and its effects on cardiac output and blood pressure.2 In our discussion, we did point out that the difference in results could have been due to the different volumes of fluid given to the two groups. However, we believe that the more stable hemodynamic status observed after colloid plus saline administration probably relates to the colloid remaining in the intravascular compartment longer than crystalloids alone. Our results are consistent with the findings of Riley et al.3 who compared the efficacy of 500 mL 6% hetastarch plus 1000 mL lactated Ringer’s solution (LR) with an even larger volume of LR (2000 mL) given alone before spinal anesthesia to patients undergoing nonurgent cesarean delivery. They concluded that 6% 288


Andre´ Riesmeier, MD Joachim Boldt, MD, PhD Department of Anesthesiology and Intensive Care Medicine Klinikum der Stadt Ludwigshafen Ludwigshafen, Germany [email protected]

Unfortunately, none of the devices cited as accurate by Hannenberg and Sessler2 are suitable for all (notably awake) patients in routine clinical practice. It may, however, be worth reassessing the temporal artery thermometer. Although Suleman et al.3 found them insufficiently accurate during rewarming, our preliminary findings, using a different model (TAT5000, Exergen, MA) over the narrower range seen in general surgery, are more encouraging.4 Infrared aural temperatures are not fit for the purpose of either clinical care or quality incentives. There is an urgent need for a noninvasive thermometer that is accurate, or at least to establish which one is the “least inaccurate.” C. Mark Harper, FRCA

REFERENCES 1. Duman A, Apiliogullari S. Patients undergoing TURP should receive minimal amount of intravenous fluids. Anesth Analg 2009;109:287– 8 2. Riesmeier A, Schellhaass A, Boldt J, Suttner S. Crystalloid/colloid versus crystalloid intravascular volume administration before spinal anesthesia in elderly patients: the influence on cardiac output and stroke volume. Anesth Analg 2009;108:650 – 4 3. Riley ET, Cohen SE, Rubenstein AJ, Flanagan B. Prevention of hypotension after spinal anesthesia for cesarean section: six percent hetastarch versus lactated Ringer’s solution. Anesth Analg 1995;81:838 – 42 DOI: 10.1213/ane.0b013e3181a7fef1

The Need for an Accurate Noninvasive Thermometer To the Editor: Modell et al.’s1 study on temperature changes during ECT does not emphasize clearly enough the fundamental problems arising from the lack of an accurate noninvasive thermometer. We found postoperative infrared aural canal temperatures to be, on average, 0.74°C less than those from the temporal artery meaning that in 60% (versus 9.1%) of our patients the temperatures were ⬍36°C. We subsequently audited a small sample of ASA PS I and II patients as soon as they arrived in our ambulatory surgery unit and found 20% had, somewhat implausibly, aural temperatures ⬍36.0°C with some as low as 35.4°C.

Brighton Anaesthetic Research Forum Royal Sussex County Hospital Brighton, UK [email protected]

REFERENCES 1. Modell JH, Gravenstein N, Morey TE. Body temperature change during anesthesia for electroconvulsive therapy: implications for quality incentives in anesthesiology. Anesth Analg 2008;107:1618–20 2. Hannenberg AA, Sessler DI. Improving perioperative temperature management. Anesth Analg 2008;107:1454 –7 3. Suleman MI, Doufas AG, Akca O, Ducharme M, Sessler DI. Insufficiency in a new temporal-artery thermometer for adult and pediatric patients. Anesth Analg 2002;95:67–71, table of contents 4. Harper CM, Crook D. A study to compare the accuracy and suitability of two methods of temperature measurement in the peri-operative setting. Eur J Anaesthesiol 2008;25:3AP1– 4 DOI: 10.1213/ane.0b013e3181a7b0c2

In Response: In response to Dr Harper’s comment1 that our “study on temperature changes during electroconvulsive therapy does not emphasize clearly enough the fundamental problems arising from lack of an accurate noninvasive thermometer,”2 we would point out that the purpose of our study was not to evaluate the accuracy of various types of equipment but, rather, to demonstrate, in the population we studied, that many variables in temperature, whether ANESTHESIA & ANALGESIA


real or by inaccurate measurement, make the adoption of the ASA Performance Guidelines for hypothermia as not meeting the standard of a single-threshold temperature in the postanesthetic period inappropriate.2 To address the issue of accuracy of thermometers used in clinical practice, we did state, “A number of studies have documented that infrared tympanic thermometers are not reliable. Our data suggest that reliance on an inaccurate device measurement, as is routinely done in clinical practice, could yield indications that performance guidelines were not being met, largely based on spurious data.”2 Harper1 brings up several valid points, suggesting that, in our clinical practice, the various types of equipment that we use to measure temperature can be fraught with error, thus further supporting our conclusion. Jerome H. Modell, MD, DSc (Hon) Nikolaus Gravenstein, MD Timothy E. Morey, MD Department of Anesthesiology University of Florida College of Medicine Gainesville, Florida [email protected]

REFERENCES 1. Harper CM. The need for an accurate noninvasive thermometer. Anesth Analg 2009;109:288 2. Modell JH, Gravenstein N, Morey TE. Body temperature changes during anesthesia for electroconvulsive therapy: implications for quality incentives in anesthesiology. Anesth Analg 2008;107:1618 –20 DOI: 10.1213/ane.0b013e3181a90858

Beware of the “Serpentine” Inferior Thyroid Artery While Performing Stellate Ganglion Block To the Editor: The frequency of retropharyngeal hematoma after stellate ganglion block is reported to be 1 in 100,000 cases with resulting airway compromise and obstruction.1 However, Kapral et al.2 reported a greater incidence of asymptomatic hematoma with the blind technique (3 of 12 patients), with no hematoma occurring when ultrasound guidance was used. Hematoma was attributed to injury to the thyroid gland or the vertebral artery. I believe that the inferior thyroid vessels may be the major source of the retropharyngeal hematoma because of their vulnerable and variable anatomy. The inferior thyroid artery originates from the thyrocervical trunk of the subclavian artery and ascends anteriorly to the vertebral artery and the longus coli muscle and then curves medially behind the carotid sheath to enter posteriorly the inferior part of the thyroid lobe. It is vulnerable to injury as it lies anterior to the vertebral artery at C7 level or more commonly when it crosses (at C6 –7 level) behind the carotid artery from lateral to medial to end in the thyroid gland. This is the most critical portion of the vessel to be injured during performing the procedure with the blind technique

or even with fluoroscopic guidance whether the approach is medial or lateral to the carotid artery. As the artery has a variable unpredicted anatomy3 and has a very tortuous serpentine course,4 it may be injured by the needle and this should be prevented using ultrasound guidance. The best way to avoid injury or injection into this artery is to perform an ultrasonographic prescan of the neck in the short-axis view to identify the artery and to follow its course from its origin at the thyrocervical trunk of the subclavian artery to the thyroid gland. As the artery curves medially behind the carotid artery, one usually can obtain a longitudinal view of the artery at this level extending along a critical area that can be easily injured using the anterior paratracheal approach whether medial or lateral to the carotid (Fig. 1). If the artery has a very tortuous course, more than one cross-section of the artery may be seen (Fig. 2). Using real-time ultrasonography, once the artery is identified, the transducer can be moved slightly cephalad or caudad until the artery is not seen. The needle can be placed either out of plane, obliquely, or preferably in plane (if technically feasible) to target the cervical sympathetic chain anterior to the longus coli muscle. In conclusion, ultrasound-guided stellate ganglion block may improve the safety of the procedure by direct visualization of the related critical anatomical structures especially the

Figure 1. (A) Ultrasonographic shortaxis image at C7 using a linear (L12-3) transducer. (B) Corresponding illustration. SCM ⫽ sternocleidomastoid muscle; Th ⫽ thyroid; symp Ch ⫽ sympathetic chain; LC ⫽ longus coli muscle; ITA ⫽ inferior thyroid artery; CA ⫽ carotid artery; IJV ⫽ internal jugular vein (compressed); Scal A ⫽ scalenus anterior muscle; VV ⫽ vertebral vein; VA ⫽ vertebral artery; N ⫽ nerve root.

Vol. 109, No. 1, July 2009


289


Figure 2. (A) Ultrasonographic shortaxis image at C7 using a curved (C8-4) transducer. (B) Corresponding illustration. Tr ⫽ trachea; Es ⫽ esophagus; Th ⫽ thyroid; symp Ch ⫽ sympathetic chain; LC ⫽ longus coli muscle; ITA ⫽ inferior thyroid artery; CA ⫽ carotid artery; VA ⫽ vertebral artery.

inferior thyroid artery and accordingly the risk of retropharyngeal hematoma may be minimized. Samer Narouze, MD, MS Program Director, Pain Medicine Fellowship Pain Management Department Cleveland Clinic Cleveland, OH [email protected]

290


REFERENCES 1. Higa K, Hirata K, Hirota K, Nitahara K, Shono S. Retropharyngeal hematoma after stellate ganglion block. Anesthesiology 2006;105:1238 – 45 2. Kapral S, Krafft P, Gosch M, Fleischmann D, Weinstabl C. Ultrasound imaging for stellate ganglion block: direct visualization of puncture site and local anesthetic spread. Reg Anesth 1995;20:323– 8

3. Toni R, Casa CD, Castorina S, Roti E, Ceda G, Valenti G. A meta-analysis of inferior thyroid artery variations in different human ethnic groups and their clinical implications. Ann Anat 2005; 187:371– 85 4. Standring S. Gray’s anatomy: the anatomical basis of clinical practice. 39th ed. New York: Elsevier Churchill Livingstone, 2005 DOI: 10.1213/ane.0b013e3181a20197


Section Editor: Paul White

Book, Multimedia, and Meeting Reviews Obstetric Anaesthesia, 1st ed. (Oxford Specialist Handbooks in Anaesthesia) Clyburn P, Collis R, Harries S, Davies S, eds. New York: Oxford University Press, 2008. ISBN-10: 0199208328; ISBN13: 978 – 0199208326. 768 pages (paperback), $79.95. bstetric Anaesthesia is the first edition of an obstetric anesthesia offering in the well-known series: Oxford Specialist Handbooks of Anaesthesia. This obstetric anesthesia manual is designed to appeal to trainees in anesthesiology, but would also be useful for non-MD providers and more senior anesthetists who do not frequently attend on the labor and delivery ward. Despite its 768 pages, the handbook’s reasonable size, weight, and plasticized cover make it easy to carry in a pocket. In addition, it is visually well-designed, with bulleted, boxed, and peach-colored sections lending an ease of reading not found in similar handbooks with either strict outline or tiny-print paragraph formats. Not surprisingly, Obstetric Anaesthesia has a particularly British slant. Examples of this include a section on the U.K. maternal morbidity and mortality data found in CEMCH (Confidential Enquiry into Maternal and Child Health), discussions of labor pain relievers unfamiliar to U.S. practitioners (Nitrous Oxide Inhalers), use of European units of measure and abbreviations (although a list at the front of the text helps with this), and hints for how to make the most of the U.K. midwifery system. The handbook nicely covers the obligatory topics of maternal physiology, labor pain relief, and management of cesarean section by both regional and general anesthesia. The chapter on maternal pathophysiology is also well organized, but heavily weighted toward cardiac problems, with 30 of 52 pages devoted to this area. In general, coexisting disease is covered less extensively in this handbook than in other similar handbooks. The chapters covering obstetric conditions and postpartum issues are extensive and particularly useful for the anesthesiology trainee trying to cope with obstetric-related problems on postpartum rounds. Particularly useful is the

O

Vol. 109, No. 1, July 2009

handbook’s approach to maternal disasters, emphasizing team-based management of both maternal hemorrhage and cardiac arrest. Unique to this book is how it addresses the nonmedical issues on the OB floor – a particularly difficult and often-ignored topic for trainees. The handbook begins with the A–Z of “surviving the labour ward,” with sage bits of advice on topics from “bleeps” (pagers) through “plan Z.” Additionally sprinkled throughout the handbook, the reader will find suggestions for coping with fetal death, needle phobia, the Jehovah’s Witness parturient, and other potential pitfalls. Missing from the handbook is a comprehensive discussion of obesity-related obstetrical complications. This everenlarging problem in industrialized nations gets only brief mention in a few sections. Another topic that deserved more space was neonatal resuscitation, which is a source of great consternation for many OB anesthesiologists who are called upon to provide care for the distressed newborn. This handbook is comparably priced to other similar manuals and would make a great reference for a UK trainee. Because of its emphasis on British techniques, it cannot be recommended as the sole handbook for a U.S. anesthesiology resident, but would make an excellent additional text, especially for its approach to nonmedical issues. Elizabeth H. Ellinas, MD Assistant Professor, Medical College of Wisconsin Director of OB Anesthesiology, WFHC St. Joseph Wheaton Franciscan Healthcare St. Joseph, Perinatal Services Milwaukee WI [email protected]

Peripheral Nerve Blocks: A Color Atlas, 3rd ed. Chelly J. Philadelphia: Lippincott Williams & Wilkins, 2009. ISBN-10: 0781768764; ISBN-13: 978-0781768764. 496 pages, $119.00. he 3rd edition of Peripheral Nerve Blocks: A Color Atlas by Dr. Chelly contains 72 chapters organized into seven separate sections. The textbook covers the entire spectrum of peripheral nerve blocks, from single injection to continuous catheter techniques, pediatric blocks, chronic pain blocks, ultrasound guidance, and

T

even nerve mapping. There are numerous illustrations throughout the text that are extremely helpful in understanding the anatomy of nerve block procedures. Although some of the images could be improved, Dr. Chelly has done an excellent job of incorporating schematic, gross anatomic, surface landmark, and ultrasound images into this textbook. The bullet point style of the text enables readers to quickly extract the essentials of a particular nerve block procedure making this text suitable as a manual for performing nerve blocks in the perioperative holding area. The first section of the book discusses general concepts related to performance of peripheral blocks including basic equipment and local anesthetic solutions, as well as nerve mapping techniques. A few minor errors occur including labeling prilocaine as an ester, as well as incorrectly interchanging “mg/kg with ml/kg” when discussing intralipid administration for local anesthetic toxicity (page 13). The second section discusses single injection peripheral blocks of the upper and lower extremity and miscellaneous anatomical blocks. Each application discusses basic considerations for performing the block and anatomical landmarks prior to discussing the block technique. The third section logically follows the second and includes advanced techniques for continuous regional blocks organized into general considerations followed by discussions of various applications, and a chapter on local anesthetics utilized for continuous infusion techniques. The fourth section is dedicated to a detailed discussion of ultrasound. This section has been greatly expanded from previous editions and is prefaced by an “in depth” discussion of basic ultrasonography followed by specific applications of ultrasound guided blocks. Another important new feature of the third edition of this book is the addition of sections describing pediatric peripheral blocks and continuous peripheral pediatric blocks, respectively. The fact that these chapters are written independent of the adult chapters emphasizes that “children are not little adults” and discuss factors important to the safe practice of pediatric blocks such as estimated distances to target nerves. The only real discontinuity is that pediatric ultrasound blocks were discussed in the basic chapter on ultrasonography. The last chapter is dedicated to a discussion of blocks for chronic pain. As with other concisely 291

Book and Multimedia Reviews

written, practical textbooks on regional anesthetic techniques, many of the sections should be considered as a very basic atlas. For more detailed information regarding the evidence basis for performance of pain-related block techniques, as well as other issues related to chronic pain, the reader should refer to more comprehensive textbooks on pain management. In summary, this latest edition of Dr. Chelly’s peripheral nerve block book enhances previous editions by including recent evidence for previ-

292

Book and Multimedia Reviews

ously discussed techniques, as well as presenting up-to-date information on novel techniques and new methods of block performance. Although the book is very well written and easy to read, one notable omission is a chapter covering postoperative management of continuous catheters in the presence of postoperative low molecular weight heparin medications. The impact of performing regional blocks in the presence of these medications has become increasingly controversial and additional discussion of this, along with a discussion of acute

pain services would enhance future editions of this textbook. We feel that this practically oriented book would be of interest to both the anesthesia resident (who is just learning the basics of regional anesthesia), as well as the experienced regional anesthesiologist. Roy Greengrass [email protected]

Christopher Robards [email protected] Mayo Clinic Jacksonville, FL


Anesthesia & Analgesia Vol 109 Issue 01 2009

Anesthesia & Analgesia Vol 110 Issue 1 Jan 2010

Manual of Equine Anesthesia and Analgesia

Anesthesia and Analgesia in Laboratory Animals

Spektrum der Wissenschaft Januar 2009 issue 01

Year 01 Issue 01

Year 01 Issue 01

Analgesia

Spektrum der Wissenschaft Oktober 2006 issue 109

The South Pole Vol 01&2

Mathematics Magazine—Vol. 82, 2009 issue 1-5

Medical Clinics of North America Vol 93 Issue 01 2009 Osteoarthritis

PET Clinics Vol 3 Issue 03 2009 Pediatric PET Imaging

Журнал Q 01 2009

Food & Function. Vol 01, No 01, October 2010

Pautas De Analgesia

Авиация и время 2009 01

Авиация и космонавтика 2009 01

Opera Philosophica Vol. 01

Grantville Gazette Vol 01

Selected Writings Vol 01

Opera Philosophica Vol. 01

Harvard Theological Review 2009-01

Erotiek Omnibus Vol. 01

Opera Theologica Vol. 01

Erotiek Omnibus Vol. 01

Kevin's Stories vol 01

Legendwar Vol 01

Law & Society Review (2008 Vol 42, Issue 2)

Spektrum der Wissenschaft Oktober 2009 issue 10

Spektrum der Wissenschaft Juni 2009 issue 06

Anesthesia & Analgesia Vol 109 Issue 01 2009

Anesthesia & Analgesia Vol 110 Issue 1 Jan 2010

Manual of Equine Anesthesia and Analgesia

Anesthesia and Analgesia in Laboratory Animals

Spektrum der Wissenschaft Januar 2009 issue 01

Year 01 Issue 01

Year 01 Issue 01

Analgesia

Spektrum der Wissenschaft Oktober 2006 issue 109

The South Pole Vol 01&2

Mathematics Magazine—Vol. 82, 2009 issue 1-5

Medical Clinics of North America Vol 93 Issue 01 2009 Osteoarthritis

PET Clinics Vol 3 Issue 03 2009 Pediatric PET Imaging

Журнал Q 01 2009

Food & Function. Vol 01, No 01, October 2010

Pautas De Analgesia

Авиация и время 2009 01

Авиация и космонавтика 2009 01

Opera Philosophica Vol. 01

Grantville Gazette Vol 01

Selected Writings Vol 01

Opera Philosophica Vol. 01

Harvard Theological Review 2009-01

Erotiek Omnibus Vol. 01

Opera Theologica Vol. 01

Erotiek Omnibus Vol. 01

Kevin's Stories vol 01

Legendwar Vol 01

Law & Society Review (2008 Vol 42, Issue 2)

Spektrum der Wissenschaft Oktober 2009 issue 10

Spektrum der Wissenschaft Juni 2009 issue 06

Recommend Documents