Management of Intraoperative Complications

Operative Techniques in General Surgery Vol 9, No 4, December 2007 Management of Intraoperative Complications Russell...

Author: Russell J. Nauta

199 downloads 802 Views 9MB 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

Operative Techniques in

General Surgery Vol 9, No 4, December 2007

Management of Intraoperative Complications Russell J. Nauta, MD, Guest Editor 139

Editorial Walter A. Koltun

140

Introduction Russell J. Nauta

142

Technical Considerations in the Difficult Colorectal Anastomosis Thomas J. Stahl

152

Control of Bleeding from the Portal/Superior Mesenteric Vein Joshua S. Hill, Sridhar Shankar, Douglas B. Evans, and Jennifer F. Tseng

160

The Control of Intraoperative Hemorrhage During Pelvic Operations Horst S. Filtzer

167

Management of Ureteral Injury Paul LaFontaine

175

Management of Common Bile Duct Injuries John C. Haney and Theodore N. Pappas

Volume 9, Number 4

December 2007

Editorial

T

here are many texts, articles, and journals that describe step by step, the way to perform a surgical procedure under elective circumstances. This journal is one such example of that fact. However, it is much more difficult to describe and address the surgical management of the unexpected and usually very idiosyncratic surgical complication. Too frequently, such an occurrence is because of a variation in patient anatomy, some unusually placed tumor or other unforeseen factor that makes the operating room the only place to then develop and implement a corrective surgical plan. One of my personal rules of surgical performance hinges on the maxim that the operating room is a very dangerous place to be making decisions. It is too stressful with too many distractions and most of the surgeon’s effort should be focused on being technically perfect, using practiced motions and following a game plan thought out well in advance of the actual operative procedure. An intraoperative complication is a monkey wrench thrown into the works of that performance paradigm. One has to redirect the operative course and frequently change the procedure originally planned. One way to mitigate a complication’s potentially dangerous consequences is to similarly plan (and mentally practice) ones responses to the multitude of conceivable complications possible. This issue is dedicated to that purpose. It would be a backhanded compliment to say that Russ Nauta knows how to handle complications. I have worked with him operating on patients that others have avoided because of the risk and challenge involved in undertaking the toughest cases that still needed to be done. The successful management of surgical complications separates the ordinary

1524-153X/07/$-see front matter © 2007 Elsevier Inc. All rights reserved. doi:10.1053/j.optechgensurg.2008.01.001

surgeon from the truly great one and Russ falls into the latter group. Many a time I have seen him come to the rescue of a colleague’s surgical misadventure, sometimes by being called into the operating room on short notice or alternatively, taking in transfer the patient that has been languishing after a postoperative complication at another institution. His ability to recognize the alternatives on how to deal with a complication and then making the often-hard decision on how to dig oneself (and the patient) out of a deep hole is unparalleled. He has assembled an experienced and knowledgeable group of contributors who have brought together a great collection of operative techniques, tricks, and lessons on how to manage surgical complications. The present issue is largely dedicated to the management of intraoperative complications. However, as usual, Russ has gone the extra mile and will also be editor on the subsequent issue addressing the management of postoperative complications. These two issues will be resources infrequently found in the literature at large. It will be the prudent general surgeon who will keep them close at hand when planning his or her next major operative procedure. Walter A. Koltun, MD Professor of Surgery Peter and Marshia Carlino Professor of Inflammatory Bowel Disease Chief, Section of Colon and Rectal Surgery Penn State College of Medicine Milton S. Hershey Medical Center Editor-in-Chief

139

Volume 9, Number 4

December 2007

Introduction

L

ike participants in a 12-step program, surgeons stand once a week, acknowledge their shortcomings, are affirmed by the community, move on, and attempt to do better. The default position of the Surgical Morbidity and Mortality Conference is that something about the care rendered could have been done better than it was. C. Walton Lillehei, MD has noted that the mature surgeon’s good judgment comes from experience, and that experience comes from modifications of clinical practice derived from prior imperfect judgment.1,2 Though on occasion a complication is attributed to course of disease, monographs on the introspection of surgeons note that they forgive yet remember, and attempt to internalize and process the lessons learned from judgment errors and technical mishaps; this is their primary mechanism for improving patient care.3,4 In this spirit, it has been my pleasure to call on the authors in this issue to discuss management of common surgical complications. Each brings to the task a sense of humility and empathy with the operating surgeon. Each hopes that a technique or temporizing measure might help the reader to conduct an operation more safely. Thomas Stahl, MD is a colorectal surgeon in academic practice in Washington, DC. I posed a series of questions to him as to how to react to EEA misfire in the conduct of low anterior anastomoses. He correctly noted the absence of consensus for several of the questions and has formulated a response rooted in experience-based pragmatism to the typical oral board type question, “Tom, what would you do?” His thoughtful reflections are those of a technically gifted surgeon considering the management of this difficult intraoperative circumstance. John Mannick, MD, for whom I worked as a resident, once told me that a surgeon could be distinguished from a nonsurgeon by simply asking him one question. The question was, “What do you fear most—venous bleeding or arterial bleeding?” The answer, as every surgeon knows, is bleeding from a major vein (Mannick JM, personal communication, 1983). While Jennifer Tseng, MD, of The University of Massachusetts, and her colleagues have correctly observed that the cardinal rule for how to handle venous bleeding is never to get into it in the first place, they present an elegant anatomical map for how to do so. They describe the confluence of the portal venous tributaries as they coalesce under the surgical neck of the pancreas, as well as maneuvers used to control blood loss from veins in that region. Variations in the venous anatomy in this region are described, as are temporizing measures and techniques of obtaining the operative 140


exposure necessary to definitively deliver oneself from catastrophic blood loss in this region. Horst Filtzer, MD has recently left his position at Harvard Medical School to take a clinical position in Arizona. As with many of our contributors, he has often served as the secondary or salvage surgeon to a younger and grateful colleague who has learned definitive control of hemorrhage or damage control under his supervision. Few surgeons are more worthy of the name general surgeon. An innovative surgeon in his own right and one of the first in metropolitan Boston to adopt minimally invasive techniques for both general and vascular surgery, Dr. Filtzer is at home in any anatomic space or body cavity. In this issue, he considers the difficult topic of pelvic hemorrhage and the options available to control it. Paul LaFontaine, MD, a urologist in affiliated academic practice in Cambridge, MA, addresses with extraordinary completeness and economy of syntax the difficult topic of ureteral reconstruction during volitional sacrifice of the ureter or repair after accidental injury. The five major reconstructive options are reviewed in a systematic way; as it is typical for this publication, Rob Gordon’s drawings beautifully illustrate the salvage techniques Dr. LaFontaine describes. The career of Theodore Pappas, MD at Duke began at roughly the same time as surgical laparoscopy for biliary disease. In his current role as a senior surgeon at that referral center, he has had the opportunity to perform the reconstructive salvage surgery on bile ducts divided at all levels in the porta hepatis and to work out how to alter that reconstruction based on the degree to which the sentinel operation devascularized these structures. His systematic approach to these injuries, distinguishing those that can be repaired primarily from those requiring hepato-jejunal anastomosis, is documented in his contribution to this issue, the latest in a long list of sequential publications on the substantial Duke experience with this problem. Matthew Hutter, MD and colleagues have asserted that the competence of the general surgeon can be best inferred not by the cases logged in specific categories, but the ability to invoke and assimilate basic general surgical principles into the care of individual patients.5 The contributions to this issue have derived their wisdom from work in the field, and their enthusiasm from a desire to help both their patients and their colleagues. Like all good surgeons, they were flattered to be asked to help, took their charge seriously, and know full well that all operating surgeons inevitably face complications of the surgical illness or of their approach to it. This principle

Nauta

141

has been articulated in many ways, such as “The only surgeon with no complications is the one who does not operate,” or “There but for the grace of God go I.” In addition, although it may be true that surgeons cannot always control the clinical cards they are dealt, I believe that these authors’ contributions to this issue will help us all to play our hand more wisely. I am grateful to Walter Koltun, MD for asking me to edit the issue and to these surgeons for sharing their thoughts. Russell J. Nauta, MD Guest Editor

References 1. Lillehei CW: New ideas and their acceptance as it has related to preservation of chordae tendinae and certain other discoveries. J Heart Valve Dis 4:S106-S114, 1995 (suppl 2) 2. Toledo-Pereyra LH: Innovation according to C. Walton Lillehei. J Invest Surg 20:205-209, 2007 3. Bosk CL: Forgive and Remember: Managing Medical Failure (2nd ed.) Chicago: University of Chicago Press; 2003 4. Gordon L: Gordon’s Guide to the Surgical Morbidity and Mortality Conference. Philadelphia: Hanley & Belfus Inc; 1994 5. Hutter M, Kellogg C, Ferguson C, Abbott M, Warshaw AL: The resident impact of the 80 hour resident workweek on surgical residents and attending surgeons. Ann Surg 243:864-875, 2006

Technical Considerations in the Difficult Colorectal Anastomosis Thomas J. Stahl, MD, FACS, FASCRS

T

here is nothing that provokes greater anxiety and consternation to the gastrointestinal surgeon than the prospect of a leak from a colorectal anastomosis. The consequences to the patient from such a complication can be significant and not infrequently life threatening. The leak rate is progressively greater as the level of the anastomosis proceeds from the upper to distal rectum, with the greatest leak rate occurring for anastomoses at or below 7 cm from the anal verge. The collective experience of general and colorectal surgeons over the years has reduced this leak rate significantly, aided in no small part by ever improving circular stapling devices. Nevertheless, recently published studies demonstrate an anastomotic leak rate of 3% to 15%, with a corresponding mortality of 2% to 7%. The surgical goals are both to prevent such a complication, and to minimize the adverse impact on the patient if a leak occurs.

General Considerations in Anastomotic Construction Up until the 1970s, the limits of a colorectal anastomosis were determined by how low a suture line could be visualized and hand-sewn. This often precluded a suture line below 8 to 10 cm from the anal verge, with resections below this level requiring abdominoperineal resection with permanent colostomy. With the emergence of the circular stapling device, that limit is now the pelvic floor. Experience with anastomoses below 8 to 10 cm has shown that these are uniquely treacherous reconstructions with the highest propensity for complications. Because the premise for this type of surgery is often the surgical treatment of rectal cancer, the ever-increasing use of preoperative chemoradiation introduces an additional healing hazard to the equation. A surgeon can only control some of the many variables in anastomotic construction. The fundamental principles of

Department of Surgery, Georgetown University Medical Center, Washington Hospital Center, Washington, DC. Address reprint requests to Thomas Stahl, MD, FACS, FASCRS, Assistant Clinical Professor, Department of Surgery, Georgetown University Medical Center, Washington Hospital Center, 106 Irving Street, NW, Suite 2100N, Washington, DC 20010. E-mail: thomas.j.stahl@medstar. net

142


preservation of an adequate blood supply, total absence of tension on the suture line, healthy bowel for both the proximal and distal ends without thickening or inflammation, and adequate bowel preparation have remained constant. The necessity of bowel preparation is, however, now a topic of considerable debate, and its necessity is being actively evaluated. The technical requirements include the creation of an airtight suture line, in some circumstances protected by a proximal diverting procedure, and/or omental wrap. Whether the anastomosis is hand-sewn in one or two layers, performed with interrupted or running suture technique, or constructed with a stapling device has no impact on leak rates. Factors often beyond the surgeon’s control are immutable comorbidities and the patient’s body habitus. Some of the more prominent patient risk factors that correspond to a higher than average risk of a leak are a history of smoking, excessive alcohol consumption, malnutrition, chronic steroid use, chronic obstructive pulmonary disease, bowel obstruction at the time of surgery, and male gender. A narrow pelvis and the presence of obesity in a man can present a maddeningly difficult technical challenge, requiring a frank preoperative discussion with the patient as to the limitations this creates.

Anastomotic Techniques For colorectal or coloanal anastomoses, there are essentially six technical alternatives: 1. 2. 3. 4. 5. 6.

Hand-sewn end to end (single or double layered) Circular stapled end to end Circular stapled side to end Hand sewn coloanal Circular stapled coloanal Circular stapled with colonic J-pouch

The choice of a given anastomosis is to some degree at the discretion and preference of the surgeon, but also based on the level of transection of the rectum and the anatomy of the patient. Most anastomoses above 7 to 8 cm can be either hand sewn or stapled (Figs 1 and 2). In patients with a fat mesentery and/or short sigmoid vasculature, the side to end technique can provide a more com-

Difficult colorectal anastomosis

143

Figure 1 Hand-sewn anastomosis between the transected proximal colon and the rectum after sigmoid colectomy.

fortable reach of the sigmoid colon into the deep pelvis, and can reduce the amount of mesenteric mobilization required (Fig 3). It also allows the anastomosis to be made through the antimesenteric border of the sigmoid colon, which in an obese patient may be the only area free of significant attached fatty tissue. The suture line should be set up to be sure the distal aspect of the anastomosis is 1 cm. proximal to the closed sigmoid end. For the coloanal anastomosis, the dissection to remove the rectum is the same whether the anastomosis is hand-sewn or stapled. The hand-sewn alternative requires a transanal approach for placing the sutures, and is usually performed in a single layer (Fig 4). The stapled alternative can either be a straight coloanal approach, or incorporate the use of a small colonic J-pouch (Fig 5A and B). Although most studies demonstrate a convergence of function between the straight coloanal and J-pouch 1 year postoperatively, the latter usually provides better immediate bowel function. There are other decisions the surgeon has to make once the anastomosis has been constructed: Whether to perform a temporary diverting stoma and what type of stoma this should be, the placement of drains, and the reinforcement of the anastomosis with an omental flap. There is a significant consensus that routine diversion is indicated in all coloanal anastomoses, as these carry the highest leak rate (10-15%). Since these low anastomoses can very rarely be revised or redone, a coloanal reconstruction is essentially a “one shot” endeavor; all reasonable measures to protect it are warranted. Most surgeons would also agree that temporary diversion is reasonable for any anastomosis at or below 7 to 8 cm, or when preoperative radiation therapy was administered. Diversion can be accomplished with either a proximal loop colostomy or ileostomy. A temporary loop ileostomy has the

advantage of being easily delivered through the abdominal wall, and being straightforward to close, usually through a peri-stomal approach. This saves the patient from having a second major laparotomy incision, and is typically a brief procedure with a short hospital stay. Several recent studies have also demonstrated that a loop ileostomy accomplishes essentially 100% fecal diversion. The long length of defunctionalized colon have never been shown to generate an adverse outcome. Although diversion has not been shown to diminish the rate of anastomotic leak, the consequences of a leak are significantly mitigated by proximal diversion. The use of pelvic drains has been shown in most studies to either have no benefit or to actually increase the incidence of leaks. Nevertheless, most surgeons are understandably uncomfortable with the notion of significant fluid and debris accumulation in the dependent pelvis after a very low or coloanal reconstruction, and the specific circumstances of the individual procedure should determine whether drains are used or not. If the operation is technically difficult and/or generates significant blood loss or oozing, a drain may be reasonably used. There is probably no justification, however, for the use of drains in an uneventful procedure where the anastomosis is above 7 or 8 cm. Employing an omental wrap around an anastomosis in the pelvis carries a common sense appeal, provided there is sufficient omentum to reach the pelvis. This is usually accomplished by creating an omental pedicle off of either the right or left gastroepiploic arcade (Fig 6). The limited clinical studies assessing this maneuver have shown results similar to what is accomplished with a proximal diversion; the leak rate is not significantly reduced, but the omental wrap confines the leak to the pelvis and diminishes the likelihood of a more widespread peritonitis.

T.J. Stahl

144

Figure 2 End-to-end anastomosis (EEA) with the circular stapler, joining the transected proximal colon to the rectum after sigmoid colectomy.


Figure 3 Anastomosis joining the side of the transected colon to the end of the transected rectum after sigmoid colectomy.

145

T.J. Stahl

146

Figure 4 Single layer transanal anastomosis of colon to anus.

Intraoperative Anastomotic Difficulties and Their Management Intraoperative Assessment of the Integrity of an Anastomosis After an anastomosis has been performed, it can be visually inspected. The security of the suture line can be assessed in a variety of ways. If there is no tension and the bowel appears healthy and viable, the suture line can be assessed by air or fluid insufflation, proctoscopic visualization, or digital examination in suture lines below 6 cm. Most surgeons employ one or more of these techniques. Proctoscopic visualization has to be done very carefully to minimize mechanical stress or injury to the anastomosis; only a fairly large defect in the suture line will be visualized with a proctoscope. Air insufflation is a very sensitive method of testing the integrity of a suture line (Fig 7). This is performed by filling the pelvis with sufficient saline to cover the suture line, gently occluding the bowel with the fingers just above the anastomosis, and then insufflating air transanally until the anastomosis is distended. If no air bubbles are seen to extravasate, this correlates very strongly with a low subsequent anastomotic leak rate. If the surgeon wishes to visually examine the anastomosis,

air can then be insufflated through the proctoscope to achieve the same assessment. This is a very simple and useful precautionary technique, and is strongly encouraged after the completion of any pelvic anastomosis. If air bubbles are seen to escape during this maneuver, then the suture line has some measure of disruption. If the anastomosis can be visualized, the location of this leak can often be identified. If the location of the leak can be seen and is accessible, then interrupted suture reinforcement and closure of the defect is required (Fig 8). If a second air leak test is normal, and there are no other concerns, then this may be adequate with no need for further precautions. If attempted repair does not control the leakage of air, and there is no major disruption of the suture line that can be seen, then a proximal diverting loop ileostomy is required. An omental wrap could be added, but probably not as an alternative to proximal diversion. For suture lines that are very low in the pelvis that are not accessible and cannot be visualized, the presence of an air leak mandates diversion. The air leak test for coloanal anastomoses is unnecessary as they usually cannot be revised, and are routinely diverted. A digital examination that reveals a circumferentially intact suture line is sufficient. If possible, a palpable defect in the suture line for a coloanal anastomosis should be transanally repaired.


Figure 5 (A) J pouch coloanal anastomosis. The linear stapler is used to create a reservoir in the previously mobilized colon. (B) J pouch coloanal anastomosis. On removal of the linear stapler, the circular stapler is introduced through the rectum and into the colostomy, where a side-to-end anastomosis is created between colon and rectum.

147

T.J. Stahl

148

Figure 6 Omental wrap of low pelvic anastomosis.


149

Figure 7 Air is insufflated into the rectum after anastomosis of the colon to the rectum after the pelvis is filled with saline. Bubbles indicate the lack of integrity of the anastomosis and the need for repair or revision.

Incomplete Anastomotic Rings in a Circular Stapled Anastomosis The phenomenon of an incomplete or missing donut after extraction of a circular stapler is one of the great mysteries of surgery. I have encountered this conundrum on several occasions, and each time have been unable to ascertain what, if any, significance this occurrence has. Occasionally, the distal donut is imbedded in the proximal portion of the stapler (Fig 9). An incomplete donut may also be dislodged when the stapler is being withdrawn, and can subsequently be identified and retrieved with the proctoscope. More often than not, however, neither of these explanations applies. This is a circumstance where proctoscopic visualization of the suture line is probably wise, in conjunction with air testing of the anastomosis. If neither reveal a problem, then no further precautions seem required. Nevertheless, in this situation, no surgeon would likely be criticized for oversewing the staple line, or performing a temporary diverting ileostomy. Unfortunately, I know of no literature that sufficiently addresses this specific problem.

Difficult Circular Stapler Extraction with Anastomotic Injury Very few circumstances generate greater frustration than the difficult extraction of a circular stapler, particularly if the anastomosis is deep in the pelvis and unreachable by the surgeon. In a more proximal suture line, the withdrawal of the stapler can sometimes be aided by milking the bowel off of the device with simultaneous gentle pulling from below. It is incumbent on the surgeon to be completely familiar with the proper use of the devices, however; not infrequently these problems occur with failure to do so. This is particularly true if the surgeon is using

a stapler from a new vendor, or a new version of a previously familiar device. The appropriate circular stapling diameter must also be chosen. Circular sizers should always be available, and a stapler should never be chosen that has a significantly larger diameter than a sizer that passes comfortably into the proximal bowel. If a difficult stapler extraction is encountered, then the above maneuvers to assess the staple line must be employed. If the suture line is able to be visualized, and there is a visible disruption, then it must be adequately oversewn and air tested. If more than 25% of the staple line has been disrupted, then the anastomosis should be taken down and redone. If the anastomosis is deep in the pelvis and cannot be reached or repaired, then proximal diversion is mandatory. The additional placement of an omental wrap into the pelvis would also be reasonable.

Conclusions It is easy to forget sometimes just how far we have come as surgeons in the ability to perform complicated pelvic gastrointestinal procedures. The ever diminishing need for complete removal of the rectum, and the impressively low rates of significant complications are enormous testimony to our efforts. These goals have been accomplished by acquiring the techniques and maneuvers itemized above, and with the skillful use of ever improving stapling devices. However, all of the available techniques and technology are for naught if they are not used. It is incumbent on the surgeon who takes on these challenging rectal cases to be fully informed of all appropriate precautions, evolving surgical techniques, and new technologies. If accomplished, the preservation of function with minimal harm to the patient will be increasingly realized.

T.J. Stahl

150

Figure 8 Inspection or insufflation may reveal that the stapling device has incompletely apposed the colon and the rectum, indicating the need for revision or repair. Suture repair of small defects may be accomplished if such a lesion can be adequately visualized and if the region has not been devascularized.


151

Figure 9 Inspection of the anastomotic “donuts.” Concentric staple lines and clean division between them should impale complete anastomotic “donuts” on the anvil of the circular stapling device. The verification that the circular pieces of tissue are intact implies that the tissue was well apposed and transected in the appropriate place. Incomplete donuts do not always imply lack of circumferential anastomotic integrity, but suggest the need for careful testing and inspection of the anastomosis.

Suggested Reading Bokey EL, Chapuis DS, Fung C, et al: Postoperative morbidity and mortality following resection of the colon and rectum for cancer. Dis Colon Rectum 38:480-487; 1995 Cavaliere F, Pemberton JH, Cosimelli M, et al: Coloanal anastomosis for rectal cancer: Long-term results at the Mayo and Cleveland Clinics. Dis Colon & Rectum 38:807-812, 1995 Hedrick TL, Sawyer RG, Foley EF, et al: Anastomotic leak and the loop ileostomy: Friend or foe? Dis Colon & Rectum 49:1167-1176, 2006 Hirsch CJ, Gingold BS, Wallack MK: Avoidance of anastomotic complications in low anterior resection of the rectum. Dis Colon & Rectum 40:42-46, 1997 Machado M, Hallbook O, Goldman S, et al: Defunctioning stoma in low

anterior resection with colonic pouch for rectal cancer. Dis Colon & Rectum 45:940-945, 2002 Marusch F, Koch A, Schmidt U, et al: Value of a protective stoma in low anterior resections for rectal cancer. Dis Colon & Rectum 45:1164-1171, 2002 Platell C, Barwood N, Dorfmann G, et al: The incidence of anastomotic leaks in patients undergoing colorectal surgery. Colorectal Disease 9:71-79, 2006 Pollard CW, Nivatvongs S, Rojanasakul A, et al: Carcinoma of the rectum: Profiles of intraoperative and early postoperative complications. Dis Colon & Rectum 37:866-874, 1994 Tocchi A, Mazzoni G, Lepre L, et al: Prospective evaluation of omentoplasty in preventing leakage of colorectal anastomosis. Dis Colon & Rectum 43:951-955, 2000 Wexner SD, Taranow DA, Johansen OB, et al: Loop ileostomy is a safe option for fecal diversion. Dis Colon & Rectum 36:349-354, 1993

Control of Bleeding from the Portal/Superior Mesenteric Vein Joshua S. Hill, MD,* Sridhar Shankar, MD,† Douglas B. Evans, MD,‡ and Jennifer F. Tseng, MD, MPH*

I

ntraoperative injuries to the portal vein (PV), superior mesenteric vein (SMV), and tributaries including the inferior mesenteric vein (IMV), splenic vein, and first jejunal branch of the SMV pose significant challenges to the safe and complete surgical removal of part or all of the pancreas. Major hemorrhage from these vessels can lead to patient morbidity or mortality. The PV and SMV course posterior to the pancreas limiting their exposure. Frequent variations in the branching pattern of the SMV have been described.1 Avulsion of the portal vein during pancreatic resection because of adherence of the vein is the most common cause of significant intraoperative hemorrhage. The best “treatment” of intraoperative injuries to the PV and SMV is prevention. With careful preoperative evaluation of the major vessels’ relationship to the primary pathology, careful planning, including assessment for possible vascular involvement and/or resection, can usually avoid or minimize massive intraoperative hemorrhage. If uncontrolled bleeding should occur despite adequate preoperative preparation, then prompt action must be taken because of the substantial blood flow present in the portal and mesenteric veins. Classical principles of hemostasis such as obtaining proximal and distal control apply where feasible, but there are also regionspecific techniques that can be employed.

Anatomy The portal vein is formed by the confluence of the splenic and the superior mesenteric veins posterior to the neck of the pancreas and anterior to the inferior vena cava (Fig 1). Coursing through the lesser omentum toward the liver in a cephalad and oblique angle, the PV travels behind the first portion

*Department of Surgery, Surgical Outcomes Analysis & Research, University of Massachusetts Medical School, Worcester, MA. †Department of Radiology, University of Massachusetts Medical School, Worcester, MA. ‡Department of Surgical Oncology, The University of Texas M.D. Anderson Cancer Center, Houston, TX. Address reprint requests to Jennifer Tseng, MD, MPH, Assistant Professor of Surgery, Director, Surgical Outcomes Analysis & Research, University of Massachusetts Medical School, UMass Memorial Medical Center, 119 Belmont Street, Swift House, Worcester, MA 01605. E-mail: tsengj@ ummhc.org

152


and medial to the second portion of the duodenum. Here the PV is the posterior most structure of the portal triad (hepatoduodenal ligament). Other structures that compose the portal triad are the common bile duct (anterolaterally) and the hepatic artery (anteromedially). After coursing cephalad approximately 8 cm, the portal vein enters the hilum of the liver. Here it divides into left and right branches supplying Couinaud’s segments II-IV and V-VIII respectively. The right branch receives blood from the cystic vein and enters into the liver parenchyma. This branch then bifurcates into the right anterior portal vein and the right posterior portal vein. The left portal vein joins the ligamentum teres and venosum before entering the liver where it provides medial branches to segment IV and distal branches to segments II and III. Although this is the usual course of the portal vein, 35% of patients will show some variation in branching.1 The anatomy of the SMV and its tributaries can also vary. Angiographic investigations have demonstrated that approximately one half of patients will have a short SMV trunk of variable length that then divides into a jejunal and ileal branch. The length of the common SMV trunk typically ranges from 5 to 50 mm. In those patients without a so-called SMV main trunk the jejunal and ileal branches drain directly into the splenic vein at a common confluence.2 The clinical correlate of this is that what appears to the surgeon as the SMV is actually the ileal branch, and the jejunal branch enters posterior, close to the level of the splenic vein. Failure to identify this high-entering first jejunal branch can lead to continued hemorrhage after apparent proximal and distal control of the portal vein and SMV. The first jejunal branch drains the duodenojenunal junction and the proximal jejunum. Its close proximity with the uncinate process of the pancreas predisposes this vessel to tumor invasion near its confluence with the SMV making it an important consideration during pancreatic tumor resection.3 The proximal course of this branch is highly variable and depends primarily on the length of the SMV main trunk.2 Kim and co-workers (2006) demonstrated, using computed tomography, that the first jejunal branch was found anterior to the SMA in 19% of patients, a variant which provides easier exposure of the vein. In 83% of patients, the tributaries of the first jejunal branch coursed to the patient’s left side; however, in 17% of patients these turned toward the patient’s right.4 The first major branch from the superior mesenteric vein

Control of bleeding from the P/SMV

153

Figure 1 Location of the PV/SMV/splenic vein confluence behind the neck of the pancreas. (© 2004 The University of Texas M.D. Anderson Cancer Center.) (Color version of figure is available online.)

system is the gastrocolic trunk. Depending on the length of the length of the SMV main trunk, the gastrocolic trunk may drain into the SMV or the 1st jejunal branch. It is present in 70% to 87% of the population and is formed by the confluence of the right gastroepiploic vein and anterior superior pancreaticoduodenal vein.2,5 Occasionally, either the right colic or the middle colic veins form a component of the gastrocolic trunk.6,7 An autopsy study performed by Ignjatovic and co-workers (2004) found that the mean distance between the gastrocolic trunk and the inferior boarder of the neck of the pancreas (as the uncinate process extends caudal to the GC trunk) to be 2.2 cm with an average total length of 16.1 mm. The mean caliber of the vessel was 5.2 mm.

Preoperative Evaluation The ease and effectiveness of pancreatic resections relate to the gland’s relationship to the portal and/or superior mesenteric vein and the principles outlined below for pancreatic resection can be applied to other operations done in the region. Because surgical resection remains the only treatment modality for pancreatic cancers that offers the possibility of a cure, patients who are appropriate candidates should be considered for resection.8 However, pancreatic resection is associated with significant morbidity and mortality. Therefore, surgeons must limit the application of pancreatic resection to only those patients for whom it will benefit. The newest revision of the American Joint Commision on Cancer (AJCC) staging guideline defines unresectable pancreatic tumors as those that invade into the SMA.9 Evaluation of the SMA on preoperative imaging therefore is imperative as it will limit unwarranted resections.

At pancreaticoduodenectomy, intraoperative palpation of the SMA is an insensitive measure of tumor spread. Even in experienced hands, it is possible to mistakenly believe that vascular involvement is absent when in fact tumor invaded the SMA (Fig 2). Moreover, the majority of injuries to the portal vein during pancreatic resection occur when attempting to dissect an adherent pancreas from the portal vein. Thus, preoperative imaging modalities should be utilized to assess the extent of vascular invasion. Various algorithms have been developed, including those that utilize computed tomography (CT), CT with angiography, magnetic resonance imaging (MRI), and/or endoscopic ultrasound (EUS).10-14 In general, there is a consensus that the best modality for imaging is high-quality thin-slice CT scanning with arterial, venous, delayed and noncontrast phases (“pancreatic protocol”).15,16

Operative Technique Bleeding from the PV or the SMV most commonly occurs when attempts are made to dissect adherent tumor and/or pancreas from the PV-SMV confluence causing an avulsion of the vein because of traction. In a series of seven patients with PV injuries reported by Oderich and co-workers (2004), six of these injuries occurred during PD and one injury occurred during a takedown of an enterocutaneous fistula. There were four partial lacerations and three complete transections of the portal vein all of which led to significant blood loss. Three patients were repaired through ePTFE interposition graft, three by lateral venorraphy and one by end-to-end anastomosis. There was a 28% mortality rate in this series, with one patient expiring in the operating room because of uncontrollable hemorrhage.17 Thus, prevention of portal venous hem-

J.S. Hill et al

154

Figure 2 (A) Intraoperative palpation of SMA is not accurate for determining SMA involvement, and does not address involvement of SMV/PV; thus, the need for careful preoperative assessment.18 (© 2004 The University of Texas M.D. Anderson Cancer Center.) (B) Angio-CT (arterial phase) showing resectable tumor; normal SMA as evidenced by the presence of the SMA/pancreas fat plane. No venous involvement. (C) Angio-CT (venous phase) showing tumor involvement of SMV. (Color version of figure is available online.)

orrhage remains important once in the operating room. Systematic stepwise performance of PD can aid in the careful controlled exposure of the portal vein, SMV, and SMA, as has been extensively described18: a brief description of critical steps follows. First the surgeon should proceed with a right medial visceral rotation (Cattell-Braasch maneuver). This maneuver allows for identification of the superior mesenteric vein. At this time, the middle colic vein should be ligated before its junction with the SMV to prevent iatrogenic traction injury. A Kocher maneuver will then allow for further dissection of the porta hepatis with exposure of the portal vein after ligation of the common bile duct and the gastroduodenal artery. Next,

the bowel and stomach should be transected. Finally, transection of the pancreas occurs with special attention paid to the presence of adherent tumor. Careful and early exposure of the relevant vascular anatomy is critical in the prevention of hemorrhage. The steps outlined above allow for proper visualization of the PV, SMV and first jejunal branch while minimizing the risk of injury to these structures. If unexpected early bleeding is encountered, the above principles still apply. If a full Kocher maneuver has not been performed, the head of the pancreas and duodenum need to be expeditiously freed from the retroperitoneum to allow for more effective compression. The stomach should then be divided with fires of a linear cutting stapler or sharply


155

Figure 3 The stomach and the pancreas divided as in pancreaticoduodenectomy. The portal vein/SMV is being gently retracted to the patient’s left, revealing small branches from the head of the pancreas to the portal vein.18 (© 2004 The University of Texas M.D. Anderson Cancer Center.) (Color version of figure is available online.)

between clamps, and the neck of the pancreas divided carefully with sharp dissection or electrocautery providing full exposure of the portal vein (Fig 3). In difficult cases, which should be ascertained preoperatively based on patient history, examination and imaging, it may be prudent to (1) identify the SMA early, and (2) dissect the PV and SMV medially (before removal of the specimen) as well as laterally, thus obtaining proximal and distal venous control, as well as control of arterial inflow if needed. As described above, the first jejunal branch of the SMV supplies the proximal jejunum, traveling posterior to the superior mesenteric artery (SMA) and entering the SMV along its posterolateral wall. The first jejunal branch usually gives off 1 or 2 branches directly to the uncinate process. If necessary, the first jejunal branch can be divided. Venous congestion of the bowel is unlikely to occur if only the first jejunal branch is ligated. However, if additional venous structures (such as the ileal branch) are concurrently ligated, then reconstruction of the venous system is necessary.3 The most feared complication during this part of the procedure is a tangential laceration of the first jejunal branch of the SMV extending posterior to the SMA. In an attempt to control such an injury, poorly placed sutures may result in an injury to the

SMA. Dissection of the distal SMV at the level of the uncinate process, and specifically, dissection of the first jejunal branch requires proper intraoperative education and an advanced level of experience. If portal vein hemorrhage is encountered at operation, surgical compression is indicated. This allows for the assessment of the patient’s general condition and provides time for the anesthesiologist to appropriately resuscitate the patient, obtain blood products and assistance. The hand of the senior surgeon in the room is frequently the best instrument for this purpose. If that surgeon is on the patient’s left side, the surgeon’s left hand (thumb anterior, fingers wrapped posterior) can compress the mesenteric inflow as well as the venous tear (Fig 4). Consultation with vascular surgical or experienced upper gastrointestinal surgeons is highly recommended for additional surgical expertise. With great care, if the hemorrhage cannot be controlled by the above maneuvers, the area of injury can be exposed, with proximal and distal compression (with a peanut or sponge-stick, or the surgeon’s opposed thumb and fingers). Immediate steps must be taken to establish inflow and outflow control, including the portal vein, superior mesenteric vein, and splenic vein. First the surgeon’s left hand should be placed posterior to the portal vein

156

J.S. Hill et al

Figure 4 (A) Manual compression of the SMV. (B) Using sponge-sticks to provide proximal and distal control during repair with Judd-Allis clamps. (C) Occlusion of SMV to repair lacerations of the 1st jejunal branch. (D) Primary closure and patch angioplasty of longitudinal injuries. (E) The use of occlusion catheters may aid in the maintenance of hemostasis during repair.


157

Figure 5 Methods for resection of a laceration that involves the portal-splenic vein confluence. These include primary repair, patch angiopastly, interposition grafts and ligation of the splenic vein. (© 2004 The University of Texas M.D. Anderson Cancer Center.) (Color version of figure is available online.)

allowing for compression of the vein against the firm neoplasm as described above. If this technique does not sufficiently reduce the hemorrhage, then inflow occlusion of arterial inflow, in the form of rapid identification of the SMA and clamping with an atraumatic vascular clamp such as a small bulldog clamp should occur. A small laceration may be immediately controlled with sutures (eg, 5-0 prolene). If the laceration is too large to be immediately sutured, but is easily visualized, placing sutures at the corners of the laceration can allow for gentle upward traction and subsequent placement of accurate repair stitches. Atraumatic clamps can be useful to approximate the venotomy, although great care must be taken not to extend the tear further, including spiraling posteriorly or extending the tear under the body of the pancreas. If a large avulsion is identified, the surgeon should prepare for the possibility of portal vein resection with interposition graft or patch angio-

plasty. The techniques for this have been previously described (Fig 5).3,19 Primary ligation of the portal vein should be avoided as this is associated with high mortality. The splenic vein can be ligated if necessary, although this can be associated with postoperative splenic vein hypertension and gastric varices.20

Trauma A brief discussion of the management of traumatic injuries to the PV/SMV may shed some light on the management of iatrogenic injuries. Patients with trauma involving the portal vein and SMV are at high risk for serious morbidity and mortality with a reported mortality between 39% and 71%.21 Most deaths are caused by uncontrolled bleeding during the peri-operative period. Asenio and co-workers (2007) demonstrated in a review of 51 patients over a 13-year period that

J.S. Hill et al

158 75% of patients with SMV injuries were admitted secondary to penetrating gunshot wounds. Traumatic portal venous injuries are most commonly associated with damage to surrounding structures, including the inferior vena cava and hepatic arteries. This must be considered during laparotomy for trauma.22 The surgeon must prepare for all forms of venous injuries at the time of operation. Adequate preparation includes readying the blood bank for possible large volume resuscitation and arranging for an auto-transfusion machine. Furthermore, instruments used for vascular control should be made available. Inflow occlusion may be performed by manual compression of the supra-celiac and infra-renal aorta. This limits blood flow through the intestine and thus the portal vein. Great care must be taken not to completely occlude the blood flow to the intestine for extended periods of time as this can lead to ischemia of the entire midgut. Stable retroperitoneal hematomas may be present in up to one half of all patients with portal venous injuries. Associated vena caval injuries are the rule, rather than the exception, when retroperitoneal hematoma is identified.21 Significant bleeding will occur once a hematoma is unroofed, leading to rapid deterioration in the patient. Therefore, adequate exposure using a wide Kocher maneuver should be performed. This will allow for the compression and vascular control of the portal vein by the surgical assistant while other venous injuries are investigated. Bleeding from the IVC must be addressed before addressing any bleeding from the portal vein. Specific methods to repair portal venous injuries depend on the location of the injury. The supra-pancreatic portal vein is relatively accessible, allowing for direct visualization of the vein and for the use of interposition grafts in its repair. Attempts to perform primary anastomoses at this location may worsen exposure and should generally be avoided. To further aid in exposure, proximal and distal control can be obtained using vascular occlusion catheters instead of vascular clamps (Fig 4).21 Exposure is more significantly limited when the injury occurs behind the pancreas. This area is difficult to approach because of the adherence of the portal vein to the pancreas and because of the multiple venous tributaries present in this area. Therefore, lateral venorraphy has been described as the preferred technique when managing injuries in this location. The outcomes reported after portal vein ligation are varied ranging from 30% to 90% mortality. Stone and co-workers (1982) described a large series of patients spanning 22 years who either underwent portal vein ligation as a salvage procedure or as a primary method of venous control. When performed as a salvage procedure, the survival rate was 13%. When ligation was performed as the primary method of repair, the survival was 80%.23 The authors conclude that the difference in mortality was because of prolonged operative time leading to lengthened periods of under-resuscitation in the ligation as a salvage therapy group. Jurkovich and coworkers (1995) described 10 patients who underwent portal vein ligation after trauma; 90% of these patients subsequently died.24 Asensio et al. compared primary repair of the SMV to ligation and found improved survival with primary repair (63% vs. 40%).25 One important consideration is that ligation of the portal vein cannot be performed if the hepatic artery is also injured. If ligation were to be performed in this

case, then the liver would lack inflow and the patient would quickly expire.

Conclusion In conclusion, the best strategy for controlling bleeding from the portal and superior mesenteric veins is first and foremost the prevention of this disastrous complication through thorough preoperative planning and meticulous operative technique. If hemorrhage from the portal vein or SMV is encountered then an expeditious determination of the source of bleeding must occur. Through a stepwise approach to pancreatic resection, adequate exposure, careful visualization of vascular structures, and attention to the general principles of vascular control, intraoperative hemorrhage from the PV/ SMV complex can be largely avoided, and if such hemorrhage should occur, be more effectively treated.

Acknowledgment This work was supported by a Howard Hughes Medical Institute Early Career Grant and by the Pancreatic Cancer Alliance (J.F.T.).

References 1. Covey AM, Brody LA, Getrajdman GI, et al: Incidence, patterns, and clinical relevance of variant portal vein anatomy. AJR Am J Roentgenol 183:1055-1064, 2004 2. Graf O, Boland GW, Kaufman JA, et al: Anatomic variants of mesenteric veins: Depiction with helical CT venography. AJR Am J Roentgenol 168:1209-1213, 1997 3. Tseng JF, Tamm EP, Lee JE, et al: Venous resection in pancreatic cancer surgery. Baillieres Best Pract Res Clin Gastroenterol 20:349-364, 2006 4. Kim HJ, Young TK, Joo WL, et al: Radiologic anatomy of the superior mesenteric vein and branching patterns of the first jejunal trunk: Evaluation using multi-detector row CT venography. Surg Radiol Anat 29: 67-75, 2007 5. Ignjatovic D, Stimec B, Finjord T, et al: Venous anatomy of the right colon: Three-dimensional topographic mapping of the gastrocolic trunk of henle. Tech Coloproctol 8:19-21, 2004 6. Yamaguchi S, Kuroyanagi H, Milsom JW, et al: Venous anatomy of the right colon: Precise structure of the major veins and gastrocolic trunk in 58 cadavers. Dis Colon Rectum 45:1337-1340, 2002 7. Ito K, Blasbalg R, Hussain SM, et al: Portal vein and its tributaries: Evaluation with thin-section three-dimensional contrast-enhanced dynamic fat-suppressed MR imaging. Radiology 215:381-386, 2000 8. Tsuchiya R, Noda T, Harada N, et al: Collective review of small carcinomas of the pancreas. Ann Surg 203:77-81, 1986 9. Green F, Page D, Irvin D, et al: AJCC Cancer Staging Manual, 6th ed. New York, Springer, 2002 10. Clarke DL, Thomson SR, Madiba TE, et al: Preoperative imaging of pancreatic cancer: A management-oriented approach. J Am Coll Surg 196:119-129, 2003 11. Lu DS, Reber HA, Krasny RM, et al: Local staging of pancreatic cancer: Criteria for unresectability of major vessels as revealed by pancreaticphase, thin-section helical CT. AJR Am J Roentgenol 168:1439-1443, 1997 12. Shoup M, Hodul P, Aranha GV, et al: Defining a role for endoscopic ultrasound in staging periampullary tumors. Am J Surg 179:453-456, 2000 13. DeWitt J, Devereaux B, Chriswell M, et al: Comparison of endoscopic ultrasonography and multidetector computed tomography for detecting and staging pancreatic cancer. See summary for patients in Ann Intern Med 16;141, 2004; I46; PMID: 15545671. Ann Intern Med 141:753-763, 2004 14. Soriano A, Castells A, Ayuso C, et al: Preoperative staging and tumor resectability assessment of pancreatic cancer: Prospective study comparing endoscopic ultrasonography, helical computed tomography,


15.

16.

17.

18.

magnetic resonance imaging, and angiography. Am J Gastroenterol 99:492-501, 2004 Boland GW, O’Malley ME, Saez M, et al: Pancreatic-phase versus portal vein-phase helical CT of the pancreas: Optimal temporal window for evaluation of pancreatic adenocarcinoma (see comment). AJR Am J Roentgenol 172:605-608, 1999 Long EE, Van Dam J, Weinstein S, et al: Computed tomography, endoscopic, laparoscopic, and intra-operative sonography for assessing resectability of pancreatic cancer. Surg Oncol 14:105-113, 2005 Oderich GS, Panneton JM, Hofer J, et al: Iatrogenic operative injuries of abdominal and pelvic veins: A potentially lethal complication. J Vasc Surg 39:931-936, 2004 Evans DB, Lee JE, Tamm EP, et al: Pancreaticoduodenectomy (whipple operation) and total pancreatectomy for cancer, in Fischer JE, Bland KI, Callery MP, et al (eds): Mastery of Surgery (Vol 1, 5th ed). Philadelphia, Lippincott Williams & Wilkins, 2006, pp 1299-1317

159 19. Tseng JF, Raut CP, Lee JE, et al: Pancreaticoduodenectomy with vascular resection: Margin status and survival duration. J Gastrointest Surg 8:935-950, 2004 20. Bold RJ, Hess KR, Pearson AS, et al: Prognostic factors in resectable pancreatic cancer: p53 and bcl-2. J Gastrointest Surg 3:263-277, 1999 21. Buckman RF, Pathak AS, Badellino MM, et al: Portal vein injuries. Surg Clin North Am 81:1449-1462, 2001 22. Pearl J, Chao A, Kennedy S, et al: Traumatic injuries to the portal vein: Case study. J Trauma 56:779-782, 2004 23. Stone HH, Fabian TC, Turkleson ML: Wounds of the portal venous system. World J Surg 6:335-341, 1982 24. Jurkovich GJ, Hoyt DB, Moore FA, et al: Portal triad injuries. J Trauma 39:426-434, 1995 25. Asensio JA, Petrone P, Garcia-Nunez L, et al: Superior mesenteric venous injuries: To ligate or to repair remains the question. J Trauma 62:668-675, 2007

The Control of Intraoperative Hemorrhage During Pelvic Operations Horst S. Filtzer, MD, FACS

T

he anatomy of the pelvis has few unusual variations. Although the tissue planes in which one performs dissection and mobilization of structures in the pelvis are readily identified in the normal state, prior surgery, malignancy, and longstanding inflammation may obscure them in disease. Urologists, gynecologists, vascular surgeons, orthopedists, and colorectal surgeons all operate within the pelvis, and sometimes with each other, as in the case of pelvic exenteration.1 Familiarity with the normal anatomy is essential for all, and techniques developed in one specialty for pelvic hemorrhage control after pelvic fracture, obstetrical injury, cesarean delivery, or soft tissue trauma surgery may be adapted by another to control unexpected bleeding encountered in more elective circumstances.2,3 Patients submitted to pelvic operations should be queried preoperatively and appropriately evaluated for blood dyscrasia or bleeding disorders.4 If found, clotting factors should be administered to normalize the coagulation milieu. A Cell Saver should be made available in cases in which there is a history of prior bleeding during surgical procedures and in cases in which von Willebrand’s disease is present. There are also subtle factors that alter coagulation and may impact intraoperative hemorrhage. A high alcohol intake on the part of the patient despite the presence of normal liver functions may lead to problems with subsequent bleeding. Herbal and natural remedies may exacerbate intraoperative coagulation difficulties; vitamin E, St. John’s Wort, ginseng, garlic, ginkgo biloba, fish oil, and beta-carotene have been implicated in altering the coagulation cascade and should be discontinued for at least 7 days before surgery. Drugs such as Plavix alter platelet function and should be discontinued 2 weeks before surgery, if the underlying cardiovascular illness permits. Although the most effective way to control bleeding is a direct occlusion of a visualized bleeding vessel, fibrin glue, Gelfoam, and surgicel are available to every surgeon and can be utilized when significant diffuse bleeding is encountered during pelvic surgery (Table 1). Platelet transfusions are an integral part of the resuscitative effort during transfusion of the bleeding patient; 10 to 20 units of platelets and supple-

Harvard Medical School, Boston, MA. Address reprint requests to Horst Filtzer, MD, FACS, Assistant Professor of Surgery, Harvard Medical School, 330 Mount Auburn Street, Cambridge, MA 02138. E-mail: [email protected]

160


mental calcium should be given after each blood volume replacement with transfused products. Factor VIII may also be administered in those patients even without a history of factor VIII deficiency because after massive transfusion, availability of factor VIII drops precipitously.4,5

Factors Impacting Blood Loss Cornelius Sedgwick, of the Lahey Clinic, advocated the “Rule of Twos” for success in surgery; completion of the operation within 2 hours with blood loss of less than two units. Operating within such parameters exacts minimal physiologic toll on the patient. Major blood loss, however, defined as replacement transfusion of one total blood volume, can set up a cascade of physiologic derangements leading not only to hypothermia, coagulopathy, and hypoxia with acidosis (the lethal triad), but also myocardial dysfunction, arrhythmia and death. Hypothermia occurs because of a combination of blood loss and resuscitation efforts with cold fluids, combined with prolonged exposure of the open abdomen to the environment. If the body temperature goes below 36°C for greater than 4 hours, cardiac arrhythmia, decreased cardiac output, increased systemic vascular resistance, and a leftward shift of the oxygen hemoglobin dissociation curve may occur. Coagulopathy then occurs because of inhibition of the coagulation cascade; hemodilution and hypothermia have an additive effect, resulting in diminution of circulating coagulating factors and platelets. As hypoperfusion continues, anaerobic metabolism occurs, acidosis worsens and a vicious circle ensues that may lead to death if not reversed. The trauma literature suggests that with uninterrupted hypoperfusion and acidosis, an “irreversible shock” state is reached and survival becomes unlikely unless specific and expeditious measures are taken. The so-called bail-out or “damage control” approach championed in the trauma and emergent surgery literature clearly has application to elective pelvic operations in which exsanguinating hemorrhage has occurred.2,3

The Bail-Out Decision In the majority of cases in which hemorrhage is encountered during pelvic operations, expeditious control can be undertaken by standard methods such as suture ligation or oversewing of injured blood vessels. Direct vascular repair of large

Intraoperative hemorrhage during pelvic operations

161

Table 1 Topical Intraperitoneal Hemostatic Agents Agent

What It Is

How It Is Applied

Avitene ultrafoam Fibrin glue Coseal Floseal Tisseal Gelfoam Surgicel

Absorbable collagen hemostat Equal amounts of cryoprecipitate and thrombin

Comes in powder; sprinkle on area Spray on affected area with double-barrel syringe or device supplied by Baxter Healthcare

Absorbable gelatin sponge Oxidized regenerated cellulose

Cut in strips of appropriate size and apply to area Cut in strips of appropriate size and apply to area

vessel injury or ligation of feeding vessels to a tumor may also be utilized. There are no hard and fast rules about when to consider the bail-out option during pelvic procedures; however, hypothermia of 34°C, a pH of 7.2 or less, a serum bicarbonate level of 15 mg/L, transfusions of greater than 4,000 mL of blood and blood products, and intraoperative volume replacement with greater than 10 liters of crystalloid are conditions that foster arrhythmia and acidosis and emphasize the preferability of a bail-out procedure to death from irreversible shock. Sometimes, an operative procedure may have to be abandoned at a stage where resection is not complete; the abdomen is packed and temporarily closed. The patient is returned, intubated, to the Intensive Care Unit for resuscitation. Hemoglobin and clotting factors are restored, and pH, intravascular volume and temperature as normalized to the extent possible. The interval between damage control and reoperation can be somewhat variable, but as longer intervals make septic complications more likely, reoperation should occur promptly after hemodynamic stability has been established and re-warming has occurred. Under most circumstances, this can be accomplished within 24 to 36 hours after the first surgery. During the resuscitative effort, as transfusion of blood and blood products seek to return oxygen delivery and clotting to normal, a predictor of success is the durability of the response of the platelet count. In surgical wounds bleeding after massive transfusion, platelet counts of greater than 100,000 are sought so that ongoing oozing and bleeding do not occur; multiple platelet transfusions may be necessary to accomplish this.

Anatomic Considerations The relatively constant course of the ureter should be known and its dissection begun high on either side of the hollow of the pelvis and developed caudally to a point even with the tip of the coccyx. This approach should occur whether one is doing a low anterior or abdominoperineal resection. There are in the normal state very few dramatic aberrations of pelvic vascular anatomy, but atherosclerotic disease and previous venous thrombosis do impact the ability to control pelvic bleeding. The most important aspects of pelvic operations are the avoidance of hemorrhage by a sound knowledge of the anatomy and those tissue planes in which dissection can be expeditious and safe. Prevention of venous bleeding demands that the surgeon know the points at which the major pelvic veins are adhesed to major structures. Distal to the bifurcation of the aorta, the iliac veins are frequently adherent to the common iliac arteries (Fig 1); likewise, the hypogastric artery needs to be carefully dissected away from both the common iliac and hypogastric veins in order not to

injure them when the hypogastric artery must be ligated or divided. Such ligation should seriously be considered during gynecologic operations if hemorrhage control is ineffective by a direct approach to distal arterial branches or their venous drainage. There are numerous reports of successful bilateral hypogastric ligations and hemorrhage control in the gynecological literature; where possible, it should be avoided because of hypothetical concerns of infarction of pelvic organs.6 The hypogastric artery itself divides into a superior and an inferior branch. The superior branch frequently requires ligation during gynecologic operations when hemorrhage from the area of the cervix and vaginal junction occurs. In mobilizing the rectum for resection, the surgeon should be aware that incision in the peritoneal reflection just above the sacrum can usually be made with little bleeding. That incision can be developed on either side of the rectum and extended in the direction of the pelvic floor and lower genitourinary tract. The areolar tissue in these areas can be gently swept away and the hollow of the sacrum can be readily entered with minimal blood loss. Only when one half of the circumference of the rectum is mobilized posteriorally are the so-called rectal stalks approached. Their dissection is begun anterior to the rectum in the space between the vagina or prostate and continued until either Denonvillier’s fascia is reached in the male or the floor of the perineum is reached in the female. A rich cavernous plexus of veins also extends anteriorally near bone and prostatic tissue; unsecured, it can cause significant hemorrhage and require packing or direct suture. A dissection plane in this anterior area may be found that is free of any significant major blood vessel and should be identified before the stalks themselves are approached and controlled. There are numerous blood vessels within the stalks that are best dealt with by stretching the tissue between two fingers and creating a two-dimensional structure that can be clamped and divided all the way down to the perineal floor. The inferior hemorrhoidal arteries are encountered in a constant location and can be visualized, mobilized, and divided in a similar manner. A unique anatomic consideration in the pelvis relates to the location of the presacral venous plexuses beneath the fascia of the sacrum. This anatomic arrangement assures that the veins are kept open by bony attachments when they are injured. Thus, when they bleed, they tend to bleed massively. When one is posterior to the desired dissection plane and venous hemorrhage is encountered, for example, in the mobilization of the rectum, the unfavorable geometric apposition of a small pelvis and a big tumor may necessitate expeditious further mobilization and specimen removal before the bleeding site can be visualized and attempts at control can be undertaken. Some venous bleeders can be con-

H.S. Filtzer

162

Figure 1 Dissection of the normally occurring adhesions between the branches of the iliac artery and the branches of the iliac veins distal to the aortic bifurcation, facilitating identification of the bleeding vessel in that region.


Figure 2 A thumbtack placed directly into the cortex of the sacral bone occludes venous bleeding from the subfascial sacral venous plexus.

163

H.S. Filtzer

164

Figure 3 “Tacks and packs”: Plain abdominal radiograph showing several thumbtacks in the sacrum and marked Mikulicz pads. The patient was transported to the surgical intensive care unit, where fluid resuscitation, administration of blood products, correction of clotting abnormalities, and warming of the patient continued. Packs were removed at a subsequent laparatomy. (Photo courtesty of Russell J. Nauta, MD.)


Figure 4 The Bakri balloon is introduced through the transected rectal stump after removal of the proximal rectum and it is inflated to apply direct pressure to adjacent bleeding points. Alternatively, as the clinical situation permits, it could be placed through an open vaginal cuff to control posthysterectomy bleeding.

165

H.S. Filtzer

166 trolled by a combination of digital pressure and subsequent suture ligation with a stout needle and a Gelfoam or Surgicel pack. More advanced techniques may be required if these maneuvers fails. An ingenious device for control of hemorrhage from the presacral space and sacral foramina has been developed by Wortrich and consists of a long handle into which a thumbtack is placed to facilitate its placement into the sacrum with an unobstructed view.7-9 It is tapped into place by the surgeon with a small mallet or secured with digital pressure (Fig 2). Other methods for controlling and tamponading hemorrhage within the pelvis have been advocated, including the use of expandable breast prostheses, tissue expanders, or tissue expander sizers placed into the pelvis, which can then be inflated and secured in place. Packs made of Surgicel, Bonewax, Gelfoam, and gauze have also been employed (Fig 3).10,11 Direct injection of fibrin glue into or near the bleeding venous branches near the hollow of the sacrum has also been advocated.12 Suffice it to say whatever has been imagined to be able to provide effective compression or to facilitate closure has been attempted.13 The Bakri balloon, an obstetrics device that is inserted into the atonic bleeding postpartum uterus transcervically and then inflated to control hemorrhage, can be analogously used to tamponade hemorrhage in the pelvis by introducing it through an open vaginal cuff or rectal stump and then inflating it (Fig 4). Traction may be utilized to put pressure on the pelvic sidewalls to stop bleeding.14-17 Once hemostasis is obtained by whatever method, the decision needs to be made whether to continue to attempt to achieve the original goals of the operation or whether to accept that damage control has been achieved, but at a physiologic cost. In some circumstances it may be appropriate to wait in the operating room for 10 or 15 minutes with packs in place and all bleeding controlled to catch up on the blood loss and fluid requirements of the patient and then to decide to go ahead and systematically remove the hemostatic packs or pressure balloons. Under these circumstances, the operation should be continued only if it is possible to do so safely, quickly, and with minimum further blood loss. When a huge amount of blood loss has already occurred and when hypothermia and acidosis are established, damage control may be the most prudent approach, with a return trip to the operating room planned within 24 to 36 hours. The surgeon may avail himself of further adjuncts to packing and transfusion of blood products in the SICU or the interventional radiology suite. Effective control of posttraumatic pelvic bleeding in cases of fractures and other injuries has been achieved by interventional radiologic methods including vessel occlusion, Gelfoam embolization of arterial bleeding, balloon occlusion, and deployment of covered stents.18-20 The intraoperative environment does not readily lend itself to the application of these methods.

However, once packs are in place and the patient is stabilized in the Intensive Care Unit, they should be considered as part of the resuscitative effort before return to the operating room. As with approaches to gastrointestinal bleeding, sheaths placed into the iliac vessels for angiography may be left in place for easier access in case subsequent bleeding suggests that a transluminal approach would be helpful.

References 1. Gallup DG: Catastrophic intraoperative hemorrhage: 5-step action plan. OBG Management 17:1-7, 2005 2. Mohr AM, Asensio JA, Garcia-Nunez LM, et al: Guidelines for the institution of damage control in trauma patients. Intl TraumaCare 15: 185-189, 2005 3. Moeng MS, Loveland JA, Boffard KD: Damage control: Beyond the limits of the abdominal cavity. A review. Intl TraumaCare xx:189, 2005 4. Hardy JF, De Moerloose P, Samama M: Massive transfusion and coagulopathy: Pathophysiology and implications for clinical management. Can J Anaesth 51:293-310, 2004 5. Counts RB, Haisch C, Simon TL, et al: Hemostasis in massively transfused trauma patients. Ann Surg 190:91-99, 1979 6. Papp Z, Toth-Pal E, Papp C, et al: Bilateral hypogastric artery ligation for control of pelvic hemorrhage, reduction of blood flow and preservation of reproductive potential. Experience with 117 cases. Orv Hetil 146:1279-1285, 2005 7. Timmons MC, Kohler MF, Addison WA: Thumbtack use for control of presacral bleeding, with description of an instrument for thumbtack application. Obst Gynecol 78:313-315, 1991 8. Stolfi VM, Milsom JW, Lavery IC, et al: Newly designed occluder pin for presacral hemorrhage. Dis Colon Rectum 35:166-169, 1992 9. Stolfi VM, Milsom JW, Lavery IC, et al: Newly designed occluder pin for presacral hemorrhage. Dis Colon Rectum 35:166-169, 1992 10. Finan MA, Fiorica JV, Hoffman MS, et al: Massive pelvic hemorrhage during gynecologic cancer surgery: “Pack and go back.” Gynecol Oncolol 62:390-395, 1996 11. Civelek A, Yeg˘en C, Özdemir Aktan A: The use of bonewax to control massive presacral bleeding. Surg Today 32:944-945, 2002 12. Losanoff JE, Richman BW, Jones JW: Cyanoacrylate adhesive in management of severe presacral bleeding. Dis Colon Rectum 45:1118-1119, 2002 13. Remzi FH, Oncel M, Fazio VW: Muscle tamponade to control presacral venous bleeding. Dis Colon Rectum 45:1109-1111, 2002 14. Howard RJ, Straughn JM Jr, Huh WK, et al: Pelvic umbrella pack for refractory obstetric hemorrhage secondary to posterior uterine rupture. Obstet Gynecol 100:1061-1063, 2002 15. Dildy GA, Scott JR, Saffer CS, et al: An effective pressure pack for severe pelvic hemorrhage. Obstet Gynecol 108:1222-1226, 2006 16. Braley SC, Schneider PD, Bold RJ, et al: Controlled tamponade of severe presacral venous hemorrhage. Dis Colon Rectum 45:140-142, 2002 17. Bakri YN, Amri A, Abdul Jabbar F: Tamponade-balloon for obstetrical bleeding. Int J Gynaecol Obstet 74:139-142, 2001 18. Miller FJ Jr, Mortel R, Mann WJ, et al: Selective arterial embolization for control of hemorrhage in pelvic malignancy: Femoral and brachial catheter approaches. Am F Roenigenol 126:1028-1032, 1976 19. Weeks SM, Stroud TH, Sandhu J, et al: Temporary balloon occlusion of the internal iliac arteries for control of hemorrhage during cesarean hysterectomy in a patient with placenta previa and placenta increta. J Vasc Interv Radiol 11:622-624, 2000 20. Jander HP, Russinovich NAE: Transcatheter gelfoam embolization in abdominal, retroperitoneal and pelvic hemorrhage. Radiology 136: 337-344, 1980

Management of Ureteral Injury Paul LaFontaine, MD

I

njury to the ureter may occur as a result of civilian or military trauma, as well as in the conduct of elective surgery. The anatomic location of the injury, degree of devascularization, and tissue damage are all determinants of the most appropriate reconstructive procedure.

Uretero-Ureterostomy Uretero-ureterostomy repairs discrete short strictures of the midureter or intraoperative injury to the mid or upper ureter above the bifurcation of the iliac vessels. A number of incisional approaches permit access to the ureter and ipsilateral kidney; midline abdominal, Pfannanstiel, Gibson, or flank incisions have all been used. The ureter can normally be identified reliably as it crosses over the bifurcation of the iliac vessels to enter the pelvis. It should be dissected free proximally and distally for as short a length as possible to permit a tension free anastomosis. Care should be taken to not skeletonize it, but to mobilize the peri-advential tissue with it to preserve its blood supply. Access to the kidney allows renal mobilization if required to avoid anastomotic tension. To avoid anastomotic stricture, the ureteral ends are first spatulated proximally and distally and then oriented with the spatulations 180 degrees apart from each other. The anastomosis is performed with multiple interrupted 4-0 or 5-0 vicryl sutures with the knots tied so that they rest outside of the ureteral lumen. The anastomosis is facilitated by placing the sutures through the apex of each spatulated end initially. After placement of the apical suture, the remainder of the interrupted sutures are placed to complete the repair. Care is taken not to pick up the back wall of the ureter with these sutures. Before anastomotic completion, a 6FR ureteral stent is guided with a wire to a position where its proximal end rests in the renal pelvis and its distal end rests in the bladder.

Transuretero-Ureterostomy (TUU) A TUU is utilized for management of long defects of the lower and mid ureter where adequate proximal ureteral length ex-

Department of Surgery, Harvard Medical School, Cambridge Urological Associates, Cambridge, MA. Address reprint requests to Paul LaFontaine, MD, Clinical Instructor in Surgery, Harvard Medical School, Cambridge Urological Associates, 300 Mount Auburn Street, Suite 519, Cambridge, MA 02138. E-mail: [email protected]


ists on the injured side to reach the recipient contralateral ureter without tension. The injured donor ureter is identified by incising the posterior peritoneum above the level of the iliac vessels. The donor ureter should be dissected for a length sufficient to allow it to reach the recipient ureter without tension. When dissecting out the donor ureter, it is important to dissect a generous amount of peri-ureteral advential tissue to assure its blood supply. The contralateral recipient ureter is identified by incising the posterior peritoneum above the iliac vessels. The recipient ureter should be dissected for as short a length as will permit a well-visualized anastomosis (Fig 1). A retroperitoneal tunnel anterior to the great vessels is created between the two incisions. It should accommodate passage of the surgeon’s finger and pass cephalad to the IMA to prevent compression of the ureter between the IMA and aorta. The donor ureter is then passed atraumatically through the retroperitoneal tunnel and spatulated. Caution must be exercised so as not to twist or kink the donor ureter as it is passed through the retroperitoneal tunnel. Ureterotomy is then created on the medial aspect of the recipient ureter. Placement of fine stay sutures on the recipient ureter proximal and distal to the ureterotomy will allow elevation of the ureteral wall and facilitates successful completion of the anastomosis. The uretero-ureterostomy is created with interrupted 4-0 or 5-0 vicryl sutures. Knots are tied extramurally so that they rest outside of the lumen (Fig 2). Ideally, both ureters should be stented to prevent obstruction. Multiple side holes are then created in two ureteral stents. Before completion of the anastomosis, at least one stent is placed over a guide wire. If the diameter of the recipient ureter below the anastomosis prevents its stenting, only the donor ureter should be stented.

Uretero-Neocystostomy If an injury to a ureter is distal enough to allow the proximal ureter to be brought to the dome of the bladder without tension, uretero-neocystostomy without psoas hitch may be the appropriate procedure. There is rarely a need to create a tunneled, or nonrefluxing, uretero-neocystostomy as reflux of urine in the adult patient is rarely of clinical significance. After determining that adequate ureteral length is present for a tension-free direct anastomosis to the bladder, the bladder is distended with saline infused through an indwelling urethral catheter. A 1 to 2 cm incision is made transversely on 167

168

P. LaFontaine

Figure 1 Transureteral ureterostomy. The retroperitoneum over each ureter is incised longitudinally. The distal end of the donor ureter is ligated. The proximal end of the donor ureter is passed through the retroperitoneal tunnel cephalad to the inferior mesenteric artery.

Ureteral injury

169

Figure 2 Transureteral ureterostomy. The donor ureter is spatulated, and end-to-side anastomosis is created.

P. LaFontaine

170

Figure 3 Ureterocystostomy. The bladder is distended with saline and a cystotomy is made. The ureter is spatulated and the anastomosis is created.

the side of the bladder ipsilateral to the injury. The detrusor is divided with electrocautery until the bladder mucosa is indentifiable by its shiny appearance. The bladder mucosa is then incised sharply and everted in a rosebud technique with multiple interrupted 5-0 chromic sutures. The distal end of the ureter is then spatulated and anastomosed to the bladder using 4-0 or 5-0 vicryl sutures, beginning the anastomosis at its apex (Fig 3). The sutures on the nonapical portion of the ureter may incorporate detrusor muscle. Before completion of the anastomosis, the ureter is stented with a 6FR stent placed over a guidewire and passed proximally to the level of the renal pelvis. The detrusor muscle is then loosely reapproximated over the anastomosis to avoid obstruction of the ureter.

Psoas Hitch A psoas hitch provides mobilization and elevation of the bladder to a level above the bifurcation of the common iliac artery. It is used for repairing injuries to the lower ureter where a direct uretero-ureterostomy is not feasible or where inadequate ureteral length does not permit a tension free uretero-neocystostomy to a bladder left in its pelvic location.

The initial maneuver used to begin mobilization of bladder is incision of the peritoneum posterior to it. The space between the bladder and the rectum is then developed using blunt dissection. On the noninjured side, the superior vesical artery is ligated and divided in the lateral bladder pedicle. After mobilizing the bladder’s attachments in the pelvis on the noninjured side, the bladder is filled through a urethral catheter. A transverse cystotomy is then made in the anterior wall of the bladder at its point of widest diameter. The cystotomy should be extended for at least half the circumference of the bladder. A shorter cystotomy may not provide enough cephalad displacement of the dome of the bladder. Using two fingers in the bladder, the dome of the bladder is elevated to a point above and lateral to the iliac artery on the psoas major muscle. The injured ureter can then be proximally freed to ensure enough length is available for a tension-free anastomosis. If enough mobility of the bladder has not been achieved with this maneuver for its cephalad mobilization to a point above the iliac vessels at the psoas major muscle on the injured side, the contralateral inferior vesical artery is ligated and divided and the contralateral endopelvic fascia is divided. It is rarely necessary to divide the superior vesical artery on the injured side. If necessary, the ipsilateral

Ureteral injury

Figure 4 Psoas hitch. The large incision in the bladder permits the cephalad mobilization of the bladder, whose superior aspect is secured to the psoas muscle before creating the anastomosis between ureter and bladder. The cystotomy is then closed.

171

P. LaFontaine

172

Figure 5 Boari flap. A proximally based flap is mobilized to bridge the gap between bladder and ureter.

Ureteral injury

Figure 6 Boari flap. The flap is fashioned into a tube and an anastomosis is created between it and the proximal ureter.

173

174 kidney can be mobilized caudally to close the gap between ureter and bladder. Once this has been accomplished, the dome of the bladder is secured to either the psoas major muscle or the tendon of the psoas minor muscle using 4 to 6 interrupted 2-0 PDS sutures (Fig 4). Attention must here be paid to two nerves. First, the genitofemoral nerve coursing over the psoas muscle in this location should be identified, and care taken to not injure or incorporate it the sutures. Second, the femoral nerve runs posterior to the psoas major muscle. Care should be taken to avoid deep sutures in the psoas major so that this structure is not damaged. A refluxing, nontunneled ureteroneocystoscopy is appropriate in most adults and technically easier to perform than a tunneled re-implant. After selecting a site on an immobile part of the bladder dome, a cystotomy is made. The ureter is then brought through the bladder wall, assessed to assure adequate length, and spatulated. An anastomosis is performed with interrupted 4-0 or 5-0 vicryl sutures starting at the apex. As the distal ureteral tip of the ureter is approached, sutures may incorporate both bladder detrusor and mucosa. A 6FR ureteral stent is guided with a wire to rest proximal to the renal pelvis. The bladder is then closed in a HenickeMickulicz fashion in two layers with a running inner layer of 3-0 chromic mucosa and a running 2-0 chromic for the detrusor. Placement of a supra-pubic tube before bladder closure may be appropriate in some patients but is not mandatory in all if the bladder suture lines are thought to be secure.

Boari Flap A Boari flap is used for repair and reconstruction of long ureteral strictures in those rare clinical circumstances where a psoas hitch and renal caudal mobilization are insufficient to provide enough ureteral length for a tension free anastomo-

P. LaFontaine sis. When doing a Boari flap a Psoas hitch should be performed concurrently. After dividing the contralateral superior and inferior vesical arteries and incising the contralateral endopelvic fascia, the surgeon pulls the bladder into a tube up to the psoas major muscle. If there is not enough ureteral length to reach the estimated position of the bladder with a psoas hitch and renal mobilization then a Boari flap should be considered. Instead of a transverse incision on the anterior wall of the bladder at its maximum diameter as for a psoas hitch, the Boari flap incision is marked out on the anterior wall of the bladder after filling the bladder to roughly half its volume through a urethral catheter. The length of the Boari flap should correspond to the distance between the distal end of the ureter and the psoas major muscle where a psoas hitch would place the bladder. This incision should be 4 cm wide at its base and 3 cm wide at its apex to ensure adequate vascularity of the flap (Fig 5). Longer flaps may require a wider base. Stay sutures are placed at the marked base and apex of the flap and the bladder is divided with electrocautery. The flap should then be elevated to reach the ureter so that are several centimeters of overlap between ureter and the flap. The ureter should be spatulated and then attached to the flap in a nontunneled refluxing end to end anastomosis with 4-0 or 5-0 vicryl sutures. The sutures on the distal aspect of the spatulated ureter should incorporate some detrusor muscle to prevent retraction of the ureter. A 6FR stent should then be placed over a guidewire proximally into the renal pelvis. The Boari flap should then be tubularized in two layers with running 3-0 vicryl mucosal suture and a running 2-0 vicryl detrusor muscle suture (Fig 6). The bladder cystotomy is closed in a similar fashion with running 3-0 vicryl mucosal sutures and running 2-0 vicryl detrusor muscle sutures after placement of a supra-pubic drain through a separate stab wound. A perivesical drain is left at the end of the procedure.

Management of Common Bile Duct Injuries John C. Haney, MD, and Theodore N. Pappas, MD

E

xtrahepatic biliary duct injury is a rare but potentially devastating condition associated with significant morbidity and mortality. The vast majority of these injuries occur as rare complications of the 750,000 cholecystectomies performed annually in the United States.1 Iatrogenic injury may also occur during gastrectomy, pancreatectomy or ERCP. Trauma and duodenal ulcer are less common causes.1-4 Improvements in length of stay, postoperative pain, and cosmetic results with the laparoscopic approach to cholecystectomy have been well documented; equally well documented has been the persistent incidence of iatrogenic extrahepatic bile duct injury. Population-based studies consistently cite an incidence of cholecystectomy-associated bile duct injury between 0.3% to 0.6% for the laparoscopic approach and 0.1% to 0.3% for open cholecystectomy.5-7 A significant amount of literature has been dedicated to the identification of patient characteristics and intraoperative factors associated with increased risk. Conditions that confer additional risk are those that increase the degree of acute inflammation within the triangle of Calot or those that obstruct visualization and impair access to the “critical view.” These include variables such as obesity and periportal fat, the presence of complex biliary disease such as choledocholithiasis, gallstone pancreatitis, or frank cholangitis, and intraoperative bleeding interfering with visualization.8 Atypical anatomy, including aberrant right hepatic duct or complex cystic duct insertion, also predispose to intraoperative injury.9

Classification Bile duct injuries may be classified by mechanism and type of injury, location of injury, effect on biliary continuity, and timing of identification. Each of these factors plays a significant role in determining the appropriate operative repair and management.

Location Identifying the location of ductal injury and the availability of healthy proximal duct is critical; successful repair requires healthy, nonischemic duct without tension or loss of length. Bismuth and colleagues classified biliary strictures in 1982

Department of Surgery, Duke University, Duke University Medical School, Durham, NC. Address reprint requests to Theodore Pappas, MD, Professor of Surgery, Duke University, Duke University Medical School, DUMC 3479, Durham, NC 27710. E-mail: [email protected]


based on their proximity to the duct confluence. Strasberg elaborated on this scheme in 1995, classifying injuries by location, mechanism, and consequence to the continuity of the system, separating leaks from occlusions (Figs 1 and 2).10,11

Mechanism Laparoscopic injuries may be caused by inadvertent duct laceration or sharp transection, excessive traction or cautery injury, partial or complete clip ligation, or ligation and transection with loss of duct length. The “classic” laparoscopic injury has been described previously and mistakes the common bile duct for the cystic duct, leading the surgeon to clip and resect the common duct (Fig 3).12 Proximal dissection and division often leads to injury of the right hepatic artery, thus the classic injury involves loss of length and occlusion of the proximal biliary tree with possible concomitant right hepatic ischemia. Other laparoscopic injuries are possible. Clip ligation of the distal common bile duct with proximal ligation and division of the cystic duct results in bile obstruction and leak. Avulsion injury of the common bile duct during distal cystic duct transection leads to bile leakage. Thermal injury, clip application, or duct transection are all likely to cause injury to the duct or its blood supply, thus causing a loss of viable length.15 More insidious presentations of ductal injury may occur with late biliary stricture, caused either by duct ischemia secondary to thermal or traction injury or by the various chronic inflammatory or infectious causes. Traumatic injuries include partial or complete extrahepatic duct transection or duct avulsion, occasionally in close proximity to the pancreas.13 The mechanism and type of injury play critical roles in patient presentation and surgical management.

Clinical Presentation and Timing of Identification The type of biliary tract injury, as well as the timing of injury identification, determines the patient’s clinical presentation. Intraoperatively, injury may be recognized by the exposure of unexpected ductal structures during the operation or by leakage of bile into the operative field from the lacerated or transected duct. If unrecognized intraoperatively, the manner of postoperative presentation will be determined by the continuity of the bile ducts and the presence or absence of a biliary leak. Patients with a significant duct leak may present with bilious drainage from an intraoperatively placed drain. More commonly, the bile leak will cause a localized biloma or free 175

176

J.C. Haney and T.N. Pappas

Figure 1 Bismuth classification of biliary strictures (Types 1-5) based on level relative to the hepatic confluence.

Managing common bile duct injuries

Figure 2 Strasberg classification of injury based on anatomic location and mechanism. Injuries are classified Type A-E, with Type E subdivided E1 to E5 based on Bismuth level (see Fig 1).

177


178

Figure 3 Classic laparoscopic bile duct injury. Misidentification leads to clipping and resection of the common bile duct as well as transection of the hepatic ducts proximally. Dissection also frequently leads to injury to the right hepatic artery.

bile ascites causing bilious peritonitis in the undrained patient. Such patients classically present with diffuse abdominal pain and persistent ileus several days postoperatively. If a biloma becomes secondarily infected, the patient may progress to fever and leukocytosis as with any intraabdominal abscess.14 Diffuse abdominal pain after laparoscopic cholecystectomy should be met with a high index of suspicion for possible unrecognized biliary tract injury. If the biliary injury involved clip ligation or resection of the common bile duct producing duct obstruction, the patient may present with clinical jaundice and possibly cholangitis. Aggressive decompression of an obstructed biliary tree is critical for successful management of such an injury.

Diagnostic Workup Given the low incidence of bile duct injury after what is a very common operation, the prompt diagnosis of a postoperative biliary tract injury requires a high index of suspicion. If injury is suspected intraoperatively and early repair is to be enter-

tained, the laparoscopic approach is converted to an open procedure and a cholangiogram is performed. If the principal surgeon is not experienced in complex hepatobiliary repair then the cholangiogram may be performed laparoscopically and, provided repair is to be deferred, adequate drains may be placed without open conversion.15 The patient is then referred to an appropriate hepatobiliary specialist without delay. When injuries are not diagnosed at the time of surgery, the patient’s clinical presentation provides clues to the differential diagnosis, which may in turn guide diagnostic workup. Patients classically present with either persistent abdominal pain or jaundice. The presence of postoperative abdominal pain is most consistent with a biloma from duct leak, but may rarely be a result of a retained stone in the common duct or the cystic duct stump. Initial laboratory workup for every patient includes a complete blood count to evaluate for leukocytosis and liver function panel to evaluate for signs of concomitant liver injury and biliary obstruction. Transabdominal ultrasound may be utilized as an initial diagnostic


Figure 4 Primary repair of bile duct injury. (A) Transected bile duct ends are freshened and are approximated with the aid of fine stay sutures. The ducts are reanastomosed in interrupted fashion using 4-0 or 5-0 absorbable suture, taking care to approximate mucosa to mucosa. (B) Before completion of the anterior wall, a separate longitudinal choledochostomy is created, preferably downstream, for insertion of a T tube across the anastomosis. The T tube is modified by removing the back wall, reducing obstruction. (C) The choledochostomy is closed with interrupted sutures around the T tube, and the anastomosis is completed. The T tube is then brought out through a stab incision in the abdominal wall.

179


180

Figure 5 Side-to-side anastomotic approach for Bismuth I injuries with a sufficient common duct stump (A), or for Bismuth II or III injuries in which the anastomosis is extended into the left or right duct for sufficient length (B).

modality, but more commonly, patients will receive an abdominal computed tomography (CT) scan. Survey of the abdomen by either modality will allow for diagnosis of either free or contained biloma and may guide percutaneous drainage. Once a biliary injury is suspected and all localized or free bile is percutaneously drained, a complete cholangiogram must be performed to define the anatomical injury.16 This may be obtained using magnetic resonance imaging or ERCP. Technetium-99 month-HIDA scintigraphy may be used as screening test for diagnosing postoperative bile leaks or as a tool for determining the adequacy of drainage.17 When patients present with jaundice, the differential diagnosis includes biliary obstruction from retained common bile duct stone or extrahepatic duct occlusion. In this case, ultrasound or CT scan may also be used as the initial test. If a common duct stone is suspected then ERCP is the appropriate next step, but in the event of occlusive injury endoscopic cholangiogram will be unable to provide adequate visualization of the entire biliary tract secondary to the loss of continuity.3 In this case, percutaneous transhepatic cholangiography, which provides the most complete picture of proximal ductal anatomy and has the advantage of allowing for placement of decompressive biliary catheters as well should be performed.18

Preoperative Management Once a bile duct injury has been diagnosed, a well-informed decision must be made regarding the timing and type of

repair. First and foremost, this requires full appreciation for the injury and the resultant anatomy. Multiple studies have shown that early repair of biliary injury may be performed, but repair in the setting of overt inflammation is unlikely to be successful.1,19 In addition to defining anatomy, a major principle of preoperative therapy is the treatment of existing ductal inflammation. This entails biliary decompression for complete obstruction, aggressive drainage of bilomas, and intravenous antibiotics guided by culture of bile if cholangitis is present. Drains may be placed in the open biliary tract intraoperatively if the injury is diagnosed immediately and repair is deferred. When complete obstruction is diagnosed postoperatively, decompression should be accomplished by the placement of percutaneous biliary catheters through a transhepatic approach. When the injury is located at least 2 cm distal to the duct confluence a single drain is sufficient, but if the injury is within 2 cm of the confluence, repair is likely to involve the confluence and thus drains should be left in both the left and right sides. If the tract is in continuity with a biloma, the transhepatic biliary catheters should be placed across the injury into the collection to facilitate drainage and identification of the ducts. Preoperative, temporary stenting of strictures is somewhat controversial, but may be utilized to restore continuity and decompress the tract.20 In cases with a dilated duct (⬎1 cm) no stent is required. Pre-reconstruction management usually involves aggressive multidisciplinary therapy and usually does not require an operation to control bile drainage. The amount of time required to quiet existing inflammation or infection varies depending on the severity of


181

Figure 6 End-to-side choledochojejunostomy. The common bile duct is spatulated and anastomosed to the antimesenteric enterotomy using interrupted 4-0 or 5-0 mucosa-to-mucosa sutures with knots tied on the inside.

the presenting illness, which may range from a mild inflammatory response when detected early to overwhelming ascending cholangitis and sepsis in the event of complete obstruction and infection.5

Repair Nonoperative Management In select cases, nonoperative interventions may be able to provide durable and definitive biliary drainage, and multiple groups report high success rates with either percutaneous or endoscopic techniques.21-23 Definitive nonoperative management requires intact bilioenteric continuity. Frequently this occurs when a small bile leak is detected postoperatively and treated with stent placement. Short biliary strictures may also be appropriate for a trial of nonoperative therapy. In general, strictures ⬍2 cm in length within an intact bile duct may be balloon dilated and stented.24 This management often requires multiple reinterventions over several months. Percutaneous management often requires at least three attempts at dilation whereas endoscopic management may prompt the placement of 2 to 3 side-by-side plastic stents left in place for many months.

Operative Management When an injury is diagnosed at the time of initial surgery, the decision of whether or not to attempt repair, be it primary or bilioenteric, depends primarily on the experience and comfort level of the surgeon with what is a technically challenging

procedure in a nondilated biliary system. Multiple studies supporting early tertiary referral have shown improved outcomes with repair by experienced hepatobiliary specialists.16,25 When operative repair is performed, the options include primary repair over a T tube and bilioenteric diversion. The critical principles producing a durable repair are the creation of a tension-free, mucosa-to-mucosa anastomosis with healthy, nonischemic bile duct.19,20,26,27 Although the blood supply to the hilar and retropancreatic bile ducts is substantial, the remainder of the extrahepatic system is supplied by two tenuous axial arteries at the 3 o’clock and 9 o’clock positions, and this supply is easily compromised.28 As a result, it is imperative to minimize unnecessary dissection whereas at the same time removing any nonviable duct length back to healthy tissue. Primary repair of partial or complete bile duct transection may be performed but should be viewed as the exception rather than the rule.10,15 The presence of any tension on the anastomosis is a contraindication to primary repair, and the use of a Kocher maneuver to gain length on the duodenal side, while advisable, is generally not sufficient to relieve tension.15 Irregular edges or ischemic tissue should be trimmed but the duct should not be routinely cleaned in either direction. Clamps should be avoided, and fine temporary stay sutures may be used to approximate the ends.29 Simple, noncircumferential lacerations may be repaired in interrupted fashion using fine (4-0 or 5-0) absorbable suture. Care should be taken to approximate mucosa to mucosa, taking small, precise bites of the tissue. If

182


Figure 7 Roux-en-Y hepaticojejunostomy. Biliary stent(s) are brought out through the anterior incision in the bile duct. A 3 to 4 cm antimesenteric enterotomy is created in the Roux limb and the stents are passed into the Roux limb. Stay sutures are placed at either end of the anastomosis. The edges are approximated in interrupted fashion with fine absorbable suture, tying the knots within and approximating the biliary and jejunal mucosa.

Managing common bile duct injuries simple primary repair is attempted a T tube may be placed through the repair site. If the primary involves any complexity the T tube should be placed through a separate choledochostomy, preferably below the site of injury, with the proximal end of the tube across the repair (Fig 4). A closedsuction drain is placed in Morrison’s pouch and brought out through the skin. After completion, the T tube is flushed and an intraoperative cholangiogram is taken to determine adequacy of the repair. By far the more common reconstructive technique is to restore biliary continuity by way of a new bilioenteric anastomosis. This may be accomplished with a choledochoduodenostomy, choledochojejunostomy, or hepaticojejunostomy. Although choledochoduodenostomy can be an acceptable repair, its utility is limited. For the vast majority of cases, a Roux-en-Y choledochojejunostomy or hepaticojejunostomy is the most durable reconstructive option. A 40 to 50 cm retrocolic Roux limb is brought in proximity to the common bile duct or common hepatic duct in anticipation of the anastomosis. Side-to-side anastomosis is preferable, as it has been shown to maintain better patency than end-to-side reconstruction.3 For Bismuth I injuries this may involve the common bile duct only, but for injuries adjacent to or involving the confluence, the anastomosis may need to be extended into one of the hepatic duct limbs to achieve adequate length (Fig 5). The anterior bile duct wall is opened sharply in longitudinal direction to a distance of 3 to 4 cm. The preoperatively placed biliary catheter (used routinely in elective reconstructions) is brought through the opening. For end-toside anastomoses, the distal end of the common bile duct is freshened and spatulated to increase its cross-section (Fig 6). A 3 to 4 cm enterotomy is made in the antimesenteric side of the Roux limb several centimeters from the terminal staple line. The biliary drain is passed through the enterotomy into the bowel lumen several centimeters. Several stay sutures may be placed to approximate the bile duct and jejunum. The anastomosis is then sewn with fine (4-0 or 5-0) absorbable suture in interrupted fashion, taking care to approximate mucosa to mucosa (Fig 7). Following completion of the anastomosis, the Roux limb is anchored to the underside of the liver with several seromuscular interrupted sutures on either side, relieving any tension from the repair. A closed suction drain is left in the Morrison’s pouch and brought out through the skin through a stab incision. A brief comment should be made regarding traumatic injury and repair. Several studies have reported good results with primary repair of bile duct transections provided that there is a lack of tension and presence of viable tissue. In the case of tissue loss a bilioenteric bypass should be created in standard fashion. In instances of distal avulsion injuries involving the retropancreatic bile duct, the proximal duct should be managed in standard fashion, and no efforts should be made to expose the pancreatic side of the transected duct; these very distal ends do not produce enteric leaks.13

Postoperative Care Patients undergoing biliary reconstruction should be managed as with any upper gastrointestinal reconstruction, with diet resumed following resolution of postoperative ileus. Pa-

183 tients should be monitored with liver function tests. T tube or percutaneous biliary drains are internalized once diet is resumed. Operative drains should remain in place until PBD or T tube internalization and then, in the absence of output after internalization, should be removed. Drains are typically left across the anastomosis for 6 weeks, at which time a cholangiogram is performed. If the anastomosis appears widely patent the drain is removed at that time; if the anastomosis appears tight, the PBD or T tube is left longer. In extreme circumstances where the anastomosis was made to a strictured segment of the duct, the stent should be left in place for approximately 8 to 10 months to allow scar tissue to form. Complications after biliary reconstruction are not uncommon. In the largest single-center experience from Johns Hopkins, postoperative complications were experienced by 42% of patients and consisted most frequently of wound infection, biloma, biliary stent complication, or anastomotic leak.1 Of note, all complications were managed nonoperatively, and it has been the experience of other authors that, while complications such as cholangitis or PBD occlusion are not uncommon, they rarely require reoperation.30 Over the long term, patients must be monitored for stricture at the anastomosis, most commonly heralded by a rising alkaline phosphatase. As many of these patients have a chronic, low-level elevation in alkaline phosphatase, the more concerning feature is a consistent upward trend. In the event of suspected anastomotic stricture, cholangiogram and possible balloon dilation and stenting are preferred. If the bile duct is in continuity with the duodenum, ERCP and stenting is the preferred approach. In the case of jejunal reconstructions where the bile duct is not in communication with the duodenum, all interventions must be accomplished via percutaneous transhepatic access.

Conclusion In summary, bile duct injury is a rare complication of laparoscopic cholecystectomy. Pain or jaundice after laparoscopic cholecystectomy should prompt an evaluation of liver function tests (bilirubin and alkaline phosphatase) followed by computerized imaging to rule out extra-biliary tract bile. Bile duct imaging with MRCP or ERCP is required if the index of suspicion is high or if common duct appears obstructed. If the duct is intact but injured, ERCP plus stenting may control the process (after all extra-biliary bile is drained). If the injury occludes or divides the duct then bile flow should be controlled percutaneously (transhepatically). Reconstruction should be accomplished in a minimally inflamed field, most commonly with a Roux-en-Y hepaticojejunostomy. Successful operative repair requires a tension-free, mucosa-to-mucosa anastomosis using viable bile duct. These reconstructions are best performed by experienced hepatobiliary surgeons. While the complication rate including anastomotic stricture remains fairly high, most can be managed nonoperatively with a low reoperation rate and good long-term patency.

References 1. Sicklick JK, Camp MS, Lillemoe KD, et al: Surgical management of bile duct injuries sustained during laparoscopic cholecystectomy: Perioperative results in 200 patients. Ann Surg 241:786-792, 2005


184 2. Sawaya DE, Johnson LW, Sittig K, et al: Iatrogenic and noniatrogenic extrahepatic biliary tract injuries: A multi-institutional review. Am Surg 67:473-477, 2001 3. Saber K, El-Manialawi M: Repair of bile duct injuries. World J Surg 8:82-84, 1984 4. Hall JG, Pappas TN: Current management of biliary strictures. J Gastrointest Surg 8:1098-1110, 2004 5. Z’graggen K, Wehrli H, Metzger A, et al: Complications of laparoscopic cholecystectomy in Switzerland. A prospective 3-year study of 10,174 patients. Surg Endosc 12:1303-1310, 1998 6. Krähenbühl L, Sclabas G, Wente MN, et al: Incidence, risk factors, and prevention of biliary tract injuries during laparoscopic cholecystectomy in Switzerland. World J Surg 25:1325-1330, 2001 7. Flum DR, Cheadle A, Prela C, et al: Bile duct injury during cholecystectomy and survival in medicare beneficiaries. JAMA 290:2168-2173, 2003 8. Davidoff AM, Pappas TN, Murray EA, et al: Mechanism of major bile duct injuries during laparoscopic cholecystectomy. Ann Surg 215:196208, 1992 9. Meyers WC, Peterseim DS, Pappas TN, et al: Low insertion of hepatic segmental duct VII-VIII is an important cause of major biliary injury or misdiagnosis. Am J Surg 171:187-191, 1996 10. Bismuth H, Majno PE: Biliary strictures: Classification based on the principles of surgical treatment. World J Surg 25:1241-1244, 2001 11. Strasberg SM, Hertl M, Soper NJ: An analysis of the problem of biliary injury during laparoscopic cholecystectomy. J Am Coll Surg 180:101125, 1995 12. Branum G, Schmitt C, Baillie J, et al: Management of major biliary complications after laparoscopic cholecystectomy. Ann Surg 217:532540, 1993 13. Melton SM, McGwin G, Cross JM, et al: Common bile duct transection in blunt abdominal trauma: Case report emphasizing mechanism of injury and therapeutic management. J Trauma 54:781-785, 2003 14. Bauer TW, Morris JB, Lowenstein A, et al: The consequences of a major bile duct injury during laparoscopic cholecystectomy. J Gastrointest Surg 2:61-66, 1998 15. Ahrendt SA, Pitt HA: Surgical therapy of iatrogenic lesions of biliary tract. World J Surg 25:1360-1365, 2001 16. Stewart L, Way LW: Bile duct injuries during laparoscopic cholecystec-

17.

18.

19. 20.

21.

22.

23.

24.

25. 26. 27. 28. 29. 30.

tomy. Factors that influence the results of treatment. Arch Surg 130:1123-1128, 1995 Shinhar S, Nobel M, Shimonov M, et al: Technetium-99m-HIDA scintigraphy versus endoscopic retrograde cholangiopancreatography in demonstrating bile leaks after laparoscopic cholecystectomy. J Nucl Med 39:1802-1804, 1998 Lillemoe KD, Melton GB, Cameron JL, et al: Postoperative bile duct strictures: Management and outcome in the 1990s. Ann Surg 232:430441, 2000 Mercado MA: Early versus late repair of bile duct injuries. Surg Endosc 20:1644-1647, 2006 Robinson TN, Stiegmann GV, Durham JD, et al: Management of major bile duct injury associated with laparoscopic cholecystectomy. Surg Endosc 15:1381-1385, 2001 Misra S, Melton GB, Geschwind JF, et al: Percutaneous management of bile duct strictures and injuries associated with laparoscopic cholecystectomy: A decade of experience. J Am Coll Surg 198:218-226, 2004 Misra S, Melton GB, Geschwind JF: Percutaneous management of bile duct strictures and injuries associated with laparoscopic cholecystectomy: A decade of experience. J Am Coll Surg 198:218-226, 2004 Kassab C, Prat F, Liguory C, et al: Endoscopic management of postlaparoscopic cholecystectomy biliary strictures. Long-term outcome in a multicenter study. Gastroenterol Clin Biol 30:124-129, 2006 Gazzaniga GM, Filauro M, Mori L: Surgical treatment of iatrogenic lesions of the proximal common bile duct. World J Surg 25:1254-1259, 2001 Thompson BNJ, Parks RW, Madhavan KK, et al: Early specialist repair of biliary injury. Br J Surg 93:216-220, 2006 Mercado MA, Chan C, Orozco H, et al: Acute bile duct injury. Surg Endosc 17:1351-1355, 2003 Quintero GA, Patino JF: Surgical management of benign strictures of the biliary tract. World J Surg 25:1245-1250, 2001 Terblanche J, Allison HF, Northover JMA: An ischemic basis for biliary strictures. Surgery 94:52, 1983 Zollinger RMJ, Zollinger RMS: Zollinger’s atlas of surgical operations (8th ed). New York, McGraw Hill, 2003 Walsh RM, Vogt DP, Ponsky JL, et al: Management of failed biliary repairs for major bile duct injuries after laparoscopic cholecystectomy. J Am Coll Surg 199:192-197, 2004

Management of Intraoperative Complications

Management of Laparoscopic Surgical Complications

Oral Complications of Cancer and its Management

Management of Complications in Refractive Surgery

Management of Infectious Complications in Cancer Patients

Complications

Complications

Pediatric Urology: Surgical Complications and Management

Complications of Regional Anesthesia

Pediatric Cataract Surgery: Techniques, Complications, and Management

Complications of Dialysis

Intraoperative Neurophysiological Monitoring

Complications of Regional Anesthesia

Pitfalls in Cervical Spine Surgery: Avoidance and Management of Complications

unexpected complications

MELANOMA IN THE CLINIC – DIAGNOSIS, MANAGEMENT AND COMPLICATIONS OF MALIGNANCY

Complications of Urologic Surgery and Practice: Diagnosis, Prevention, and Management

Unexpected Complications

Unexpected Complications

Unexpected Complications

Complications of Gynecologic Endoscopic Surgery

Medical Complications of Kidney Transplantation

Complications of Urologic Laparoscopic Surgery

Complications of Urologic Laparoscopic Surgery

Intraoperative Neurophysiological Monitoring, Third Edition

Complications in Pediatric Surgery

Complications in Pediatric Otolaryngology

Infectious Pregnancy Complications

Complications In Pediatric Surgery

Complications in Cutaneous Surgery

Management of Intraoperative Complications

Management of Laparoscopic Surgical Complications

Oral Complications of Cancer and its Management

Management of Complications in Refractive Surgery

Management of Infectious Complications in Cancer Patients

Complications

Complications

Pediatric Urology: Surgical Complications and Management

Complications of Regional Anesthesia

Pediatric Cataract Surgery: Techniques, Complications, and Management

Complications of Dialysis

Intraoperative Neurophysiological Monitoring

Complications of Regional Anesthesia

Pitfalls in Cervical Spine Surgery: Avoidance and Management of Complications

unexpected complications

MELANOMA IN THE CLINIC – DIAGNOSIS, MANAGEMENT AND COMPLICATIONS OF MALIGNANCY

Complications of Urologic Surgery and Practice: Diagnosis, Prevention, and Management

Unexpected Complications

Unexpected Complications

Unexpected Complications

Complications of Gynecologic Endoscopic Surgery

Medical Complications of Kidney Transplantation

Complications of Urologic Laparoscopic Surgery

Complications of Urologic Laparoscopic Surgery

Intraoperative Neurophysiological Monitoring, Third Edition

Complications in Pediatric Surgery

Complications in Pediatric Otolaryngology

Infectious Pregnancy Complications

Complications In Pediatric Surgery

Complications in Cutaneous Surgery

Recommend Documents