Management of Infectious Complications in Cancer Patients

Management of Infectious Complications in Cancer Patients

Cancer Treatment and Research Steven T. Rosen, MD, Series Editor

Nathanson L (ed): Malignant Melanoma: Genetics, Growth Factors, Metastases, and Antigens. 1991. ISBN 0-7923-0895-6. Sugarbaker P H (ed): Management of Gastric Cancer. 1991. ISBN 0-7923-1102-7. Pinedo HM, Verweij J, Suit H D (eds): Soft Tissue Sarcomas: New Developments in the Multidisciplinary Approach to Treatment. 1991. ISBN 0-7923-1139-6. Ozols R F (ed): Molecular and Clinical Advances in Anticancer Drug Resistance. 1991. ISBN 0-7923-1212-0. Muggia FM (ed): New Drugs, Concepts and Results in Cancer Chemotherapy 1991. ISBN 0-7923-1253-8. Dickson RB, Lippman ME (eds): Genes, Oncogenes and Hormones: Advances in Cellular and Molecular Biology of Breast Cancer. 1992. ISBN 0-7923-1748-3. Humphrey G Bennett, Schraffordt Koops H, Molenaar WM, Postma A (eds): Osteosarcoma in Adolescents and Young Adults: New Developments and Controversies. 1993. ISBN 0-7923-1905-2. Benz CC, Liu ET (eds): Oncogenes and Tumor Suppressor Genes in Human Malignancies. 1993. ISBN 0-7923-1960-5. Freireich EJ, Kantarjian H (eds): Leukemia: Advances in Research and Treatment. 1993. ISBN 0-7923-1967-2. Dana BW (ed): Malignant Lymphomas, Including Hodgkin's Disease: Diagnosis, Management, and Special Problems. 1993. ISBN 0-7923-2171-5. Nathanson L (ed): Current Research and Clinical Management of Melanoma. 1993. ISBN 0-7923-2152-9. Verweij J, Pinedo HM, Suit H D (eds): Multidisciplinary Treatment of Soft Tissue Sarcomas. 1993. I SBN 0-7923-2183-9. Rosen ST, Kuzel TM (eds): Immunoconjugate Therapy of Hematologic Malignancies. 1993. ISBN 0-7923-2270-3. Sugarbaker PH (ed): Hepatobiliary Cancer. 1994. ISBN 0-7923-2501-X. Rothenberg ML (ed): Gynecologic Oncology: Controversies and New Developments. 1994. ISBN 0-7923-2634-2. Dickson RB, Lippman ME (eds.): Mammary Tumorigenesis and Malignant Progression. 1994. ISBN 0-7923-2647-4. Hansen H H (ed): Lung Cancer. Advances in Basic and Clinical Research. 1994. ISBN 0-7923-2835-3. Goldstein LJ, Ozols R F (eds.): Anticancer Drug Resistance. Advances in Molecular and Clinical Research. 1994. ISBN 0-7923-2836-1. Hong WK, Weber RS (eds.): Head and Neck Cancer. Basic and Clinical Aspects. 1994. ISBN 0-7923-3015-3. Thall P F (ed): Recent Advances in Clinical Trial Design and Analysis. 1995. ISBN 0-7923-3235-0. Buckner C D (ed): Technical and Biological Components of Marrow Transplantation, 1995. ISBN 0-7923-3394-2. Winter JN (ed.): Blood Stem Cell Transplantation. 1997. ISBN 0-7923-4260-7. Muggia FM (ed): Concepts, Mechanisms, and New Targets for Chemotherapy. 1995. ISBN 0-7923-3525-2. Klastersky J (ed): Infectious Complications of Cancer. 1995. ISBN 0-7923-3598-8. Kurzrock R, Talpaz M (eds): Cytokines: Interleukins and Their Receptors. 1995. ISBN 0-7923-3636-4. Sugarbaker P (ed): Peritoneal Carcinomatosis: Drugs and Diseases. 1995. ISBN 0-7923-3736-3. Sugarbaker P (ed): Peritoneal Carcinomatosis: Principles of Management. 1995. ISBN 0-7923-3727-1. Dickson RE, Lippman ME (eds.): Mammary Tumor Cell Cycle, Differentiation and Metastasis. 1995. ISBN 0-7923-3905-3. Freireich EJ, Kantarjian H (eds.): Molecular Genetics and Therapy of Leukemia. 1995. ISBN 0-7923-3912-6. Cabanillas F, Rodriguez MA (eds.): Advances in Lymphoma Research. 1996. ISBN 0-7923-3929-0. Miller AB (ed.): Advances in Cancer Screening. 1996. ISBN 0-7923-4019-1. Hait WN (ed.): Drug Resistance. 1996. ISBN 0-7023-4022-1. Pienta KJ (ed.): Diagnosis and Treatment of Genitourinary Malignancies. 1996. ISBN 0-7923-4164-3. Arnold AJ (ed.): Endocrine Neoplasms. 1997. ISBN 0-7923-4354-9. Pollock R E (ed.): Surgical Oncology. 1997. ISBN 0-7923-9900-5. Verweij J, Pinedo HM, Suit HD (eds.): Soft Tissue Sarcomas: Present Achievements and Future Prospects. 1997. ISBN 0-7923-9913-7. Walterhouse DO, Cohn SL (eds.): Diagnostic and Therapeutic Advances in Pediatric Oncology. 1997. ISBN 0-7923-9978-1. Mittal BB, Purdy JA, Ang KK (eds.): Radiation Therapy. 1998. ISBN 0-7923-9981-1. Foon KA, Muss H B (eds.): Biological and Hormonal Therapies of Cancer. 1998. ISBN 0-7923-9997-8. Ozols RF (ed.): Gynecologic Oncology. 1998. ISBN 0-7923-8070-3.

Management of Infectious Complication in Cancer Patients edited by

GARY A. NOSKIN Associate Professor of Medicine Northwestern University Medical School Medical Director, Infection Control Healthcare Epidemiology Northwestern Memorial Hospital Chicago, Illinois

KLUWER ACADEMIC PUBLISHERS BOSTON/DORDRECHT/LONDON

'w Ld

Distributors for North, Central and South America: Kluwer Academic Publishers 101 Philip Drive Assinippi Park Norwell, Massachusetts 02061 USA Distributors for all other countries: Kluwer Academic Publishers Group Distribution Centre Post Office Box 322 3300 AH Dordrecht, THE NETHERLANDS Library of Congress Cataloging-in-Publication Data Management of infectious complications in cancer patientsledited by Gary A. Noskin. cm. - (Cancer treatment and research; v. 96) p. Includes bibliographical references and index. ISBN 0-7923-8150-5 (alk. paper) 1. Infection - Treatment. 2. Cancer - Complications - Treatment. 3. Immunosuppression. I. Noskin, Gary A. 11. Series. RC112.M35 1998 616.99'4 - dc21 98-16254 CIP Copyright O 1998 by Kluwer Academic Publishers All rights reserved. No part of this publication may be reproduced, stored in a retrieval system or transmitted in any form or by any means, mechanical, photocopying, recording, or otherwise, without the prior written permission of the publisher, Kluwer Academic Publishers, 101 Philip Drive, Assinippi Park, Norwell, Massachusetts 02061

Printed o n acid-,free paper. PRINTED IN THE UNITED STATES OF AMERICA

Contents

Contributors . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . vii

......................................................

xi

1. Host impairments in patients with neoplastic diseases . . . . . . . . . . . BEN E. DE PAUW, J. PETER DONNELLY, and BART-JAN KULLBERG

1

2. Epidemiology of infectious complications in cancer patients. . . . . . TERESA ZEMBOWER

33

3. Approach to fever in the neutropenic host . . . . . . . . . . . . . . . . . . . . ATHENA STOUPIS and STEPHEN H. ZINNER

77

Preface

4. Infections associated with solid tumors . . . . . . . . . . . . . . . . . . . . . . . . 105 SARAH H. SUTTON and JOHN P. FLAHERTY

5. Role of the clinical microbiology laboratory in the diagnosis of infections . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 143 RICHARD B. THOMSON, JR. and LANCE R. PETERSON 6. Recent advances in the management of fungal infections. . . . . . . . . 167 JASON SANCHEZ and GARY A. NOSKIN 7. Recent advances in the management of viral infections. . . . . . . . . . 183 JOHN R. WINGARD 8. Cytokines and biological response modifiers in the treatment ofinfection . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 201 BRIGITTA U. MUELLER and PHILLIP A. PIZZO 9. Prevention of infection in immunocomprised hosts. GARY A. NOSKIN

. . . . . . . . . . . . . 223

10. Pharmacologic considerations with antimicrobials used inoncology . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 247 MICHAEL POSTELNICK and SARA R. HALBUR 11. Economic impact of infections in patients with cancer.. . . . . . . . . . 283 DAVID J. SHULKIN and LAWRENCE J. ANASTASI Index

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 291

Contributors

Lawrence J. Anastasi, DO University of Pennsylvania Medical Center 3400 Spruce Street Philadelphia, Pennsylvania 19104-4283 Ben E. De Pauw, MD, PhD Departments of Haematology and Blood Transfusion University Hospital Nijmegen Geert Grooteplein 8 PO Box 9101 6500 HB Nijmegen The Netherlands J. P. Donnelly, MD Department of Haematology University Hospital Nijmegen Geert Grooteplein 8 PO Box 9101 6500 HB Nijmegen The Netherlands John P. Flaherty, MD University of Chicago Hospitals 5841 S. Maryland Avenue Mail Code 5065 Chicago, Illinois 60637 Sara R. Halbur, Pharm D Department of Pharmacy Northwestern Memorial Hospital 250 E. Superior Street Chicago, Illinois 60611

B. J. Kullberg, MD Departments of Internal Medicine and Infectious Diseases University Hospital Nijmegen Geert Grooteplein 8 PO Box 9101 6500 HB Nijmegen The Netherlands Brigitta U. Mueller, MD Pediatric Branch National Cancer Institute Building 10, Room 13N240 Bethesda, Maryland 20892 Gary A. Noskin, MD Division of Infectious Diseases Northwestern University Medical School 710 N. Fairbanks Court-216 Chicago, Illinois 60611 Lance R. Peterson, MD Director, Clinical Microbiology Northwestern University Medical School 250 East Superior Street, Wesley 565 Chicago, Illinois 60611 Philip A. Pizzo, MD Pediatric Branch National Cancer Institute Building 10, Room 13N240 Bethesda, Maryland 20892 Michael Postelnick, R. Ph Department of Pharmacy Northwestern Memorial Hospital 250 E. Superior Street Chicago, Illinois 60611 Jason Sanchez, MD Department of Medicine Northwestern University Medical School 250 E. Superior Street Chicago, Illinois 60611

Athena Stoupis, MD Brown University School of Medicine Rhode Island Hospital Providence, Rhode Island 02903 David J. Shulkin, MD Chief Medical Officer University of Pennsylvania Medical Center 3400 Spruce Street Philadelphia, Pennsylvania 19104-4283 Sarah Sutton, MD MacNeal Hospital 3231 S. Euclid Avenue Suite 405 Berwyn, Illinois 60402 Richard B. Thompson, Jr., PhD Department of Pathology Northwestern University Medical School Evanston Hospital 2650 Ridge Road Evanston, Illinois John R. Wingard, MD Division of HematologylOncology University of Florida College of Medicine P.O. Box 100277 Gainesville, Florida 32610-0277 Teresa Zembower, MD Division of Infectious Diseases Northwestern University Medical School 303 E. Superior Street, 8E Chicago, Illinois 60611 Stephen Zinner, MD Brown University School of Medicine Rhode Island Hospital Roger Williams Medical Center 825 Chalkstone Avenue Providence, Rhode Island 02908-4728

Preface

Infection is a major cause of morbidity and mortality in patients with neoplastic disease because of compromised host defenses. These defects result in an increased risk of infection and its complications. The nature of the underlying malignancy, the immunodeficiencies associated with it, and the treatments directed against it are all important determinants of infection. In recent years the introduction of more intensive chemotherapeutic regiments and the widespread use of bone marrow and peripheral stem cell transplantation have changed the pattern of infection in many patients. Furthermore, the increasing use of central venous access devices and antimicrobial prophylaxis has changed the epidemiology of infection in these patients as well. To complicate matters, the microbiology of infections is continuously evolving. Whereas in the past the majority of serious infections were caused by gram-negative bacilli, this has changed to favor more gram-positive pathogens and fungi. This has resulted in the emergence of vancomycinresistant enterococci, glycopetide-insensitive Staphylococcus aureus, and Candida krusei. The objective of this volume in the Cancer Treatment and Research series is to emphasize that whereas the management of infection in cancer patients is common, it is constantly changing. With the increasing complexity of these patients, optimal management requires a multidisciplinary approach. Ultimately, it is hoped that this book will assist clinicians in the diagnosis, management, and prevention of infection in order to optimize care for patients with cancer.

1. Host impairments in patients with neoplastic diseases Ben E. De Pauw, J. Peter Donnelly, and Bart-Jan Kullberg

1. Introduction In the course of evolution nature has provided the normal human individual with an impressive and effective defense system against microbial enemies that eclipses even Star Wars, arguably the most advanced and ingenious defense program ever designed by human beings. On its own, the normal defense system recognizes foreign invaders, alerts the relevant protective mechanisms, launches counterattacks, ceases hostilities as soon as the job is done, and clears up the battlefield, causing only negligible collateral damage. An intact system offers protection against most microbial aggressors through a complex interrelationship of protecting surfaces, cells, and soluble factors. Optimal nutritional status and normal organ function form the basis of resistance to potentially dangerous microorganisms; therefore, it is somewhat artificial to further delineate the separate lines of defense [I] because all components are more or less dependent upon each other in attaining maximum efficacy. For instance, the skin and mucosal membranes are ranked amongst the first line of defense but they can only exert optimal activity in conjunction with the immunoglobulin (Ig) A and other secretory substances. Moreover, the surfaces of the human body exhibit a clear propensity to interact with colonizing microorganisms. The so-called commensal resident flora are normally avirulent, do not cause infection, and protect against more aggressive pathogens by competing for binding sites on the surfaces and for the available nutrients. Therefore, white blood cells (granulocytes, macrophages, and lymphocytes), platelets, soluble factors of the immunoglobulins, complement, lymphokines, and other cytokines, as well as the physical barriers, have to be considered as integral and virtually indispensable components of a unitary defense system (Figure 1). Given its complexity, it is not surprising that such a finely tuned system is subject to profound perturbation by therapeutic manipulation and hematologic malignancies. Any qualitative or quantitative defect in one of the components of the human defense system may predispose to infection, which remains a major Gary A . Noski11 (en), M A N A G E M E N T O F INFECTIOUS C O M P L I C A T I O N S IN C A N C E R PATIENTS. 0 1998. Kl~iwerAcadenllc Ptrblrshr~rs,Boston. All rights reserved.

THROMBOCYTES

MUCOUS MEMBRANES

MACROPHAGE

COMMENSAL FLORA

CELLULAR IMMUNITY

RGAN FUNCTION Figure I. Normal defense systems.

cause of morbidity and mortality among patients undergoing treatment for malignancy. However, isolated deficiencies are rarely encountered because malfunction of one part of the system exerts an impact on several other parts. Moreover, therapeutic interventions and the underlying disease conspire to afflict a range of defense mechanisms. The effects of the various different noxious events that occur while treating malignancy differ in severity as well as in primary targets (Figure 2). To complicate things further, hazardous events do not remain static but rather exert their impact dynamically as the degree of disturbance varies with time during or after a course of treatment (Figure 3). The human defense system is capable of coping with a tremendous number of insults before it finally begins to show the first signs of surrender. It should therefore be obvious to physicians that their activities put the entire defense system of patients at considerable risk. At the very least the patient should receive proper instruction on the prevention of infection and optimal hygiene should be maintained at all times. Physicians also have to select the most appropriate treatment regimen, thereby avoiding potentially dangerous medications, and to limit invasive procedures to those that are absolutely warranted. This complex interaction between host defenses and therapeutic modalities has a profound effect on patient outcome.

I SKIN PENETRATION:1

Figure 2. Factors influencing the human defense systems.

2. Basic clinical condition and organ function

2.1 Nutritional status Weight loss correlates inversely with survival in patients with cancer. This occurs whether or not intensive treatment is given because the integrity of host defenses can be endangered by the catabolic state induced by cachexia and malnutrition, resulting in a quantitatively deficient intake of calories and protein, with insufficient vitamin levels and trace metal concentrations [2,3]. Cachexia will be exacerbated by anorexia, chemotherapy-induced nausea and vomiting, gastrointestinal obstructions, and metabolic derangements. The final extent of the damage to the defense system depends upon the degree of cachexia and malnutrition. This may result in delayed wound healing, mucosal atrophy with a decrease in the secretions of lysozyme and secretory IgA, as well as impairment of both the classical and alternative complement pathways. A deficiency of vitamin A may also have a detrimental effect on the cellular immune system [4]. Furthermore, deficiencies of trace elements may undermine host defense in compromised patients. Zinc deficiency, as has been observed during total parenteral nutrition, generates a disturbed function of phagocytes and T cells,

CELLULAR IMMUNITY

HUMORAL IMMUNITY Figzlr-e 3. Evolution of impairment of defense systems after treatment for malignancy.

which can be neutralized by the addition of this mineral [5]. The in vitro microbicidal capacity of neutrophils and T-lymphocyte function are reduced in patients with iron deficiency, but it is uncertain whether this has any clinical significance. On the hand, it is well known that iron overload, a conceivable consequence of multiple blood transfusions, may lead to an increased susceptibility to infection, which is possibly related to a direct interaction between the iron available and the fungus Mucor. A phosphate deficit, which may occur during episodes of starvation and insufficient parenteral nutrition, is associated with a decrease in the chemotactic, phagocytic, and microbicidal functions of granulocytes in vitro, and clinically with bacterial and fungal infections [61. In elderly patients, the atrophy and dryness of the skin and mucosal membranes that occurs with age may lead to an increased susceptibility to infections. In addition, the primary and secondary humoral responses, as well as the oxidative metabolism of neutrophils and T-cell functions, decline with age, but the exact role of these regularly found abnormalities with regard to susceptibility to infection is unclear [7]. Concomitant chronic illnesses enhance the risk of infection in many patients. Even mild graft-versus-host disease is deleterious to the integument [8], and patients with a pre-existing immune disturbance, such as HIV infection or

a congenital immunodeficiency syndrome, are placed in double jeopardy. Much more common, however, is the detrimental effects of smoking, particularly in patients with primary lung tumors, due to colonization of their airways with virulent microorganisms and impaired clearance of secretions

PI.

Patients with poorly controlled diabetes mellitus are more likely to develop wound infections after all kinds of skin penetrations, and they frequently suffer from concurrent vascular disease and neuropathy. High concentrations of glucose in the urine and oral secretions promotes colonization by Candida species and other pathogens [lo]. Diabetes mellitus has been associated with notorious infections, such as rhinocerebral mucormycosis and malignant external otitis [Ill, which is not difficult to explain in view of several other aberrations that are associated with diabetes, such as impaired opsonization and decreased chemotactic activity of granulocytes and monocytes. Reduction of phagocytic adherence and defective phagocytosis, as well as bactericidal function of granulocytes, have been shown during episodes with high glucose concentrations and a low pH, putatively due to an impaired glucose metabolism of the phagocytes. A remarkable observation in this context is the relation between myeloperoxidase deficiency and serious fungal infections in patients with diabetes [12]. 2.2 Physiological and psychological status Psychological stress is thought to suppress host defense mechanisms. This general assumption has been corroborated by the observations that psychological stress has a negative influence on the function of T cells and NK cells. Indeed, stress appears to be connected with an increased risk of acute viral respiratory illness, a risk that is related to the amount of stress. This is most likely mediated by endogenous opioids, hormones from the hypothalamicpituitary-adrenal axis, catecholamines, and cytokines [13]. Tumors themselves may also predispose to infection by local organ dysfunction. In patients with solid tumors, obstruction of natural passages can lead to inadequate drainage of secretory or excretory fluids from nasal sinuses, bronchi, and bile ducts. Furthermore, tissue invasion may create connections between normally sterile spaces and the environment through disruption of epithelial surfaces. Examples include perforation of the esophagus by mediastinal tumors, invasive gynecological malignancies with local pelvic abscesses caused by gram-negative rods and anaerobes, skin ulcerations with cellulitis and deep soft-tissue infections, and invasion of the bowel wall by tumors of the lower gastrointestinal tract, resulting in bacteremia. Localizations in the central nervous system, spinal cord compression, and paraneoplastic neuropathy are associated with an increased risk of infection due to lethargy and, for instance, a diminished ability to cough and swallow, and incomplete emptying of the bladder [9]. Of course, hematologic malignancies are notorious for infectious com-

plications because the neoplasm resides within the immune system itself and interferes directly and indirectly with its function. In patients undergoing splenectomy, the risk that they will develop overwhelming sepsis at some time during their life is approximately 5 % . Encapsulated bacteria such as Streptococcus pneumoniae and Haemophilus in.uenzne are the prevalent pathogens, but Neisseria meningitidis and staphylococci are occasionally encountered [14]. Several factors might explain this well-established increased susceptibility to microbial infection. Encapsulated bacteria are able to elude phagocytosis because specific opsonizing antibodies are necessary for efficient phagocytosis. Furthermore, a reduced level of the complement factor properdin, which may lead to suboptimal opsonization, and a decrease in functional tuftsin have both been demonstrated after splenectomy [I].The spleen is the principal organ for eliminating particles that are not opsonized, and it is left to the macrophages that occupy strategic positions within the organ to remove them. The primary immunoglobulin response also takes places in the spleen, and low levels of circulating IgM have been observed after splenectomy in children. Because of the risk of pneumococcal infection after splenectomy, immunization with a polyvalent pneumococcal vaccine is recommended, preferably prior to splenectomy to ensure a better immune response during later life. However, the protection from vaccination is probably limited to several years, and, although vaccination has been shown to be effective, infection still may occur [15]. Therefore, for children it is recommended that vaccination be supplemented by antibiotic prophylaxis. In splenectomized adults suffering from a hematologic malignancy, patient-initiated treatment with oral amoxicillin at the onset of fever should be considered because the response to vaccines is usually suboptimal in patients with pre-existing immune deficiencies.

3. Integument and commensal microflora The integument comprises the skin, respiratory tract - including the nasal cavity, ears, and conjunctiva - the alimentary tract, and the genitourinary tract (Figure 4) and provides the first line of defense against microbial invasion. In physical terms the only difference between the skin and the other parts of the integument is that it is dry, whereas the others are bathed in mucins and therefore continually moist. Thus, while both surfaces are normally colonized with a variety of microorganisms, including many different genera of bacteria and yeasts, the range and number of species and the biomass associated with mucosal surfaces is much greater than those of the skin. However, the resident microbial flora of each surface play an integral role in helping maintain the function and integrity of these first lines of defense. Moreover, when intact and healthy, both the mucosa and skin are capable of resisting colonization with foreign or allochthonous organisms found in the immediate environment and

NASAL SECRETIONS 10%organlsmdrn DENTAL PLAQUE 10'' organism9 SALIVAlO' organlsrnstrnL

SKIN (axilla 8 groin) 10" organ~smskm~ SKIN (other sites) lo3 organ~srnslcrn; ONJUNCTIVA I00

Hypogammaglobulinemia; functional asplenia with decreased IgG and IgM levels for 6-8 months; decreased IgA levels for 1 year or longer If GVHD develops, immunologic function recovers more slowly and patients are susceptible to opportunistic infections for longer periods of time

VZV; S, pneumoniae; chronic viral hepatitides, such as HCV acquired in the early pre-engraftment period

HCV = hepatitis C virus; GVHD = graft-versus-host disease; CMV herpes simplex virus; PCP = Pnel~mocystiscarinii pneumonia; VZV

= =

cytomegalovirus; HSV varicella zoster virus.

=

The third time period, after day 100, is the late postengraftment period. It is characterized by low levels of circulating immunoglobulins. IgG and IgM levels may remain depressed for 6-8 months or longer, and IgA levels may remain depressed for at least 1 year. Patients who develop chronic GVHD recover their immunologic function more slowly and are susceptible to opportunistic infections for longer periods of time. Characteristic infections occurring after day 100 include VZV and bacteremic pneumococcal infection. Nearly 40% of bone marrow transplant patients develop VZV infection, and the median time of onset is 5 months after transplantation. The incidence of pneumococcal bacteremia relates to chronic GVHD-induced hyposplenism with loss of opsonizing antibodies to the encapsulated gram-positive organisms. After day 100, infection with hepatitis C virus that may have been acquired through transfusions during the first 3 weeks after transplantation can become clinically manifest [SO-1521.

5. Emerging pathogens All clinicians caring for cancer patients must be aware of the emerging or newly recognized pathogens that are increasingly affecting this patient population. Infections with multidrug-resistant organisms are becoming more and more common. Liberal use of broad-spectrum antibiotics, prophylactic antimicrobial therapy, and complacency in prescribing techniques underlies the development of these resistant organisms [89,153,154]. Furthermore, organisms that were previously considered commensal or nonpathogenic have been shown to cause serious infections in these immunocompromised hosts. An awareness of emerging pathogens is essential for the diagnosis and management of patients with underlying malignancies (Table 6).

5.1 Bncteria 5.1.1 Aerobic gram-positive bacteria. Enterococci have emerged as important nosocomial pathogens. Estimates suggest that enterococci play a causal role in 12% of all hospital-acquired infections. Two species of enterococci, Enterococcus faecalis and Enterococcus faecium, account for the majority of clinical infections [90]. Vancomycin resistance among enterococci was first described clinically in 1988 [I551 and has since become a global problem. Indeed, many strains are not only vancomycin resistant but are multidrugresistant, and some multidrug-resistant strains of E. faecium are untreatable. Risk factors for invasive infection in oncology patients include the total amount and duration of antibiotic treatment, and the number of days of ceftazidime use [91]. Montecalvo and colleagues described an outbreak of multidrug-resistant E. faecium bacteremia on an oncology unit. Of the five patients who had multiple positive blood cultures, four died [156]. A report by Noskin and associates suggests that once patients are colonized with vancomycin-resistant enterococci (VRE), this colonization can last indefinitely and can later lead to invasive infection [157]. Although several antibiotics and antimicrobial combinations have been tried, standard therapeutic regimens have not been defined. Therefore, therapy may require unconventional or investigational agents [158]. Streptococci have recently received renewed interest as an important pathogen in the neutropenic host [56,159,160]. Streptococcus mitis, a viridans streptococci, has been associated with sepsis and ARDS in leukemic patients. The emergence of infection with both Streptococcus pneumoniae and the viridans streptococci has been demonstrated in cancer patients receiving Auoroquinolone prophylaxis. Fluoroquinolones, such as ciprofloxacin, have little activity against these organisms, giving them a selective advantage [161]. Streptococcus pneumoniae are becoming increasingly resistant to penicillin, and many strains are multidrug-resistant. Researchers at Northwestern University in Chicago have reported that in 1996, 35% of their strains possessed intermediate resistance to penicillin while 11% were highly penicillin resistant

[162]. Outbreaks of penicillin-resistant pneumococci on oncology wards have recently been reported. Because of this, when life-threatening infections with pneumococci are encountered or suspected, many centers are using either vancomycin or ceftriaxone as first-line therapy. Leuconostoc species, previously regarded as a commensal, have been reported to cause bacteremia in patients with intravascular catheters. These organisms can be confused microbiologically with enterococci, viridans Table 6. Emerging pathogens Pathogen

Comments -

Bacteria Aerobic gram-positive bacteria VRE Streptococci Viridans streptococci S, pneunzoniae Le~lcorlostocspp.

Aerobic gram-negative bacteria 3. ~1zr~lfop1ziLia GNB-containing ESBLs P. cepacia, M. exforqriens, A. radiobacfer, 0. rrnthropi, A. xylosoxidarzs A. plitre,facier.rs

Capnocytophagia Anaerobic bacteria C. septicrlm C. tertizirn F. nuclenninz, L, br~ccalis Mycobacteria M. tcrberculosis M. avirtm complex M. fortuirrrm, M. chelonae M. haemophil~im

No standard therapy; may require unconventional or investigational agents Increased incidence noted in patients receiving fluoroquinolone prophylaxis Associated with sepsis and ARDS in leukemic patients Increasingly penicillin and multidrug-resistant: empiric vancomycin or ceftriaxone used for life-threatening infections Inherently resistant to vancomycin; variably sensitive to penicillin and first-generation cephalosporins Mucositis is a risk factor, treat with penicillin or vancomycin Multiply drug resistant; remains vancomycin susceptible p-lactam antibiotics rarely effective in vitro: empiric therapy should consist of vancomycin or clindamycin with or without gentamicin 61 % survival with antibiotics alone; 75% for antibiotics plus surgical resection Nortoriously multidrug resistant; usually resistant to all aminoglycosides Plasmid-mediated resistance to many penicillinsithirdgeneration cephalosporins Associated with catheter-related bacteremias Fulminant syndrome of overwhelming sepsis and DIC in immunocon~promisedhosts Risk factors are mucositis and neutropenia Necrotizing enterocolitis Perirectal cellulitis Part of the normal oral flora; mucositis and pharyngitis are risk factors Combined medical and surgical approaches may be necessary in multidrug-resistant M. tuberculosis A c c o u ~ ~for t s 27% of nontuberculous mycobacteria in cancer patients Most commonly cause catheter infections Optimal treatment regimens not well defined

Table 6. (contirzued) Pathogen Viruses VZV, HSV CMV HTLV-I HTLV-I1 HIV KSHV (HHV-8) Fungi Candida spp.

Aspergillrrs spp. Fzisarircnz spp., Scoplt lariopsis spp., P. boydii

Comments Treatment with foscarnet in acyclovir-resistant strains Treatment with ganciclovir or foscarnet Associated with T-cell NHL Associated with hairy cell leukemia Associated with B-cell NHL Herpes virus associated with Kaposi's sarcoma

C. albicans still predominates, although non-albicans species are increasing At some institutions, A . ,flavus is now more common than A. filrnigafrrs Severe, often fatal infection in neutropenic hosts; recovery of neutrophil count is required for a successful outcome

Crlrv~rlariaspp.. Bipolaris spp., Exserohilum spp., Alternaria spp., Aspergillus spp.

Fungal sinusitis, which may lead to brain abscess

T, beigelii

Disseminated disease may occur; treatment is difficult and relapse is common Severe, disseminated infection in neutropenic patients Catheter-related sepsis in patients receiving parenteral lipids: sensitive to amphotericin B and the azoles; catheter removal and discontinuation of lipids Wide, deep surgical debridement or cryosurgery in early-stage disease; medical therapy has been disappointing All have been associated with catheter infections

B. capitatus M. firrfrlr

E. jeanselrnei, E. pisciphila, E. spinifera, S. inJIntzrm

R. rubra, H. anorrzalrt, G. candidrm, S. cerevi~eae. Drechslera spp., P. parasitica, Acrenzonirrn? spp., P. filrinosa P. carirzii Protozoa and parasites T. gondii

Other B. quintana. B. henselue

P, wickerhamii

Increased incidence in patients with brain tumors attributed to intensive chemotherapy and high doses of corticosteroids Primary treatment should be followed by maintenance therapy to prevent relapse in chronically immu~~osuppressed patients No reliable palliative or curative therapy exists Very high mortality in immunocompromised hosts; thiabendazole is the only effective therapy Causative agents of bacillary angiomatosis and bacillary peliosis; doxycycline or erythromycin is the treatment of choice Surgical debridement plus amphotericin B have been used successfully

ARDS = adult respiratory distress syndrome; CMV = cytomegalovirus; DIC = disseminated intravascular coagulation: ESBLs = extended-spectrum beta-lactamases; GNB = gram-negative bacteria; HHV-8 = human herpes virus 8; HSV = herpes simplex virus; HTLV = human T-cell lymphotropic virus; KSHV = Kaposi's sarcoma herpes virus; NHL = non-Hodgkin's lymphoma; VRE = vancomycin-resistant enterococci; VZV = varicella zoster virus.

streptococci, or lactobacillus. It is important to differentiate Leuconostoc because this organism is inherently vancomycin-resistant and only variably sensitive to penicillins and the first-generation cephalosporins [163]. Stomatococcus nzucilaginosus is a slime-producing, gram-positive coccus that is found in the normal oral flora of humans [164]. McWhinney and coworkers recovered S. m~ucilaginosus from eight febrile, neutropenic patients. Seven of these patients had bacteremia and one had positive cerebrospinal fluid cultures resulting in a fatal meningitis. Four of these patients had proven infections with this organism, and all eight had mucositis attributable to chemotherapy. All but one of these isolates was sensitive to penicillin, and all were sensitive to vancomycin [165]. Prior to 1976, Corynebacterium spp., bacteria normally abundant on the skin and mucous membranes, rarely caused infections and were susceptible to most antibiotics. However, in 1976, four cases of sepsis at the National Institutes of Health caused by Corynebacterium jeikeium were reported [166]. This organism is highly antibiotic-resistant but is vancomycin susceptible. Since then C. jeikeium has been increasingly recognized to cause intravascular catheter-associated bacteremias in neutropenic patients with underlying hematologic malignancies [167]. It has also caused meningitis and transverse myelitis in one neutropenic patient [168]. Cases of primary cutaneous Bacillus cereus infections have occurred in neutropenic patients with cancer or aplastic anemia. In these patients, the lesions were vesicular or pustular, only occurred on the limbs, and arose in the spring or summer. They all responded to antibiotic therapy [169]. B. cereus can also cause more severe clinical syndromes, such as necrotizing fasciitis, pneumonitis, and meningitis. In vitro data suggest that (3-lactam antibiotics are rarely effective; therefore, empiric therapy for a suspected infection with Bacillus spp. should consist of vancomycin or clindamycin with or without an aminoglycoside [170]. Rhodococcus equi is an uncommon pathogen that has been reported to cause infection in patients with impaired cellular immunity. HIV infection is the most common predisposing risk factor; however, infection in patients with other forms of cellular immunodeficiency have also been described. R. equi is most frequently associated with a cavitary pneumonia, which may mimic a fungal infection or tuberculosis. In a study by Harvey and Sunstrum, the survival rate for patients receiving antibiotics alone was 61% compared with 75% for those receiving both antibiotics and surgical resection [171].

5.1.2 Aerobic gram-negative bacteria. Although the overall incidence of gram-negative infections is decreasing in cancer patients, the gram-negative aerobic bacteria still cause significant morbidity and mortality. The use of broad-spectrum prophylactic and ernpiric antibiotics targeting the gramnegative bacteria has also led to increased antimicrobial resistance among these organisms.

Stenotrophomonas (Xanthomonas) maltophilia is an organism that is frequently isolated from the environment, particularly from water supplies. Both colonization and infection among immunocomprornised patients is increasing, especially in those receiving broad-spectrum antibiotics, particularly imipenam. S. maltophilia causes pneumonia, urinary tract infections, bacteremia, and wound infections in debilitated patients and is notoriously multidrug-resistant, making treatment difficult [I721. Enterobacteriaceae are also becoming more drug resistant. For example, a study by the International Antimicrobial Therapy Cooperative Group (IATCG) of the EORTC between 1983 and 1993 showed that the number of patients receiving fluoroquinolones had increased from 1.4% to 45%. During this same period, E. coli resistance to these drugs increased from 0% between 1983 and 1990 to 27% between 1991 and 1993 [17]. Certain Enterobacteriaceae, namely, E. coli and Klebsiella spp., have developed extended-spectrum beta-lactamases (ESBLs), enzymes that render them resistant to many penicillins and cephalosporins, particularly the third-generation cephalosporins. Although first- and second-generation cephalosporins may appear susceptible in vitro, clinical failures have been observed when these antibiotics are used. Therefore, imipenem, meropenem, or possibly a filactaml(3-lactamase inhibitor combination such as piperacillinltazobactam is now the treatment of choice for patients who develop life-threatening infections with these organisms. This antibiotic resistance is plasmid mediated and thus can be spread from patient to patient, leading to nosocomial outbreaks [173-1751. Therefore, for patients infected or colonized with bacteria expressing ESBLs, contact isolation is appropriate. Several unusual gram-negative pathogens have been associated with catheter-related bacteremia in cancer patients. An outbreak of Pseudomonas cepacia occurred on an oncology ward related to a contaminated heparin solution that was used to flush central venous catheters [176]. Methylobacterium extorquens bacteremia occurred in three acute leukemics, two of whom had undergone bone marrow transplantation and one who was receiving consolidation chemotherapy [177]. Catheter-associated bacteremia has also been reported with Agrobacteriuvrz radiobacter. This organism has also been associated with peritonitis, urinary tract infection, and endocarditis [178]. The first case of Ochrobactrum anthropi infection was reported in a 3year-old girl undergoing chemotherapy for retinoblastoma [179]. The experience with Achromohacter xylosoxidans bacteremia at M . D. Anderson Cancer Center from 1983 to 1988 has been reviewed. During this period, 10 cancer patients had positive blood cultures for this organism. In four, the infection was believed to be catheter related. An additional four had associated gastrointestinal pathology, and in the remaining two the predisposing factor could not be determined. Interestingly, neutropenia did not appear to be a risk factor [180]. Alteromonas (Pseudornonas) putrefaciens is an unusual organism that causes two distinct syndromes of bacteremic infection. One syndrome is

associated with chronic lower extremity infection, is mild, and responds to antimicrobial therapy. The second, more fulminant syndrome is associated with severe underlying illnesses such as liver disease and malignancy. These patients are more likely to experience overwhelming sepsis and disseminated intravascular coagulation [181]. Capnocytophagia is a facultative anaerobe that constitutes part of the normal oral flora and is an unusual cause of systemic infection. Capnocytophagia bacteremia has been reported following autologous bone marrow transplantation for Hodgkin7sdisease. Infection followed pretreatment conditioning that was associated with severe oral mucositis and neutropenia [182].

5.1.3 Anaerobic bacteria. Traditionally, anaerobes have accounted for relatively few primary infections in cancer patients, except as part of mixed infections, such as necrotizing gingivitis, perianal cellulitis, or perirectal abscesses. However, recently, a few anaerobes have been recognized as important pathogens in these immunocompromised hosts. Two species of Clostridium, C. septicurn and C. terti~1n.1,can cause significant infections in cancer patients. C. septicum has been known to cause necrotizing enterocolitis, and C. tertiurn has been isolated in neutropenic patients with perirectal cellulitis or another presumed gastrointestinal tract source [183,184]. Fusobacterium nucleat~tnzwas recovered from a leukemic patient with ulcerative pharyngitis and nodular pulmonary infiltrates suggestive of septic emboli [185,186]. Leptotrichia buccalis, part of the normal oral flora, has caused bacteremia in neutropenic patients whose only predisposing factor was mucositis [1871.

5.1.4 Mycobacteria. Strains of Mycobacterium tuberc~llosis resistant to isoniazid and many other first-line agents have recently emerged. When infected with tuberculosis, the immunocompromised cancer patient is at increased risk for disseminated or miliary disease. The second-line agents that must be used, in many instances, are less effective; thus, combined medical and surgical approaches may be necessary in these patients. Mycobacteriurn avium complex (MAC) is most commonly seen in AIDS patients but has also been reported in patients with underlying malignancies. MAC accounts for 27% of nontuberculous mycobacterial infections in cancer patients, especially patients with lung cancer. The most common clinical presentation is pulmonary disease; however, dissemination can occur [188]. M. fortuiturn and M. chelonae, two rapidly growing strains of mycobacteria, are known to cause catheter-related infections. They cause disseminated disease less commonly than MAC, but often result in cellulitis, skin abscesses, or subcutaneous nodules. Treatment of the catheter infections requires antimicrobial therapy, catheter removal, and surgical excision if tunnel infection is present. For other skin and soft tissue infections, a combination of antimicrobial therapy and surgical debridement may be required [189,190].

Reports of M. haernophil~lminfection are increasing in lymphoma patients, bone marrow and renal transplant patients, and AIDS patients. Infection with this mycobacteria most commonly presents as cutaneous ulcerations, joint effusions, or osteomyelitis. Optimal treatment regimens are not well defined [I911.

5.2 Viruses Varicella-zoster virus, HSV, and CMV are well-known pathogens in cancer patients. However, prophylactic use of acyclovir and ganciclovir has led to both acyclovir and ganciclovir-resistant strains, which must then be treated with foscarnet [192]. Recently, cidofivir became available for therapy of CMV retinitis in AIDS patients [193,194]. Retroviruses have been increasingly recognized as important pathogens in cancer patients over the past decade. Human T-cell leukemia virus-I (HTLV-I) is associated with adult T-cell non-Hodgkin7slymphoma, HTLV-I1 with hairy cell leukemia, and HIV with adult B-cell non-Hodgkin's lymphoma. Recently, a member of the gamma-herpesvirus family, referred to as Kaposi's sarcoma-associated herpesvirus (KSHV) or human herpesvirus 8 (HHV-8), has been associated with four types of Kaposi's sarcoma: classic KS, AIDS-associated KS, post-transplantation KS, and KS that occurs in endemic areas for the disease. It has also been associated with body-cavity-based lymphomas [195,196].

5.3 Fungi Candida spp. and Aspergillus spp. are the most common fungal infections in immunocompromised cancer patients, and recently several important trends have been noted. The incidence of nosocomial candidal bacteremia rose sharply in the late 1980s and early 1990s. At some institutions, Candidn bacteremia now surpasses that of the Enterobacteriaceae, Pseudornonas spp., and Entvrococc~lsspp. [197]. Candida albicans is still the most common species, accounting for more than half of fungal isolates from cancer patients, although the incidence of non-albicans species is increasing. Among these are C. tropicalis, C. parpsilosis, C. krusei, and C. (Torulopsis) glnbmta [198,199]. Central venous catheters, total parenteral nutrition, and the increasing use of azoles for antifungal prophylaxis are some of the presumed mechanisms thought to account for this rising trend. Of the Aspergillus species, Aspergillus fumigatus is the most commonly isolated species to cause invasive disease. However, at some institutions, A. flavus has supplanted A. fumigatus as the most common cause of aspergillosis [200]. Many fungi are ubiquitous in nature and were thought, until recently, to be commensal or nonpathogenic. Several of these have now been proven to cause serious infections in immunocompromised cancer patients. Fusarium spp., Scopulariopsis spp., and Pseudallescheria boydii, members of the

hyalohyphomycoses group, are important pathogens. Fusarium causes severe, often fatal, infections in neutropenic patients, particularly those who have undergone bone marrow transplantation. Fusariurn is highly resistant to conventional antifungal drugs, and rising neutrophil counts are required for a successful response. Subsequent neutropenic episodes are associated with a high incidence of recurrence [201,202].Scopulariopsis spp. and P. boydii have likewise been shown to cause serious disseminated disease in these patients [200]. Among the phaeohyphornycoses, Cttrvt~laria spp., Bipolaris spp., Exserohilum spp., and Alternaria spp. are known pathogens. These darkwalled molds are responsible for allergic fungal sinusitis. Patients may present with an indolent onset of sinus pain or painless proptosis. These infections can extend from the ethmoid or frontal sinuses into the frontal lobe causing brain abscesses. Differentiation of these fungi from Aspergillus is important. Other clinical syndromes, such as subcutaneous abscesses, cutaneous granuloma, disseminated disease, pneumonitis, prosthetic valve endocarditis, osteomyelitis, and septic arthritis, have been reported [203]. Trichosporon beigelii (T. cutaneum) can be part of the normal human flora. Patients with leukemia or bone marrow aplasia may develop disseminated trichosporonosis. In patients with fungemia due to T. beigelli, multiple red papular skin lesions may develop. An infection of hair shafts, known as white piedra, has also been described. Treatment of these infections is difficult and relapse is common [204-2061. Blastoschizornyces capitatus (Trichosporon capitatum) has caused severe, disseminated infection in at least 25 patients, 19 of whom were neutropenic. Six of these patients had papulonecrotic skin lesions similar to those seen in disseminated candidiasis in leukemic patients [207]. Malassezia furfur, a lipophilic yeast, colonizes normal skin and is also the causative agent of tinea versicolor. In cancer patients, it has been associated with catheter-related sepsis in patients receiving parenteral lipids. In vitro, it is susceptible to amphotericin B and irnidazoles. Therapy requires both catheter removal and discontinuation of parenteral lipids [208]. The dematiaceous soil fungi, Exophiala jeanselrnei, E. pisciphila, E. spinifera, and Scedosporiurn inflat~~rn, have caused infection in neutropenic patients [209]. In particular, S. inflaturn has been associated with catheterrelated fungemia, retinal lesions, esophagitis, and hepatosplenic infections [210]. In vitro, these organisms are resistant to amphotericin B and fluconazole. In early stage disease, wide and deep surgical excision or cryosurgery are most effective. Medical therapy has been disappointing. In small series, long-term itraconazole therapy has shown some success; however, relapse after treatment has been reported [211]. Several other fungi have been associated with infections in immunocompromised patients. Some of these include Rhodotorula rubra, Hansen~tla anomala, Geotrichurn candidurn, Saccharornyces cereviseae, Drechslera spp., Phialophora parasiticia, Acrernoniurn spp., and Pichia farinosa. Infections

with these organisms are primarily associated with central venous catheters [212,213]. Recently, Pneumocystis carinii has been reclassified as a fungus based on DNA homology. This organism results in pneumonia and is the most common opportunistic infection in HIV-infected patients. Although infrequent, the incidence is increasing in patients with underlying malignancies. In the 1970s, most cancer patients who developed Pneumocystis carinii pneumonia (PCP) had hematologic malignancies. However, in the 1980s the incidence of PCP increased most among patients with solid organ tumors. Indeed, 31% of patients who developed PCP had primary or metastatic brain tumors. The increase in patients with brain tumors has been attributed to more intensive chemotherapeutic regimens and the use of higher doses of corticosteroids [45,214]. The concomitant decrease in patients with hematologic malignancies and those undergoing bone marrow transplantation reflects the use of PCP prophylaxis in this patient population.

5.4 Protozoa and parasites

Toxoplasma gondii is an intracellular protozoan parasite. Reactivation of 7'. gondii has been reported in cancer patients. In fact, more than one third of the cases of toxoplasmosis in non-HIV patients occurs in patients with Hodgkin's disease or leukemia. Toxoplasmosis is associated with corticosteroid use and can present with pneumonitis, uveitis, or central nervous system disease. The diagnosis of reactivation disease includes an assay for specific toxoplasma IgG antibodies. However, in an immunocompromised patient, the IgG titer is often low and IgM assays may be negative. In this situation, a definitive diagnosis of toxoplasmosis depends on the demonstration of the tachyzoites on tissue histology or on specific T. gondii DNA by the polymerase chain reaction (PCR). Chronic (latent) asymptomatic infection in immunodeficient patients is not treated; however, immunodeficient patients with acute infection should always be treated. The first-line therapy includes pyrimethamine and folinic acid plus either sulfadiazine or clindamycin for 4-6 weeks after resolution of all signs and symptoms of disease (often for 6 months or longer). For chronically immunosuppressed patients, particularly AIDS patients, acute treatment is followed by life-long maintenance therapy to prevent relapse of disease [215]. Cryptosporidiurn is an intracellular protozoan parasite first described in 1907 [216]. The first cases of human diarrhea were described in 1976, and today it is most commonly associated with diarrhea in patients with HIV [217]. It has also been detected with increasing frequency in patients with underlying hematologic malignancies, those who have undergone bone marrow transplantation, and those who are receiving corticosteroid therapy. In one review of 20 patients with hematologic malignancies and cryptosporidiosis, 5 patients had severe diarrhea, 10 had moderate diarrhea, and 5 were asymptomatic carriers. Extraintestinal cryptosporidiosis with pulmonary involvement was

observed in one case. Resolution occurred spontaneously, although relapse was common, occurring in four patients [218]. Diagnosis is typically made by recognizing the characteristic oocysts in the material sampled, usually stool, duodenal aspirates, bile, or respiratory secretions. There is no reliable palliative or curative treatment for cryptosporidiosis. In immunocompetent or temporarily immunocompromised patients, the disease is self-limited. In patients who are irreversibly immunocomprornised, therapy is symptomatic and supportive. Strongyloides stercoralis is a nematode that can cause an overwhelming fatal infection in immunocompromised patients. S. stercoralis is increasingly recognized in patients with underlying hematologic malignancies, in patients treated with corticosteroids, and in patients with HIV. Humans are usually infected through skin contact with soil containing the infective filariform larvae or through autoinfection via the lower GI tract or perianal region by larvae that transform into infective organisms during their passage with feces. After infection and tissue invasion, the larvae enter the bloodstream and travel to the lungs. Here they break into the alveolar spaces and ascend to the glottis, where they are swallowed to reside in the small intestines. Contrary to most other worm infections, the patient's worm burden in strongyloidiasis is dependent not only on the size of the larval inoculum but also on the degree of autoinfection, which may be enhanced in immunocompromised hosts. This "autoinfection" cycle can result in an overwhelming larval invasion seen in strongylides hyperinfection syndrome, and these patients can develop massive larval invasion of the lungs and other tissues. This syndrome of hyperinfection strongyloidiasis is characterized by severe generalized abdominal pain, diffuse pulmonary infiltrates, ileus, shock, and meningitis or sepsis from gram-negative bacilli. Eosinophilia may be absent in the immunocompromised host. Definitive diagnosis is made by demonstrating the larvae in feces or in duodenal fluid. Thiabendazole is the only effective therapy; however, mortality remains very high in immunocomprornised patients. Therefore, patients with a past history of exposure to S. stercoralis should be thoroughly examined and treated prior to receiving any immunosuppresive therapy [219-2241.

5.5 Other pathogens 5.5.1 Bartonella spp. Bartonella (formerly Rickettsia) quintana and Bartonella (formerly Rochalimaea) henselae are slow-growing, motile, curved, gram-negative bacilli that have recently been recognized as human pathogens. Originally described in HIV patients, they have now been documented in cancer patients, including those with neutropenia. Both B. quintana and B. henselae can cause bacteremia or localized tissue infections, such as bacillary angiomatosis or bacillary peliosis. Bacillary angiomatosis was originally described in HIV patients as a neovascular proliferative disorder involving the

skin and regional lymph nodes. Characteristically, the lesions are cutaneous, can number from few to hundreds, bleed copiously when incised, and may resemble Kaposi's sarcoma, from which they need to be differentiated. Bacillary angiomatosis involving the liver, spleen, bone, and brain has also been documented. Bacillary peliosis is a distinct clinical syndrome that is less dramatic than bacillary angiomatosis because its lesions are exclusively visceral and are only associated with nonspecific symptoms. These organisms can be detected in blood cultures; however, they require prolonged periods of incubation. When they are suspected, the laboratory should be instructed to hold the blood cultures for up to 30 days. These organisms can also be detected in tissue biopsy specimens by Warthin-Starry staining or by electron microscopy. The initial treatment of choice is either oral doxycycline or oral erythromycin, often for prolonged periods of time. Tetracycline, minocycline, chloramphenicol, azithromycin, and clarithromycin have also been used successfully. Relapses can occur, and chronic suppressive therapy with doxycycline or erythromycin may need to be considered in these cases [225].

5.5.2 Prototheca wickerhamii. This unicellular, achloric algae is found in a wide range of environmental sites and has been isolated as the cause of infection in at least 61 cases. One was a child with Hodgkin's disease who had a central venous catheter in place and had P. wickerhamii isolated from the bloodstream [226]. The typical clinical presentation is a single lesion of the skin or subcutaneous tissue that gradually enlarges over weeks to months and may ulcerate. Characteristically, these lesions do not heal spontaneously. P. wickerhnnzii is resistant to fluconazole. Surgical debridement and amphotericin B have been used successfully [227-2291.

6. Summary Patients with underlying malignancies are at risk for a wide array of infectious diseases that cause significant morbidity and mortality. To develop a clear etiologic understanding of the infectious agents involved first requires a knowledge of the factors that predispose to infection. Neutropenia is clearly the single most important risk factor for infection in the cancer patient. However, a variety of both host and treatment-associated factors act together to predispose these patients to opportunistic infections. Approaching the individual malignancies with a knowledge of the underlying risk factors helps logically guide diagnosis and therapy. The astute clinician must also be aware of new and emerging infections in this patient population. As new pathogens are discovered and established pathogens become increasingly drug resistant, they will continue to present challenges for physicians caring for these patients in the years ahead.

References 1. Chang H, Rodriguez V, Narboni G, et al. Causes of death in adults with acute leukemia. Medicine 1976;55:259-268. 2. Hughes WT. Fatal infections in childhood leukemia. Am J Dis Child 1971;122:283-287. 3. Hersh EM, Bodey GP, Nies BA, et al. Causes of death in acute leukemia. JAMA 1965;193:99-103. 4. Morrittu L, Earl HM, Souhami RL, et al. Patients at risk of chemotherapy-associated toxicity in small cell lung cancer. Br J Cancer 1989;59:801-804. 5. Viscoli C, Castagnola E. Factors predisposing cancer patients to infection. In: Klastersky J, ed. Infectious Complications of Cancer. Boston: Kluwer Academic, 1995, pp. 1-30. 6. Bodey GP, Buckley M, Sathe YS, et al. Quantitative relationships between circulating leukocytes and infection in patients with acute leukemia. Ann Intern Med 1966;64:328340. 7. Feld R, Bodey GP. Infections in patients with malignant lymphoma treated with combination chemotherapy. Cancer 1977;39:1018-1025. 8. Glauser MP, Zinner SH. Mechanisms of acquisition and development of bacterial infections in cancer patients. In: Klastersky J, ed. Infections in Cancer Patients. New York: Raven Press, 1982, pp. 13-30. 9. De Jongh CS, Joshi JH, Newn~anKA, et al. Antibiotic synergism and response in gramnegative bacteremia in granulocytopenic cancer patients. Am J Med 1986;80:96-100. 10. Wade JC, Schimpff SC. Epidemiology and prevention of infection in the compromised host. In: Rubin RH, Young LS, eds. Clinical Approach to Infection in the Compromised Host. New York: Plenum Medical, 1988, pp. 5-40. 11. Dale DC, Guerry DI, Wewerka JR, et al. Chronic neutropenia. Medicine 1979;58:128-144. 12. Viscoli C, Bruzzi P, Castagnola E, et al. Factors associated with bacteraemia in febrile, granulocytopenic cancer patients. Eur J Cancer 1994;30A:430-437. 13. The EORTC International Antimicrobial Therapy Pro,ject Group. Three antibiotic regimens patients with cancer. J Infect Dis in the treatment of infection in febrile gra~~ulocytopenic 1978;137:14-29. 14. Schimpff SC, Young VM, Greene WH, et al. Origin of infection in acute nonlymphocytic leukemia: Significance of hospital acquisition of potential pathogens. Ann Intern Med 1972;77:707-714. 15. The EORTC International Antimicrobial Therapy Project Group. Ceftazidime combined with a short or long course of arnikacin for empirical therapy of gram-negative bacteremia in cancer patients with granulocytopenia. N Engl J Med 1987;317:1692-1698. 16. Johanson WG, Pierce AK, Sanford JP. Changing pharyngeal bacterial flora of hospitalized patients. Emergence of gram-negative bacilli. N Engl J Med 1969;281:1137-1140. 17. Cometta A, Calandra T, Bille J, et al. Escherichin coli resistant to fluoroquinolones in patients with cancer and neutropenia. N Engl 3 Med 1994;330:1240-1241. 18. Levine SS, Schimpff SC, Graw RG Jr., et al. Hematologic malignancies and other marrow failure states: Progress in the management of complicating infections. Semin Hematol 1974;11:141-202. 19. Chanock S. Evolving risk factors for i~~fectious con~plicationsof cancer therapy. Hematol Oncol Clin North Am 1993;7:771-793. 20. Koll BS, Brown AE. Changing patterns of infection in the immunocompromised patient with cancer. Hematol Oncol Clin North Am 1993;7:753-769. 21. Guiot HFL, Fibbe WE, van't Wout JW. Risk factors for fungal infection in patients with malignant hematologic disorders: Implications for empirical therapy and prophylaxis. Clin Infect Dis 1993;18:525-532. 22. Warren JR, Scarpelli DG, Reddy JK, et al. Chronic inflammation and repair. In: Essentials of General Pathology. New York: Macmillan, 1987, pp. 27-46. 23. Johnston RB, Jr. Monocytes and macrophages. N Engl J Med 1988;318:747-752.

24. Young RC, Corder MP, Haynes HA, et al. Delayed hypersensitivity in Hodgkin's disease. A study of 103 untreated patients. Am J Med 1972;52:63-72. 25. Seligmann M, Chess L, Fahey JL, et al. AIDS - An immunologic reevaluation. N Engl J Med 1984;311:1286-1292. 26. Aubertin J, Lacut JY, Hoerni B, et al. Factors favoring infections in cancer patients. In: Armstrong D, ed. Opportunistic Infections in Cancer Patients. New York: Masson, 1978, pp. 16-28. 27. Krick J, Remington J. Opportunistic fungal infection in patients with leukemia and Iymphoma. Clin Hematol 1976;5:249-310. 28. Lum LG. Immune recovery after bone marrow transplantation. Bone Marrow Transplant Hematol Oncol Clin North Am 1990;4:659-675. 29. Whittaker K, Rees K, Clark CG. Reduced lymphocyte transformation in breast cancer. Lancet 1971;1:892-893. 30. Han T, Takita H. Immunologic impairment in bronchogenic carcinoma: A study of lymphocyte response to phytohemagglutinin. Cancer 1972;30:616420. 31. Brooks WH, Netsky MG, Normansell DE, et al. Depressed cell-mediated immunity in patients with primary intracranial tumors. J Exp Med 1972;136:1631-1647. 32. Orita K, Miwa H, Fukada H, et al. Preoperative cell-mediated immune status of gastric cancer patients. Cancer 1976;38:2343-2348. 33. Schellhammer PF, Bracken RB, Bean MA, et al. Immune evaluation with skin testing. A study of testicular, prostate and bladder neoplasms. Cancer 1976;38:149-156. 34. Mackall CL, Fleisher TA, Brown MR, et al. Lymphocyte depletion during treatment with intensive chemotherapy for cancer. Blood 1994;84:2221-2228. 35. Mackall CL, FIeisher TA, Brown MR, et al. Age, thymopoiesis, and CD4+ T-lymphocyte regeneration after intensive chemotherapy. N Engl J Med 1995;332:143-149. 36. Kim JH, Perfect JR. Infection and cyclosporine. Rev Infect Dis 1989;11:677-690. 37. Bishop JF, Schimpff SC, Diggs CH, et ai. Infections during intensive chemotherapy for nonHodgkin's lymphoma. Ann Intern Med 1981;95:549-555. 38. Bookman MA, Longo DL. Concomitant illness in patients treated for Hodgkin's disease. Cancer Treat Rev 1986;13:77-111. 39. Hardy I, Gershon AA, Steinberg SP, et al. The incidence of zoster after immunization with Iive attenuated varicella vaccine. N Engl J Med 1991;325:1545-1550. 40. Hughes WT, Townsend TR. Nosocomial infections in immunocompromised children. Am J Med 1981;70:412-416. 41. Ampel NM, Wing EJ. Legionellosis in the compromised host. In: Rubin RH, Young LL, eds. Clinical Approach to Infection in the Compromised Host. New York: Plenum Medical, 1988, pp. 305-319. 42. Luft BJ, Remington JS. Toxoplasmosis of the central nervous system. Curr Clin Top Infect Dis 1985;6:315-358. 43. Longworth DL, Weller PF. Hyperinfection syndrome with strongyloidiasis. Curr Clin Top Infect Dis 1986;7:1-26. 44. Varthalitis I, Meunier F. Pneurnocystis carinii pneumonia in cancer patients. Cancer Treat Rev 1993;19:387-413, 45. Sepkowitz KA, Brown AE, Telzak EE, et al. Pneumocystis carinii pneumonia among patients without AIDS at a cancer hospital. JAMA 1992;267:832-837. 46. Sepkowitz KA. Pneumocystis carinii pneumonia in patients without AIDS. Clin Infect Dis 1993;17(Suppl. 2):S416-S422. 47. Hughes WT, Feldman S, Aur RJA, et al. Intensity of immunosuppressive therapy and the incidence of Ptzerlmocystis carinii pneumonitis. Cancer 1975;36:2004-2009. 48. Salmon SE. In~munoglobulinsynthesis and tumor kinetics of multiple myeloma. Semin Hematol 1973;10:135-147. 49. Fahey JL, Scoggins R, Utz JP, et al. Infection, antibody response and gamma globulin components in multiple myeloma and macroglobulinemia. Am J Med 1963;35: 698-707.

50. Jacobson DR, Zolla-Pazner S. Immunosuppression and infection in multiple myeloma. Semin Oncol 1986;13:282-290. 51. Mordasini RC, Keller H. Schlumpf E, et al. Humorale Infektabwherschwache bei paraproteinamischen Erkrankungen. Schewiz Med Wschr 1972:102:625-635. 52. Homberg JC, Cartron J, Ropars C, et al. Anomalies immunologiques observees au cours des leucemies lymphoides chroniques hyperlymphocytaires. Nouv Rev Fr Hematol1971;11:476483. 53. Sheridan JF, Tutschka PJ, Sedmak DD, et al. Immunoglobulin G subclass deficiency and pneumococcal infection after allogeneic bone marrow transplantation. Blood 1990;75:15831586. 54. Atkinson K. Reconstruction of the haemopoietic and immune systems after bone marrow transplantation. Bone Marrow Transpl 1990;5:209-226. 55. Axelrod PI, Lorber B, Vonderheid EC. Infections complicating mycosis fungoides and Sezary syndrome. JAMA 1992;267:1354-1358. 56. Awada A, van der Auwera P, Meunier F. et al. Streptococcal and enterococcal bacteremia in patients with cancer. Clin Infect Dis 1992;15:33-48. 57. Hosea SW, Brown EJ, Hamburger MI, et al. Opsonic requirements for intravascular clearance after splenectomy. N Engl J Med 1981;304:245-250. 58. Van Der Meer JWM. Defects in host-defense mechanisms. In: Rubin RH, Young LS, eds. Clinical Approach to Infection in the Compromised Host. New York: Plenum Medical, 1988, pp. 41-63. 59. Gopal V, Bison AL. Fulminant pneumococcal infections in "normal" asplenic hosts. Arch Intern Med 1977;137:1526-1530. 60. Advisory Committee on Itnmunization Practices. Pneumococcal polysaccharide vaccine. MMWR 1989;38:64-68, 73-76. 61. Kahls P, Kier P, Lechner K. Functional asplenia after bone marrow transplantation. Ann Intern Med 1990;114:805-806. 62. Schimpff SC. Infections in the cancer patient - Diagnosis, prevention, and treatment. In: Mandell GL, Bennett JE, Dolin R, eds. Principles and Practices of Infectious Diseases. New York: Churchill Livingstone, 1995. pp. 2666-2675. 63. Schwartz SN, Dowling JN. Benkovic C, et al. Sources of gram-negative bacilli colonizing the trachea of intubated patients. J Infect Dis 1978:138:227-231. 64. Kelly CP, Pothoulakis C, Lamont JT. Clostridi~~rn dificle colitis. N Engl J Med 1994;330:257262. 65. Anand A, Glatt A. Clostridiurn dificle infection associated with antineoplastic chemotherapy: A review. Clin Infect Dis 1993;17:109-113. 66. Fan-Havard P, Capano D, Smith SM, et al. Development of resistence in Cnndida isolates from patients receiving prolonged antifungal therapy. Antimicrob Agents Chemother 1978:35:2302-2305. 67. Meunier F, Aoun M. Bitar N. Candidemia in immunocompromised patients. Clin Infect Dis 1992;14(Suppl. 1):S120-S125. 68. Boken DJ, Swindells S, Rinaldi MG. Fluconazole-resistant Candidn albicans. Clin Infect Dis 1993;17:1018-1021. 69. Aisner J, Schmipff SC, Sutherland JC, et al. Torz~lopsisglabrata infections in patients with cancer. Am J Med 1976;61:23-28. 70. Merz WG, Karp JE, Schron D, et al. Increased incidence of fungemia caused by Cnndida krusei. J Clin Microbiol 1986;24:581-584. 71. Wingard JR, Merz WG, Rinaldi MG, et al. Increase in Cal~didakrrrsei infection among patients with bone marrow transplantation and neutropenia treated prophylactically with fluconazole. N Engl J Med 1991;325:1274-1277. 72. van der Waaij DD, Berghuis J, Lekkerkerk JEC. Colonization resistance of the digestive tract of mice during systemic antibiotic treatment. J Hyg (Camb) 1972;70:605-610. 73. Wilson BD, Surgalla MJ. Yates JW. Aerobic and anaerobic surgical wound contamination in patients with cancer. Surg Gpnecol Obstet 1974;139:329-332.

74. Rolandelli R, Roslyn JJ. The colon and rectum. Surgical anatomy and operative procedures. In: Sabiston DCJ, Lyerly HK, eds. Textbook of Surgery. The Biological Basis of Modern Surgical Practice. Philadelphia: W.B. Saunders, 1997, pp. 971-975. 75. Gongaware RD, Slanetz CAJ. Hartmann procedure for carcinoma of the sigmoid and rectum. Ann Surg 1973;178:28-30. 76. Naito H, Toya S, Shizawa H, et al. High incidence of acute post operative meningitis and septicemia in patients undergoing craniotomy with vertriculoatrial shunt. Surg Gynecol Obstet 1973;137:810-812. 77. Geelhoed GW, Ketcham AS. Pseudomonas meningitis complicating radical resection for radiorecurrent cancer of paranasal sinuses: Report of two patients successfully treated with intrathecal polymyxin. J Surg Oncol 1973;5:365-374. 78. Harris M, Gumport S, Berman I, et al. Ilioinguinal lymph node dissecton for melanoma. Surg Gynecol Obstet 1973;136:33-39. 79. Villar HV, Warneke JA, Peck MD, et al. Role of surgical treatment in the management of complications of the gastrointestinal tract in patients with leukemia. Surg Gynecol Obstet 1987;165:217-222. 80. Wade DS, Nava HR, Douglass HOJ. Neutropenic enterocolitis. Clinical diagnosis and treatment. Cancer 1992;69:17-23. 81. Schimpff SC, O'Connel MJ, Greene WH, et al. Infections in 92 splenectomized patients with Hodgkin's disease. Am J Med 1975;59:695-701. 82. Say CC, Donegan W. A biostatistical evaluation of complications from mastectomy. Surg Gynecol Obstet 1974;138:370-376. 83. Smith MGM, Golding PL, Eddleston ALWF, et al. Cell-mediated immune responses in chronic liver diseases. Br Med J 1972;1:527-530. 84. Castagnola E. Garaventa A , Conte M, et al. Survival after fungemia due to Frlsariurn monilifornte in a child with neuroblastoma [letter]. Eur J Clin Microbiol Infect Dis 1993;12:308-309. 85. Wade JC, Schimpff SC, Hargadon MT. et al. A comparison of trimethoprimsulfamethoxazole plus nystatin with gentamicin plus nystatin in the prevention of infections in acute leukemia. N Engl J Med 1981:304:1057-1062. 86. Schimpff SC, Greene WH, Young WM, et al. Infection prevention in acute nonlymphocytic leukemia. Ann Intern Med 1975;82:351-358. 87. Hathorn JW. Critical appraisal of antimicrobials for prevention of infections in immunocompromised hosts. Hematol Oncol Clin North A m 1993;7:1051-1099. 88. McQuiHen DP, Zingman BS, Meunier F, et al. Invasive infections due to Candida krusei: Report of ten cases of fungemia that include three cases of endophthalmitis. Clin Infect Dis 1992;14:472-478. 89. Shlaes DM, Binezewski B, Rice BL. Emerging antimicrobial resistance and the immunocornpromised host. Clin Infect Dis 1993;17(Suppl. 2):S527-536. 90. Stosor V , Noskin G A , Peterson LR. The management and prevention of vancomycinresistant enterococci. Infect Med 1996;13:488,493-498. 91. Livornese LL, Dias S, Same1 C, et al. Hospital-acquired infection with vancomycin-resistant Enterococcrls fi~eciur?? transmitted by electronic thermometers. Ann Intern Med 1992;117:112-116. 92. Groeger JS, Lucas AB, Thaler HT, et al. Venous access devices in patients with cancer. Ann Intern Med 19933 19:1168-1174. 93. Tenney J H , Moody MR. Newman KA, et al. Adherent microorganisms on lumenal surfaces of long-term intravenous catheters. Arch Intern Med 1986;146:1949-1954. 94. Pizzo PA. Management of fever in patients with cancer and treatment-induced neutropenia. N Engl J Med 1993;328:1323-1332. 95. Clarke DE, Raffin TA. Infectious complications of indwelling long-term central venous catheters. Chest 1990;97:966-972. 96. Abi-Nader JA. Peripherally inserted central venous catheters in critical care patients. Heart Lung 1993;22:428-434.

97. Lam S, Scannell R, Roessler D, et al. Peripherally inserted central catheters in an acute-care hospital. Arch Intern Med 1994;154:1833-1837. 98. Loughran SC, Borzatta M. Peripherally inserted central catheters: A report of 2506 catheter days. J Parent Ent Nut 1995;19:133-136. 99. Gottesdiener KM. Transplanted infections: Donor-to-host transmission with the allograft. Ann Intern Med 1989;110:1001-1016. 100. Barbara JAJ, Contreras M. Infectious complications of blood transfusion: Bacteria and parasites. Br Med J 1990;300:386-389. 101. Soulier JP. Diseases tramsmissible by blood transfusion. Vox Sang 1984;47:1-6. 102. Alter HJ, Purcell RH, Shih JW, et al. Detection of antibody to hepatitis C virus in prospectively followed transfusion recipients with acute and chronic non-A, non-3 hepatitis. N Engl J Med 1989;321:1494-1500. 103. Sayers MH, Anderson KC, Goodnough LT, et al. Reducing the risk for transfusiontransmitted cytomegalovirus infection. Ann Intern Med 1992;116:55-62. 104. Mayer KH, Opal SM. Unusual nosocomial pathogens. Infect Dis Clin North Am 1989;3:883899. 105. Wittner M, Rowin KS, Tanowitz HB, et al. Successful chemotherapy for transfusion babesiosis. Ann Intern Med 1982;96:601-604. 106. Tabor E, Garety RJ. Five cases of Psetldomonas sepsis transmitted by blood transfusions. Lancet 1984;1:1403. 107. Wright DC, Selss IF, Vinton KJ, et al. Fatal Yersinia enferocolitica sepsis after blood transfusion. Arch Path01 Lab Med 1985;109:1040-1042. 108. Yomtovian R, Lazarus LT, Goodnough NV, et al. A prospective microbiologic surveillance program to detect and prevent the transfusion of bacterially contaminated platelets. Transfusion 1993;33:902-909. 109. Schimpff SC. Therapy of infection in patients with granulocytopenia. Med Clin North Am 1977;61:1101-1118. 110. Bodey GP, Rodriguez V, Chang H, et al. Fever and infection in leukemic patients. A study of 494 consecutive patients. Cancer 1978;41:1610-1622. 111. Sanders JW, Powe NR, Moore RD. Ceftazidime monotherapy for empiric treatment of febrile neutroenic patients: A metaanalysis. J Infect Dis 1991;164:907-916. 112. Sanders CC. New B-lactams: New problems for the internist. Ann Intern Med 1991;115:65& 651. 113. Whimbey E, Kiehn TE, Brannon P, et al. Bacteremia and fungemia in patients with neoplastic disease. Am J Med 1987;82:723-730, 114. Singer C, Kaplan MH, Armstrong D. Bacteremia and fungemia complicating neoplastic disease. Am J Med 1977;62:731-742. 115. The EORTC International Antimicrobial Therapy Cooperative Group. Gram-positive bacteremia in granulocytopenic cancer patients. Eur J Cancer 1990;26:569-574. 116. Viscoli C, Van der Auwera P, Meunier F. Gram-positive infections in granulocytopenic patients: An important issue? J Antimicrob Chemother 1988;21(Suppl. C):149-156. 117. Winston DJ, Dudnick DV, Chapin M, et al. Coagulase-negative staphylococcal bacteremia in patients receiving immunosuppressive therapy. Arch Intern Med 1983;143: 32-36. 118. Karp JE, Dick J. Angelopulos C, et al. Empiric use of vancomycin during prolonged treatment-induced granulocytopenia: Randomized, double-blind, placebo-controlled clinical trial in patients with acute leukemia. Am J Med 1986;81:237-242. 119. Arning M, Gehrt A, Aul C, et al. Septicemia due to Streptococcrls n~itisin neutropenic patients with acute leukemia. Blut 1990;61:364-368. 120. Neu HC. The crisis in antibiotic resistance. Science 1992;257:1064-1073. 121. Anaissie E. Opportunistic mycoses in the immunocompromised host: Experience at a cancer center and review. Clin Infect Dis 1992;14(Suppl. 1):S43-53. 122. Meunier-Carpentier F, Kiehn TE, Armstrong D. Fungemia in the immunocompromised host: Changing patterns, antigenemia, high mortality. Am J Med 1981;71:363-370.

123. Lecciones JA, Lee JW, Navarro EE, et al. Vascular catheter-associated fungemia in patients with cancer: Analysis of 155 episodes. Clin Infect Dis 1992;14:875-883. 124. Wey SB, Mori M. Pfaller MA, et al. Hospital-acquired candidemia: The attributable mortality and excess length of stay. Arch Intern Med 1988;148:2642-2645. 125. Gerson SL, Talbot GH, Hurwitz S, et al. Prolonged granulocytopenia: The major risk factor for invasive pulmonary aspergillosis in patients with acute leukemia. Ann Intern Med 1984:100:345-351. 126. Fischer BD, Armstrong D, Yu B, et al. Invasive aspergillosis: Progress in early diagnosis and treatment. Am J Med 1981;71:571-577. 127. De Gregorio M, Lee W, Linker C, et al. Fungal infections in patients with acute leukemia. Am J Med 1982;73:543-548. 128. Hawkins C, Armstrong D. Fungal infections in the immunocompron~ised host. Clin HaematoI 1984;3:599-630. 129. Bodey G. Bueltmann B, Duguid W, et al. Fungal infections in cancer patients: An international autopsy survey. Eur J Microbiol Infect Dis 1992:11:99-109. 130. Core! L, Spear PG. Infections with Herpes simplex virus. N Engl J Med 1986;314:686-691, 749-757. 131. Rand KH, Kramer B, Johnson AC. Cancer chemotherapy associated symptomatic stomatitis: Role of Herpes simplex virus (HSV). Cancer 1982;50:1262-1265. 132. Buss DH, Scharyj M. Herpesvirus infection of the esophagus and other visceral organs in adults: Incidence and clinical significance. Am J Med 1979;66:457-462. 133. Ramsey PG, Fife KH, Hackman RC, et al. Herpes simplex virus pneumonia: Clinical, virologic. and pathologic features in 20 patients. Ann Intern Med 1982;97:813-820. 134. Straus SE, Ostrove JM, Inchauspe G, et al. VaricelIa-zoster virus infections: Biology, natural history, treatment, and prevention. Ann Intern Med 1988;108:221-237. 135. Meyers JD, Flournoy N, Thomas ED. Risk factors for cytomegalovirus infection after human marrow transplantation. J Infect Dis 1986;153:478-488. 136. Kaplan LJ, Daum RS, Smaron M. et al. Severe measles in immunocompromised patients. JAMA 1992;267:1237-1241. 137. Zahradnik JM, Spencer MJ, Porter DD. Adenovirus infection in the immunoco~npromised patient. Am J Med 1980;68:725-732. 138. Meunier F. Infections in patients with acute leukemia and lymphoma. In: Mandell GL, Bennett JE, Dolin R, eds. Principles and Practices of Infectious Diseases. New York: Churchill Livingstone, 1995, pp. 2675-2686. 139. Cotran RS. Kumar V, Robbins SL. Diseases of white cells, lymph nodes, and spleen. In: Robbins Pathologic Basis of Disease. Philadelphia: W.B. Saunders, 1989, pp. 703-754. 140. Goyal R, Bajpai S, Chopra HK, et al. Hairy cell leukemia - an unusual presentation. Leuk Res 1995;19:485-487. 141. Mackowiak PA, Demian SE, Sutker WL, et al. Infections in hairy cell leukemia: Clinical evidence of a pronounced defect in cell-mediated immunity. Am J Med 1980:68:718-723. 142. Bennett C, Vardinlan J, Golomb H. Disseminated atypical rnycobacterial infection in patients with hairy cell leukemia. Am J Med 1966;80:891-896. 143. Attal M. Schlaifer D, Rubie H. et al. Prevention of gram-positive infections after bone marrow transplantation by systemic vancomycin: A prospective. randomized trial. J Clin Oncol 1991;9:865-870. 144. Goodinan JL, Winston DJ, Greenfield RA, et al. A controlled trial of fluconazole to prevent fungal infections in patients undergoing bone marrow transplantation. N Engl J Med 1992;326:845-851. 145. Goodrich JM, Reed EC, Mori M, et al. Clinical features and analysis of risk factors for invasive candidal infection after marrow transplantation. J Infect Dis 1991;164:731-740. 146. Derouin F, Devergie A, Auber P, et al. Toxoplastnosis in bone marrow-transplant recipients: Report of seven cases and review. Clin Infect Dis 1992;15:267-270. 147. Miller W, Flynn P, McCulough J, et al. Cytomegalovirus infection after bone marrow transplantation: An association with acute graft-v-host disease. Blood 1986;67:1162-1167.

148. Wingard JR, Mellits D, Sostrin MB, et al. Interstitial penumonitis after allogeneic bone marrow transplantation: Nine year experience at a single institution. Medicine 1988;67:175-186. 149. Pannuti CS, Gingrich RD, Pfaller MA, et al. Nosocomial pneumonia in adult patients undergoing bone marrow transplantation: A 9 year study. J Clin Oncol 1991;9:77-84. 150. Meyers JD, Thomas ED. Infection complicating bone marrow transplantation. In: Rubin RH, Young LS, eds. Clinical Approach to Infection in the Compromised Host. New York: Plenum Medical, 1988, pp. 525-556. 151. Hiemenz JW, Greene JN. Special considerations for the patient undergoing allogeneic or autologous bone marrow transplantation. Hematol Oncol Clin North Am 1993;7:961-1002. 152. Rubin RH, Ferraro MJ. Understanding and diagnosing infectious complications in the immunocompromised host. Hematol Oncol Clin North Am 1993;7:795-812. 153. Cohen ML. Epidemiology of drug resistance: Implications for a post-antimicrobial era. Science 1992;257:1050-1055. 154. Kunin CM. Resistance to antimicrobial drugs - A worldwide calamity. Ann Intern Med 1993;118:557-561. 155. Leclercq R, Derlot E, Duval J, et al. Plasmid-mediated resistance to vancomycin and teicoplanin in Enterococc~lsfcrecilirn. N Engl J Med 1988;319:157-161. 156. Montecalvo MA, Horowitz H, Gedris C, et al. Outbreak of vancomycin-, ampicillin-, and aminoglycoside-resistant E~~ferococczls faeciurn bacteremia in an adult oncology unit. Antimicrob Agents Chemother 1994;38:1363-1367. 157. Noskin GA. Cooper I, Peterson L. Vancomycin-resistant Enterococcrrs faecium sepsis following persistent colonization. Arch Intern Med 1995;155:1445-1447. 158. Infectious Diseases Society of America. IDSA disseminates the latest on infectious diseases. Infect Dis Alert 1996;16:38-39. 159. Bouchud PY, Calandra T. Francioli P. Bactere~nia due to viridans streptococci in neutropenic patients: A review. Am J Med 1994;97:256-264. 160. Kern W, Kurrle E, Sclimeiser T. Streptococcal bacteremia in adult patients with leukemia undergoing aggressive chemotherapy. A review of 55 cases. Infection 1990;18:138-145. 161. Elting LS, Bodey GP, Keefe BH. Septicimia and shock syndrome due to viridans streptococci: A case-control study of predispoisng factors. Clin Infect Dis 1992;14:1201-1207. 162. Borek AP, Dressel DC, Hussong J, et al. Changing susceptibility of Streptococclrs pneumoniae to antimicrobial agents in Chicago, Illinois between 1993 and 1996:The need for accurate laboratory identification and susceptibility testing. Diagn Microbiol Infect Dis 1997;29:209-214. 163. Handwerger S, Horowitz H, Coburn K, et al. Infection due to Le~lconostocspecies: Six cases and review. Rev Infect Dis 1990;12:602-610. 164. Ascher DP, Zbick C, White C, et al. Infections due to Stor~zatococcusn~~icilaginosi~s: 10 cases and review. Rev Infect Dis 1991;13:1048-1052. 165. McWhinney PH, Kibbler CC, Gillespie SH, et al. Stonzatococous nzucilnginosus: An emerging pathogen in neutropenic patients. Clin Infect Dis 1992;14:641-646. 166. Young VM, Meyers WF. Moody MR, et al. The emergence of coryneform bacteria as a cause of nosocomial infections in compromised hosts. Am J Med 1981;70:646-650. 167. Riebel W, Frantz N. Adelstein D, et al. Coryizebacte~.irmiJK: A cause of nosocomial devicerelated infection. Rev Infect Dis 1986;8:42-49. 168. Johnson A, Hulse P, Oppenheim BA. Corynebacterir~mjeikeilim meningitis and transverse meylitis in a neutropenic patient. Eur J Clin Microbiol Infect Dis 1992;11:473474. 169. Henrickson K, Shenep J, Flynn P, et al. Primary cutaneous Bncill~lscereus infection in neutropenic children. Lancet 1989;1:601-603. 170. Tuazon CU. Other Bacillus species. In: Mandell GL, Bennett JE, Dolin R, eds. Principles and Practice of Infectious Diseases. New York: Churchill Livingstone, 1995, pp. 1890-1894. 171. Harvey RL, Sunstrum JC. Rhodococc~iseqrii infection in patients with and without human immunodeficiency virus infection. Rev Infect Dis 1991;13:139-145. 172. Khardori N, Elting L, Wong E, et al. Nosocomial infections due to X~mthomonrisn.zaltophilia ( P s e ~ ~ d o m o nmaltophilicl) as in patients with cancer. Rev Infect Dis 1990;12:997-1003.

173. Garau J. Beta-lactamases: Current situation and clinical importance. Intensive Care Med 1994;2O:S5-S9. 174. Sirot D. Extended-spectrum plasmid-mediated B-lactamases. J Antimicrob Chemother 1995;36(Suppl. A):19-34. 175. Meyer KS. Urban C, Eagan JA, et aI. Nosocomial outbreak of Klebsiella infection resistant to late-genertation cephalosporins. Ann Intern Med 1993;119:353-358, 176. Pegues DA, Carson LA, Anderson RL, et al. Outbreak of Pseudonzonas cepacin bacteremia in oncology patients. Clin Infect Dis 1993;16:407-411. 177. Kaye KM, Macone A, Kazanjian PH. Catheter infection caused by Methylobcrcterilcnz in immunocompromised hosts: Report of three cases and review of the literature. Clin Infect Dis 1992;14:1010-1014. 178. Edmond MB. Riddler SA, Baxter CM, et al. Agrobacterium radiobacter: A recently recognized opportunistic pathogen. Clin Infect Dis 1993;16:388-391. 179. Cieslak TJ, Robb ML, Drabick CJ, et al. Catheter-associated sepsis caused by Ochrobactr~im anthropi: Report of a case and review of related nonfermentative bacteria. Clin Infect Dis 1992;14:902-907. 180. Legrand C. Anaissie E. Bacteremia due to Achromobacter xylosoxidans in patients with cancer. Clin Infect Dis 1992;14:479-484. 181. Kim JH, Cooper RA, Welty-Wolf KE, et al. Pseudonzonas prrtrefciciens bacteremia. Rev Infect Dis 1989:11:97-104. 182. Bilgrami S, Bergstrom SK, Peterson DE, et al. Capnocytoyhcigcr bacteremia in a patient with Hodgkin's disease following bone marrow transplantation: Case report and review. Clin Infect Dis 1992;14:1045-1049. 183. Lorber B. Gas gangrene and other Clostridium-associated diseases. In: Mandell GL. Bennett JE. Dolin R. eds. Principles and Practice of Infectious Diseases. New York: Churchill Livingstone. 1995, pp. 2182-21 95. 184. Valtonen M, Sivonen A, Elonen E. A cluster of seven cases of Clostridium tertitim septicemia in neutropenic patients. Eur J Clin Microbiol Infect Dis 1990;9:40-42. 185. Zinner SH. New and unusual infections in neutropenic patients. In: Klastersky J, ed. Infectious Complications of Cancer. Boston: Kluwer Academic, 1995. pp. 173-184. 186. Huyghebaert MF, Dreyius F. Paul G, et al. Septicemie a Frlsobacteri~~rrznucleatz~nz, producteur de beta-lactamase chez un sujet neutropenique. Ann Med Interne (Paris) 1989:140:225-226. 187. Weinberger M, Wu T. Rubin M, et al. Leptotriclzn bllccalis bacteremia in patients with cancer: Report of four cases and review. Rev Infect Dis 1991;13:201-206. 188. Simon HB. Mycobacterial and nocardial infections in the compromised host. In: Rubin RH, Young LS. eds. Clinical Approach to Infection in the Compromised Host. New York: Plenum. 1988, pp. 221-251. 189. McWhinney PH. Yates M, Prentice HG, et al. Infection caused by Mycobacteri~rmchelonae: A diagnostic and therapeutic problem in the neutropenic patient. Clin Infect Dis 1992;14:1208-1212. 190. Raad I.I., Vartivarian S, Khan A, et al. Catheter-related infections caused by the Mycobcrcferirrin forfuitrliu complex: 15 cases and review. Rev Infect Dis 1991;13:1120-1125. 191. Kristjansson M, Bieluch VM, Byeff PD. Mycobacferzrim haernophilunz infection in immunocompromised patients: Case report and review of the literature. Rev Infect Dis 1991;13:906-910. 192. Chatis PA, Crumpacker CS. Resistance of Herpesviruses to antiviral drugs. Antimicrob Agents Chemother 1992;36:1589-1595. 193. Lalezari J, Stagg R, Kupperman B, et al. Intravenous cidofovir for peripheral Cytomegalovirus retinitis in patients with AIDS. Ann Intern Med 1997;126:257-263. 194. Studies of ocular complications of AIDS Research Group and the AIDS Clinical Trials Group. Parenteral cidofovir for Cytomegalovirus retinitis in patients with AIDS: The HPMPC peripheral Cytomegalovirus retinitis trial. Ann Intern Med 1997;126:264274.

195. Foreman KE, Friborg J, Kong W, et al. Propatation of a human Herpesvirus from AIDSassociated Kaposi's sarcoma. N Engl J Med 1997;336:163-171. 196. Strauchen JA, Hauser D, Burstein D, et al. Body cavity-based malignant lymphoma containing Kaposi sarcoma-associated Herpesvirus in an HIV-negative man with previous Kaposi sarcoma. Ann Intern Med 1996;125:822-825. 197. Pfaller M, Wenzel R. Impact of the changing epidemiology of fungal infections in the 1990's. Eur J Clin Microbiol Infect Dis 1992;11:287-291. 198. Walsh TJ, Lee J, Lecciones J, et al. Empiric therapy with amphotericin B in febrile granulocytopenic patients. Rev Infect Dis 1991;13:496-503. 199. Wingard J. Merz W, Sara1 R. Canditin fropicalis: A major pathogen in immunocompromised patients. Ann Intern Med 1979;91:539-543. 200. Anaissie E, Bodey G. Nosocomial fungal infections: Old problems and new challenges. Infect Dis Clin North Am 1989;3:867-882. 201. Gamis AS, Gudnason T, Giebink GS, et al. Disseminated infection with Fzdsnriurn in recipients of bone marow transplants. Rev Infect Dis 1991;13:1077-1088. 202. Nucci M, Spector N, Lucena S. et al. Three cases of infection with F~isnriunzspecies on neutropenic patients. Eur J Clin Microbiol Infect Dis 1992;11:1160-1162. 203. Bennett J. Miscellaneous fungi, and prototheca. In: Mandell G, Bennett J, Dolin R, eds. Principles and Practices of Infectious Diseases. New York: Churchill Livingstone, 1995, pp. 2389-2393. 204. Nahass GT, Rosenberg SP, Leonardi CL, et al. Disseminated infection with Trichosporon beigelii. Arch Dermatol 1993;129:1020-1023. 205. Walsh TJ, Newman KR, Moody M, et al. Trichosporonosis in patients with neoplastic disease. Medicine 1986;65:268-279. 206. Kalter DCA, Tschen JA, Cernoch PL, et al. Genital white piedra: Epidemiology, microbiology and therapy. J Am Acad Dermatol 1986;14:982-993. 207. Martino P, Venditti M, Micozzi A , et al. Blastoschizoirlyces cnpitntzls: An emerging cause of invasive fungal disease in leukemia patients. Rev Infect Dis 1990;12:570-582. 208. Danker W, Spector S, Fierer J. Malnssezia fungemia in neonates and adults: Complication of hyperalimentation. Rev Infect Dis 1987;9:743-837. 209. Sudduth E, AJ CI, WE F. Phaeohyphomycosis due to Exophinla species: Clinical spectrum of disease in humans. Clin Infect Dis 1992;15:639-644. 210. Wood GM, McCormack JG. Muir DB, et al. Clinical features of human infection with Scedospori~inziizflnttinz. Clin Infect Dis 1992;14:1027-1033. 211. Wagner KF. Agents of chro~nomycosis.In: Mandell GL, Bennett JE, Dolin R, eds. Principles and Practice of Infectious Diseases. New York: Churchill Livingstone, 1995, pp. 2324-2327. 212. Anaissie E, Bodey GP, Kantarjian H, et al. New spectrum of fungal infections in patients with cancer. Rev Infect Dis i989;11:369-378. 213. Kiehn TE, Gorey E, Brown AE, et al. Sepsis due to Rhodotorr,la related to use of indwelling central venous catheters. Clin Infect Dis 1992;14:841-846. 214. Henson JW, Jalaj JK, Walker RW, et al. Pnerrnzocystis cnrirlii pneumonia in patients with primary brain tumors. Arch Neurol 1991;48:406-409. 215. Beaman MH, McCabe RE, Wong S, et al. Toxoplasrna gondii. In: Mandell GL, Bennett JE. Dolin R. eds. Principles and Practice of Infectious Diseases. New York: Churchill Livingstone, 1995, pp. 2455-2475. 216. Ledingham JCG, Penfold WJ, Woodcock HM. Recent bacteriological experiences with typhoidal disease and dysentery: With notes on the protozoan parasites in the excreta. Br Med J 1915;2:704-711. 217. Faust EC, Giraldo LE, Caicedo G. Human isosporosis in the western hemisphere. Am J Trop Med Hyg 1961;10:343-349. 218. Gentile G, Venditti M, Micozzi A, et al. Cryptosporidiosis in patients with hematologic malignancies. Rev Infect Dis 1991;13:842-846. 219. Plumelle Y, Edouard A. Strongyloides stercornlis in T-cell leukemia/lymphoma in adults and acquired immunodeficiency syndrome. Rev Med Interne 1996;17:125-129.

220. Nucci M, Portugal R, Pulcheri W, et al. Strongyloidiasis in patients with hematologic malignancies. Clin Infect Dis 1995;21:675-677. 221. Savage D, Foadi M, Haworth C, et al. Marked eosinophilia in an immunosuppressed patient with strongyloidiasis. J Intern Med 1994;236:473-475. 222. Aydin H, Doppl W, Battmann A, et al. Opportunistic Strongyloides stercoralis hyperinfection in lymphoma patients undergoing chemotherapy and/or radiotherapy - report of a case and review of the literature. Acta Oncol 1994;33:78-80. 223. Simpson WG, Gerhardstein DC, Thompson JR. Disseminated Strongyloides stercoralis infection. South Med J 1993;86:821-825. 224. Wilkinson R, Leen CL. Chronic lymphocytic leukaemia and overt presentation of underlying Stronglyloides stercoralis infection. J Infection 1993;27:99-100. 225. Slater LN, Welch DF. Rochalimaea species (recently renamed Bartonella). In: Mandell GL, Bennett JE, Dolin R, eds. Principles and Practice of Infectious Diseases. New York: Churchill Livingstone, 1995, pp. 1741-1747. 226. Heney C, Greeff M, Davis V. Hickman catheter-related Prorothecal algaemia in an immunocompromised child. J Infect Dis 1991;163:930-931. 227. Kwon-Chung KJ, Bennett JE. Medical Mycology. Philadelphia: Lea & Febiger, 1992. 228. Kaminski ZC, Kapila R, Sharer LR, et al. Meningitis due to Prototheca wickerhamii in a patient with AIDS. Clin Infect Dis 1992:15:704-706. 229. Iacoviello VR, DeGirolami PC, Lucarini J, et al. Protothecosis complicating prolonged endotracheal intubation: Case report and literature review. Clin Infect Dis 1992;15:959-967.

3. Approach to fever in the neutropenic host Athena Stoupis and Stephen H. Zinner

1. Introduction The approach to fever in the neutropenic host in large part depends on an understanding of the basic immune defect. White blood cells are critically important in maintaining an infection-free state. Nearly 30 years ago, the risk of serious bacterial infections in patients with a low number of circulating leukocytes was first recognized by Bodey and coworkers [I]. Diminished polymorphonuclear leukocyte numbers or function and impaired humoral immunity result in the development of serious infections, with the attendant risk of morbidity and mortality. The risk of infection progressively increases as the neutrophil count decreases below 1000 cells/mL. The most severe risk for bacteremia occurs when the granulocyte count falls below 100 cellslmL [I]. The depth of neutropenia, as well as the duration, are also considered to be important parameters that influence the development of infection [2]. One study revealed that in patients with neutropenia lasting less than 1 week ("short neutropenia"), fewer than 30% developed fever or other evidence of infection, whereas almost all patients with neutropenia lasting longer than 1week ("long neutropenia") developed fever or evidence of infection [2-51. In patients with malignancy, neutropenia may result from the use of cytotoxic chemotherapy, marrow irradiation, or neoplastic infiltration of the bone marrow. Another coexisting factor associated with malignancy that affects host defense is loss of the integrity of the skin and mucosa, which can occur by direct extension of solid tumors or by erosive mucosal lesions often encountered in patients with leukemia. The placement of intravenous or urinary catheters, and of endotracheal tubes, and the presence of decubitus ulcers often encountered in the hospital setting also violate physical defense barriers and may result in infections caused predominantly by nosocomial pathogens, particularly staphylococci, streptococci, or aerobic gram-negative rods that colonize the skin and gut of these patients. Impairment of humoral immunity with resultant deficits in antibody responses that normally potentiate phagocytosis can be found in patients who have undergone splenectomy for staging purposes or have developed splenic Gory A . Noskin (ed), M A N A G E M E N T O F INFECTIOUS C O M P L I C A T I O N S IN C A N C E R PATIENTS. O 1998. Klrlwer Academic Pnblishers, Bostorz. Ail rights reserved.

infiltration with malignant cells, as in patients with Hodgkin's disease, multiple myeloma, or leukemia [6,7]. These patients are at high risk for the development of overwhelming infection due to encapsulated bacteria such as Streptococcus pneumoniae, Haernophilus influenzae, Neisseria meningitidis, and others. The problem is further complicated by an inability to mount an appropriate antibody response to prophylactic polysaccharide vaccines following splenectomy [7,8]. Cell-mediated immunity also may be impaired in many patients who receive cytotoxic chemotherapy, radiation, and corticosteroids. This results in infection with intracellular organisms such as Listeria monocytogenes, Legionella, and Salmonella spp., and less commonly Mycobacterium species [9,10]. These patients also may be infected with herpes viruses (herpes simplex, varicella zoster, cytomegalovirus) and are subject to fungal infections, including disseminated candidiasis, cryptococcal meningitis, and aspergillosis, among others [7,10]. Other risk factors for infection, such as the patient's clinical instability, advanced malignant disease, and overt organ dysfunction, also contribute to the risk of serious infection in patients with cancer [I I]. 2. Risk factor assessment

In recent years, several retrospective studies have evaluated neutropenic cancer patients with fever in an attempt to identify subsets of patients at high and low risk for the development of bacteremia and other serious infectious complications [12]. Talcott and coworkers [13,14] studied various clinical variables that could be assessed within the first 24 hours of presentation of fever and neutropenia, with the aim of developing a predictive model that could stratify patients according to risk of complications. The majority of patients were projected at high risk on the basis of "comorbidity factors" such as cardiac or renal disease, uncontrolled cancer, or existing clinical status at onset of admission that included hypotension, respiratory failure, altered mental status, dehydration, and inadequate oral intake. Of the 444 patients studied, those with any of these high risks were more likely to have serious medical complications (35%) than patients without any risk factors (5%). These studies have important implications for the selection of patients amenable to oral or outpatient therapy. Pappo and Buchanan [I51 also retrospectively examined multiple clinical and laboratory criteria to identify low-risk features at the onset of fever in pediatric patients. Over a 3-year period, 600 hospital admissions were assessed and 93 episodes of bacteremia were noted and analyzed. The analysis revealed that all but 7 of the 93 positive blood cultures (7.5%) occurred in patients defined as "high risk" by the presence of one or more of these factors: (1) primary disease not in remission; (2) age 500 STOP IN 4-5 DAYS: REEVALUATE

ANC < 500 CONTINUE FOR 14D

a CONTIhIVE UNTIL

Figure 3. Duration of antibiotic therapy according to recent guidelines from the Infectious Diseases Society of America (88a). ANC = absolute neutrophil count; D = days; DC = discontinue.

patients. The most commonly used agents are the hematopoietic growth factors, granulocyte macrophage-colony stimulating factor (GM-CSF) and granulocyte-colony stimulating factor (G-CSF). These cytokines are glycoproteins that stimulate the proliferation and maturation of bone marrow stem cell lines, resulting in an increase in peripheral granulocyte counts. Several studies [101-1031 have revealed a reduction in the duration and severity of neutropenia in patients receiving chemotherapy. However, in these studies hematopoeitic growth factors were administered prophylactically with the intent of shortening the period of risk and lowering the incidence of infectious complications. More recently, investigations exploring the interventional rather than prophylactic administration of CSFs demonstrated an equivocal effect on neutropenia and no reduction in the frequency of infections. Also, with respect to their use in patients with acute myelocytic leukemia, CSFs have not stimulated regrowth of leukemic cells but also did not result in an increase of disease free survival [104]. Hence, hematopoietic growth factors are currently recommended only for prophylactic use in patients with prolonged and severe neutropenia and at high risk of infectious complications. According to the American Society of Clinical Oncology (ASCO) guidelines [105], CSFs are indicated for primary prophylaxis if the expected incidence of neutropenic fever is greater that 40% and for secondary prophylaxis if chemotherapy doses cannot be reduced to minimize infection. The ASCO guidelines do not advocate the use of these cytokines routinely in all patients receiving chemotherapy. Although CSFs clearly have decreased the duration of neutropenia, beneficial results of the incidence of severe infection and the long-term effects on antileukemia treatment remain to be determined definitively [106]. Other less commonly used adjunctive therapies also have been utilized to improve the host defenses in neutropenic patients. Passive immunization with antiendotoxin antibodies is still undergoing clinical evaluation. In an early clinical trial using antibody to core glycolipid of the Enterobacteriaceae (J5 antiserum), mortality was decreased in patients with gram-negative bacteremia, except in patients who had neutropenia [107]. This early trial ignited further investigational work on the use of core glycolipid for passive immunization. More recent trials using monoclonal core glycolipid antibodies have produced inconclusive results. Clinical studies of anti-lipid A monoclonal antibodies [108,109], specifically E5 and HA-lA, seemed promising because both antibodies appeared to protect subsets of patients with sepsis syndrome. E5 appeared to improve the survival of patients with gram-negative sepsis with refractory shock, but only when they were bacteremic. Questions concerning the clinical efficacy and cost effectiveness of these agents in patients with neutropenia and their ultimate impact on survival mandate the need for further investigation. Antibody to tumor necrosis factor (TNF) and interleukin-1 receptor antagonist have been disappointing. Recently, the efficacy and safety of antiTNF-alpha monoclonal antibody (Mab) was studied in a randomized

double-blind placebo-controlled trial [110]. A total of 971 patients were prospectively stratified into shock or nonshock groups and then randomized to receive a single infusion dose of 15mgJkg of TNF-alpha Mab, 7.5mg/kg of TNF-alpha Mab or placebo. There was no reduction in mortality between placebo and TNF-alpha Mab in all infused patients. Since the cloning of the interleukin-1 receptor antagonist (IE-lra) there has been intensive research on the genetics and potential therapeutic value of this protein; however, the role of IL-lra in normal physiology or in host defense mechanisms remains unclear. Preliminary results of clinical trials in animal models and in humans indicated a possible benefit of IL-lra in sepsis syndrome, rheumatoid arthritis, and graft-versus-host disease, but clinical studies were disappointing [I 11-1131. Intravenous immunoglobulins have been found to reduce the frequency of respiratory infections [I071 in patients with chronic lymphocytic leukemia in whom antibody production is deficient. CMV-antibody-rich IV immunoglobulin to prevent CMV infection in CMV-seronegative blood recipients has also been investigated [114].

References 1. Bodey GP, Buckley M, Sathe YS, Freireich EJ. Quantitative relationships between circulating leukocytes and infection in patients with acute leukemia. Ann Intern Med 1966;64:328340. 2. Lee JW, Pizzo PA. Management of the cancer patient with fever and prolonged neutropenia. Hematol Oncol Clin North Am 1993;7:937-960. 3. Pizzo PA, Commers JR, Cotton D, et al. Approaching the controversies in the antibacterial management of cancer patients. Am J Med 1984;76:436439. 4. Schimpff SC. Empiric antibiotic therapy for granulocytopenic cancer patients. Am J Med 1989;8O(S~ppl.5C):13-20. 5. Pizzo PA, Robichaud KJ, Gill FA, Witebsky PG. Empiric antibiotic and antifungal therapy for cancer patients with prolonged fever and granulocytopenia. Am J Med 1982; 72:101-111. 6. Desser RK. Ultman JE. Risk of severe infection in patients with Hodgkin's disease or lymphoma after diagnostic laparotomy and splenectomy. Ann Intern Med 1972;77:143-146. 7. Mayer KH, Opal SM. Antibacterial therapy in patients with malignancies. Cancer Metastas Rev 1987;5:271-293. 8. Siber GR, Weitzman SA, Aisenberg AC, et al. Impaired antibody response to pneumococcal vaccine after treatment for Hodgkin's disease. N Engl J Med 1978;299:442-448. 9. Lamb D, Pilney F, Kelly WD, Good RA. A comparative study of the incidence of anergy with carcinoma, leukemia, Hodgkin's disease and other lymphomas. J Immunol1962;89:555558. 10. Zinner SH. Treatment and prevention of infection in immunocompromised hosts. In: Gorbach SL, Bartlett JG, Blacklow NR, eds. Infectious Diseases. Philadelphia: W.B. Saunders, 1992, pp. 1057-1064. 11. Buchanan GR. Approach to treatment of the febrile cancer patient with low risk neutropenia. Hematol Oncol Clin North Am 1993;7:919-935. 12. Freifeld AG, Pizzo PA. The outpatient management of febrile neutropenia in cancer patients. Oncology 1996;10:599-612.

13. Talcott JA, Finberg R, Mayer RJ, et al. The medical course of cancer patients with fever and neutropenia. Clinical identification of a low risk subgroup at presentation. Arch Intern Med 1988;148:2561-2568. 14. Talcott JA, Siege1RD, Finberg R, Goldman L. Risk assessment in cancer patients with fever and neutropenia: A prospective, two-center validation of a prediction rule. J Clin Oncol 1992:10:316-322. 15. Pappo AS, Buchanan GR. Predictors of bacteremia in febrile neutropenic children with cancer (abstr.). Proc Annu Meet Am Soc Clin Oncol 1991:10:A1167. 16. Patrick CC. Coagulase-negative staphylococci: Pathogens with increasing clinical significance. J Pediatr 1990;116:497-507. 17. Rubin M, Hathorn JW. Marshall D, et al. Gram-positive infections and the use of vancomycin in 550 episodes of fever and neutropenia. Ann Intern Med 1988;108:30-35. 18. Anaissie EJ. Vadhan-Raj S. Is it time to redefine the management of febrile neutropenia in cancer patients? Am J Med 1995:98:224-231. 19. Klastersky J. Febrile neutropenia. Support Care Cancer 1993:1:233-239. 20. Zinner SH, Gaya H. Glauser M, et al. The International Antimicrobial Therapy Group of the European Organization for Research and Treatment of Cancer. Con~binationof amikacin and carbenicillin with or without cefazolin as empirical treatment of febrile neutropenic patients. J Clin Oncol 1983:1:597-603. 21. Giamarellou H. Empiric therapy for infections in the febrile, neutropenic, compromised host. Med Clin North Am 1995:79:559-580. 22. Meunier F. Infections in patients with acute leukemia and lymphoma. In: Mandell GL, Douglas RG, Bennett JE. eds. Principles and Practice of Infectious Disease, 4th edi. New York: Churchill Livingstone, 1995, pp. 2675-2686. 23. Sickles EA, Greene WH, Wiernick PH. Clinical presentation of infection in granulocytopenic patients. Arch Intern Med 1975:135:715-719. 24. Rubin RH, Green R. Clinical approach to the compromised host with fever and pulmonary infiltrates. In: Rubin RH, Young LS, eds. Clinical Approach to Infection in the Compromised Host. New York: Plenum, 1994. pp. 121-161. 25. Cockerill FR 111, Wilson WR, Carpenter HA, et al. Open lung biopsy in immunocon~promisedpatients. Arch Intern Mcd 1985;145:1398-1404. 26. Cheson BD. Samlowski WE, Tang TT, Spruance SL. Value of open lung biopsy in 87 immunocompromised patients with pulmonary infiltrates. Cancer 1985:55:453-459. 27. Peterson DE, Minah GE, Overholser CD, et al. Microbiology of acute periodontal infection in myelosuppressed cancer patients. J Clin Oncol 1987;5:1461-1468. 28. Gerard M, Defresne N, Daneau D, et al. Incidence and significance of Clostridi~ir?~ dificile in hospitalized cancer patients. Eur J Clin Microbial Infect Dis 1988:7:274-278. 29. Milligan DW, Kelly JK. Pseudomembranous colitis in a leukemia unit: A report of fatal cases. J Clin Path01 397932: 1237-1243. 30. Varki AP, Armitage JO, Feagler JR. Typhlitis in acute leukemia: Successful treatment by early surgical intervention. Cancer 1979;43:695-697. 31. Anaissie E. Opportunistic mycoses in the immunocompromised host: Experience at a cancer center and review. Clin Infect Dis 1992;14(Suppl. 1):S43-S53. 32. Bodey GP. Jadeja L, Elting L. Pseudomonas bacteremia. Retrospective analysis of 410 episodes. Arch Intern Med 1985:145:1621-1629. 33. Hiemenz J. Skelton J, Pizzo PA. Perspective on the management of catheter-related infections in cancer patients. Pediatr Infect Dis 1986;5:6-11. 34. Johnson PR, Decker MD, Edwards KM, et al. Frequency of broviac catheter infections in pediatric oncology patients. J Infect Dis 1986:154:570-578. 35. Kralner BS, Pizzo PA, Robichaud KJ. et al. Role of serial microbiologic surveillance and clinical evaluation in the management of cancer patients with fever and granulocytopenia. Am J Med 1982;72:561-568. 36. The International Antimicrobial Therapy Cooperative of the European Organization for Research and Treatment of Cancer (Writing Committee: Calandra T. Zinner SH, Viscoli C,

37. 38. 39.

40. 41. 42. 43. 44. 45. 46. 47.

48. 49.

50.

51. 52. 53. 54. 55.

56. 57.

de Bock R, Gaya H, Meunier F, Klastersky J, Glauser M). Efficacy and toxicity of single daily doses of amikacin and ceftriaxone versus multiple daily doses of amikacin and ceftazidime for infection in patients with cancer and granulocytopenia. Ann Int Med 1993;119:584593. Elting LS, Bodey GP, Keefe BH. Septicemia and shock syndrome due to viridans streptococci: A case-control study of predisposing factors. Clin Infect Dis 1992:14:1201-1207. Handwerger S. Horowitz H, Coburn K, Kolokathis A , Wormser GP. Infection due to Lerlconostoc species: Six cases and review. Rev Infect Dis 1990;12:602-610. Swartz MN. Cellulitis and subcutaneous tissue infections. In: Mandell GL, Douglas RG. Bennett JE, eds. Principles and Practice of Infectious Disease, 4th ed. New York: Churchill Livingstone, 1995. pp. 909-929. Zinner SH. New and unusual infections in neutropenic patients. In: Klastersky J, ed. Infectious Complications of Cancer. Boston: Kluwer Academic. 1995, pp. 173-184. Anaissie E. Kantarjian H. Ro J , et al. The emerging role of Fusarium infections in patients with cancer. Medicine 1988;67:77-83. Martino P, Gastaldi R, Raccah R, Girmenia C. Clinical patterns of Fzunrir~nzinfections in immunocompromised patients. J Infect 1994:28(Suppl. 1):7-15. Raad I, H a c h e ~ nR. Treatment of central venous catheter-related fungemia due to F~rstrriunz oxysponinr. Clin Infect Dis 1995:20:709-711. Walter E A , Bowden R A . Infection in the bone marrow transplant recipient. Infect Dis Clin North Am 1995;9:823-847. Samonis G , Bafaloukos D. Fungal infections in cancer patients: An escalating problem [review]. In Vivo 1992;6:183-193. Gokaslan A, Anaissie E. A novel murine model of disseminated trichosporonosis. Infect Immun 1992;60:3339-3344. Schimpff SC. Satterlee W, Young VM. Serpick A. Empiric therapy with carbenicillin and gentamicin for febrile patients with cancer and granulocytopenia. N Engl J Med 1971; 284: 1061-1 065. Gibson J. Johnson L, Snowdon L. et al. Trends in bacterial infections in febrile neutropenic patients: 1986-1992. Aust N Z 9 Med 1994:24:374-377. Coinetta A, Zinner S, de Bock R, et al. and the International Antimicrobial Therapy Cooperative Group of the European Organi~ationfor Research and Treatment of Cancer. Piperacillin-tazobactam plus alnikacin versus ceftazidime plus arnikacin as empiric therapy for fever in granulocytopenic patients with cancer. Antimicrob Agents Chemother 1995: 39:445-452. Fergie JE, Shema SJ, Lott L. Crawford R. Patrick CC. Pseudonzotlns cieruyiizosti bacteremia in immunocompromised children: Analysis of factors associated with a poor outcome. Clin Infect Dis 1994;18:390-394. Schimpff SC, Greene WH, Young VM. et al. Significance of Pseudon~onasrreniginosn in the patient with leukemia and lymphonla. J Infect Dis 1974;130(Suppl.):S24-S31. Hersch EM. Bodey GP, Nies BE, Freireich EJ. Causes of death in acute leukemia. JAMA 1965:L93:105-109. Rubin M. Hathorn JW, Pizzo PA. Controversies in the management of febrile neutropenic cancer patients. Cancer Invest 1988;6:167-1 84. Klastersky .I.Zinner SH. Synergistic combination of antibiotics In gram-negative bacillary infections. Rev Infect Dis 1982;4:294-298. Klastersky J. Schimpff SC. Gaya H. Glauser MP. Zintler SH. EORTC Intertlational Antimicrobial Therapy Cooperative Group. Prospective randomized comparison of three antibiotic regimens for empirical therapy of suspected bacteren~icinfection in febrile granulocytopenic patients. Antimicrob Agents Chemother 1986:29:263-270. Tattersall MH. Spiers ASD. Darrell JH. Initial therapy with combination of five antibiotics in febrile pat~entswith leukaemia and neutropenia. Lancet 1972:1:162-165. Zinner SH, Gaya H. Glauser M, Klastersky J, Schimpff SC. Tattersall MHN. EORTC International Antimicrobial Therapy Cooperative Group. Combination of amikacin and

58.

59.

60.

61. 62.

63. 64. 65.

66. 67.

68. 69.

70.

71.

72.

73.

74. 75.

76.

carbenicillin with or without cefazolin as empirical treatment of febrile neutropenic patients. J Clin Oncol 1983;1:597-603. Calandra T, Klasterksy J, Gaya H, Glauser MP, Meunier F, Zinner SH. The EORTC International Antimicrobial Therapy Cooperative Group. Ceftazidime combined with a short or long course of amikacin for empirical therapy of gram-negative bacteremia in cancer patients with granulocytopenia. N Engl J Med 1987;317:1692-1698. Gibson J, Johnson L, Snowdon L, et al. Single daily ceftriaxone and tobramycin in the empirical management of febrile neutropenic patients: A randomised trial. Int J Hematol 1993;58:63-72. Martino P, Girmenia C, Raccah R, et al. Single daily dose ceftriaxone plus amikacin treatment of febrile episodes in neutropenic patients attending day hospital for hematologic malignancies. Oncology 1992;49:49-52. Gutmann L, Williamson R,Kitzis M, et al. Synergism and antagonism in double beta-lactam antibiotic combinations. Am J Med 1986;80(Suppl.5C):21-29. Joshi JH, Newman KA, Brown BW, et al. Double beta-lactam regimen compared to an aminoglycoside/beta-lactam regimen as empiric antibiotic therapy for febrile granulocytopenic cancer patients. Support Care Cancer 1993;1:186-194. Young LS. Double beta-lactam therapy in the im~nunocompron~ised host, J Antimicrob Chemother 1985;16:4-6. Pizzo PA, Thaler M, Hathorn J, et al. New beta-lactam antibiotics in granulocytopenic patients: New options and new questions. Am J Med 1985;79(Suppl. 2A):75-82. De Pauw BE, Deresinski SC, Feld R, Lane-Allman EF, Donnelly JP for The Interncontinental Antimicrobial Study Group. Ceftazidime compared with piperacillin and tobramycin for the empiric treatment of fever in neutropenic patients with cancer. A multicenter randomized trial. Ann Intern Med 1994;120:834-844. Bizette GA, Brooks BJ Jr., Alvarez S. Ceftazidime as monotherapy for fever and neutropenia: Experience in a community hospital. J La State Med Soc 1994;146:448-452. Pizzo PA, Hathorn JW, Hiemenz J, et al. A randomized trial comparing ceftazidime alone with combination antibiotic therapy in cancer patients with fever and neutropenia. N Engl J Med 1986;315:552-558. Sanders JW, Powe NR, Moore RD. Ceftazidime monotherapy for empiric treatment of febrile neutropenic patients: A meta-analysis. J Infect Dis 1991;164:907-916. Naumovski L, Quinn JP, Miyashiro D, et al. Outbreak of ceftazidime resistance due to a novel extended-spectrum beta-lactamase in isolates from cancer patients. Antimicrob Agents Chemother 1992;36:1991-1996. Eggimann P, Glauser MP, Aoun M, Meunier F, Calandra T. Cefepime monotherapy for the empirical treatment of fever in granulocytopenic cancer patients. J Antimicrob Chemother 1993;32(Suppl. B):151-163. Leyland MJ, Bayston KF, Cohen J, et al. A comparative study of imipenem versus piperacillin plus gentamicin in the initial management of febrile neutropenic patients with haematological malignancies. J Antimicrob Chemother 1992;30:843-854. Rolston KV. Berkey P, Bodey GP, et al. A comparison of imipenem to ceftazidime with or without amikacin as empiric therapy in febrile neutropenic patients. Arch Intern Med 1992;152:283-291. Freifeld AG, Walsh T, Marshall D, et al. Monotherapy for fever and neutropenia in cancer patients: A randomized comparison of ceftazidime versus imipenem. J Clin Oncol 1995; 13:165-176. Donnelly JP, Horrevorts AM, Sauerwein RW, De Pauw BE. High-dose meropenem in meningitis due to Pseirdon~onasaerrlginosa [letter]. Lancet 1992;339:1117. The Meropenern Study Group of Leuven, London and Nijmegen. Equivalent efficacies of meropenem and ceftazidime as empirical monotherapy of febrile neutropenic patients. J Antimicrob Chemother 1995;36:185-200. Cometta A, Calandra T, Gaya H, et al. and the International Antimicrobial Therapy Cooperative Group of the European Organization for Research and Treatment of Cancer and

77.

78.

79.

80.

81. 82.

83.

84. 85.

86.

87.

88.

88a. 89.

90. 91. 92.

The Gruppo Italiano Malattie Ematologiche Maligne Dell'Adulto Infection Program. Monotherapy with meropenem versus combination therapy with ceftazidime plus amikacin as empiric therapy for fever in granulocytopenic patients with cancer. Antimicrob Agents Chemother 1996;40:1108-1115. Meunier F, Zinner SH, Gaya H, Calandra T, Viscoli C, Klastersky J, Glauser M (EORTCIATCG). Prospective randomized evaluation of ciprofloxacin versus piperacillin plus amikacin for enlpiric antibiotic therapy of febrile granulocytopenic cancer patients with lymphomas and solid tumors. Antimicrob Agents Chemother 1991;35:873-878. Bow EJ, Loewen R, Vaughan D. Reduced requirement for antibiotic therapy targeting gram-negative organisms in febrile, neutropenic patients with cancer who are receiving antibacterial chemoprophylaxis with oral quinolones. Clin Infect Dis 1995;20:907-912. Talcott JA, Whalen A, Clark J, Rieker PP, Finberg R. Home antibiotic therapy for low-risk cancer patients with fever and neutropenia: A pilot study of 30 patients based on a validated prediction rule. J Clin Oncol 1994;12:107-114. Mullen CA, Buchanan G R . Early hospital discharge of children with cancer treated for fever and neutropenia: Identification and management of the low-risk patient. J Clin Oncol 1990;8:1998-2004. Rubenstein EB. Rolston K, Benjamin RS, et al. Outpatient treatment of febrile episodes in low-risk neutropenic patients with cancer. Cancer 1993;71:3640-3646. Kolston K, Rubenstein EB, Elting LS. et al. Ambulatory management of febrile episodes in low-risk neutropenic patients - an acceptable alternative to hospital based therapy. In: Program Proceedings and Abstracts of the 6th International Symposium Supportive Care in Cancer, 1994. Malik IA. Khan WA, Karim M, Aziz Z, Khan MA. Feasibility of outpatient management of fever in cancer patients with low-risk neutropenia: Results of a prospective randomized trial. A m J Med 1995;98:224-231. Gillis S, Dann EL. Rund D . Selective discharge of patients with acute myeloid leukemia during chemotherapy-induced neutropenia. A m J Hemtaol 1996;51:26-31. D e Marie S, Van den Broek PJ. Willemze R, Van Furth R. Strategy for antibiotic therapy in febrile neutropenic patients on selective antibiotic decontamination. Eur J Clin Microbiol Infect Dis 1993;122397-906. E O R T C International Antimicrobial Therapy Cooperative Group and the National Cancer Institute of Canada CIinical Trials Group. (Writing Committee: Calandra T. Zinner SH, Glauser MP, Meunier F, Viscoli C, Feld R, Gaya H, and Klastersky J.) Vancomycin added to empirical combination antibiotic therapy for fever in granulocytopenic cancer patients. J Infect Dis 1991;163:951-958. Ramphal R, Bolger M. Oblon DJ, et al. Vancomycin is not an essential component of the initial e~npirictreatment regimen for febrile neutropenic patients receiving ceftazidime: A randomized propsective study. Antimicrob Agents Chemother 1992;36:1062-1067. E O R T C International Antimicrobial Therapy Cooperative Group. (Writing Committee: Meunier F, Gaya H, Calandra T. Glauser M, Klastersky J, Sylvester R, and Zinner S.) Empiric antifungal therapy in febrile granulocytopenic patients. Am J Med 1989;86:668672. Hughes WT. Armstrong D, Bodey GP, et al. 1997 guidelines for the use of antimicrobial agents in neutropenic patients with unexpected fever. Clin Infect Dis 1997;551-573. Pizzo PA. Empirical therapy and prevention of infection in the immunocompromised host. In: Mandell G, Douglas RG. Bennett JE, eds. Principles and Practice of Infectious Diseases. New York: Churchill Livingstone, 1995, pp. 2686-2696. AnttiIa VJ, Ruutu P, Bondestam S, et al. Hepatosplenic yeast infection in patients with acute leukemia: A diagnostic problem. Clin Infect Dis 1994;18:979-981. Anaissie E , Pinczowski H. Invasive candidiasis during granulocytopenia. Recent Results Cancer Res 1993;132:137-145. Flannery MT, Simmons DB. Saba H. Altus P, Wallach PM, AdeLman HM. Fluconazole in the treatment of hepatosplenic canclidiasis. Arch Int Med 1992;152:406-408.

93. D e Pauw BE. Raemaekers JM, Donnelly JD, Kullberg BJ, Meis JF. An open study on the safety and efficacy of fluconazole in the treatment of disseminated Candida infections in patients treated for hematologic malignancy. Ann Hemato1 1995;70:83-87. 94. Van't Wout JW, Novakova I, Verhagen CA, et al. The efficacy of itraconazole against systemic fungal infections in neutropenic patients: A randomised conlparative study with amphotericin B. J Infect 1991;22:45-52. 95. D e Marie S, Janknegt R, Bakker-Woudenberg IA. Clinical use of liposomal and lipidcomplexed amphotericin B [review]. J Antimicrob Chemother 1994;33:907-916. 96. Janknegt R , D e Marie S, Bakker-Woudenberg IA. Liposornal and lipid formulations of amphotericin B. Clin Pharmacokinet 1992;23:279-291. 97. Adler-Moore J. AmBisoine targeting to fungal infections. Bone Marrow Transplant 1994; 14(Suppl. 5):S3-S7. 98. Lopez-Berestein G , Bodey GP, Fainstein V. et al. Treatment of systemic fungal infections with liposomal amphotericin B. Arch Intern Med 1989;149:2533-2536. 99. Pizzo PA, Robichaud KJ, Gill FA, et al. Duration of empiric antibiotic therapy in granulocytopenic patients with cancer. A I J~Med 1979;67:194-200. 100. Young LS. Neutropenia: Antibiotic combinations for empiric therapy. Eur J Clin Microbiol Infect Dis 1989;8:118-122. 101. Gabrilove JL, Jakubowski A, Scher H, et al. Effect of granulocyte colony-stimulating factor on neutropenia and associated morbidity due to chemotherapy for transitional cell carcinoma of the urothelium. N Engl J Med 1988;318:1414-1422. 102. Dibenedetto SP, Ragusa R, Ippolito AM. et al. Assessment of the value of treatment with granulocyte colony-stimulating factor in children with acute lymphoblastic leukemia: A randomized clinical trial. Eur J Haernatol 1995;55:93-96. 103. Stone RM, Berg DT, George SL, et al.. for The Cancer and Leukemia Group B. Granulocyte-macrophage colony-stimulating factor after initial chemotherapy for elderly patients with primary acute inyelogenous leukemia. N Engl J Med 1995;332:1671-1677. 104. Maher DW, Lieschke GJ. Green M, et al. Filgrastim in patients with chemotherapy-induced febrile neutropenia. A double-blind, placebo-controlled trial. Ann Intern Med 1994;121:492501. 105. American Society of Clinical Oncology. American Society of Clinical Oncology recommendations for the use of hematopoietic colony stimulating factors: Evidence based. clinical practice guidelines [special article]. J Clin Oncol 1994;12:2471-2508. 106. Roilides E , Pizzo PA. Perspectives on the use of cytokines in the management of infectious complications of cancer. Clin Infect Dis 1993:17(Suppl. 2):S385-S389. 107. Pizzo PA. Management of fever in patients with cancer and treatment-induced neutropenia. N Engl J Med 1993;328:1323-1332. 108. Ziegler EJ, Fisher CJ Jr.. Sprung CL. et al. Treatment of gram-negative bacteremia and septic shock with HA-1A human n~onoclonalantibody against endotoxin: A randomized. double-blind, placebo-controlled trial. N Engl J Med 1991:324:429-436. 109. Greenman RI, Schein RMH, Martin MA, et al. and The Xoma Sepsis Study Group. A controlled clinical trial of ES murine monoclonal IgM antibody to endotoxin in the treatment of gram-negative sepsis. JAMA 109 1 :266:1097-1 102. 110. Abraham E , Wunderink R, Silverman H, et al. for The TNF-u MAb Sepsis Study Group. Efficacy and safety of monoclonal antibody to human tumor necrosis factor a in patients with sepsis syndrome. A randomized, controlled, double-blind, multicenter clinical trial. J A M A 1995;273:934-941. 11 1. Dinarello CA. Blocking interleukin-1 receptors. Int J Clin Lab Res 1994:24:61-79. 112. Arend WP. Interleukin-1 receptor antagonist. Adv Iinrnunol 1993;54:167-227. 113. Lennard AC. Interleukin-l receptor antagonist. Crit Rev Irnmunol 1995;15:77-105. 114. Winston DJ, H o WG, Bartani K, Champlin RE. Intravenous immunoglobulin and CMVseronegative blood products for prevention of CMV infection and disease in bone marrow transplant recipients. Bone Marrow Transplant 1993;12:283-288.

4. Infections associated with solid tumors Sarah H. Sutton and John P. Flaherty

1. Introduction According to American Cancer Society estimates, leukemias and lymphomas accounted for only 7% of male and 6% of female cancers diagnosed and 8% of cancer deaths; the remaining cancers were solid tumors [I]. Although solid tumors account for the vast majority of cancer in adults, severe infectious complications in these patients are far less common than in patients with hematologic malignancies. For instance, Elting et al. [2] reported that polymicrobial sepsis was 16 times less common per patient admission in patients with solid tumors than patients with acute leukemia. Similarly, Mayo and Wenzel [3] found that nosocomial bloodstream infections were 15 times less likely in solid tumor patients than leukemia patients. Nevertheless, patients with solid tumors complicated by infection are not unusual. We reviewed the last 100 cancer patients seen by our infectious disease inpatient consultation service and the majority (62) had underlying solid tumors (J. P. Flaherty, unpublished observation). Common scenarios included wound infection, pneumonia, intravascular catheter-related sepsis, and fever and neutropenia following intensive chemotherapy. The function of the immune system is a major factor in determining the spectrum of infections to which cancer patients are vulnerable. Unlike solid tumor patients, those with henlatologic malignancies have a malignant leukocyte clone that does not function correctly, replaces the marrow, or interferes with specific immune functions. The presence of a solid tumor may have indirect deleterious effects upon the immune system, but these are poorly understood. For example, depressed CD4 and CD8 lymphocyte counts have been measured in solid tumor patients [4], but the clinical significance of these abnormalities is uncertain. A subset of solid tumor patients receive intensive chemotherapy that is complicated by neutropenia. Patients with hematologic malignancies undergo prolonged, often uninterrupted, courses of combination chemotherapy, resulting in extended periods of neutropenia. Because the periods of neutropenia that most solid tumor patients experience tend to be brief, this patient population has a lower risk of infection than neutropenic patients with hematologic Grrry A . Noskit1 (en), M A N A G E M E N T O F INFECTIOUS COIMPLICATIONS IN C A N C E R PATIENTS.

0 1998. Kllr~vrrAcndernic Publishers, Boston. All rights reserved.

malignancies. When febrile and neutropenic, the solid tumor patient does have a high rate of infection. Pizzo and coworkers [5] documented a specific infectious source in 52% of febrile, neutropenic children and young adults with solid tumors, which is similar to the 55% of those with leukemia and 46% of those with lymphoma. While the majority of solid tumor patients do not experience severe treatment-related immunosuppression, they cannot be considered normal hosts. First, because solid tumors are more likely to occur in the elderly, many solid tumor patients have changes in the immune system associated with the aging process. Also, because solid tumor patients often have indolent presentations, malnutrition and cancer cachexia may be severe. Several comorbidities, such as chronic obstructive pulmonary disease (COPD) and poor dentition, may increase infection risk at presentation and throughout therapy. Solid tumor patients are at risk of acquiring nosocomial infections because they often undergo invasive diagnostic and therapeutic procedures, intravenous line placement, and hospitalization. As a result of antibiotic therapy, they are at increased risk of acquiring Clostridium difJicile colitis and infection with resistant organisms such as vancomycin-resistant Enterococci (VRE). Infections in of the non-neutropenic solid tumor patient contribute to significant morbidity and mortality. For example, the non-neutropenic counterparts in Pizzo's study mentioned earlier had identifiable sources of infection in 17 of 112 febrile episodes (15%) for solid tumor patients, which compared with 21 O/O for leukemic patients and 17% for lymphoma patients. In a review of bacteremic and fungemic episodes in patients with solid and liquid tumors at Memorial Sloan-Kettering Cancer Center (New York City), investigators reported a 26.6% mortality rate for 192 non-neutropenic episodes of sepsis in solid tumor patients [6]. As treatment regimens rate for these malignancies become more aggressive, the role of infection in these patients is likely to grow.

2. Infectious complications of lung cancer Lung cancer is the leading cause of cancer deaths in the United States, contributing to 141,285 deaths in 1990 [I]. In 1994, lung cancer accounted for 16% of all new cancer cases among males and 13% among females [I]. Infection plays a potentially critical role in the outcome of patients with lung cancer. The lung must continue to function as an organ of gas exchange during and after cancer therapy. Exposure to the external environment, and the attendant risk of exposure to infectious pathogens, must be maintained for gas exchange to occur. In addition, the lung serves as a vast vascular bed, and hematogenous seeding by bacteria and fungi can occur. The lung's defenses against these insults may be impaired by lung cancer and its treatment. The lung cancer patient has altered host defenses in several ways. Lung cancer commonly occurs in older adults and host defenses diminish with age.

Age-dependent changes that predispose to respiratory infections include an increased tendency to aspirate, decreased cough reflex, decreased mucociliary clearance, and increased oropharyngeal colonization by aerobic gramnegative bacilli [7]. Increasing age is also associated with decreased cellular and humoral immunity. In addition, there are changes that are associated with the cancer itself. The most apparent is local bronchial obstruction by a tumor mass, leading to impaired clearance of respiratory secretions. Lung cancer, particularly with advanced disease, has also been associated with diminished delayed cutaneous hypersensitivity reactions [8f.The lung cancer patient may present with serious comorbidities such as COPD or malnutrition, which can prolong postsurgical or postpneumonia ventilatory support or require tracheostomy. Such complications can increase pulmonary and nonpulmonary infection risks. Finally, treatment of lung cancer can increase infection risk. Most patients with stage I and I1 non-small cell carcinoma of the lung undergo surgery as initial therapy. In a study of 103 such patients, the risk of postthoracotomy pneumonia was 22% [9]. Small cell lung cancer patients, and recently some non-small cell lung cancer patients, have disease that can be responsive to chemotherapy and radiation; therapy-related immunosuppression and injury can increase infection risk, particularly in the lung. 2.1 Bacterial infections

Lung cancer patients are predisposed to develop focal lung infections secondary to bronchial obstruction by the tumor. Several older studies documented the simultaneous diagnoses of lung cancer and bacterial pneumonia, lung abscess, or empyema. In a review of 579 hospitalized patients with lung cancer at a Japanese university hospital over 15 years, 139 (24%) developed respiratory infections, most of which were bacterial [lo]. Patients with extensive disease were more likely to develop pulmonary infection than those with cancer at early stages. Twenty-seven percent of the pneumonias were postobstuctive by chest radiograph. Several older studies documented the simultaneous diagnoses of lung cancer and bacterial pneumonia, lung abscess, or empyema. Strang and Simpson [11] reported 70 patients with lung abscesses among 1,930 patients with a lung cancer diagnosis in Great Britain over an 11year period, an incidence of 3.6%. In many cases, partial obstruction by tumor led to postobstructive atelectasis and pneumonia. In a minority of cases, infection occurred secondarily within an area of tumor necrosis. Rarely, when an abscess was found distant to the tumor, aspiration was thought to be its etiology. Patients generally presented with the abrupt onset of cough productive of sputum, fever, and chest pain. A subset presented in a more indolent manner, with weight loss and anorexia as prominent symptoms. Sputum cultures were usually polymicrobial. In contrast to most patients with simple abscesses who improved with penicillin, most patients with cancer and abscess did not show clinical benefit or radiologic improvement after penicillin therapy alone.

Fig~rreI . A 7L-year-old Vietnamese male with squamous cell carcinoma of the soft palate and primary adenocarcinolna of the lung iollowing surgical resection of both tumors, radiation therapy. and chemotherapy with taxol. 5-fluorouracil, and hydroxyurea presented with recurrent fever. chills, and productive cough. The chest x-ray (A) showed a right upper lobe deformity and pleural thickening attributed to postoperative changes following right upper lobectomy. The C T scan (B) showed a large, thick-walled cavitary lesion in the right apex. Thc sputum Gram stain showed many polyn~orphonuclearleukocytes. The sputum culturc grcw moderate Alcoligertes .~ylasoxirlnns.Fever persisted despite ceftazidime and tobramycin. Because of his upbringing in Southeast Asia, the possibility of tuberculosis was considered. The sputum smear was positive for acid-fast bacilli. The sputum culture grew Mycobncreriutn luberc.l~los.is,and he responded to antitubercukous therapy.

Postobstructive pneumonias are thought to develop secondary to partial obstruction of an airway with overgrowth of bacteria distal to the obstruction; however, this may be an overly simplistic. Other factors likely contribute to the risk of pneumonia associated with an endobronchial tumor. The organisms

recovered from lung abscesses secondary to obstructing tunlor are frequently more virulent than those recovered from primary lung abscesses. In a review of 97 lung abscesses, Perlman and associates [12] found that Stpl~ylococcus nzrrezis and gram-negative enteric organisms were recovered from patients with underlying lung cancer more often than those without lung cancer. Cultures from primary abscesses (noncancer patients) were more likely to reflect normal upper respiratory flora, especially alpha-hemolytic streptococci. This shift to more virulent organisms in the Iung abscesses of cancer patients likely results from aspiration of altered oropharyngeal flora. Oropharyngeal colonization changes during illness, probably due to alterations in epithelial cell surface receptors, resulting in increased proliferation of aerobic gram-negative rods. Empyema occasionally complicates postobstructive pneumonia. In a review of 105 cases of empyema, only 7 (6.7%) were associated with postobstructive pneumonia secondary to bronchogenic carcinoma [13]. Kohno and

others [lo] reported only 2 empyemas among 148 episodes of pulmonary infection in hospitalized patients with lung cancer. Empiric therapy for postobstructive pneumonia or abscess should emphasize coverage of anaerobes, Staphylococcus aureus, and aerobic gram-negative bacilli. A variety of antibiotic regimens may be appropriate and should be guided by the results of sputum gram stains (and later cultures), previous antibiotic exposure (especially recent), and knowledge of local (community and institutional) antibiotic susceptibility patterns. Usual lung abscess treatment (e.g., clindamycin) may be adequate if cultures fail to identify aerobic gram-negative bacilli. Prolonged therapy may be required when bronchial obstruction prevents adequate drainage of the infected lung. 2.2 Mycobacterial infection The frequency of mycobacterial disease may be increased in patients with cancer. In a retrospective review at M. D. Anderson in Houston, Texas, the incidence of mycobacterial disease among their cancer patients was 65 cases per 100,000 persons, in comparison with 45 cases per 100,000 among Texans age 45-65 years old [14]. Kaplan et al. [15] reviewed 201 cases of tuberculosis that developed in cancer patients at Sloan-Kettering Cancer Center over 20 years. Lung cancer patients had the highest prevalence, 920 per 100,000, among solid tumor patients, which was second only to Hodgkins' disease overall. Lung cancer and head and neck cancer patients were more likely to present with tuberculosis at the time of cancer diagnosis; patients with the other neoplasms were more likely to develop tuberculosis while receiving cancer therapy [IS]. Historically, two mechanisms of tuberculosis reactivation in lung tumor patients have been invoked: first, tumor can break down granulomas harboring sequestered mycobacteria, or, second, malignancy-associated cachexia may impair cell-mediated immunity, resulting in reactivation [16]. As therapy for lung cancer has become more aggressive, chemotherapy-related immunosuppression may contribute to reactivation (Figure 1). In a population with high baseline rates of tuberculosis, autopsy data revealed that corticosteroids plus antineoplastic agents increased the incidence of mycobacterial infection compared with antineoplastic agents alone [17]. Of 304 lung cancer patients who came to autopsy, four died of tuberculosis. Each of the four patients with fatal tuberculosis had abnormalities on prechemotherapy chest x-rays consistent with old tuberculosis, suggesting that reactivation had occurred following chemotherapy. Three of four cases occurred within 3 weeks of institution of corticosteroid therapy. A high index of suspicion for tuberculosis in a cancer patient is indicated if the patient's history or epidemiological background suggests prior exposure or if unexplained or rapidly progressive pulmonary symptoms, signs, or chest xray findings develop. At diagnosis of solid tumor disease, we recommend that a tuberculin skin test be placed. Regardless of tuberculin skin test status, at

lung cancer diagnosis, sputum samples and lung biopsy samples should be sent for acid-fast bacillus (AFB) stain and culture. If the tuberculin skin test is positive and there is no evidence of active tuberculosis, prophylaxis with daily isoniazid for 6-12 months is indicated. If a patient has been previously adequately treated for a positive tuberculin skin test, a repeat course of isoniazid is not recommended. Whenever acid-fast bacilli are identified on smears or histopathology, or when mycobacteria are identified in respiratory cultures, therapy for presumed active pulmonary tuberculosis is indicated. Some of these smears or cultures will prove to represent contamination of specimens by nonpathogenic mycobacteria (e.g., Mycobacterium gordonae) and therapy can be discontinued. Some others may prove to represent true infection caused by atypical organisms (e.g., Mycobacterium kansasii or Mycobacterium aviumintracellulare) and therapy can be altered appropriately.

2.3 Fungal infection Individual case reports of relatively immunocompetent lung cancer patients with focally invasive Aspergillus infection are scattered throughout the literature. In these cases, necrotic tumor itself serves as the substrate in which Aspergillus germinates, colonizes, andlor invades. Saprophytic colonization nearby or within the tumor appears to be the most common presentation in relatively immunocompetent lung cancer patients. Smith and Bveneck [18] noted that these focal Aspergillus infections rarely cause life-threatening hemorrhage and uncommonly form fungal balls, in contrast to post-tuberculous aspergillomas. Symptoms from a growing tumor may result in an earlier recognition of Aspergill~lsinfection than post-tuberculous cases, before complications associated with more long-standing infection can develop. As a result, aspergillomas forming in the presence of lung cancers have rarely been described. In one case, misdiagnosis contributed to development of a fungal ball over several months. A 61-year-old male presented with hemoptysis and a multiloculated cystic lung lesion; over the next several months, while the patient was treated empirically for tuberculosis, a fungal ball developed within the cystic cavity [19]. At lobectomy, a mass of Aspergillus f~lmigatuswas found within a previously undiagnosed necrotic, cavitating adenocarcinoma. No evidence of tuberculosis was identified. Only rarely is focal fungal disease detected at the site of the tumor prior to any cancer therapy. Most Aspergillus infections in solid tumor patients are much more aggressive, developing in the setting of intensive immunosuppression associated with chemotherapy andlor radiation therapy. Aspergill~ispneumonia has become increasingly common in a subset of solid tumor patients. This increase appears to be associated with intensification of chemotherapy and prolonged neutropenia. At Memorial SloanKettering Cancer Center, a retrospective study noted twice as many Aspergillus infections during 1969-1970 as during 1964-1965 [20]. Of the 93 collected cases of Aspergillus infection in cancer patients, 14 involved solid

Figure 2. A 62-year-old woman with recurrent breast cancer following surgery, radiation, and chemotherapy was admitted to the hospital with several days of malaise, myalgias. and blurred vision. She was febrile to 39.4"C and pus could be expressed from her percutaneously inserted central venous catheter (PICC) site. Conjunctival hemorrhages (A) and purpuric skin lesions (B) were evident. Retinal exam also identified multiple chorioretinal abscesses. PICC line site, catheter tip, and blood cultures were positive for rnethicillin-resistant Sftrphylococclrs aureus. A transesophageal echocardiogra~nshowed no evidence of cardiac valvular vegetations. Nevertheless, a presumptive diagnosis of endocarditis was made, and she responded to 6 weeks of intravenous vancomycin.

tumor patients. Like the affected leukemia and lymphoma patients, the solid tumor patients who developed invasive Aspergillus were more likely to have leukopenia or a history of recent chemotherapy or corticosteroid therapy. A common presentation was the abrupt onset of unremitting fever and pulmonary infiltrates that did not respond to broad-spectrum antibacterial therapy.

The lung was the most common organ involved with Aspergillus infection, and bronchopneumonia and hemorrhagic infarction were the most common manifestations. The role of corticosteroids, whether exogenous or endogenous, in the development of invasive Aspergillus is illustrated by the following cases. Borkin et al. [21] reported a case of a 53-year-old male with history of adenocarcinoma of the left lung, following resection, who presented with brain metastasis. He was place on dexamethasone; 5 weeks later he developed fever and right chest pain while hospitalized for brain irradiation. Chest x-ray showed a dense right lower lobe infiltrate. He developed respiratory distress over 2 days and subsequently died after massive hemoptysis. Sputum cultures among other pathogens. Autopsy revealed grew Aspergillus f~~mignt~ls, necrotizing pneumonia of the right lung; microscopically, vascular invasion by fungal hyphae was seen. No evidence of malignancy was found in the lung. Smith et al. [22] reported a case of a 47-year-old male who presented with a 3-month illness associated with a 19-kg weight loss and a 2-week history of cough. He was found to be grossly cushingoid in appearance. A chest x-ray demonstrated a right hilar mass, a focal infiltrate, and lymphadenopathy. Hemoptysis prompted a transbronchial biopsy, which revealed invasive pulmonary aspergillosis. Small cell carcinoma was found on bone marrow examination. The Cushing's syndrome was attributed to ectopic hormone secretion by tumor. Following his death on the ninth day of hospitalization, autopsy showed widespread fungal abscesses. Animal models demonstrated the impact of corticosteroids on clearance of an aerosol challenge of AspergilIus spores [23]. The macrophages of untreated control mice effectively phagocytized the spores and the animals remained healthy. The macrophages of mice receiving corticosteroids failed to effectively phagocytize spores. The Aspergill~~s spores germinated and produced invasive hyphae; hemorrhagic bronchopneumonia developed and the majority of animals died. Sputum cultures that demonstrate AspergiEEus species in an immunocompetent patient with deteriorating pulmonary status or new infiltrates should prompt a more thorough investigation for evidence of invasive disease. In severely immunosuppressed individuals with deteriorating pulmonary status or new infiltrates, recovery of AspergiEE~.lsfrom the respiratory tract should prompt empiric antifungal therapy. If possible, immunosuppressive therapy should be discontinued. Amphotericin B is the usual therapy for invasive aspergillosis, but itraconazole has also proven effective and may be a reasonable compromise when the suspicion of invasive disease is not high or the risk of amphotericin B-related toxicity is considered substantial. Other fungal infections have been noted rarely in patients with lung cancer, but no clear association has been made. These have included blastomycosis [24], candidal infections, and cryptococcosis. In a survey of 170 Veterans Affairs hospitals over 12 years, 198 cases of blastomycosis were found; only 3 had underlying bronchogenic carcinoma and 2 had metastatic lung

lesions [25]. Isolation of Cnndida species from the lung has proved to be airway colonization in most cases. Thirty-one cases of Candida pneumonia, however, were documented by autopsy over a 20-year period at the M. D. Anderson Cancer Center [26]. Sixteen (52%) of these cases had underlying solid tumors; the remainder had hematologic malignancies. Associated with the development of Candida pneumonia were broad-spectrum antibiotics (28 patients), corticosteroid therapy (15 patients), and neutropenia (9 patients). 2.4 Pneumocystis carinii pneumonia

The risk of acquiring Pneumocystis carinii pneumonia (PCP) in non-AIDS patients appears to be highest in patients with hematologic malignancies and those with severe T-cell depression. PCP is a rare but recognized risk in immunosuppressed solid tumor patients. Yale and Limper [27], in a review of 116 non-AIDS patients with PCP at the Mayo Clinic, reported that 13% had underlying solid tumors, 30% had hematologic malignancies, 25% had received organ transplant~,and 22% had inflammatory disorders. The underlying solid tumors included brain tumors, lung carcinoma, breast carcinoma, colon carcinoma, renal cell carcinoma, and melanoma. Accumulated evidence from human and animal studies documents that chronic corticosteroid administration is a significant risk factor for PCP. In the aforementioned [27], 91% of non-AIDS patients with PCP had recently received several weeks of systemic corticosteroids. The median steroid dose associated with PCP in solid tumor patients was equivalent to 30mg of prednisone per day (range, 6.6-240mg per day); the median duration of corticosteroid use was 12 weeks (range, 4-14 weeks). A review of 142 PCP cases in non-AIDS cancer patients at Memorial Sloan-Kettering Cancer Center reported that 31% had underlying solid tumors and 67% had hematologic malignancies [28]. Eighty-seven percent of these patients had been on corticosteroids within 3 months of PCP diagnosis. A small, retrospective study among patients with Cushing's syndrome and opportunistic infections identified the highest morning cortisol levels in those with PCP [29]. The animal model of PCP is based on the fact that rats routinely develop progressive PCP when treated with corticosteroids [30,31]. In some patients, the development of symptomatic PCP may be associated with tapering of chronic steroids. Henson et al. [32]reported 10 patients with primary brain tumors complicated by PCP. All of these patients were receiving dexamethasone for greater than or equal to 1.5 months at the time of PCP diagnosis. Eight of the 11 cases presented during steroid taper. Similarly, when Poplin and colleagues [33] reported two cases of PCP in solid tumor patients; a patient with metastatic prostate carcinoma developed PCP while tapering his oral steroids. The mechanism by which chronic steroid use predisposes to PCP is unclear. Chronic corticosteroid therapy causes CD4 cell depletion and impaired macrophage function, which may allow the development of P. cnrinii

infection while limiting the host inflammatory response. Withdrawal of steroids may remove the antiinflammatory effects before the immunosuppressive effects have resolved, leading to clinical exacerbation. In general, the presentation of PCP in non-AIDS patients is more acute than in AIDS patients. Kovacs et al. [34] found that prior to presentation, non-AIDS patients had a median duration of pulmonary symptoms of 5 days (range, 1-42 days), whereas AIDS patients had a median duration of 28 days (range, 1-270 days). Non-AIDS patients were more likely to have fever and severe hypoxemia, and showed a wider range of respiratory rates. In the report by Henson et al. [32],brain tumor patients with PCP had pulmonary symptoms a median of 7.4 days before admission (range, 1-30 days). Of 10 patients, 8 had dyspnea and 6 had fever. Chest x-rays upon admission varied from normal (I), to an isolated focal infiltrate (21, to diffuse or bilateral infiltrates (7). Lactate dehydrogenase levels were elevated, ranging from 336 to 1284UlL (median 510UlL). Non-AIDS patients with PCP, and specifically those with solid tumors, have a 10-fold lower organism load [35]. While the low organism load may reduce the diagnostic yield of bronchoscopy, it does not appear to reduce the severity of illness. In the series reported by Yale and Limper [27], 7 of 15 patients with PCP and underlying solid tumors developed respiratory failure and died. In Kovac et al. [34], the survival of AIDS and non-AIDS patients presenting with PCP was not significantly different (57% and 41%, respectively). Solid tumor patients with severe cell-mediated immunosuppression, for example, those who are severely malnourished, bone marrow transplant recipients, and those who receive chronic corticosteroids, are candidates for PCP prophylaxis (trimethoprim-sulfamethoxazole, dapsone, or monthly aerosolized pentamidine). Individuals receiving chronic steroid therapy, especially high-dose therapy, should continue on PCP prophylaxis during steroid tapering [36].

2.5 Viral infection Viral pneumonia is a rare cause of infectious complications in solid tumor patients. Camazine et al. [37] reported three cases of herpes simplex virus (HSV) type 1 pneumonia occurring in patients who had undergone recent thoracotomy for carcinoma involving the lung. A 52-year-old female presented with fever and hypoxemia on postoperative day 2 following thoracotomy for pulmonary metastasis of rectal carcinoma, a 72-year-old male developed fever and hypoxemia on postoperative day 3 after thoracotomy for squamous cell carcinoma of the lung, and a 72-year old male with mesothelioma presented with fever, hypoxemia, and respiratory failure on postoperative day 4. Two of the three patients developed diffuse interstitial infiltrates; one developed a focal, progressive infiltrate. Respiratory cultures were positive for herpes simplex virus (HSV) and, in one case, cytomegalovirus (CMV) as well; no other pathogens were isolated. Bronchial washings

Figzrre3. A 49-year-old woman with recurrent squamous cell cascino~naof the cervix, treated with radiation therapy and multiple wide excisions, presented with persistent pelvic pain. On admission, she was febrile and had a WBC coutlt of 5300/mm3and an ESR ol119. CT scan of the pelvis showed a fluid collection within the pelvis with an enhancing rim of soft tissue. Inlraoperative cultures grew many beta-hemolytic streptococci. group F. Bone scan and MRI showed evidence of osteomyelitis of the left ileum.

revealed intranuclear inclusions consistent with HSV infection. All three patients responded promptly to acyclovir. The authors recommend delaying cardiothoracic surgery in the presence of perioral HSV lesions. Significant immunosuppression, similar to that experienced by organ transplant patients and bone marrow transplant patients, appears to be necessary to develop pneumonia secondary to CMV, respiratory syncytial virus (RSV), parainfluenza virus, and adenovirus. It is likely that solid tumor patients undergoing autologous bone marrow transplantation are at risk for these infections. 2.6 Noninfectio~cscauses of pulr?zonary bzfiltrrrtes in solid tlinzor patients

Radiation pneumonitis typically presents 3-4 months after irradiation with insidious onset of nonproductive cough, fever, and shortness of breath [38]. Symptoms frequently become apparent when corticosteroids are tapered [39]. Physical findings are uncommon, but include pulmonary consolidation, pleural

friction rub, or pleural effusion [39]. When patients become symptomatic, radiologic changes are evident and are almost always precisely limited by the edges of the radiation field. During the period 2-4 months after irradiation, many more individuals develop abnormalities on chest x-ray than develop symptoms. Ground glass opacification or haze is common in the early stages [39], possibly followed by dense infiltrates. Because radiation pneumonitis can respond dramatically to steroids, diagnosis is clinically important. Some chemotherapeutic agents, such as actinomycin D and adriamycin, have been associated with reactivation of radiation pneumonitis [40]. Other noninfectious causes of pulmonary infiltrates in patients with lung cancer include congestive heart failure, pulmonary emboli, pulmonary hemorrhage, adult respiratory distress syndrome, drug toxicity, chemical aspiration, and progression of tumor.

2.7 Evnluntion The urgency with which diagnosis of pulmonary infiltrates in a lung cancer patient must be made depends upon the level of immunosuppression. A lung cancer patient who is severely malnourished, neutropenic, or receiving highdose corticosteroid therapy who develops pulmonary infiltrates should undergo an urgent diagnostic procedure. In addition, empiric broad-spectrum antimicrobial therapy should be considered. If a patient is relatively immunocompetent, on the other hand, a more methodical approach can be undertaken, although invasive procedures are often indicated. The relatively immunocompetent patient with lung cancer is more likely to develop focal infiltrates. These patients may benefit from a CT scan of the chest with intravenous contrast to better define the extent of disease and facilitate obtaining a biopsy. Sputum for gram stain, culture, AFB smear, mycobacterial culture, and fungal culture should be obtained. A tuberculin skin test should be placed. Bronchoscopy with bronchoalveolar lavage (BAL), transbronchial biopsy (TBB), or percutaneous transthoracic needle biopsy (if the lesion is located perphierally) are options for rapid diagnosis. If tissue is obtained, aerobic, anaerobic, fungal, and rnycobacterial studies, as well as cytology, are indicated. Because fungi can colonize the airways of relatively immunocornpetent individuals, biopsy to detect tissue invasion is often indicated. Lung cancer patients who develop diffuse pulmonary infiltrates typically occur in those with severe immunosuppression. Diffuse pulmonary infiltrates are more likely to be secondary to infection when the patient has received chemotherapy, when fever develops with the radiographic changes, and when the radiographic changes occur rapidly [41]. Nevertheless, clinical presentation is not predictive enough to direct therapy with confidence. In the setting of diffuse pulmonary infiltrates and immunosuppression, an invasive diagnostic procedure (BAL, TBB, or open lung biopsy) should be considered early. Meanwhile, specimens of blood and sputum (if accessible) should be sent for

bacterial, fungal, and mycobacterial culture. Because viruses can cause diffuse pulmonary infiltrates in the setting of severe immunosuppression, a nasopharyngeal or oropharyngeal swab may be sent for viral culture and, if available, direct staining for influenza or RSV. Skin should be inspected daily in the immunocompromised patient, and new skin lesions should be biopsied promptly. Bronchoscopy and BAL is often utilized as the first-line diagnostic procedure in the immunosuppressed patient with new diffuse pulmonary infiltrates. Specific studies of solid tumor patients have not been performed, so recommendations must be derived from reports in which leukemia and lymphoma patients predominate 1421. When the patient is severely immunosuppressed, specimens should be sent for gram stain and quantitative culture, acid-fast stain and mycobacterial culture, fungal stain and culture, Pneumocystis stains, Legionella direct fluorescent antibody (DFA) staining and culture, viral culture and DFA or in situ hybridization for HSV and CMV, possibly cultures for Chlamydia pneumoniae and Mycoplasma, and DFA and culture for respiratory viruses (e.g., influenza virus, parainfluenza virus, and RSV). Because BAL generally dilutes lower respiratory tract secretions by a factor of 1:10 to 1:100, the diagnostic threshold for bacterial culture is 1 0 4 C F U / m ~However, . some organisms, including Mycobacterium tuberculosis, Legionella species, and Nocardia species, should be considered pathogens whenever isolated. A sample should also be sent for cell count and differential, and for cytopathology. The cell count and differential may be used to assess specimen adequacy. For example, squamous and bronchial epithelial cells accounting for >1% of the total cells suggests contamination by oropharyngeal flora. Detection of plentiful hemosiderin-laden macrophages by direct microscopy suggests pulmonary hemorrhage. BAL appears to have a good yield in the diagnosis of PCP, tuberculosis, CMV, and aspergillosis [43]. TBB is more invasive than BAL alone, but improves the diagnostic yield in some infectious and noninfectious processes. However, TBB is associated with an increased risk of pneumothorax and hemorrhage. Open lung biopsy (OLB) is associated with greater morbidity than either BAL or TBB, but may be considered when BAL and TBB fail to provide a diagnosis or are contraindicated because of respiratory instability or bleeding risk.

3. Infectious complications of breast cancer Breast cancer accounts for approximately one third of the new cancer cases diagnosed annually in females in the United States [I]. Almost all of these patients undergo some surgical procedure of the involved breast and ipsilatera1 lymph nodes. Most infectious complications in breast cancer patients involve skin and soft tissue, which has been altered by surgery and/or irradia-

tion. These infections occur both immediately postoperatively and very late, up to several years after cancer therapy. Increasingly, patients with a wide range of disease stages are now receiving some combination of chemotherapy, radiation, and surgery. Limited information is available regarding the impact of combined-modality therapy on infection rates, although chemotherapy and radiation can theoretically delay wound healing and impede wound drainage. The subset of patients with advanced breast cancer or metastatic disease who undergo bone marrow transplantation or other intensive chemotherapy protocols are at risk of additional complications, similar to those that leukemia and lymphoma patients experience (Figure 2). 3.1 Postoperative infectio~lscomplications of breast cancer patients Despite the fact that most breast cancer surgery is classified as clean, acute wound infection rates in the range of five times the rate of other clean surgeries has been reported [44]. Disruption of the skin integrity and contamination by skin flora accounts for most wound infections within the first 6 weeks after breast surgery [44]. By far the most common organisms involved in these wound infections are streptococcal species, Staphylococcus nureus, and coagulase-negative staphylococci [44,45]. A recent surveillance study of surgical wound infections revealed that variable wound infection rates appear to be dependent on the type of breast cancer procedure performed [44]. More extensive surgery resulted in higher wound infection rates: simple breast biopsies had a rate of 2.3%, lumpectomy with lymph node dissection had a wound infection rate of 6.6%, and mastectomies had a rate of 19% (P < .05). In addition, variation in surgical technique, such as drain type (e.g., closed suction, Jackson-Pratt) and placement (e.g., new skin incision separate from the cancer incision vs. through the wound), also influenced infection rates [44]. Alternatively, other investigators have reported similar wound infection rates for modified radical mastectomy and lumpectomy. For example, Vinton et al. [46] reviewed 387 modified radical mastectomies and 173 lumpectomies between 1983 and 1989. The wound infection rates were not significantly different for modified radical mastectomy (15%) and lumpectomy (13%). The impact of the shift toward more conservative surgery on breast cancer wound infection rates is unclear.

3.2 Late-onset infectious conzplications of breast cancer

A subset of breast cancer patients experience skin and soft tissue infections months to years after therapy concludes. Historically, the most common late infection in postmastectomy patients was upper extremity cellulitis ipsilateral to the breast cancer surgery and lymph node dissection. A classic presentation is sudden onset of a painful and erythematous rash, which spreads rapidly on the involved upper extremity, with or without fever. Blood cultures, positive in only a minority of cases, may yield skin flora. Many, but not all, of the patients

who experience one or more episodes of cellulitis had obvious pre-existing chronic swelling and lymphedema of the ipsilateral upper extremity. Lymphedema was a common complication after radical and, less frequently, modified radical mastectomy. The incidence of chronic lymphedema following these surgeries has been estimated at 15% [47]. Risk factors for the development of lymphedema included the following: a greater number of lymph nodes removed, and delayed postoperative wound healing, cellulitis, radiodermatitis, hematoma, seroma formation, and skin flap necrosis [47]. The rates of upper extremity cellulitis following lumpectomy, radiation therapy, and chemotherapy are unavailable, but appear to be lower than those following radical or modified radical mastectomy. Clinical lymphedema appears to be a marker for increased risk of cellulitis, but subclinical abnormalities of lymphatic drainage can also predispose to infection. Bertelli et al. [48] reported that 7 of 21 patients with ipsilateral upper extremity cellulitis following breast cancer surgery did not have clinically detectable lymphedema. Breast cancer patients who experience a single episode of cellulitis appear to be at risk of recurrence; in one series, 11 of 15 patients with cellulitis after breast cancer surgery had more than one episode [49]. A breast cancer patient, especially one who experiences difficulties with wound healing postoperatively, should be considered at life-long risk of developing upper extremity cellulitis. Osteomyelitis and septic arthritis of the ipsilateral shoulder can present years after breast cancer therapy; this complication is apparently also linked to preexisting lymphedema. Chaudhuri et al. [50] described five cases in which ipsilateral humeral osteomyelitis and septic arthritis presented 2-12 years after breast cancer surgery. All five patients had received a radical or modified radical mastectomy followed by radiation therapy. A striking characteristic of these cases was indolent presentation - ipsilateral shoulder pain and restricted movement for 4 or more months, without fever. Four patients had an erythrocyte sedimentation rate (ESR) of >100mm/h. Radiologic studies confirmed the diagnosis of osteomyelitis: all five patients had positive bone scans, three of the five patients had findings of osteomyelitis on plain radiographs, and none had findings consistent with radiation necrosis of adjacent bones within the radiation field. It is unclear whether these serious late complications, osteomyelitis and septic arthritis, will occur following lumpectomy. Despite the shift in management of breast cancer to more conservative surgery, skin and soft tissue infections continue to be the major delayed complication of breast cancer. Instead of ipsilateral arm cellulitis, focal cellulitis of the involved breast has been reported following lumpectomy with axillary lymph node dissection. Rescigno and coworkers [51] documented 20 episodes of breast cellulitis in 11 patients. The authors estimated the incidence of this complication among patients after lumpectomy to be 2.5-3.0%. Time from completing radiation to first episode of cellulitis ranged from 9 days to 4 years (median, 4.3 months). Each of these patients presented acutely with rapidly spreading erythema, warmth, and tenderness of the breast. The origi-

Figirre 4. A 52-year-old man with squamous cell carcinoma of the tongue was treated with surgery, radiation, and administration of cisplatin. 5-fluorouracil. and hydroxyurea. H e p1,esented t o the hospital with fever and drainage from the site of a lormer lcft subclavian Port-A-Cath. C T scan of thc chest showed air or gas collections in the left chest wall and anterior mediastinum. and a soft tissue mass in the left superior rncdiastinum. At surgery. infection was identified to have tracked from the port site along the catheter tract into the metliastinuni. Intraoperative cultures grew Klcbsiell~ptzertri?oriioe, alpha-hemolytic streptococci, and Eikenellrr c o r r o r l e ~ ~ . ~ .

nal site of erythema was often removed from the surgicaI scar. In eight cases, erythema spread beyond the breast tissue to the back, shoulder, or arm. Fever or breast swelling was present in some, but not all, patients. Six cases required hospitalization. Two patients experienced repeated episodes of chronic recurrent celIulitis. Episodes of acute and recurrent breast cellulitis responded to antibiotics directed against gram-positive cocci (e.g., oxacillin, cefazolin). Finally, four patients experienced chronic persistent cellulitis, which did not fully resolve despite prolonged antibiotic therapy. This apparent noninfectious erythema may be to radiation-induced inflammation [52].This latter condition is treated with corticosteroids. Late development of breast abscesses may be unique to the breast cancer patients who received lumpectomy and radiation. Keidan et al. [53] reviewed

112 lumpectomies, finding 7 breast abscesses that developed 1.5-8 months following surgery. This 6% abscess incidence was higher than expected for clean surgery. Aspirated fluid grew Staphylococcus aureus in 3 cases and S. epidermidis in 3 cases. Postsurgical manipulation, such as previous seroma aspiration, was a risk factor for abscess development. Notably, each of the seven patients with breast abscess had received local irradiation. Irradiation may result in impaired lymphatic drainage and local ischemia, which may increase the risk of infection. Axillary node dissection likely contributes to poor lymphatic drainage.

3.3 Infectious complications in breast cancer patients undergoing bone marrow transplantation Breast cancer patients who receive high-dose chemotherapy and autologous bone marrow transplant are at risk for complications similar to those of leukemia and lymphoma patients. The risk of infection is related to the duration of severe neutropenia, and clinical manifestations are similar to other patients with chemotherapy-related neutropenia. Mudad et al. [54] reported that 12 of 206 patients with breast cancer following autologous bone marrow transplantation developed CT scan findings consistent with hepatosplenic candidiasis 6-12 weeks after transplantation. Only 4 of the 12 breast cancer patients were febrile when hepatosplenic candidiasis was diagnosed by CT scan. Only three patients had positive blood cultures for Candida. As with other patients, the most common laboratory abnormality was an increased alkaline phosphatase. A subset of patients underwent purging of their bone marrow complicated by prolonged neutropenia (median 24 days vs. 12 days in nonpurged bone marrow patients). Seven (22%) of the 32 patients on the purge protocol developed hepatosplenic candidiasis. In contrast, only 5 (2.8%) of 174 patients receiving nonpurged marrow developed hepatosplenic candidiasis. Outcome may be better for the breast cancer patient than patients with hematologic malignancies. Two patients received no antifungal therapy, yet survived. The remaining patients were treated with fluconazole or amphotericin B with or without Aucytosine. No deaths were attributed to the fungal infections, although five died from cancer progression and one died from drug toxicity.

3.4 Management of infections in breast cancer patients New onset of erythema in the breast tissue or ipsilateral upper extremity in a patient with breast cancer is most likely cellulitis. However, noninfectious causes of skin erythema should also be considered, especially when erythema occurs within the radiation field and when it persists after a trial of antibiotic therapy. Radiation changes alone can cause chronic, sometimes progressive, inflammation. In addition, "radiation recall" of the chest has been reported when patients receive chemotherapy agents long after their radiation course

[55].Differential diagnosis of a focal breast abscess includes fat necrosis, which often presents as a tender hard mass with or without appreciable inflammation [56]. If fluid is not detected, biopsy may be necessary to differentiate tumor recurrence, fat necrosis, and abscess. Stewart-Treves syndrome, a very rare type of angiosarcoma that occurs exclusively in the ipsilateral arm of patients following mastectomy, develops subacutely in the setting of chronic lymphedema; skin biopsy should be considered if swelling and erythema do not resolve with a seemingly appropriate course of antibiotic therapy. The underlying pathophysiology of most infectious complications of breast cancer patients is damaged local host defense secondary to surgery and radiation therapy, namely, an impaired skin barrier, injured microcirculation, and impaired lymphatic drainage. Patients usually present with acute onset of erythema, pain, possibly with fever and leukocytosis. The infections tend to be local, involving skin, underlying dermis, and possibly fat. Because the infections tend to be focal, blood cultures are rarely positive, but should be performed. Tissue samples for culture are also rarely positive; unless there is fluctuance, tissue aspiration is not recommeded, so as to avoid further damage to the skin barrier. When cultures are positive, isolated pathogens are almost always gram positive; thus, in most cases anti-staphylococcal coverage is adequate. The patient should be treated with anti-staphylococcal coverage, such as nafcillin, cefazolin, clindamycin, or vancomycin. Oral antibiotics, such as cephalexin and dicloxacillin, may be used to complete a course once swelling and erythema have been reduced. Any patient who has undergone lymph node dissection should be educated about protecting the ipsilateral arm from trauma, pressure, or damage. Patients are advised to avoid phlebotomy, intravenous catheters, or blood pressure monitoring in the upper extremity ipsilateral to previous breast cancer surgery [57]. The use of compression garments and elevation of the affected arm can reduce lymphederna. Early aggressive therapy of cutaneous fungal infections can minimize local skin breakdown [58]. Those who experience poor wound healing initially, develop clinical lymphedema, or experience an episode of cellulitis should take measures to prevent repeated episodes. Selected patients with recurrent cellulitis may benefit from chronic suppressive antibiotic therapy. 4. Infectious complications of abdominal and pelvic cancer Malignant tumors do not honor mucosal barriers. As a consequence, infectious complications of solid tumors in the abdomen and pelvis appear to occur mainly secondary to mucosal invasion of tumor, with subsequent local abscess formation or dissemination. The gastrointestinal tract serves as the dominant source of bacteria within the abdomen and pelvis. When the gastrointestinal tract is intact, a sterile peritoneum lays within millimeters of colonic luminal contents bearing a bacterial density of 10" per gram of dry weight [59].

Similarly, an intact ileocecal valve and proper peristalsis separate the mainly aerobic small bowel luminal contents, bearing only 10' to lo6 bacteria per milliliter, from colonic contents, most of which is anaerobic [59]. Tumor invasion of gastrointestinal structures causes contamination of previously sterile structures, spaces, or fluids. As a result, the infections that result from cancers in the abdomen and pelvis are often polymicrobial. In a 10-year review of polymicrobial septicemia in cancer patients, Elting and associates [2] found three types of underlying malignancies most frequently: hematologic malignancy in 47%, genitourinary cancers in 16%, and gastrointestinal cancers in 13%. In addition, bacterial flora within the gastrointestinal lumen may shift toward more virulent organisms secondary to chemotherapy, irradiation, or in association with the malignancy itself. 4.1 Streptococcus bovis colonization and endocarditis

The relationship of Streptococcus bovis bacteremia and colon carcinoma was uncovered during the 1970s. Klein et al. [60] reported two patients with adenocarcinoma of the colon who developed S. bovis endocarditis. They showed that fecal carriage of S. bovis in patients with colon carcinoma was significantly greater (56%) than in healthy controls (lo%), patients with nongastrointestinal carcinomas, and patients with other gastrointestinal disorders. The authors recommended evaluation for colon carcinoma in all cases of S. bovis endocarditis. Subsequently, these same investigators prospectively evaluated 29 patients with 30 episodes of S. bovis endocarditis [61]. Twelve patients were found to have colonic neoplasms, 3 of which were malignant; 3 other patients had undiagnosed colonic masses and 10 had diverticulosis. Of note, the majority of the patients with colonic neoplasms and endocarditis did not have neoplasms that invaded the muscularis mucosa: the presence of a friable intraluminal mass or diverticuli, which also bleed easily, may allow bacteria access to the bloodstream. Following the reported association of S. bovis and colon carcinoma, there have been case reports of patients with endocarditis involving other streptococci (e.g., S. snlzguis, S. agnlncfiae) who were found to have colonic neoplasms [62,63] and gastric carcinoma [64]. The diagnosis of Streptococc~isbovis bacteremia and endocarditis is based on recovery from blood cultures. Once a blood culture isolate is identified as S. bovis, it is necessary to establish whether endocarditis is likely by performing serial blood cultures and echocardiography. In addition, the lower gastrointestinal tract should be evaluated for the presence of disease. Colonoscopy or a barium enema may be performed; if a barium enema is negative, however, a colonoscopy is indicated. 4.2 Clostridium septicum myonecrosis and bacterenzia

An association between malignancy and Clostridi~imsepticutn myonecrosis has been noted. Particularly vulnerable to this acute, life-threatening bacterial

infection are those with leukemia or occult colon or rectal carcinoma [65-671. In a literature review covering 42 years, Kornbluth et al. [67] identified 162 cases of C. septicum infection; 34% had colorectal carcinoma and 47% had hematologic malignancies. Occult malignancies were found in 37%. C. septicur?~,a sporulating gram-positive, toxin-producing rod, is an uncommon pathogen in humans. In a review of 114 cases of clostridial infection by Gorbach and Thadepalli [68], only 3 were caused by C. septicurn. In a similar review, Kornbluth and colleagues [67] review four studies of clostridial infection: of a total of 612 clinical isolates of CIostriditirn, only 6 (1.3%) were C. septicurn. The relative risk of C. septicurn infections in cancer patients is unknown, but an increased risk of Clostridiurn septiciirn infections among patients with colorectal cancers seems apparent. Typically, a patient with C. septicurn myonecrosis presents with a history of acute onset of severe focal pain, usually of an extremity, fever, and a toxic appearance. The painful site may initially look normal, but within hours the involved skin becomes discolored and edematous, bullae form, and the discolored area enlarges rapidly [67]. As a late finding, the subcutaneous tissue becomes crepitant [69]. Some patients present with diffuse abdominal pain as the most prominent initial symptom [67]. Patients progress rapidly from a toxic presentation to shock and ultimately death, often within 48 hours. The site of myonecrosis, such as the shoulders, limbs, or chest wall, is usually is distant from the carcinoma; therefore, hematogenous spread of Clostridi~~m presumed. The most common sites of underlying adenocarcinoma in patients with C. septicum are the cecum and distal ileum. Tumor invasion into the mucosa is thought to supply access of the organism to the blood stream. In a review of the published literature, Kaiser et al. [70] described 23 cases of distant clostridial myonecrosis, 12 of which had underlying colon or rectal carcinoma. C. perfringens was isolated in approximately half of these cases, causing a syndrome indistinguishable from that caused by C. septicurn. The 12 patients with underlying colorectal cancer had mucosal breakdown at the ileum, colon, or rectum documented at surgery or autopsy. Seventeen of the 23 patients died, many within a few hours of admission. Of 59 cases with C. septicurn bacteremia reviewed by Koransky et al. [66], 21 had solid tumors, 14 of which were colon cancers. Of the 28 patients autopsied, a colonic lesion was documented in 17. Seven autopsies demonstrated evidence of "fecal peritonitis from bowel perforation or gangrene." Additional evidence of hematogenous spread comes from case reports of colon cancer patients with C. septicum septic arthritis [71,72], bacteremia, and septic shock [73], and with bacteremia and polymicrobial abscess within a hepatic metastasis [74]. Host factors such as diabetes meIlitus [67,75], granulocytopenia [66], and atherosclerosis [66] may increase the risk of developing C. septicurn infection. Nineteen of 100 patients with C. septicurn infection in the review by Kornbluth et al. [67] had diabetes rnellitus. The diabetic patients were significantly more likely to have an occult malignancy than nondiabetic patients. Kudsk [75] reported five cases

of C. septicurn myonecrosis in diabetic patients, each of whom was found to have an occult malignancy. The cellulitis, bullae formation, and myonecrosis are thought to occur secondary to several locally produced toxins. The diagnosis of C. septicurn myonecrosis should be made rapidly, based on clinical presentation in those with and without known malignancies. Blood cultures should be obtained before initiating antibiotics. Gram stain and culture of percutaneous tissue aspirates and bullae aspirates should be performed emergently [70]; identification of short, plump gram-positive rods suggests clostridial infection in the appropriate clinical setting . Because gas on x-ray may be a late finding, performing such studies should not delay debridement. If the patient presents with abdominal symptoms or signs, exploratory laparoscopy may be considered as well. Patients may require multiple surgeries over several days following presentation. High-dose penicillin G is the traditional drug of choice for C. septicum infection. Clindamycin is favored by some investigators because its mechanism of action may inhibit clostridial toxin production and more effectively halt the progression of established disease. Initial antibiotic management may be broadened to include gram-negative, staphylococcal, and anaerobic coverage, until the diagnosis is confirmed by culture results and the extent of intraabdominal disease is known. The patient should be aggressively treated for sepsis and monitored closely for hemodynamic deterioration. It is unclear whether patients benefit from adjunctive hyperbaric oxygen therapy.

4.3 Pyogenic abscesses Pyogenic liver abscess is another extremely rare entity that can complicate gastrointestinal malignancies. There appear to be two typical presentations for pyogenic liver abscesses in patients with gastrointestinal malignancies. First, abscesses may herald the discovery of a previously undiagnosed, usually advanced, luminal or pancreaticobiliary malignancy. Secondly, pyogenic liver abscesses occur in patients with known malignancies, many of whom have undergone recent gastrointestinal procedures. In a review of 20 cases of pyogenic liver abscess, 5 had underlying gastrointestinal carcinomas [76]. Patients frequently presented with fever of unknown origin. Most pyogenic liver abscesses associated with colon carcinomas were polymicrobial and included anaerobes and enteric gram-negative rods. Organisms presumably spread from areas of mucosal breakdown to the liver via the portal circulation. Pancreaticobiliary malignancy may obstruct the biliary tract, resulting in ascending cholangitis, and then in multiple hepatic abscesses. The development of one or more liver abscesses in a known cancer patient is extremely rare. A review of liver abscess in cancer patients at the National Cancer Institute yielded only 37 patients over 35 years [77].The etiology of these abscesses was bacterial in 17 and fungal in 20. Twelve of the 17 patients

with bacterial liver abscesses had solid tumors; the remaining 5 had hematologic malignancies. Most bacterial abscesses were polymicrobial, with gramnegative and anaerobic organisms recovered on culture. Marcus and associates [77] found that recent gastrointestinal instrumentation was a strong risk factor for liver abscess development. Ten of the 17 patients with bacterial abscesses had undergone either a surgical or radiologic procedure on the gastrointestinal system, such as surgical resection of liver metastases and biliary stent placement. Recent neutropenia did not appear to be a risk factor for development of bacterial liver abscesses; only one such patient had been neutropenic within 60 days of presentation. Hepatic and splenic abscesses have occurred after invasive procedures for hepatocellular carcinoma. Okada et al. [78] reported a case of a 56-year-old woman with hepatocellular carcinoma and distant cholecystoduodenostomy who presented with fever and leukocytosis while hospitalized after her third percutaneous ethanol injection procedure. One of her two injected liver lesions had become more hyperechoic, consistent with gas formation. This lesion was drained percutaneously, yielding Klebsiella pneumoniae. Isobe et al. [79] reported a case of probable splenic abscess in a cirrhotic woman with hepatocellular carcinoma who became febrile with new left upper quadrant pain 1 day after percutaneous ethanol injection. By ultrasound, her spleen had multiple hyperechoic lesions. These lesions resolved by 10 days with antibiotics alone. Whereas breakdown of the mucosal barrier of the gut or pancreaticobiliary system is the basis of pyogenic liver abscess in the undiagnosed colon carcinoma patient, in the known cancer patient undergoing therapy, recent gastrointestinal instrumentation appears to be the primary risk factor. Pyogenic hepatic abscesses should be suspected in a patient with fever, malaise, right upper quadrant pain, andlor jaundice or in a patient with persistent unexplained fever. Liver abscesses may be identified by CT scan or by ultrasound. Blood cultures should be obtained upon admission arid when the patient is febrile. If a patient with a pyogenic abscess has no clear pancreaticobiliary obstruction, a search for a lesion within the gastrointestinal tract is indicated. Treatment of pyogenic liver abscesses is controversial and evolving. Many continue to rely on surgical or percutaneous drainage [SO]. Recently, some patients have done well with prolonged antibiotic therapy alone, following percutaneous aspiration and identification of the infecting organisms and their antimicrobial susceptibilities [Sl]. Empiric broadspectrum antibiotics, including gram-negative and anaerobic coverage, is indicated initially; choices such as piperacillinltazobactam, ticarcillinl clavulanate, imipenem-cilistatin, or meropenem are among the many appropriate initial regimens. Use of an aminoglycoside can often be avoided or used only in the initial days of therapy while awaiting culture results. Antibiotic coverage should be tailored to the results of antimicrobial susceptibility testing. Because polymicrobial infections are common in these infections, retaining

broad coverage agent gram-negative and anaerobic bacteria is usually indicated on the assumption that more organisms may be involved than can readily be identified. Periodic imaging studies can be used to monitor the resolution of the abscesses; up to 4 months of antibiotic therapy may be warranted. 4.4 Salmonella infections

Salmonella species infect humans via the gastrointestinal tract. In the healthy host, Salrnorzella infections are associated with ingestion of a large inoculum that manages to evade the acid barrier of the stomach. Individuals who have added risk of salmonellosis include those with cellular immune dysfunction, the elderly, and those with achlorohydria. An increase in Salmonella infections have been associated with malignancies, especially metastatic disease. In a retrospective study of 95 patients with salmonellosis and neoplastic disease at Memorial Cancer Center [82], there was a clear relationship between Salrrzonella spp. infections and leukemias and lymphomas. The hematologic malignancies accounted for 46 cases; the most common organism isolated from In contrast, nontyphimurium Salmonella these patients was S. fyphirnl~ri~mt. were responsible for most infections among the 40 solid tumor patients. Of solid tumors, gastrointestinal malignancies were most commonly associated with SnIrnonella infection. Intraabdominal malignancy appears to be a risk factor for development of serious Salrr.zonella infections. In a 7-year review of Salr?zonella infections, Han et al. [83] found 4 such infections among 1258 admissions with underlying colon cancer (0.3 %); 3 cases among 2891 admissions with bladder, uterine, or ovarian cancers (0.1%); and 9 cases among 1235 admissions with hematologic malignancies (0.7%). The rate of salrnonellosis for patients admitted with all types of malignancy was approximately 13 times the rate for those without malignancy. Manifestations of infection in patients with intraabdominal cancers varied from gastroenteritis to focal infections to sepsis. Other solid tunloss may rarely be associated with salmonellosis. Two case reports of elderly patients with advanced, undiagnosed lung cancer presenting with Salt~zonellapulmonary infections have been reported [84,85]. Host factors that appear to predispose cancer patients to Snlrqaonella infections include increasing age, impaired cellular immunity, altered gastrointestinal function (e.g., postoperative ileus), acl~lorohydria (secondary to atrophic gastritis or H2 blocker or antacid therapy), chemotherapy, and corticosteroid therapy. Salrr~onella infections can be treated with quinolones, trimethoprimsulfan~ethoxazole,ampiciIlir1, cliloramphenicol, or tetracycline. Antibiotic choice should be tailored to the sensitivities obtained because S~rlrr.zonelln species have high resistance rates. Uncomplicated gastroenteritis in a healthy host is generally not treated with antibiotics. Because of the increased risk of dissemination in solid tumor patients, antibiotic therapy should be considered.

4.5 Gynecologic cancers In a retrospective study of infectious morbidity on a university gynecologic oncology service, Brooker and others [86] found 20 (6%) of 494 patients had a serious infection on admission and 54 patients (11%) developed serious infections during hospitalization. The infection rate per admission varied by cancer origin: 8% for cervical cancer, 7% for uterine cancer, 3% for ovarian cancer, and 21 % for vulvar cancer. Bacteremia in gynecologic cancers may be caused by a single organism or be polymicrobial, with the primary tumor the likely portal of entry [Z]. In general, the organisms involved in gynecologic infections are normal flora of the vagina, gastrointestinal tract, and skin (Figure 3). Infectious complications of gynecologic cancers at diagnosis highlight underlying problems that may occur secondary to changes in endogenous flora. When infection complicates stage I cervical cancer, the infection is typically limited to the vagina, covering only the surfaces of the tumor itself [87].The abnormal neoplastic tissue allows bacterial overgrowth of normal flora to take place. Streptococcal species are usually isolated from the purulent debris. Rose and Wilson [88]presented a case of toxic shock syndrome in a patient with previously undiagnosed advanced cervical cancer. The cervical cancer was believed to be the portal of entry for staphylococcal toxins to reach the bloodstream. Obstruction probably contributes to adnexal infections in patients with advanced cervical disease. Barton et al. [89] presented three unusual cases in which patients presented with cervical cancer complicated by tuboovarian abscesses. In two cases, the patients were initially overstaged secondary to inflamed adnexal masses, which were later found to be free of cancer. The third patient developed an acute abdomen secondary to a ruptured tuboovarian abscess shortly after detection of an exophytic cervical mass. Patients with cervical disease that has invaded surrounding tissues by direct extension may be more likely to develop pyometra (pus in the uterus) [87]. Some patients with pyometra present with the classic triad of purulent vaginal drainage, fever, and lower abdominal pain [90]. Pyometra, however, can often be asymptomatic, presenting without fever or pelvic pain [87]. Typical organisms are aerobic and anaerobic streptococci [87]. Not all patients with this infection have a malignancy, and the pathogenesis is not always associated with obstruction at the cervical 0s. An early study [91] demonstrated that a fixed obstruction of the cervix was not necessary to maintain a pyometra; in this collection of 52 removed uteri with pyometra, only 24 had obstructed cervices. Of note, 26 patients developed infection following radiation therapy. Of 52 patients, only 8 had malignancies of the endometrium or cervix and 3 patients had gastrointestinal carcinomas. Peritonitis can occur if a pyo~netraruptures. A collection of 15 cases in the literature of spontaneous perforation of pyometra found that one third of the patients had malignant disease [90]. All 15 patients presented with fever, 53%

with vomiting, and 20% with atypical genital bleeding. The most common organisms in peritoneal fluid were Escherichia coli, Bacteroides spp., and polymicrobial. Douvier et al. [92] reported two cases of perforation of the uterus at the site of endometrial carcinoma, resulting in peritonitis. Very advanced, usually undiagnosed, carcinoma of the cervix has been associated with spontaneous rupture into the retroperitoneum [87]. The severity of infectious complications at cancer diagnosis appears to correlate with the extent of tumor invasion and secondary obstruction; because the vast majority of gynecologic cancers in the United States are now diagnosed when disease is localized, complications such as peritonitis are extremely rare at the time of cancer detection. Surgery, chemotherapy, andlor radiation therapy contribute to infectious complications of gynecologic oncology patients. Under such abnormal conditions, a normally benign vaginal commensal can proliferate, invade the bloodstream, and cause sepsis. For example, Andriessen et al. [93] reported a nonneutropenic patient who presented with sepsis following chemotherapy for metastatic choriocarcinoma; multiple blood cultures grew LactobaciElus acidophilus. A gallium scan showed only diffuse uptake in her uterus. A second patient became septic 2 days following surgery and was found to have a Lactobacillus spp. pelvic abscess and bacteremia [94]. In a retrospective study of infections associated with gynecologic cancers, cervical cancer had the highest surgical infection rate (22%); examples of such infections were peritonitis, pelvic hematomas, groin abscesses, and drainage tube infections [86]. Similar types of infections were documented as complications of uterine and ovarian cancers. Prior radiation therapy and surgery appeared to be risk factors for infection in patients with cervical and uterine cancers. Preoperative subclinical pelvic infections, invasive diagnostic procedures, and invasive devices of supportive care (nasogastric tubes, urinary catheters, central lines) may also contribute to the development of postoperative infections [86]. Graham [95] suggested that two additional factors contributed to gynecologic surgical infections: first, removed organs and tissue create a space that fills with blood and serum, an excellent culture medium, and, second, bowel obstruction or ileus can result in poor nutritional status preoperatively and postoperatively. Ovarian carcinoma is the most common gynecologic cancer to be treated with chemotherapy; these patients are at risk for the infectious complications associated with neutropenia. Complications of pelvic irradiation for cervical cancer, such as fistula formation and small bowel obstruction or perforation, are rare and may be associated with previous pelvic inflammatory disease [96]. Infections associated with gynecologic cancers can be life-threatening. Empiric antibiotics for a febrile patient with suspected advanced gynecologic cancer should cover anaerobes and aerobic gram-negative bacilli. Candidrr spp. may play a role in some infections because they are part of normal and abnormal vaginal flora. Enterococci and group B streptococci can also be involved in gynecologic infections, particularly abscesses. Cultures of

blood, vagina, pyometra, and abscess drainage abscesses can help guide therapy.

4.6 Bacillus Calmette-Guirin dissemination Dissemination of intravesicular bacillus Calmette-Gukrin (BCG) therapy is an unusual outcome of a unique anticancer therapy commonly used in bladder cancer. The antineoplastic mechanism of action of this Mycobacterium bovis strain is thought to be as an immune modulator. Cases of disseminated BCG have been rare. Lamm et al. [97] found granulomatous hepatitis and1 or pneumonitis in 0.7% of bladder cancer patients receiving intravesicular BCG therapy. M. bovis is presumed to spread hematogenously from the bladder. The time from installation to symptomatic presentation appears to be highly variable, ranging from hours to several months after exposure. Proctor et al. [98] presented a case of an elderly male who experienced fever and rigors 5 hours after installation. A blood culture from that day grew M. bovis. The patient developed mild hepatitis, hyperbilirubinemia, and AFB-positive hepatic granulomas. The patient improved on isoniazid and rifampin for 12 months plus 2 months initial ethambutol. In contrast, Hakim and colleagues 199) reported a case in which an elderly bladder cancer patient presented with a M. bovis psoas abscess 9 months after BCG therapy. Katz et al. [loo] reported a case of a man who presented with lumbar vertebral osteomyelitis and psoas abscess approximately 4 months after completing a year of BCG therapy. Cultures from bone and abscess fluid grew M. bovis. Subsequent vertebral surgery revealed necrotic bone with AF'B-positive caseating and noncaseating granulomas. He responded well to abscess drainage and isoniazid and rifampin. Other cases of BCG infection that are consistent with hematogenous spread included: acute prosthetic knee arthritis [loll, septic arthritis of the elbow [102], and pulmonary infections [103]. Pulse field gel electrophoresis has been used to confirm that the instilled organism is identical to the infecting pathogen [103]. Infectious arthritis, which should be treated with antituberculous drugs, should be differentiated from BCG-associated reactive polyarthritis [I041 or Reiter's syndrome [105], which are treated with antiinflammatory agents. One needs to have a high index of suspicion that BCG may be the cause of infection in a febrile patient who is receiving or received BCG therapy for bladder cancer. Attempts should be made to identify the organism because prolonged, potentially toxic treatment is required. When BCG disseminates, therapy requires prolonged antituberculous medication and drainage of any abscess. M. bovis is intrinsically resistant to pyrazinamide. Most cases in the literature reported cure when patients were treated with isoniazid and rifampin for 6-9 months. A patient who suffers invasive BCG infection should not receive any additional BCG therapy.

5. Infectious complications of head and neck cancer Infections occur commonly in patients with head and neck cancer. Because patients typically become symptomatic at late stages of disease, they often present when large tumors obstruct airways or inhibit swallowing. Tumor involvement of the oral mucosa, a reservoir of substantial numbers of bacteria, provide a ready source of pathogens (Figure 4). In the late stages of disease, many patients experience profound weight loss secondary to cachexia and restricted intake. The resulting malnutrition and related immunosuppression, as well as frequent comorbidities of chronic obstructive pulmonary disease (COPD), liver disease, and poor dental hygiene, further diminish the ability of these patients to effectively fight infection. As in other solid tumor patients, surgery, radiation therapy, and chemotherapy further impair local and systemic host defenses. The combination of these factors results in a significant risk of infection. 5.1 In,fections after radiation therapy

Head and neck cancer patients commonly receive treatment with chemotherapy, radiation, and surgery. Radiation can result in delayed healing, posing an increased risk of wound infection. In addition, radiation can dramatically reduce saliva production, causing an alteration in oral flora favoring more virulent organisms [106-1081. Microbial samples from plaque and saliva of patients before and after irradiation revealed significant increases in Streptococcus mutans, Lactobacillus spp., Candida spp., Staphylococc~tsspp., enteric gram-negative bacilli, and anaerobes [108]. A progressive drop in salivary production was identified, beginning within 2 weeks of bilateral parotid gland irradiation and decreasing to 6% of initial flow rates in patients by 3 months; at that time dental caries were noted with increased frequency [109]. Fungal infections cause a great deal of morbidity in the head and neck patient following irradiation [107], including increased pain and difficulty with speech [110]. Irradiation to the mouth or the larynx changes the skin flora and the mouth flora for up to 6 months, resulting in significant overgrowth by a variety of yeast. 5.2 Postoperative woun,d infections Most surgery for head and neck cancer, because it involves the upper respiratory and gastrointestinal tract, is considered to be contaminated or clean contaminated. During surgery and throughout the healing process, most wounds are in intimate contact with the rnucosal surfaces, or secretions, of the oropharynx and respiratory tract. Salivary bacterial counts are in the range of 10~109/m [Ill]. ~ Anaerobes account for 90% of the organisms in the oral cavity; the remainder are gram-positive and gram-negative aerobic organisms. Contamination during and after surgery by oral flora is thought to contribute

greatly to the high rate of wound infections after head and neck surgery. Infection rates of over 80% have been recorded when prophylactic antibiotics were not used. The impact of anaerobic organisms in the pathogenesis of wound infections in head and neck cancer became evident during the 1980s, coinciding with improved techniques of isolating anaerobic bacteria. Sawyer [I121 clarified the need for anaerobic coverage in most head and neck cancer surgeries by demonstrating a significantly reduced infection rate following prophylaxis with metronidazole and cefazolin when compared with recent historical controls using cefazolin alone. A prospective, randomized trial in head and neck cancer patients confirmed improved infection rates with cefazolin and metronidazole (9.5% infection rate vs. 18.6% infection rate for cefazolin alone) [113]. The majority of head and neck postoperative wounds have been polymicrobial. Brook and Hirokawa [I141 cultured 24 postoperative wounds of head and neck cancer patients, finding that 88% of the wounds were mixed anaerobic and aerobic flora. Peptostreptococcus spp., Bacteroides spp. (non-frcrgilis), and Fzwobacterium spp, were the most commonly identified anaerobes. Increased risk of wound infection can be attributed to the extent of disease, duration of surgery, and technical constraints. While tumor removal is of primary importance, surgeons attempt to preserve the airway, cough reflex, diaphragm function, speech, facial muscles and nerves, hormonal function, lymphatic drainage, and saliva production [115]. Seroma and hematoma formation can contribute to the risk of abscess; effective hemostasis and use of drains can minimize bleeding and edema. Patients with advanced disease often require removal of large amounts of tissue, thereby exposing extensive wounds. Technical decisions that influence perfusion of a graft or skin flap can have significant impact on infection risk. Tandon et al. [I161 found that patients who underwent a muscle flap procedure, which reflects extensive disease, had increased risk of infection. Eight of 12 patients who underwent a pectoralis major flap developed wound infections. Robbins et al. [I171 reviewed 400 head and neck patients who underwent surgery and found a wound infection rate of 20%. Presence of advanced disease, duration of surgery greater than 6 hours, placement of a flap, as well as absence of anaerobic coverage perioperatively were found to be significant risk factors for wound infections. Brown and coworkers [1181 recorded an overall 7% wound infection rate among 245 head and neck patients receiving perioperative antibiotics; the subset of patients with stage 1%' disease had a 15% risk of wound infection. Similarly, those patients who underwent a myocutaneous flap procedure had an infection rate of 36%, whereas those receiving simpler procedures had a risk of 6%. These investigators also identified probable errors in surgical decision-making or technique in 10 of the 17 patients who developed wound infections, many of which resulted in flap or skin graft failure from ischemia, bleeding, tension, or trauma. Some risk factors for wound infection in head and neck cancer patients may be difficult to modify,

namely, the extent of disease at presentation and the technical and physical constraints that result.

5.3 Nonwound infections in head and neck patients The lower respiratory tract is the primary site of nonwound infections in head and neck cancer patients. These infections are a major cause of mortality in this patient group. Hussain et al. [I191 reviewed 12 months of admissions to a university head and neck cancer service. Eighty-six infections were documented among 102 febrile episodes in 67 patients. Forty-three percent of the infections were attributed to pneumonia or tracheobronchitis. Eighteen percent of deaths were directly attributed to pneumonia. Papac [I201 reported 78 infectious complications among 191 patients with advanced head and neck cancer hospitalized on a medical oncology service at a Veterans Affairs hospital. Of 111 reported deaths, pneumonia was the most frequent cause of death (26%), twice as common as the next leading cause of death, tumor or metastasis. Pneumonia was the most common infection in a study of perioperative morbidity among patients with head and neck cancer [121]. Twenty-two of 225 patients experienced lower respiratory tract infections. Of the 22 patients, 19 developed postoperative pneumonia and 3 had tracheobronchitis. Duration of surgery greater than 6.2 hours increased the risk of nonwound infection from 4.5% to 15.3%. Having a greater than 70 pack-year history of smoking and receiving a blood transfusion perioperatively also significantly increased the risk of pulmonary infection. Length of stay for those with pneumonia increased from a mean of 14.6 days to 23 days (P < .05). Other sources of nonwound infections were the urinary tract (3), septic phebitis (I), and acute sinusitis (1).

5.4 Management of infections in head and neck cancer At the time of diagnosis of head and neck cancer, some steps can be taken to prevent infectious complications. Because tuberculosis can become reactivated in these patients, it is prudent to administer isoniazid prophylaxis to those who are tuberculin positive, but without disease, despite their advancing age and possible underlying liver disease. Pneumococcal vaccination prior to instituting therapy and annual influenza vaccination may be protective. Anticipating problems involving the oral cavity is key to minimizing infectious complications during and after radiation therapy, and to a lesser extent, during and after surgery. Before receiving radiation therapy or undergoing surgery, a head and neck cancer patient should have a complete dental evaluation, including dental x-rays [115,122]. Carious teeth should be removed or restored. Institution of oral hygiene can reduce the risk of subsequent infectious complications. Patients should use salivary substitutes throughout the day and receive regular fluoride treatments [122]. Some investigators recommend selective decontamination with topical antibiotics during the weeks of

irradiation, but this remains unproven. The patient should be well educated regarding the benefits of rigorous oral hygiene during and after radiation. Postoperative management is critical in the efforts to reduce morbidity in these patients. In the event of a postoperative wound infection, aerobic and anaerobic cultures should be sent. Empiric antibiotic therapy covering anaerobes, gram-negative bacilli, and aerobic gram-positive cocci is appropriate. Infected fluid collections should be drained. Coleman [I151 urged prompt exploration of the surgical wounds if fistulae or necrosis (or other evidence of infection) develop. The index of suspicion of infection should be higher if the patient has received preoperative irradiation. Should a fistula form near or overlying the carotid sheath, an emergency exploration should be performed; infection of the carotid artery in a previously irradiated site can result in septic emboli or carotid artery rupture with exsanguination [115]. Attempts to prevent aspiration in a patient with advanced head and neck cancer are often futile. Pneumonia is relatively common and often responds to empiric broadspectrum antibiotics.

6. Conclusions The medical literature describing infectious complications in patients with solid tumors is limited to isolated case reports and retrospective studies dominated by patients with hematologic malignancies. There are no infections unique to solid tumor patients. Two broad categories of infection in patients with solid tumors can be described. Some infections are directly attributable to the tumor. These infections develop because the neoplastic process causes focal injury, breaks down normally intact barriers, or causes local obstruction. Examples include S. bovis endocarditis or C. septicurn myonecrosis in patients with colon cancer, postobstructive pneumonia complicating lung cancer, and pyometra complicating advanced cervical carcinoma. The second major category of infections may be attributed to the effects of cancer treatment: surgery, chemotherapy, and radiation therapy. Examples include upper extremity cellulititis complicating the lymphedema that results from surgery and radiation therapy for breast cancer, intraabdominal sepsis following attempted resection of a solitary hepatic metastasis in a patients with colon cancer, and catheter-related sepsis in a patient with head and neck cancer. Wound infection deserves particular attention in patients with solid tumors. For all types of surgery, the level of contamination and the length of the procedure are the critical determinants of wound infection risk. Neoplasia appears to have a definite, if indirect, impact on wound infection rates because the location and extent of tumor define the level of contamination and influence the duration of surgery. A recent large study at Memorial SloanKettering Cancer Center identified four significant risk factors for wound infections after cancer surgery: prolonged surgery, a poor preoperative physical assessment score, and contaminated or dirty surgeries and obesity [123].

Preoperative chemotherapy, radiation therapy, and the neoplastic process itself no doubt continue to impair wound healing and the risk of infection. Only a small subset of those with solid tumors sustain the level of chemotherapy-related immunosuppression associated with an increased risk of Pneumocystis cnrinii pneumonia, disseminated aspergillosis, hepatosplenic candidiasis, or neutropenic enterocolitis. Nevertheless, the types of cancers that are treated with intensive chemotherapy regimens and autologous bone marrow transplantation appear to be increasing. Advanced non-small cell lung cancer patients demonstrate improved 5-year survival with the addition of induction chemotherapy to radiation [124]. Ongoing trials of autologous bone marrow transplantation for advanced breast cancer are attempting to establish whether such intensive chemotherapy results in survival advantage. As more solid tumor patients receive intensive chemotherapy, we can anticipate an increase in the incidence of severe infectious complications. Recognizing the subset of solid tumor patients who are severely immunosuppressed is critical when infection is first suspected; outcome in these individuals may depend upon considering an expanded differential diagnosis at the outset, obtaining appropriate microbiologic and pathologic specimens, often via invasive procedures, and employing empiric antibiotics thoughtfully and rapidly. Management of infectious complications of solid tumors is aided by understanding their pathogenesis, minimizing nosocomial risks, and utilizing modern imaging, invasive procedures, and modern culture techniques.

References 1. Boring CC, Squires TS, Tong T. Montgomery S. Cancer Statistics, 1994. CA Cancer J Clin 1994;44:7-26. 2. Elting LS, Bodey GP, Fainstein V. Polymicrobial septicemia in the cancer patient. Medicine 1986;65:218-225. 3. Mayo JW, Wenzel RP. Rates of hospital-acquired bloodstream infections in patients with specific malignancy. Cancer 1982;50:187-190. 4. Mathe G, Meyer P, Brienza S. Gil-Delgado M, Orbach-Arbouys S, Itzhaki M. Retrospective study correlating clinical infectious history and peripheral blood T-cell subpopulations in cancer, G v H and HIV+ patients. Biomecl Pharmacother 1992:46:17-19. 5. Pizzo PA, Robichaud KJ, Wesley R, Con~mersJR. Fever in the pediatric and young adult patient with cancer: A prospective study of ZOO1 episodes. Medicine 1982;61:153-165. 6. Singer C, Kaplan MH, Armstrong D. Bacteremia and fungemia complicating neoplastic disease: A study of 364 cases. Am J Med 1977;62:731-742. 7. Krumpe PE, Knudson RJ, Parsons G, Reiser K. The aging respiratory system. Clin Geriatr Med 1985;1:143-175. 8. Brugarolas A, Takita H. Immunologic status in lung cancer. Chest 1973;64:427-430. 9. Busch E, Verazin G, Antkowiak JG, Driscoll D, Takita H. Pulnlonary complications in patients undergoing thoracotomy for lung carcinoma. Chest 1994;105:760-766. 10. Kohno S. Koga H. Oka M, Kadota J. Kaku M, Soda H, Tomono IS, Hara K. The pattern of respiratory infection in patients with lung cancer. Tohoku J Exp Med 1994;173:405-411. 11. Strang C, Simpson JA. Carcinon~atousabscess of the lung. Thorax 1953;8:11-25.

12. Perlman LV, Lerner E , D'Esopo N. Clinical classification and analysis of 97 cases of lung abscess. A m Rev Respir Dis 1969;99:390-398. 13. Snider GL, Saleh SS. Empyema of the thorax in adults: Review of 105 cases. Dis Chest 1968;54:12-11. 14. Feld R. Bodey GP, Groschel D. Mycobacteriosis in patients with malignant disease. Arch Intern Med 1976:136:67-70. 15. Kaplan MH, Armstrong D, Rosen P. Tuberculosis complicating neoplastic disease: A review of 201 cases. Cancer 1974;33:850-858. 16. Snider GL. Placik B. The relationship between pulmonary tuberculosis and bronchogenic carcinoma: A topographic study. Am Rev Resp Dis 1969:99:229-236. 17. Nagata N, Nikaido Y, Kido M, Ishibashi T, Sueishi K. Terminal pulmonary infections in patients with lung cancer. Chest 1993:103:1739-1742. 18. Smith FB, Bveneck B. Localized aspergillus infestation in primary lung carcinoma: Clinical and pathological contrasts with post-tuberculous intracavitary aspergilloma. Chest 1991; 100:554-556. 19. D e Vuyst P, D e Troyer A , Yernault J, Verhest A, Vanderhoeft P. Aspergilloma in a necrotic bronchial adenocarcinoma. Eur J Respir Dis l980;6 1:213-217. 20. Meyer RD, Young LS. Armstrong D , Yu B. Aspergillosis complicating neoplastic disease. An1 J Med 1973;54:6-15. 21. Borkin MH, Arena FP, Brown A E , Armstrong D. Invasive aspergillosis with massive fatal hemoptysis in patients with neopla5tic disease. Chest 1980;78:835-839. 22. Smith JM, Hampton JR, Webb RW. Disseminated aspergillosis, Cushing's syndrome, and oat cell carcinoma of the lung. Chest 1979:76:112-113. 23. Sidran3ky H. Friedman L. The effect of cortisone and antibiotic agent on experimental pulmonary aspergillosis. A m J Path01 1959:35:169-183. 24. Hasan FM, Jarrah T, Nassar V. The association of adenocarcinotna of the lung and blastomycosis from an unusual geographical location. Br J Dis Chest 1978;72:242-246. 25. Blastomycosis Cooperative Study of the Veterans Administration: A review of 198 collected cases in Velerans Administration hospitals. Ail1 Rev Resp Dis 1964;89:659-672. 26. Haron E , Vartivarian S. Anaissie E, Dekmezian R, Bodey GP. Primary Crindlda pneumonia: Experience at a large cancer center and review of the literature. Medicine 1993;72:137142. 27. Yale SH. Limper AH. Przeut?zoc~~stis corrnii pneumonia in patients without acquired immunodeficiency syndrome: Associated illnesses and prior corticosteroid therapy. Mayo Clinic Proc 1996;7 15-13. 28. Sepkowitz KA, Brown AE, Tolzak E E , Gottlieb S, Armstrong D . Pneunzocysti~carinii pneunlotlia among patients without AIDS at a cancer hospital. JAMA 1992;267:832-837. 29. Graham BS. Tucker WS. Opportunistic infections in endogenous Cushing's syndrome. Ann Intern Med 1984:101:334-338. 30. Frenkel JK. Good JT. Shultz JA. Latent Pnez(nzocystis infection in rats, relapse, and chemotherapy. Lab Invest 1966;15:1559-1577. 31. Walzer PD. LaBine M. Redington TJ, Cushion MT. Lymphocyte changes during chronic administration of and withdrawal from corlicosteroids: Relation to Pnerrinocyst~scarinii pneumonia. J Immunol 1984:133:2502-2508. 32. Henson JW, Jalaj JK, Walker RW, Stover DE. Fels AOS. Pt~eumocystiscar.il~iipneumonia in patients with pri~narybrain tumors. Arch Neurol 1991:48:406-409. 33. Poplin EA, Gordon CJ, Piskorowski TJ. Chandrasekar PH. Pnertnzocystis carinii pneumonia in patlents with solid tumors without acquired imtnune deficiency syndrome. Cancer 199 1:68:193-194. 34. Kovacs J A , Hiemenz JW, Macher AM, Stover D , Murray HW, Shelhamer J, Lane HC, Urmacher C, Honig C. Longo DL, Parker MM. Natanson C, Parrillo J E , Fauci AS. Pizzo PA, Masur H. Ptzerrmocystis cllrirril pneumonia: A comparison between patients with the acquired in~munodeficiencysyndrome and patients with other immunodeficiencies. Ann Intern Med 1984;100:663-671.

35. Limper AH, Offord KP, Smith TF, Martin WJ. Pneumocystis carinii pneumonia: Differences in lung parasite number and inflammation in patients with and without AIDS. Am Rev Respir Dis 1989;140:1204-1209. 36. Sepkowitz KA. Pneumocystis carinii pneumonia without acquired immunodeficiency syndrome: Who should receive prophylaxis? Mayo Clin Proc 1996;71:102-103. 37. Camazine B, Antkowiak JG, Nava MER, Lipman BJ, Takita H. Herpes simplex viral pneumonia in the postthoracotomy patient. Chest 1995;108:876-879. 38. Coggle JE, Lambert BE, Moores SR. Radiation effects in the lung. Environ Health Perspect 1986;70:261-291. 39. Gross NJ. Pulmonary effects of radiation therapy. Ann Intern Med 1977;86:81-92. 40. McInerney DP, Bullimore J. Reactivation of radiation pneumonitis by adriamycin. Br J Radiol 1977;50:224-227. 41. Schwarz MI, Whitcomb ME, Goldman AL. The spectrum of diffuse pulmonary infiltration in malignant disease. Chest 1973;64:88-93. 42. Pisani RJ, Wright AJ. Clinical utility of bronchoalveolar lavage in imnlunocompromised hosts. Mayo Clin Proc 1992;67:221-227. 43. Shure D. Transbronchial biopsy and needle aspiration. Chest 1989;95:1130-1138. 44. Rotstein C, Ferguson R, Cummings KM, Piedmonte MR, Lucey J, Banish A. Determinants of clean surgical wound infections for breast procedures at an oncology center. Infect Control Hosp Epidemiol 1992;13:207-214. 45. Sauter ER, Eisenberg BL, Hoffman JP, Ottery FD, Boraas MC. Goldstein LJ, Solin LJ. Postmastectomy morbidity after combination preoperative irradiation and chemotherapy for locally advanced breast cancer. World J Surg 1993;27:237-242. 46. Vinton AL, Traverso LW, Jolly PC. Wound coinplications after modified radical mastectomy compared with tylectomy with axillary lymph node dissection. Am J Surg 1991;161:584-588. 47. Mozes M, Papa MZ, Karasik A, Reshef A, Adar R. The role of infection in post-nlastectomy lymphedema. Surg Ann 1982;14:73-83. 48. Bertelri G, Venturini M, Forno G, Macchiavello F, Dini D. An analysis of prognostic factors in response to conservative treatment of postrnastectomy lymphedema. Surg Gynecol Obstet 1992;175:455-460. 49. Simon MS, Cody RL. Cellulitis after axillary lymph node dissection for carcinoma of the breast. Am J Med 1992;93:543-548. 50. Chaudhuri K, Lonergan D, Portek I, McGuigan L. Septic arthritis of the shoulder after mastectomy and radiotherapy for breast carcinoma. J Bone Joint Surg Br 1993;75-B:318-321. 51. Rescigno J, McCormick B, Brown AE, Myskowski PL. Breast cellulitis after conservative surgery and radiotherapy. Int J Radiat Oncol Biol Phys 1994;29:163-168. 52. Olivotto IA, Weir L, Kim-Sing C. In response to Rescigno et al. Int J Radiat Oncol Biol Phys 1995;31:203. 53. Keidan RD, Hoffman JP, Weese JL, Hanks GE, Solin LJ, Eisenbe BL, Ottery FD, Borass M. Delayed breast abscesses after lumpectomy and radiation therapy. Am Surg 1990;56:440444. 54. Mudad R, Vredenburgh J. Paulson EK, Ross M, Meisenberg B, Hussein A, Peters WP. A radiologic syndrome after high dose chemotherapy and autologous bone marrow transplantation, with clinical and pathologic features of systemic candidiasis. Cancer 1994;74:13601366. 55. Parry BR. Radiation recall induced by tamoxifen. Lancet 1992;340:49. 56. Rostom, AY, ME El-Sayed. Fat necrosis of the breast: An unusual complication of lumpectomy and radiotherapy in breast cancer. Clin Radiol 1987;38:31. 57. Bertelli G, Dini D, Forno GG, Gozza A. Preventing cellulitis after axillary lymph node dissection. Am J Med 1994;97:202-203. 58. Dixon JM. Breast infeclion. Br Med J 1994;309:946-949. 59. Finegold SM. Anaerobic bacteria: General concepts. In: Mandell GL, Bennett JE, DoIin R, eds. MandelI, Douglas and Bennett's Principles and Practice of Infectious Diseases, 4th ed. New York: Churchill Livingstone, 1995, pp. 2156-2173.

60. Klein RS, Recco RA, Catalano MT, Edberg SC, Casey JI, Steigbigel NH. Association of Streptococcils bovis with carcinoma of the colon. N Engl J Med 1977;297:800-802. 61. Klein RS, Catalano MT, Edberg SC, Casey JI, Steigbigel NH. Streptococcus bovis septicemia and carcinoma of the colon. Ann Intern Med 1979;91:560-562. 62. Fass R, Alim A, Kaunitz JD. Adenocarcinoma of the colon presenting as Streptococcus snngrlis bacteremia. Am J Gastroenterol 1995;90:1343-1345. 63. Wiseman A, Rene P, Crelinsten GL. Streptococcils agalactiae endocarditis: An association with villous adenomas of the large intestine. Ann Intern Med 1985;103:893-894. 64. Gelfand MS, Hughey JR, Sloas DD. Group B Streptococcal bacteremia associated with adenocarcinoma of the stomach. Clin Infect Dis 1994;39:364. 65. Katlic MR. Derkac WM, Coleman WS. Clostridiunz septiciinz infection and malignancy. Ann Surg 1981;193:361-364. 66. Koransky JR, Stargel MD, Dowel1 VR. Clostridirlm septicurn bacteremia: Its clinical significance. Am J Med 1979;66:63-66. 67. Kornbluth AA, Danzig JB, Bernstein LH. Clostridiilnz septiciinz infection and associated malignancy: Report of 2 cases and review of the literature. Medicine 1989;68:30-37. 68. Gorbach SL, Thadepalli H. Isolation of Clostridiilrn in human infections: Evaluation of 114 cases. J Infect 1975;131:S81-S85. 69. Jendrzejewski JW. Jones SR, Newcombe RL, Gilbert DN. Nontraumatic clostridial myonecrosis. Am J Med 1978;65:542-546. 70. Kaiser CW, Milgrom ML, Lynch JA. Distant nontraumatic clostridial myonecrosis and malignancy. Cancer 1986;57:885-889. 71. Hovenden JL, Murdoch GE, Evans AT. Non-traumatic Clostridium septicurn arthritis in a patient with caecal carcinoma. Rr J Rheumatol 1992;31:571-572. 72. Fallon SM, Guzik HJ, Kramer JE. Clostridium septicunz arthritis associated with colonic carcinoma. J Rheumatol 1986;13:662-663. 73. Alpern RJ, Dowel1 VR. Clostricliiinz septic~iminfections and malignancy. JAMA 1969; 209:385-388. 74. Kahn SP, Lindenauer SM, Wojtalik RS, Hildreth D. Clostridia hepatic abscess: An unusual manifestation of metastatic colon carcinoma. Arch Surg 1972;104:209-212. 75. Kudsk KA. Occult gastrointestinal malignancies producing metastatic Clostridiiirn septicirm infections in diabetic patients. Surgery 1992;112:765-772. 76. Cohen JL, Martin M, Rossi RL, Schoetz DJ. Liver abscess: The need for complete gastrointestinal evaluation. Arch Surg 1989:124:561-564. 77. Marcus SG. Walsh TJ, Pizzo PA, Danforth DN. Hepatic abscess in cancer patients: Characterization and management. Arch Surg 1993;128:1358-1364. 78. Okada S, Aoki K, Okazaki N, Nose H, Yoshimori M, Shimada K, Yamalnoto J, Takayama T, Kosuge T, Yamaski S, Takayasu K, Moriyama N. Liver abscess after percutaneous ethanol injection (PEI) therapy for hepatocellular carcinoma. A case report. Hepatogastroenterology 1993;40:496-498. 79. Isobe H, Sakamoto S, Nawata H. Splenic abscess as a complication of percutaneous ethanol injection therapy for hepatocellular carcinoma. J Clin Gastroenterol 1992;15:167169. 80. Teitz S, Guidetti-Sharon A, Manor H, Halevy A. Pyogenic liver abscess: Warning indication of silent colonic cancer: Report of a case and review of the literature. Dis Colon Rectum 1995;38:1220-1223. 81. Lonardo A, Grisendi A, Pulvirenti M, Della Casa G, Melini L, DiGregorio C, Nasi G, Sarti M. Tamborrino E, Lonardo F. Right colon adenocarcinoma presenting as Bactevoides fragilis liver abscesses. J Clin Gastroenterol 1992;14:335-338. 82. Wolfe MS, Armstrong D, Louria DB, Blevins A. Salmonellosis in patients with neoplastic disease: A review of 100 epidsodes at Memorial Cancer Center over a 13-year period. Arch Intern Med 1971;128:546-554. 83. Han T, Sokal JE, Neter E. SalmoneHosis in disseminated malignant diseases: A seven-year review (1959-1965). N Engl J Med 1967;276:1045-1052.

84. Kate P, Osei K, Chiemchanya S, Zatuchni J. E~npyemadue to SnL?zonella typhinzliriun~with underlying alveolar ccll carcinoma. South Med J 1984:77:234-236. 85. Berkeley D , Mangels J. Snlntoncll~~ pneutnonia in a patient with carcinoma of the lung. Am J Clin Path01 1980;74:476-478. 86. Brooker DC, Savage JE, Twiggs LB. Adcock LL, Prem KA. Sanders CC. Infectious morbidity in gynecologic cancer. An1 J Obstet Gynecol 1987;156:513-520. 87. WeIton JA. Significance of infection in carcinonia of the cervix. Obstet Gynccol1965;26:441446. 88. Rose PG, Wilson G. Advanced cervical carcinoma presenting with toxic shock syndrome. Gynecol Oncol 1994:52:264-266. 89. Barton DPJ, Fiorica JV. Hoffman MS. Roberts WS, Finan MA, Cavanagh D. Cervical cancer and tuboovarian abscesses: A report of three cases. J Reprod Med 1993;38:561-564. 90. Imachi M, Tanaka S, Ishikawa S, Matsuo K. Spontaneous perforation of pyometra preTenting as generalized peritonitis in a patient with cervical cancer. Gynecol Oncol 1993;50:384388. 91. DeVoe RW, Randall LM. Pyometra: A clinical and pathologic study. Am J Obstet Gynecol 1949;58:784-789. 92. Douvier S, Nabholtz J-M, Friedman S, Cougard P, Ferry C. Aupecle P. Infectious pneumoperitoneum as an uncommon presentation of endometrial carcinoma: Report of two cases. Gynecol Oncol 1989;33:392-394. 93. Andriessen MPHM. Mulder JG. Sleijfer DT. Lactobacillus septicemia, an unusual complication during the treatment of metastatic choriocarcinoma. Gynecol O~icol1991;40:87-89. 94. Connor JP, Buller RE. Lactobacillus sepsis with pelvic abscess. Gynecol Oncol 1994;54:99100. 95. Graham JE. Infectious morbidity in gynecologic oncology. J Reprod Med 1990;35:348-352. 96. Van Nagell JR. Parker JC, Maruyama Y, Utlep J. Hager WD, Donaldson ES, Wood EG. The effect of pelvic inflammatory disease on enteric complications following radiation therapy for cervical cancer. Am J Obstet Gynecol 1977;128:767-77 1. 97. Lamm DL, van der Meijdcn PM, A Morales A, Brosman SA. Catalona WJ. Herr HW, Soloway MS, Steg A, Debruyuc FM. Incidence and treatment of colnplications of bacillus Calmette-Guirin intravesical therapy in supcrficial bladder cancer. J Urol 1992;147:596-600. 98. Proctor DD, Chopra S, Rubenstein SC. Jokela JA. Uhl L. Mycobacterernia and granulomatous hepatitis following initial intravesical bacillus Calmette-Gukrin installation for bladdcr carcinoma. Am J Gastroenterol 1993;S8:1112-1115. 99. Hakim S, Heaney JA. Heinz T. Zwolak RW. Psoas abscess following intravesical bacillus Calmette-Guerin for bladdcr cancer: A case report. J Urol 1993;I50:188-189. 100. Katz DS, Wogalter H, D'ETposito RF. Cunha BA. Mycohacter.rlir?z bovis vertebral osteornyelitis and psoas absccss after intravcsical BCG thcrapy for bladder carcinoma. Urology 1992;40:63-66. 101. Chazerain P, Desplaces N. Mamoudy P, Leonard P, Ziza J. Prosthetic total knee infection with a bacillus Calmette Gukrin (BCG) strain after BCG therapy for bladder cancer. J Rheumatol 1993:20:2171-2172. 102. Van Thiel RJ. Nossent GD. Tjonpiangi NP. Articular complications of intravesical BCG treatment for bladder carcinoma. Br J Urol 1992:70:446-447. 103. Kristjansson M. Green P. Manning HL, Slutsky AM, Brecher SM. Fordharn von Reyn C. Arbeit RD, Maslow JN. Molecular confirmation of bacillus Calruettc-Guirin as the cause of pulmonary infect~onfollowing urinary tract installation. Clin Infect Dis 1993:17:228230. 104. Jawad ASM. Kahn L. Copland RFP, Henderson DC, Abdul-Ahad AK. Reactive arthritis associated with bacillus Calmette-Gukrin immunothcrapy for carcinoma ot the bladder: A report of two cases. Br J Rheumatol 1993:32:1018-1020. 105. Pancaldi P, Van Linthoudt D, Alboriiio D, Haefliger JM, Ott H. Reiter's syndrome after intravesical bacillus Calmette-Guerin treatment for superficial bladder carcinoma. Br J Rheuniatol 199332:1096-1 098.

106. Ariyan S. Infections following surgery for head and neck cancer. Clin Plast Surg 1979:6:523535. 107. Curtis TA. Griffith MR, Firtell DN. Complete denture prosthodontics for the radiation patient. J Prosthet Dent 1976;36:66-76. 108. Brown LR, D r e i ~ e nS, Handler S. Johnston D A . Effect of radiation-induced xerostomia on hurnan oral microflora. J Dent Res 1975:54:740-750. 109. Dreizen S. Brown LR. Handler S, Levy BM. Radiation-induced xcroslomia in cancer patients: Effect on salivary and serum electrolytes. Cancer 1976;38:273-278. 110. Martin MV, Al-Tikriti U, Bramley PA. Yeast flora of the mouth and skin during and after irradiation for oral and laryngeal cancer. J Med Microbiol 1981:14:457-467. 111. Bartlett JG, Gorbach SL. Anaerobic infections of the head and ncck. Otolarygol Clin North 1976:9:655-678. 112. Sawyer R. Clinical implications of metronidazole antianaerobic prophylaxis in major head and neck surgical procedures. Ear Nose Throat J 1988;67:655-662. 113. Robbins KT. Byers RM, Cole R. Fainstein V. Guillamondegui OM, Schantz SP, Weber RS, Wolf P, Goepfert H. Wound prophylaxis with metronidazole in head and neck surgical oncology. Laryngoscope 1988:98:803-806. 114. Brook I. Hirokawa R. Microbiology of wound infection after head and neck cancer surgery. Ann Otol Rhino1 Laryngol 198?:98:323-325. 1 15. Coleman JJ. Complications in hcad and ncck surgery. Surg Clin North Am 1986;66:149-167. 116. Tandon D A , Bahadur S. Laldirla HC, Sundaram KR. Role of prophylactic antibiotics in surgery for advanced head and neck canccr. J Laryngol Otol 1992;106:621-624. 117. Robbins KT. Favrot S. Hanna D , Cole R. Risk of wound infection in patients with head and neck cancer. Head Neck 1990;12:143-148. 118. Brown BM. john so^^ JT, Wagner RL. Etiologic factors in head and neck wound infections. Laryngoscope 1987;?7:587-590. 119. Hussain M. Kish JA. Crane L. llwayda A. Cummings G, Ensley JF. Tapazoglou E. Al-Sarraf M. The role of infection in the morbidity and mortality of patients with head and neck cancer undergoing multirnodal therapy. Cancer 1991:67:716-721. 120. Papac RJ. Medical aspects of head and neck cancer. Cancer Invest 1985;3:435-444. 121. Weber RS, Hankins P, Kosenbaum B. Rand I. Nonwound infections following head and neck oncologic surgery. Laryngoscope 1993:103:22-27. 122. Jansma J. Vissink A. Spijkervet FKL. Roodenburg JLN. Panders AK. Vermey A, Szabo BG. Johannes's-Gravcnmade E. Protocol for the prevention and treatment of oral sequelae resulting from head and neck radiation therapy. Cancer 1992;70:2171-2180. 123. Barber G R , Miransky J, Brown AE. Coit DG, Lewis FM, Thaler HT, Kiehn TE, Armstrong D. Direct observations of surgical wound infections at a con~prehensivecancer center. Arch Surg 199.5;130:1042-1047. 124. Dillman RO. Herndon J. Seagren SL, Eaton WL, Green MR. Improved survival in stage I11 non-small-cell lung cancer and leukemia group B (CALGB) 8433 trial. J Natl Cancer Inst 1996:88:1210-1215.

5. Role of the clinical microbiology laboratory in the diagnosis of infections Richard B. Thornson, Jr. and Lance R. Peterson

1. Introduction The role of the microbiology laboratory and microbiologist are especially important to the clinician caring for a patient with compromised host defenses. The microbiologist can assist in establishing a differential diagnosis and selection of laboratory tests to make an infectious diagnosis. Complete understanding of microbiology test results not only improves patient management but reduces the cost of medical care.

2. Infections associated with specific immune deficiencies Abnormalities of the immune system that affect T lymphocytes, B lymphocytes, granulocytes, or a combination of these cell functions predispose an individual to specific infectious diseases [I]. In addition, splenectomy, diminished serum complement, organ transplantation, and the use of corticosteroid or cytotoxic therapy further depress immune function and predispose one to infection. Table 1 summarizes the types of immunosuppressive impairments often observed in patients with cancer and the associated microorganisms.

3. Selection, collection, and transport of high-quality specimens, and interpretation of stain and culture results The initial evaluation of a cancer patient suspected of having an infectious disease involves collection of specimens to detect disseminated (e.g., blood) as well as localized (e.g., urine or sputum) disease. It is important to understand why a laboratory may "reject" a specimen, because this risks a delay in establishing a specific diagnosis. Especia.11~in infectious diseases, the likelihood of establishing a diagnosis is directly related to the quality of the specimen submitted for analysis. A clinical specimen that is contaminated with normal flora that overgrow important pathogens, is exposed to prolonged transport, contributing to the Gary A. Noskin (ed), M A N A G E M E N T O F INFECTIOUS COMPLICATIONS IN C A N C E R PATIENTS. 0 1998. Kluwer Acadenzic Publishers, Boston. All rights reserved.

Table 1. Selected immunosuppressive impairments and associated microorganisms

Microorganisms Defect

Bacteria -

Impaired antibody formation

Impaired cellmediated immunity

Reduced granulocyte function

Decreased serum conlplenlent

Fungi

Viruses

Parasites

Enteroviruses

Ginrdin

-

Pneumococcus H. irzflrlenzae Meningococcus Gastrointestinal pathogens Listeria Nocardia Snlnzonell~r Mycobacteria S. ailrezls P. aencgir~osa Other gram-negative bacilli Nocardill Pneumococcus H, iq41~enzae Meningococcus

C. r~eo,forr?~rmsHerpes viruses. H. crrysrdlatr/in esp. CMV B. dernzatifidis C. inznzitis Ptzei~nzocystis Candidn Asyergillzls

Toxoylasrna Strongyloide~

death of fastidious bacteria, or is representative of material overlying but not within a pathologic process generates results that mislead, delay, or prevent proper diagnosis and therapy.

3.1 Blood Blood cultures are one of the most important high-volume microbiology specimens. Collection practices that impact on obtaining reliable culture results include volume of blood per specimen, number of specimens, specimen collection from different sites, and proper skin antisepsis prior to venipuncture (Table 2) [2]. The volume of blood collected to diagnose bacteremia is a critical determinant. Every study addressing the issue of blood volume concludes that more blood is better. The usual bacteremic adult is intermittently bacteremic with fewer than one bacterium per milliliter of blood; in fact, 20% have fewer than 1 microbe per 10mL of blood. A czllture set is optimal when inoculated with at least 20mL of blood. Each culture, referred to as a culture set, represents a set of two or three bottles inoculated with the blood specimen from a single venipuncture. Higher volumes, such as 30mL, improve recovery of pathogenic bacteria but are less commonly used because of perceived difficulties with large-volume collection and the potential for overphlebotomizing patients. However, three sets containing 20mL each are equivalent in volume to two sets of 30mL, and this volume per day appears optimal for detection of

Table 2. Collection of blood for culture Type of culture Adult stat bacterial Adult routine bacterial

Geriatric bacterial Pediatric bacterial

When ordered Acute febrile episode Antimicrobials to be started or changed immediately Non-acute disease Antimicrobials will not be started or changed immediately Altered mental status vomiting, >6% bands Acute, non-acute febrile episode

No. of cultures

Volume of blood

2

20mL each culture

Collected immediately, before antimicrobials started

2-3

20mL each culture

2

20mL each culture

With 8-12 h Intervals no closer than 1 h, drawing at time of fever spike not routinely necessary Same as stat or routine above

2

1-20 mL each culture, depending on weight of patient

Timing

Immediately?

bacteremia. A review of blood volume studies concluded that each additional milliliter of blood collected increased significant positive cultures by 3% [3]. The most common contaminants of blood cultures are the coagulasenegative staphylococci, which reside on the skin and, in spite of adequate antisepsis, contaminate 1-3% of all blood culture phlebotomies. Two separate blood cultures can be used to evaluate the potential significance of an isolate of coagulase-negative staphylococcus [4]. Significant bacteremia with a coagulase-negative staphylococcus usually develops from an intravascular source, resulting in a continuous bacteremia with all cultures positive. Therefore, a single positive blood culture for a coagulase-negative staphylococcus is less likely to be significant than a strain of coagulase-negative staphylococcus isolated from two or three separate cultures. The accepted recommendation for number of blood cultures is two to three per day from separate venipuncture sites. This takes into consideration the relatively large blood volume needed to improve sensitivity and the need for more than one culture to evaluate staphylococcal contamination. More than three cultures in 1 day are rarely needed and should not be ordered unless there is a significant change in the patient's clinical status. Proper skin antisepsis is required to prevent excess blood culture contamination. At a minimum, skin antisepsis should include the use of an iodinecontaining compound that is allowed to dry before blood collection. Tincture of iodine dries more quickly than iodophor preparations and may represent a more useful antiseptic in a busy clinical environment [5]. Transport of blood collected for culture from the patient's bedside to the laboratory should be rapid to ensure immediate incubation and shortened detection times for positive samples. Blood should be collected directly

into culture bottles and held at room temperature until received by the laboratory. Test-of-cure blood cultures, or those collected after diagnosis and treatment to document clearing of the bacteremia, are rarely necessary. In general, monitoring the clinical progress and ordering additional blood cultures only when the patient does not respond properly is preferable. However, the technique of repeat blood cultures on the day following a documented episode of sepsis can be helpful in diagnosing a suspected intravascular infection (endocarditis or catheter-associated bacteremia), because only rarely does bacteremia due to an extravascular source persist beyond 24 hours. In addition, patients who have failed therapy, who have a difficult bacterium to eradicate (e.g., Brrrcella spp.), or who have disease caused by multiply resistant bacteria (e.g., vancomycin and ampicillin-resistant enterococci) may require a test-of-cure culture because data are not available to indicate what treatment regimens are likely to be successful.

3.2 Respiratory specimens Transport of all lower respiratory tract specimens must be rapid. Overgrowth of potential pathogens by contaminating bacteria can occur within 1-2 hours at room temperature and the death of fastidious pathogens, such as S. pneumonine, can occur in even less time. Specimens should be stored at room temperature and must arrive in the laboratory within 2 hours of collection. The three most common respiratory tract specimens received by the microbiology laboratory are expectorated or induced sputum, endotracheal aspirates, and bronchoalveolar lavage (BAL) fluid. Ideally, all three specimens represent bronchial and alveolar secretions that are largely free of normal oral bacterial flora. Most etiologies of community- and hospital-acquired lower respiratory tract disease are oral flora that colonize the patient at the time of aspiration and tracheal inoculation. For this reason, the mere presence of potentially pathogenic bacteria does not definitively establish the cause of disease. Respiratory specimens can provide clinically useful information from staining and culturing if screened for gross oropharyngeal contamination, indicated by elevated counts of squamous epithelial cells, and rejected if contaminated. Purulent, bloody, or tenacious portions of sputum are used to prepare a smear and to inoculate culture media. Comparison of sputum culture results with those of transtracheal aspirates and bronchoscopy specimens confirms that useful information is gained only from those sputums that contain low numbers of squamous epithelial cells. Most laboratories use either 25 squamous epithelial cells indicates that culture of the specimen is colonized with oropharyngeal flora and will likely provide no useful information. In some laboratories, acceptable sputum specimens are identified by both low numbers of squamous

epithelial cells and elevated counts of polymorphonuclear leukocytes. The increased complexity of screening based on two cell types may not be justified because studies have shown squamous epithelial cell screens are equivalent to screening procedures utilizing squamous epithelial cells and white blood cells. In addition, severely neutropenic patients may have an infected lower respiratory tract with minimal oropharyngeal contamination that is devoid of inflammatory cells. Gram stains of acceptable sputum specimens should have inflammatory cells and predominating or intracellular bacteria quantitated and reported [6]. Identification and reporting of potential pathogens detected in culture should include those predominating over normal respiratory flora or those with morphotypes noted in the Gram stain as predominant or intracellular . Endotracheal aspirates can be screened and cultured, and results interpreted, in a manner similar to sgutums, with one exception. Any endotracheal aspirate that shows no bacteria in the Gram-stained smear does not require a bacterial culture. Studies have demonstrated that cultures of Gram stain"negative" aspirates do not provide information that is clinically useful, even from neonates [7]. Bronchoalveolar lavage specimens provide useful information when processed using both the Gram-stained smear and a semiquantitative culture. A cytospin smear preparation of lavage fluid should be examined for inflammatory cells and bacteria. If more than 1% of cells present are squamous epithelial cells, excessive contamination with upper respiratory tract material has occurred and relative numbers of bacteria are not helpful. If very few ( < I % ) cells are squamous epithelial cells, the quantity of neutrophils and bacteria should be noted. Intracellular bacteria (within neutrophils) should also be reported. Potential pathogens and obvious contaminating oral flora should be quantitated and reported. Bronchoalveolar lavage smear and culture results should he interpreted according to Table 3 [S]. Using a "protected" catheter to obtain the bronchoalveolar lavage (BAL) specimen can avoid the problem of upper airway contamination and is recommended [9]. Specimens obtained using a protected bronchial brush also may be processed quantitatively. However, they are not as useful as lavage specimens due to the small amount of material obtained with this technique. One exception is in the diagnosis of invasive fungal (i.e., Aspergillus) pneumonia. In this setting a bronchial brushing is likely to be positive even when a lavage specimen is negative. 3.3 Urine specimens Urine is considered an easy specimen to collect but, because it is easily contaminated, interpretation of results may be difficult and misleading. Three common specimens received by the microbiology laboratory include midstream urines, straight (removed) catheter urines, and urine from indwelling catheters. Difficulties encountered when collecting urine by each

Table 3. Guidelines for interpretation of bronchoalveolar lavage fluid smear and culture results

Direct Gram stain

Culture

Interpretation

Elevated number of squamous epithelial cells (SEC), i.e., >1% of all cells are SECs

Any quantity of potential pathogen and respiratory flora

< I % of all cells are SECs;

100,000 potential pathogens1lnL in unconcentrated BAL fluid Quantity of respiratory flora less than potential pathogen

BAL fluid likely contaminated with orophryngeal flora or aspiration pneumonia; many PMN's with mixed intra and extracellular respiratory flora suggests aspiration Does not suggest disease by routine (cultivatable) bacterium Significance of potential pathogen indeterminate: intracellular bacteria in direct Gram stain suggest potential pathogen causing disease

PMNs may or may not be present < I % of all cells are SECs; PMNs usually present

I 0' potential pathogenslmL

>10' potential pathogens1mL >lo' potential pathogenslmL >I@ potential pathogens1mL Any number (usually >lo3) potential pathogenslmL >los potential pathogenslnll

method and the pathogenesis of urinary tract infection in a variety of hosts necessitate different interpretative criteria for urine cultures (Table 4) [lo]. Urine collected by any method requiring transit through the urethra is likely to be contaminated. Most bacteria divide rapidly in urine at room temperature. Refrigeration will stabilize colony counts for several days. Boric acid-containing tubes are commercially available for transportation at room temperature when transit and processing times are expected to be greater than 1 hour. Colony counts of bacteria in boric acid are stabilized without altering viability for culture. Urinalysis can be performed on specimens from the boric acid tubes for at least 24 hours after collection. Any urine specimen can be evaluated rapidly using a cytocentrifugation method. The cytospin is an instrument that uses high-speed centrifugation

and capillary action to concentrate cells in a smear approximately l c m in diameter. For all patient groups except outpatient women, sensitivity is great enough that culture is not necessary when a negative Gram stain is reported [Ill. 3.4 Cerebrospinal fluid specimens Detection of microorganisms in cerebrospinal fluid (CSF) is a medical emergency. Rapid interpretation of a stained smear, combined with immediate inoculation of specimen to culture media, is essential. Bacteria that infect the meninges, such as N. meningitidis, S. pneumoniae, and H. influenzae, are fastidious, requiring room temperature transportation and rapid inoculation onto culture media to ensure recovery. The issue of which tube, of the multiple tubes collected during the lumbar puncture, to use for culture is less important than sending a sufficient volume of specimen to ensure detection. It is vitally important to have sufficient fluid for all the microbiology tests ordered, bacterial (1.5mL), fungal (5mL), mycobacterial (5mL), and viral (1 mL). Many of these infections involve a basilar meningitis and unless sufficient volume is submitted for processing, a false-negative result is likely, which may complicate proper diagnosis. Antigen tests have been used for more than 15 years to detect bacteria in body fluids, but, unfortunately, have proven themselves to be of limited value [12], Antigens of H. influenzne, N. meningitidis, S. pneumonine, and group B Streptococcus can be detected with fair sensitivity and high specificity. However, evaluations have shown that performance is similar to that of the Gram stain, and when an antigen test is positive changes are infrequently made in patient management, leading only to increased medical costs without improvement in patient care. At this point, antigen test results offer very little to the diagnosis of infections in patients with cancer and should be used sparingly or not at all.

3.5 Gastrointestinn1 specinzens Fecal specimens and endoscopic bowel biopsies are common and useful specimens for the diagnosis of gastrointestinal infections in cancer patients. Spontaneously passed stool or endoscopically collected aspirates are used to document the bacterial or parasitic etiology of infectious diarrheas, whereas biopsies collected during endoscopy are used to detect viral causes of gastrointestinal disease. Infectious causes of gastrointestinal disease are listed in Table 5. Freshly passed stool should be processed and cultured within 1 hour. Clostridium dificile is nearly the exclusive cause of diarrhea acquired within the hospital, especially in patients with prior antimicrobial therapy. This exclusive association suggests that stool cultures for other bacterial causes of diarrhea are not warranted if diarrhea begins in the hospital [13].

Table 5. Usual infectious causes of gastrointestinal disease Etiologies -

-

Setting

Bacteria

Parasites

Viruses

Travel

E. coli Salmonella Carnpylobacter A eromonas Carnpylobacter Salmonella Shigella E. coli (enterohemorrhagic) C. dificile C. dificile Same as above plus M. avium

Giardia Cryptosporid ium Cyclospora Amoeba Giardia

Unknown

Outpatient

Inpatient Compromised

Same as above plus: Microsporidium Isospora Biastocystis hominis

Rotaviruses (pediatric)

Rotaviruses (pediatric) CMV

Small and large bowel biopsies are used to detect cytomegalovirus (CMV) disease. CMV-positive cultures of washes or feces have poor predictive value for active disease. Histologic confirmation, which includes finding CMV inclusions within epithelial lining cells, is strongly suggestive of significant gastrointestinal disease. Biopsy tissue for viral culture should be transported to the laboratory in sterile viral transport medium (VTM). 3.6 Wound and cutaneous specimens

Wound and cutaneous specimens are very helpful when collected aseptically from a "deep" site. Material swabbed from the skin or wound surface is frequently colonized with contaminating bacterial flora and should be avoided. Wound specimens are best collected by aspiration or biopsy, as a way to bypass the contaminated surface. Irrigating with or injecting nonbacteriostatic saline may be necessary to ensure sufficient sampling volume. If anaerobic culture is needed, transport must occur in an oxygen-free container. Small pieces of tissue and purulent fluid should be added or injected into a transport vial or tube. Swabs are difficult to maintain in an oxygen-free environment, easy to contaminate with normal flora, may actually "trap" or kill the bacteria causing an infection, and sample too little material. Therefore, swabs as collection devices for wound and cutaneous specimens should not be used. 3.7 Intravascular catheter specimens

Skin and soft-tissue infections are confirmed by culturing purulent material or tissue from the cutaneous and intracutaneous area around the outside of the

catheter. Documenting the catheter as the source of a bacteremia is more difficult. When the catheter is removed, the distal 2 inches of the intravascular portion can be cultured by one of several methods to assess the catheter as the likely source of bacteremia. The method of Maki and colleagues is the simplest, requiring that the catheter be rolled on an agar plate [14]. Less than 15 colonies of a single isolate suggests the catheter is not the source of the bacteremia. Greater than 15 colonies, and in most cases greater than 100 colonies, suggests the catheter is the source of the bacteremia, but requires a positive blood culture to confirm the association. Alternatively, sonicating the catheter tip, a method that samples both the external and internal surfaces, improves the detection of infected catheters by approximately 10%. The sonicate is then quantitatively cultured. Less than 1000microorganisms per milliliter of a single isolate suggests the isolate is not causing the bacteremia [15]. Transport of catheter segments to the laboratory should occur rapidly and in a sterile container. As the catheter surface dries, adherent bacteria can die. Bacteria also continue to multiply in the moist slime on the catheter surface. Therefore, in order to ensure accurate colony counts and viability, transport should be 5 times the counts of the venipuncture specimen [16].

4. Methods for the diagnosis of infections caused by unusual or fastidious etiologies There are many organisms that require special culture media or laboratory preparation to optimize recovery. If clinical microbiology laboratory techniques cannot be optimized, then false-negative cultures may result. 4.1 Legionella Legionella spp., especially L. pneurnophilia, are causes of both communityand hospital-acquired pneumonia. The laboratory diagnosis of pneumonia

,

Table 6. Summary of Legionella diagnostic tests

Test

Sensitivity

Specificity

Turn-around time

Culture Fluorescent antibody stain (DFA) of smear of respiratory specimen Serology

40-100% 25-75%

100% 95-99%

3-5 days Hours to 1 day

25-75%

95-99%

Urinary antigen"

40-75%

95%

Weeks to months (convalescent specimen needed) Hours to days (send out)

"Detects disease caused by L. pnez~rnophilnserogroup 1 only.

caused by the legionellae can be accomplished by detection of serum antibody, fluorescent antibody staining, and culture of respiratory secretions. A test for detection of legionella antigen in urine is also available, with comparative studies suggesting that it performs as well as fluorescent antibody staining or culture. A summary of available legionella diagnostic tests occurs in Table 6. Although a positive culture provides definitive identification of disease, it is not highly sensitive unless the sputum undergoes careful preparation (washing) with plating to several media in the laboratory. In addition, positive cultures require 3-5 days to be recognized. The direct fluorescent antibody (DFA) test is the most rapid test for the diagnosis of legionellosis but lacks both sensitivity and specificity. It should not be done in the absence of culture as an accompanying test. Serology requires testing of acute and convalescent serum specimens, necessitating a delay of several weeks to months before a significant increase in antibody is detected. Detection of urinary antigen is a rapid and very specific test but has not been widely offered because of its use of radioisotopes. Fortunately, an enzyme-based test is now commercially available [17]. DFA testing and culture of specimens obtained during bronchoscopy or by lung biopsy are the most rapid and accurate diagnostic approaches, and are recommended when legionnaires' disease is strongly suspected.

4.2 Bartonella The bartonellae are a large group of organisms, with a limited number recently recognized as causing disseminated infection in compromised hosts, including those not infected with HIV. Reports suggest that B. henselne, and possibly B. quintann and B. elizabethae, can be recovered from blood using automated blood culture methods, acridine orange staining, wet mount testing for motility, and culture on blood agar following lysis centrifugation (Isolator) treatment [18]. Figure 1 shows a culture and detection algorithm that can be used by most laboratories. Here, too, consultation with laboratory personnel is necessary to determine when blood submitted for culture should be processed by this more laborious pathway.

Blood. tissue, or fluid specimen

Automated Blood Culture System Inoculate up to 5 m L of specimen into aerobic broth medium Incubate 14 days with rockinglagitation

1

Positive growth reading

Wet mountlphase microscopy Use oil immersion objective to examine for rachety motility

Stain one drop of concentrate with acridine orange. Bartor~ellaare small pleon~orphic bacilli

Process blood from positive bottle by injecting into IsoIator (Wampole Laboratories) tube

Inoculate concentrate onto chocolate and sheep blood agar plates Incubate taped plated at 36' C 0 2for 1 month or until growth is evident

Figure I . Proposed algorithm for culture and identification of Bartonella spp. The Automated Blood Culture System is a continuous monitoring system, including BacTiALert (Organon Leknika Corp.), Bactec 9000 Series (Becton Dickinson), and ESP (Difco Laboratories).

4.3 Capnocytophngin Capnocytophagia spp. are common inhabitants of the oral cavity, and are also well-documented opportunistic pathogens in patients with impaired host responses. C. sputigena, C. ochracea, and C. gingivalis can cause sepsis in patients with granulocytopenia [19]. Detection is relatively easy using normal laboratory media, including usual blood culture broth formulations. 4.4 Gmm-positive bacilli

Gram-positive bacilli that can act as opportunists should be suspected in cancer patients with undiagnosed infections [20]. These include Listericr rnoaocytogenes, Nocardia asteroides complex, Rhodococc~lsequi, and Bacillus cereus. Although these microbes grow on usual culture media, an infectious diagnosis may be missed if they are ignored as contaminants.

4.5 Mycobacteria Mycobacteria are ubiquitous environmental bacteria and can result in serious disease in patients with cancer. In spite of the severity and continued presence of Mycobacterium tuberculosis disease, species of mycobacteria other than tuberculosis (MOTT) are responsible for more disease in most laboratory settings [21]. Table 7 summarizes the most pathogenic mycobacteria, including the type of diseases reported along with the hosts' immunocompromising conditions. Molecular detection of M. tuberculosis by polymerase chain reaction (PCR) techniques can aid in the rapid detection of infected patients. Although PCR detection is less sensitive than culture and, therefore, should not replace traditional detection methods, it can be used to rapidly confirm the presence or absence of M. tuberculosis in specimens that contain acid-fast bacilli in the direct smear. Settings where the expense of this additional diagnostic test are justified, which can run between $150 and $200 in direct laboratory cost when a single specimen is tested on an urgent basis, must be established for each practice, institution, and laboratory. Table 7. Pathogenic mycobacteria other than Mycobacterium tuberculosis Mycobacterium M. bovis

Disease

Host

Pulmonary, extrapulmonary, Normal to severely bladder, or, rarely, severe compromised complications following intravesical inoculation to treat superficial bladder cancer Leprosy Not endemic in U.S.A. M. leprae Photochromogens (colony pigment produced following exposure to Iight) M. kansasii Pulmonary with dissemination, Normal or compromised skin, bone M. rnarinlinz Skin, soft tissue, musculoskeletal Normal M. sinziae Pulmonary Normal or compromised Scotochromogens (colony pigment produced in light and dark) M. szulgai Pulmonary, musculoskeletal Normal or compomised M. xenopi Pulmonary, disseminated Generally infects patients with underlying respiratory disease M. scroj%laceum Lymphadenitis, lung, disseminated Normal or compromised M. rhernzoresistible Pulmonary, skin Normal or compromised Nonchromogens (no colony pigment) M. avium-intmcellulare Pulmonary, disseminated Normal or compromised (MA11 M. ~ ~ l c e r a n s Skin, soft tissue Not endemic in U.S.A. M. malnzoense Respiratory, skin and disseminated Normal or compromised less common M. haernophilirrm Skin and subcutaneous tissues Normal and compromised M. genavence Multisystem disease AIDS Rapid growers M. fortuitum complex Skin, soft tissue, lung, disseminated Normal or compromised M. fortuitum M. chelonae M. nbscesus

4.6 Fungi Most fungi grow readily on a wide variety of media if incubated for a sufficiently long period of time. It is important to remember that yeasts such as Candida, Torulopsis, and Cryptococcus spp. should be reported when detected from clinical specimens. Other fungi capable of causing invasive disease, such as Histoplasma, Blastor?zyces, Coccidioides, Rhizopus, and Aspergillus, are not routinely detected by bacterial culture. A specific fungal culture request should be used to ensure 3-4 weeks incubation and the use of selective media to prevent bacterial overgrowth for these latter organisms. This includes blood cultures specified for the detection of fungi. Fungi are strict aerobes and grow best on the surface of an agar plate, not in a broth medium away from oxygen. The Isolator lysis centrifugation blood culture system, which uses a blood concentrate inoculated to the surface of agar plates, has been found to be the best all-around method for detecting yeast and fungi in blood [221. However, automated, continuously monitoring bacterial blood culture instruments are reliable for the detection of Candida and Torulopsis species causing a fungemia. 4.7 Viruses The commercial availability of high-quality reagents and cell cultures for the detection of viruses allows many clinical laboratories to rapidly and reliably detect pathogenic viruses. Viral diseases in cancer patients may be thought of as those that reactivate in the immunocompromised host, such as the herpes viruses, and those that are usual community pathogens, such as influenza and adenoviruses. The quickest method for virus detection is direct staining of a clinical specimen using an antibody specific for the virus suspected that is conjugated to fluorescein isothiocyanate (FITC). In the appropriate setting, fluorescent antibody staining to detect influenza virus, respiratory syncytial virus (RSV), adenovirus, herpes simplex virus (HSV), cytomegalovirus (CMV), or varicella-zoster virus (VZV) can quickly establish a diagnosis. Viruses that are not cultivatable in normally available cell cultures can be detected by molecular methods, such as the PCR test. High detection sensitivity for viral nucleic acids by PCR must be balanced against the high cost and apparent detection of latent viral genome in situations where no clinically significant viral disease is present. At the current time it appears that PCR is the test of choice for detection of herpes group viruses (HSV, VZV, CMV) in spinal fluid. Its utility in other settings is less well defined. Isolation of viruses in cell culture is the standard laboratory method for detection [23]. Conventional cell culture, which requires the appearance of a cytopathic effect (CPE) in the infected cells, takes 1 day to 2 weeks before the presence of a virus can be confirmed. Rapid cell culture, using the shell vial method, grows and detects viruses within 1-2 days. Shell vials are more rapid

because fluorescent antibody staining is used to detect virally encoded proteins produced within hours of host cell infection. Conventional culture is the only available method where viable virus is recovered that can then be used for susceptibility testing or epidemiologic investigation. Table 8 summarizes detection methods for common viruses.

4.8 Parasites In addition to the traditional list of pathogenic animal parasites, such as Entamoeba histolytica, Giardia lamblia, and the nematodes, an increasing number of opportunistic gastrointestinal and systemic pathogens are being identified in the immunocompromised host [24]. Table 9 summarizes these opportunists, especially those that have been recognized within the last decade. To ensure that all parasitic forms are adequately preserved, a specific transport medium for parasitic analysis should be used if transportation time will exceed 1 hour. In addition, stool specimens containing barium are unacceptable for examination for 5-10 days after barium is given to the patient. Although most parasites, which are opportunistic gastrointestinal pathogens, are detected by examining a single diarrheal stool specimen, multiple specimens may occasionally be necessary. It is important to notify the laboratory of unusual pathogens that may be suspected or underlying diseases that can predispose the patient to infrequently encountered parasites. 4.9 Pneumocystis carinii Pneumocystis carinii is a pathogen of major importance in many centers treating immunocompromised hosts. While infection is common is patients with HIV, Pne~~rnocystis carinii pneumonia can also occur in patients with leukemia and following bone marrow transplantation. Recent studies have shown the high sensitivity of a DFA stain applied to properly collected, induced sputum of sufficient volume [25]. Under these circumstances, and when disease is highly suspected, the sensitivity of DFA with a single specimen exceeds 90%. A negative stain result should suggest the need for bronchoscopy to directly obtain pulmonary specimens for laboratory analysis.

5. Tests for rapid detection of microorganisms The best practice of medicine requires prudent use of medical resources, including laboratory testing. The typical turn-around times and relative laboratory costs for common microbiology tests are detailed in Table 10. It is important to note that many of the most rapid tests for diagnosing infectious disease still involve direct microscopic inspection of the specimen from a suspected site of infection. Commonly used stains are summarized in Table 11. The Gram-stained slide can provide rapid and useful information within 30-60

^ (^

u

=pE cO

0 CD

'3 03

yi

uuu0u

to T :

1 < Cui CL,

C3

X) X) X) JO

O O 0 O O T3 T3 T3 T3 JO •3 ' 3 ' 3 ' 3 ' 3 q3 C !3! T : C

)-H

Jli X> X l J3 CO

)^

U




.3

c c CO 03

•"> ^

.b

>

o

^ ?

5 > -^ i I

(U CO

p:;>

> 157

Table 9. Opportunistic parasites Parasite

Disease

Specimen

Diagnostic test

Cryptosporidium parvurn Cyclospora cayetanesis Isospora belli

Diarrhea Diarrhea Diarrhea

Stool Stool Stool

Sarcocystis spp. Microsporidia Acantharnoeba spp.

Diarrhea Diarrhea Cutaneous, neurologic, nasal Hyperinfection syndrome

Stool Stool Biopsy (tissue) Stool, sputum

Acid-fast or fluorescent stain Acid-fast stain Wet mount of stool concentrate and acid-fast stain Wet mount of stool concentrate Modified trichrome stain Hematoxylin-eosin stain, culture Microscopic examination

Strongyloides stercoralis

Table 10. Turn-around times (TAT) and relative laboratory costs for common microbiology tests Test Stains Gram, acid-fast, Calcofluor white (fungi) Fluorescent antibody Trichrome (parasites) Antigen detection Latex agglutination Enzyme immunology assay Molecular (nucleic acid) detection DNA probe PCR~LCR' Culture Bacterial Mycobacterial Fungal Viral

TAT

Relative cost"

2-6 h 6h to 2 days 1-4 d 2-4 weeks 1-3 weeks 2 d to 3 weeks

"Approximate cost to laboratory: +, $5-10; + +, $10-20; + + +, $20-30; 'PCR = polymerase chain reaction; LCR = ligase chain reaction.

+ + + +, $30-50.

minutes of specimen receipt by the laboratory. While it appears to be a simple technique, considerable experience is needed to avoid misinterpretation of what is seen. For this reason it is best to view the slide with a well-trained observer in the clinical microbiology laboratory before using the results as the primary basis for initiating specific treatment. Quantitation of inflammatory cells and recognition of bacterial morphology provide a narrowing of etiologic possibilities, and a guideline for selection of empiric antimicrobial therapy. Table 12 summarizes common bacterial microscopic morphologies and their implied etiology. Historically, fungi have been detected using a KOH wet mount examination. In many laboratories a fluorochrome, calcofluor white, is added to KOH solutions to help differentiate fungal structures from specimen debris. Table 13 summarizes common fungal morphologies seen in clinical material.

Table 11. Stains used to demonstrate microorganisms in clinical material Microorganism group

Stain

Comment

Bacteria

Gram Acridine-orange Fluorescent antibody KOH/Calcofluor white Gomori methenamine silver Carbofluchsin based Kinyoun Ziehl-Neelsen Flurochrome based Auramine-rhodamine Iodine wet mount Trichrome Fluorescent antibody Giemsa Fluorescent antibody Giemsa

Gram negative, gram positive Stains nucleic acid Must use organism-specific reagents

Fungi Mycobacteria

Parasites - intestinal

Parasites - blood Viruses

Used to stain fungi in tissue Can be modified to stain Nocardia

Best for ova and larvae Best for protozoa Must use parasite-specific reagents Must use virus-specific reagents Stains viral inclusions

Table 12, Interpretation of Gram stains Observed morphology

Interpretation

Gram-positive cocci clusters Gram-positive cocci chains Gram-positive diplococci Gram-negative diplococci Gram-negative coccobacilli Gram-positive rod diphtheroid Gram-positive rod boxcar Gram-positive rod branching Gram-positive rod Gram-negative rod

Staphylococci Streptococci Pneumococci Nei~seria!Moraxella Haernophil~is Cor~nebacteri~~rnlPropionibacteri~~nz Clostridi~smiBacill~~s Nocardia!Actinomyces Other gram-positive rods Enteric and pseudomonas-like Gram-negative rods and Cry,vtococcus) Yeast (e.g., Cnnclidn, Tor~~lopsis, Candida Mold (e.g., Aspergillus)

Yeast cells Yeast cells with pseudohyphae Hyphae

Acid-fast stains are used to demonstrate the presence of mycobacteria. The main reason for low sensitivity of stains for mycobacteria is that these infections are often present with a low number of organisms, and even with appropriate concentration of the collected material, false-negative results occur in 50% of cases. Most parasitic worms (adults, larvae, and ova) are large enough to be detected without the use of special stains. In contrast, protozoans in blood and most protozoans in stool need staining to be seen. Direct staining of clinical material for viruses makes use of the Giemsa and fluorescent antibody stains. Fluorescence staining, using antibody specific for

T a b k 13. Identification of fungi seen in stained smears of clinical material Fungus

Morphology in clinical material

Candida, Tor~ilopsis, Cryptococcus, and other yeasts Aspergillus, Fusariunz, and other nonpigmented molds Rhizopus and other related Zygomyceies Exophilia, Curvlilaria and other brown-pigmented molds

Budding yeast cells; in addition, Canrlida spp. produce pseudohyphae and cryptococci produce capsules

Histoplasma capsulatum Blastomyces dernlatitidis

Hyaline (clear), septate hyphae Wide, ribbon-like, nonseptate, hyaline hyphae Dematiaceous (brown pigmented) septate hyphae, grains or compact masses of dematiaceous hyphae, or muriform (cells divided by fusion plane) Small budding yeasts (intra- and extra-cellular) Budding yeast with broad neck between parent and daughter cell Large spherules containing smaller, nonbudding endospores

a suspected virus, is the more sensitive method and can be used for rapid, specific diagnosis of adenovirus, influenza virus, RSV, HSV, VZV, and CMV in the appropriate specimen. Close communication between the clinical and laboratory services can help optimize diagnostic results, improving speed, accuracy, and overall cost-efficient testing.

6. Determining microbial susceptibility or resistance

Standard in vitro tests used to measure antimicrobial activity are the disk diffusion and the minimum inhibitory concentration (MIC) tests. Both tests measure inhibition of microbial growth as an endpoint. Although bactericidal testing is occasionally needed, the vast majority of experience comparing clinical outcomes to laboratory test results are with inhibitory tests. The MIC test is performed in antimicrobial-containing broth and is defined as the lowest concentration of antimicrobial (expressed in micrograms per milliliter) that inhibits growth of the test organism. The disk diffusion test can be thought of as an MIC test on the surface of the agar plate. Antimicrobial concentration is greatest close to the disk and least away from the disk, where less drug has diffused. Results of both tests are correlated with the breakpoint of the antimicrobial in question, resulting in a report of susceptible or resistant. The breakpoint can be thought of as the concentration of antimicrobial that can be safely achieved after a normal dose in a patient. Breakpoints for various antimicrobials are different because of dosing and pharmacokinetic variables. If the MIC result is lower than the breakpoint, the antimicrobial is reported as susceptible. If the MIC is greater than the breakpoint, the result is resistant. MIC results that fall near the breakpoint may be reported as intermediate. Susceptible implies that the microorganism tested will be inhibited by normally achievable levels following a standard dose. Resistance implies that

inhibition should not be expected. Because intermediate results imply the MIC is near the breakpoint, they can be interpreted to mean that the microorganism will be susceptible in areas of the body where the antimicrobial in question is concentrated, such as urine, or that the isolate will be susceptible if higher than normal doses can and will be safely administered. A common error when examining a laboratory report listing many antimicrobials is to assume the lowest MIC value represents the most effective drug to use. 6.1 Antibiograms

Empiric therapy is often initiated in cancer patients, especially those with neutropenic fever. Initial therapy while awaiting culture results is dependent upon an antibiogram, which is a periodic (usually annual) compilation of susceptibility results. All common pathogens and antimicrobials tested by the laboratory are reported as percent susceptible, for example, the percentage of S. aureus susceptible to oxacillin. The antibiogram for every medical center will be different, depending on the resident bacterial flora. Knowledge of the local antibiogram is essential for the selection of proper empiric therapy. 6.2 Specinl susceptibility tests In addition to routine bacterial susceptibility testing, tests that are useful in limited situations include: antimicrobial assays, combination agent testing (synergylantagonism), and serum bacteriocidal testing. Antimicrobial susceptibility testing for microorganisms other than rapidly growing bacteria also may be appropriate in limited circumstances. These tests are those for antimycobacterial, antifungal, and antiviral agents. Antimicrobial assays measure the quantity (measured in micrograms per milliliter) of drug in a fluid, usually serum. Most often, assays are used to avoid toxic concentrations and to optimize therapeutic levels of the antibiotic [26]. Vancomycin and aminoglycoside assays comprise the majority of assays performed by hospital laboratories. Specimen collection and transport must ensure that labile drugs do not degrade, resulting in a falsely low level. Ideally, clotted blood should reach the laboratory within 4 hours of collection. If storage or transport is prolonged, the sample should be placed on ice or refrigerated at 2-8OC. Storage for greater than 3 days should be at -20 or -70°C. Information necessary for proper interpretation and dosing calculations that should be included with the assay request includes the exact time of specimen collection; the exact time of last dose; whether the specimen represents a peak, trough, or random level; what other antimicrobial agents the patient is receiving; and which drug should be assayed. Combination testing refers to measuring the inhibitory or bacteriocidal (killing) effects of testing two antimicrobials at the same time. If together they perform better than predicted by their individual effects, they are synergistic. If their combined effect is less than that predicted by their individual effects,

they are antagonistic. Traditionally, combination testing has been used to document synergy for serious infections such as endocarditis or osteomyelitis [27]. Combination testing may also be necessary if rifampin is to be added to a cell wall-active antistaphylococcal drug. In this instance, antagonism should be measured because reports have documented this combination may be synergistic, antagonistic, or indifferent [28]. In the era of emerging resistance to many of our potent antibacterial compounds, use of these tests may become more prevalent and important when new multidrug resistant microbes make our past clinical response data no longer applicable. The need for serum bacteriocidal testing remains controversial [29]. The test involves collecting serum after the antimicrobial regimen has been selected and initiated. The serum is diluted and a standard suspension of the patient's pathogen, which has been cultured previously, is inoculated to each dilution. Ideally, lower dilutions of the serum inhibit and kill the inoculum. Absence of these effects suggests the therapeutic regimen may be inadequate. Historically, testing was reserved for serious infections requiring prolonged therapy with bacteriocidal drugs such as osteomyelitis and endocarditis. More standardized in vitro tests and experience with clinical outcomes has obviated the need for most serum bacteriocidal testing. An expanded role for less conventional antimicrobial susceptibility testing may arise with the emergence of multiply antibimicrobial resistant organisms in patients with cancer. This is most evident with vancomycin-resistant enterococci, multiply antibiotic-resistant Pseudornonas aeruginosa, and azoleresistant Candida species. Because combination and bacteriocidal susceptibility tests are not fully standardized and lack a national consensus on interpretation of the results, it is important for both medical and legal purposes that trained personnel interpret the results of this testing. This can be done through written consultation by the microbiology laboratory director documenting the medical interpretation of results.

7. Summary

The proper use and interpretation of clinical microbiology test results may be complicated but critical to the care of cancer patients. The microbiology laboratory director is often available to offer advice concerning the differential diagnosis, choice of specimens, as well as the optimal stains and cultures to facilitate diagnosis. Additionally, the rapid interpretation of Gram-stained smears provides useful, occasionally lifesaving, information relative to the etiologic diagnosis and empiric antimicrobial therapy. The microbiology laboratory director should also provide further interpretation of culture and antimicrobial testing results that allow the clinical service to focus on the most critical data. Person-to-person or telephone conversations discussing important laboratory information should be followed up by a written summary

MICROBIOLOGY REPORT

Patient's Name (Last-First-Middle) Patient No:

Sex:

Room No:

Specimen Number:

DOBIAGE:

Date Specimen Received:

Attending Physician:

SUMMARY:

Filamentous fungus detected in direct Gram stain and culture. Morphology suggests a dematiaceous fungus resembling Exoohiala species. Dematiaceous fungi are responsible for a number of subcutaneous diseases. The morphology in the Gram stain suggests subcutaneous phaeohyphomycosis.

TEST REQEST:

Bacterial and fungal culture

SPECIMEN TYPE:

Aspirate of leg lesion

COMMENT:

The presence of septate, swollen hyphae in the Gram stain and a heavy growth of mold on culture plates suggests this b g u s is a "significant" isolate, rather than an incidental contaminant or colonizer. Dematiaceous fungi are a family of yeasts and molds that contain brown or black pigment in their cell walls. They cause a range of subcutaneous diseases, including; phaeohyphomycosis, mycetorna, and chromoblastomycosis. The septate hyphae seen in the direct examinatiok with no detectable grains (granules) or muriform bodies, suggests phaeohyphomycosis. The literature suggests surgical excision, especially in immunocompromised hosts, is the treatment of choice. Newer azole antifungal agents, however, have been used successfully. Case discussed with Dr. Berlin.

Figure 2. Special microbiology written reports.

report placed in the patient's chart so all services involved share the same interpretation (Figure 2). The clinical service has an important responsibility to communicate with the laboratory to optimize care of the patient with cancer. The laboratory compiles data collected from groups of patients that is available and useful to physicians. Review and discussion of test utilization is essential for costeffective, quality health care. This may include analysis of blood cultures documenting an acceptable level of contamination, appropriate number collected per day, and sufficient blood volume per culture. In addition, information about changing resistance patterns or nosocomial transmission can be provided to the clinician. As patients with malignancies become more complex and their infections increasingly difficult to treat, regular interaction between the laboratory and clinician is likely to improve patient care. References 1. Rosenow E, Wilson W, Cockerill F. Pulmonary disease in the immunocompromised host. Mayo Clin Proc 1985;60:473-487.

2. Reimer L, Wilson M, Weinstein P. Update on dectection of bacteremia and fungemia. Clin Microbiol Rev 1997;10:444-465. 3. Mermel A, Maki D. Detection of bacteremia in adults: Consequences of culturing an inadequate volume of blood. Ann Intern Med 1993;119:270-272. 4. Kloos W, Bannerman T. Update on clinical significance of coagulase-negative staphylococci. Clin Microbiol Rev 1994:7:117-140. 5. Schifman R, Pindur A. The effect of skin disinfection materials on reducing blood culture contamination. J Clin Microbiol 1993;99:536-538. 6. Murray P, Washington J. Microscopic and bacteriologic analysis of expectorated sputum. Mayo Clinic Proc 1975;50:339-334. 7. Morris A, Tanner D, Reller B. Rejection criteria for endotracheal aspirates from adults. J Clin Microbiol 1993;31:1027-1029. 8. Kahn F, Jones J. Diagnosing bacterial respiratory infection by bronchoalveolar lavage. J Infect Dis 1987;155:862-869. 9. Baselski V, Baselski VS, el-Torky M, Coalson JJ, Griffin JP. The standardization of criteria for processing and interpreting laboratory specimens in patients with suspected ventilatorassociated pneumonia. Chest 1992:102(Suppl. 1):571S-579s. 10. Hooton T, Stamm W. Diagnosis and treatment of uncomplicated urinary tract infection. Infect Dis Clin North Am 2997;11: 551-581. 11. Goswitz JJ, Willard KE, Eastep SJ, et al. Utility of slide centrifuge Gram's stain versus quantitative culture for diagnosis of urinary tract infection. Am J Clin Pathol 1993:99:132136. 12. Kiska DL, Jones MC, Mangum ME, Orkiszewski D, Gilligan PH. Quality assurance study of bacterial antigen testing of cerebrospinal fluid. J Clin Microbiol 1995;33:1141-1144. 13. Fan K, Morris A, Reller 3 . Application of rejection criteria for stool cultures for bacterial enteric pathogens. J Clin Microbiol 1993:31:2233-2235. 14. Maki DG. Weise CE, Sarafin HW. A semiquantitative culture method for identifying intravenous-catheter-related infection. N Engl J Med 1977;296:1305-1309. 15. Kelly M. Wciorka LR, McConico S, Peterson LR. Sonicated vascular catheter tip cultures, quantitative association with catheter-related sepsis and the non-utility of an adjuvant cytocentrifuge Gram stain. Am J Clin Pathol 1996;105:210-215. 16. Yagupsky P, Nolte F. Quantitative aspects of septicemia. Clin Microbiol Rev 1990;3:269-279. 17. Kashuba A, Ballow C. Legionella urinary antigen testing: Potential impact on diagnosis and antibiotic therapy. Diagn Microbiol Infect Dis 1996;24:129-139. 18. Tierno PM Jr, Inglima K, Parisi MT. Detection of Bartonella (Rochalinznen) lzenselae bacteremia using bactlalert blood culture system. Am J Clin Pathol 1995:104:530-536. 19. Warren J, Allen S. Clinical. pathogenetic, and laboratory features of Capnocytophaga infections. Am J Clin Pathol 1986;86:513-518. 20. Berkowitz FE. The Gram-positive bacilli: A review of the microbiology, clinical aspects, and antimicrobial susceptibilities of a heterogeneous group of bacteria. Pediatr Infect Dis J 1994;13:1126-1 138. 21. Woods G, Washington J. 1987. Mycobacteria other than mycobacterium tuberculosis: Review of microbiologic and clinical aspects. Rev Infect Dis 1987;9:275. 22. Wilson M, et al. Controlled comparison of the bactec high-blood-volume fungal medium, BACTEC plus 26 aerobic blood culture bottle, and a 10-milliliter isoIator blood culture system for detection of fungemia and bacteremia. J Clin Microbiol 1993;31:865-871. 23. Thomson RB. Laboratory methods in basic virology. In: Baron EJ, Peterson LR, Finegold SM, eds. Baily & Scotts Diagnostic Microbiology, 9th ed. St. Louis, MO: Mosby, 1994, pp. 634-688. 24. Sun T. Current topics in protozoal diseases. Am J Clin Pathol 1994;102:16-29. 25. Homer KS, Wiley EL, Smith A, et al. Monoclonal antibody to Pneumocystis carinii. Am J Clin Pathol 1992;97:619-624. 26. Catchpole C, Hastings JGM. Measuring pre- and post-dose vancomycin levels - time for a change? J Med Microbiol 1995;45:309-311.

27. Eliopoulos GM, Eliopoulos CT. Therapy of enterococcal infections. Eur J Clin Microbiol Infect Dis 1990;9:118-126. 28. Hackbarth C, Chambers H, Sande M. Serum bactericidal activity of rifampin in combination with other antimicrobial agents against Stnphylococcl~s aurezls. Antimicrob Agents Chemother 1986;29:611-613. 29. Peterson L, Shanholtzer C. Tests for bactericidal effects of antimicrobial agents: Technical performance and clinical relevance. Clin Microbiol Rev 1992;5:420-432.

6. Recent advances in the management of fungal infections Jason Sanchez and Gary A. Noskin

1. Introduction The treatment of cancer has undergone many advances in recent years, leading to significant increases in patient survival. Improvements have come not only in the therapies themselves, but also in our ability to sustain critically ill patients. This, in turn, has allowed more intensive cytotoxic chemotherapies and an increased application of bone marrow transplantation for a broader range of neoplasms. More aggressive chemotherapy regimens have led to more profound immunosuppression and increased risk of infection. Longer survival in more critically ill patients, as well as the relative success of antibacterial therapies, have contributed to a dramatic increase in the incidence of invasive fungal disease among oncology patients. Additional risk factors for fungal infection, common in cancer patients, include the interruption of natural protective barriers by indwelling intravascular catheters, extensive surgical procedures, and chemotherapy-induced disruption of mucosal integrity. Opportunistic fungal infections have emerged as among the leading causes of mortality in cancer patients with severe, prolonged immunocompromised states. This trend has been most evident among patients with prolonged episodes of neutropenia, underscoring the primary importance of polymorphonuclear leukocytes in preventing and containing fungal infections. Candidn and Aspergillus have historically represented the most common fungal pathogens affecting cancer patients and continue to be the most frequent offenders. As the number of severely immunocompromised patients continues to grow, however, a greater variety of fungal opportunists are emerging. The last decade has witnessed a startling increase in invasive disease from more obscure fungi or those that were previously considered harmless commensals. Clinical experience with these organisms as pathogens is limited, necessitating increased awareness of their presentation, diagnosis, and management.

Gary A . Noskin (en), M A N A G E M E N T OF INFECTIOUS COMPLICATIONS IN C A N C E R PATIENTS. 0 1998. Klrlwer Academic Prthlisher~,Bosrotz. All rights re.rervecl.

2. Candida 2.1 Epidemiology Candida species represent the largest and most important group of opportunistic fungi in oncology patients. The increased incidence of invasive candidiasis in this setting has been well documented [1,2] and is associated with a high mortality [3,4]. The attributable mortality of hematogenous candidiasis alone has been estimated to be approximately 38%, and this figure increases as organ involvement is confirmed [5]. Candida albicans is the most common; however, other candidal species are beginning to assume a greater significance, especially in cancer patients [2,6]. The advent of molecular typing techniques has revolutionized the understanding of Candidn as a pathogen in the immunocompromised host. The ability to identify not only different species but unique strains within the same species is invaluable to the epidemiology of Candida. It is now clear that a hospitalized patient's endogenous flora is not the only potential source of infection but that organisms can be spread nosocomially from patient to patient by healthcare workers. Several outbreaks of infection with a particular strain of Candida have been documented. One investigation revealed that Candida was quite prevalent (17%) on the hands of intensive care unit workers and that these same strains were responsible for patient infection [7]. These findings underscore the importance of simple hand washing, as well as other more stringent infection control measures, in specific patient populations. Candidn nlbicans is an obligate human parasite that can often be isolated from the gastrointestinal tract, female genital tract, and the skin [2,8]. Candida albicans has been shown to be one of the most virulent species in animal models and has traditionally accounted for one half to two thirds of invasive candidal infections [2,8,9]. Candida species other than albicans comprise the remainder of invasive candidal infections. This group collectively tends to occur more often among patients with neoplastic diseases than with other immunocompromised states [2,10]. A recent review examined 1591 cases of systemic infections caused by Candida in patients with cancer between 1952 and 1992, and found that 46% were caused by the non-albicans Candida species. In this review, C. tropicnlis accounted for 25 %, C. glnbratn S%, C. parapsilosis 7%, and C. krusei 4% of the species identified [2]. Cnndida tropicalis, which colonizes patients less often than C. nlbicans, has been shown to be as virulent and continues to be a major pathogen, especially in cancer patients 121. Infections resulting from C. parapsilosis are also increasing, and they have been associated with both hyperalimentation and prosthetic devices [S]. Despite its relatively low virulence and previously infrequent role as a pathogen, C. krusei has emerged as an organism of increasing concern in the last decade. The widespread use of fluconazole both as a therapeutic and prophylactic agent has been associated

with an increased incidence of C. krusei infections in some medical centers [ll-151. One retrospective study of 463 bone marrow transplant and leukemic patients revealed a sevenfold increase in the incidence of bloodstream or visceral infections caused by C. krusei in the group that received fluconazole prophylaxis compared with those that did not (8.3 % vs. 1.2%) [12]. Many factors contribute to the prevalence of candidal infections in patients undergoing treatment for cancer. Neutropenia, the breakdown of anatomic barriers, use of broad-spectrum antibacterial agents, and fungal colonization are among the most important, while other predisposing factors have also been implicated (Table 1). Polymorphonuclear cells exert their effect by phagocytizing and killing candidal blastospores, and are also capable of damaging pseudohyphae [16-181. Monocytes and eosinophils are also involved in the killing of Candida by phagocytosis. The risk of hematogenous candidiasis that is incurred with the onset of neutropenia increases substantially as the duration exceeds 1 week and is greatest when the absolute neutrophil count falls below 100cells/mL [I]. Such profound neutropenia, previously seen primarily in the treatment of leukemia, is now more common as chemotherapy regimens for solid tumors have become increasingly aggressive [19]. The survival of neutropenic patients who acquire disseminated candidiasis is primarily dependent on the recovery of their granulocyte count, even if prompt antifungal therapy is instituted. The primary importance of polyrnorphonuclear cells is also evidenced by the observation that AIDS patients, despite frequent, severe mucocutaneous candidiasis, rarely develop disseminated infection [20]. These patients often have multiple risk factors for disseminated candidiasis but are protected by their relatively preserved neutrophil numbers and function. Anatomic barriers provided by intact skin and mucosal surfaces are a vital part of protection against invasive fungal disease. The disruption of these barriers is extremely common in the treatment of cancer and plays an important role in the pathogenesis of invasive candidiasis. Intravascular catheterization has been shown to increase the incidence of hematogenous

Table I. Risk Factors for Systemic Fungal Infections -

-

Underlying host defects Neutropenia Defects in humoral or cell-mediated immunity Jmmunosuppression Diabetes mellitus Cytotoxic chemotherapy High dose corticosteroids Long-term antimicrobial therapy Prosthetic devices Vascular or bladder catheters Prolonged hospitalization Solid organ transplantation Severe burns

candidiasis [19,21]. These catheters may serve as a portal of entry for primary infection or become secondarily infected during candidemia from another source [21]. There is mounting evidence that the gastrointestinal tract is the primary route by which Candidn invades patients with cancer. The process first requires colonization of the GI tract, usually by endogenous Candida, but strains acquired nosocomially are increasingly being recognized [7,22]. One prospective study of 139 neutropenic patients with hematologic malignancies documented invasive candidiasis in 22.2% of patients who had been colonized by Cnndida species at multiple sites, 4.0% of patients colonized at a single site, and none of the patients who had not been colonized 1231. The risk of colonization with Candida increases with the duration of hospitalization [2] and is associated with the administration of broad-spectrum antibiotics, corticosteroids, H2 blockers, and antacids [24]. The stomach and esophagus are reported to be the most commonly colonized sections of the GI tract [24]. Human and animal studies have shown that candidemia will occur in normal hosts given a high enough oral inoculum; [25,26] however, most disseminated disease in cancer patients is thought to involve a combination of mucosal damage and neutropenia or neutrophil dysfunction. Disruption of the GI mucosa occurs most commonly as a direct result of cytotoxic chemotherapy, especially cytarabine, resulting in mucositis and ulceration [24]. Other conditions that lead to mucosal damage include surgery, graft-versushost disease, [25] hypotension, [27] and concomitant infection with HSV, CMV, or bacterial pathogens [28]. Once colonization and mucosal damage coincide, local invasion is facilitated and disseminated infection can rapidly occur in the absence of neutrophils. 2.2 Clinical manifestations

In addition to localized mucocutaneous infections, Candida species are capable of a variety of deep tissue infections, including primary and secondary fungemia, meningitis, peritonitis, pyelonephritis, endocarditis, myocarditis, osteomyelitis, arthritis, endophthalmitis, pneumonia, and diseases of the spleen, liver, and gastrointestinal tract. [29,30]. The clinical presentation of disseminated candidiasis in neutropenic cancer patients may vary widely, presenting as an acute illness with the initial fungemic episode or as a more indolent disease, appearing only after granulocyte recovery. These different presentations have been labeled acute disseminated candidiasis (ADC) and chronic disseminated candidiasis (CDC), respectively. ADC is characterized by the onset of sudden illness, fungemia, and occasionally shock with multiorgan failure [31]. Characteristic macronodular skin lesions occur in up to 10% of neutropenic patients with disseminated candidiasis [32]. CDC, previously referred to as hepntosplenic candidinsis, may present over several months as a progressive debilitating illness. The liver, spleen, kidneys, and lungs are the most common sites for the large candidal abscesses that characterize this condition [31,33]. Chronic disseminated candidiasis is seldom asso-

ciated with documented fungemia or hypotension, but may have more subtle findings of upper abdominal tenderness and mild liver enzyme abnormalities [3,32]. Persistent fever, despite broad-spectrum antibiotic coverage, is a common but unreliable finding, especially in patients receiving corticosteroid therapy. Computed tomography has been useful in demonstrating the abdominal lesions of CDC, which are often visible only after granulocyte recovery [33]. It is important to view ADC and CDC as different ends of a clinical spectrum of disease that may present with elements of both or neither. 2.3 Diagnosis The diagnosis of candidiasis in the immunocompromised host relies primarily on the culture of organisms from normally sterile sites. A heightened awareness of the prevalence of Candida as a pathogen in cancer patients and the recognition that cultures may remain negative despite advanced fungal disease have led to the search for a reliable serodiagnostic test or marker. Despite many novel approaches, no method currently exists that has enough sensitivity and specificity to be clinically useful. The isolation of Candida from sites such as urine, feces, sputum, and skin should be viewed in the context of the overall clinical picture because it may represent either infection of colonization. The presence of a positive blood culture, however, is significant and carries the same prognosis as multiple positive blood cultures [34,35]. Candida is rarely a contaminant [36,37], even if the one positive blood culture is obtained from a central venous catheter. Eecciones and colleagues [35] reported that the incidence of disseminated candidal infection was 75% for patients with positive cultures obtained from both a peripheral vein and a catheter, and 68% when obtained from the catheter alone. Furthermore, mortality was higher in the later group (55% vs. 46%). Awareness of the cumulative risk factors for candidal infection, careful attention to commonly affected organ systems, and frequent blood cultures remain central to a diagnostic evaluation. 2.4 Treatment

The treatment of candidiasis depends on several factors, including the location and severity of infection, the underlying host, and hospital resistance patterns. The prominence of nosocomially acquired candidemia and the excess mortality associated with this infection mandate prompt initiation of antifungal therapy. Delays in the treatment of candidemia of even 1 day may have adverse effects on mortality [35]. Amphotericin B has long been the standard of therapy, and it remains the optimal initial treatment for candidemia in patients who are immunocompromised or critically ill. The addition of 5-flucytosine (5-FC) can provide synergistic activity, but its use is complicated by a narrow therapeutic window that requires monitoring of blood levels as well as its toxic effect on the bone marrow [38]. However, the combination

of amphotericin B and 5-FC is the preferred treatment of CNS and intraocular infections. Amphotericin B should be used in cases of organ infection with Candida and in cases of endocarditis, combined with surgical intervention. The most common toxicity of amphotericin B is dose-dependent nephrotoxicity, which has prompted a search for safe and effective alternatives. The advent of the triazoles in the late 1980s provided drugs with superior safety profiles; however, questions of their efficacy in immunocompromised patients have not been completely answered. A recent study that compared the use of amphotericin B (0.5-0.6mglkglday) with fluconazole (400 mglday) in patients with candidemia in the absence of neutropenia revealed no statistically significant difference in their effectiveness [39]. The use of lipid-based amphotericin B has shown much promise in the treatment of candidal infections, with comparable efficacy and greatly reduced renal toxicity [40]. Routine use of these formulations await large, randomized trials to verify their efficacy, but they appear to be reasonable alternatives in patients with worsening renal function who are taking conventional amphotericin B. Ultimately, the decision regarding which antifungal agent to choose should also take into account the prevalence of resistant Candidn species at a particular institution. However, the role of antifungal sensitivity testing requires additional testing to determine its optimal role. The use of indwelling intravascular catheters poses a definite risk of disseminated fungal infection and is a particular problem in cancer patients who require access devices for the administration chemotherapy and blood products [19]. Once candidemia has been established, it is necessary to remove intravascular catheters whenever possible. Failure to do so has been clearly associated with increases in morbidity and mortality [21,41,42]. Despite removal of any intravascular devices, antifungal therapy must accompany management, because catheter removal alone has also been associated with worse outcome [21].

3. Aspergillus Ubiquitous in the environment, Aspergillus grows particularly well on decaying vegetation [43]. This mold reaches the majority of patients through airborne spores that are inhaled into the lower respiratory tract, but it can also invade the nose and paranasal sinuses [44,45] as well as infect patients through interruption of the skin. Pulmonary disease is the most common clinical manifestation of aspergillosis in patients with cancer, and it usually manifests as a rapidly progressive pneumonia with high fevers followed by dense pulmonary consolidation [46,47]. Aspergilloma, or fungus ball of the lung, is another form of pulmonary aspergillosis that occurs in patients with pre-existing lung disease such as sarcoidosis, tuberculosis, bronchiectasis, or previous cavitary lesions. It occurs

almost exclusively in the upper lobes of the lung [43]. Massive hemoptysis remains both a common and fatal complication of pulmonary aspergillosis and occurs most frequently shortly after recovery from granulocytopenia [48,49]. Once disseminated, Aspergill~isspp. can invade any organ and is characterized pathologically by extension along blood vessels, leading to hemorrhagic infarction and necrosis [46]. Aspergilltls fumigatus is the most common cause of aspergillosis. A. flavus, although not as common, is also an important pathogen that tends to involve the nose and paranasal sinuses of severely immunocompromised patients [50,51]. Individuals who are at the highest risk of aspergillosis include those receiving bone marrow or solid organ transplants and those who are neutropenic following cytotoxic chemotherapy [52-541.

3.1 Diagnosis

A definitive diagnosis of invasive aspergillosis requires the demonstration of tissue invasion on biopsy and is facilitated by culture of this tissue. Blood, cerebrospinal fluid, and bone marrow are rarely positive, even in patients with advanced disease [43]. Aggressive pursuit of new pulmonary infiltrates, especially in patients with neutropenia, may lead to early diagnosis. In patients at high risk, isolation of hyphae from pulmonary secretions obtained during bronchoscopy can be highly suggestive of disease, whereas computed tomographic (CT) scans can often identify suspicious patterns not seen on chest x-ray. Radiographically, the "halo" and "air crescent" signs that correspond to a central fungal nodule surrounded by a rim of coagulative necrosis, are relatively specific for aspergillosis in the appropriate clinical setting [55,56]. Computed tomographic or magnetic resonance imaging (MRI) scans of the paranasal sinuses have also proven to be quite useful in the detection of occult rhinocerebral disease. The use of serodiagnostic tests, as well as enzymatic evaluation of bronchoalveolar lavage fluid, has proven too insensitive and nonspecific for accurate diagnosis. The propensity of Aspergillus to invade blood vessels with subsequent tissue necrosis should prompt a search for this organism in patients with black nasal discharge or eschars.

Invasive aspergillosis responds poorly to medical therapy alone and requires large doses of amphotericin B in the range of 1.0-1.5 mglkgld. Because of the need for such high doses and the risk of nephrotoxicity, the use of lipid-based amphotericin B has been investigated. Currently, there are three lipid-based preparations that are approved for use in the United States: amphotericin B lipid complex (Abelcet), amphotericin B cholesteryl sulfate complex (Amphotec), and liposomal amphotericin B (AmBisome). The major advantage of these products is reduced nephrotoxicity compared with conventional amphotericin B. This is particularly beneficial in patients with pre-existing

renal impairment or receiving other concurrent nephrotoxins. Although there is cautious optimism with the lipid-based preparations, their cost, which often exceeds $500 a day, may limit their widespread use. Itraconazole, an oral triazole, has activity against Aspergill~lsand has been used with some success in more indolent cases [57,58]. For rhinocerebral aspergillosis, surgical debridement in addition to medical therapy is necessary to optimize response. Surgical resection has also been advocated in the treatment of localized pulmonary disease during neutropenia; however, the benefits are not as clear and it remains controversial. 4. Mucormycosis

Mucormycosis is the term used to describe disease caused by fungi of the order Mucorales. Rhizopus and Rhizomucor are the two genera most commonly associated with pathogenesis. These fungi exist as molds and are common throughout the environment, gaining access to the majority of patients through the inhalation of airborne spores. Cutaneous infections can also occur by direct inoculation of skin in areas of breakdown. In industrialized nations, mucormycosis occurs primarily in neutropenic cancer patients, diabetics, and victims of trauma. Mucormycosis often involves the paranasal sinuses, and it can extend into surrounding structures to invade the central nervous system [59]. Rhinocerebral invasion is observed most often in leukemia patients with prolonged neutropenia and diabetics in association with acidosis [60-621. Profoundly neutropenic patients are also at increased risk of pulmonary involvement as well as disseminated disease [63,64]. As with aspergillosis, mucormycosis has a proclivity for vascular invasion, leading to hemorrhage, infarction, and necrosis, and can produce the same findings of black nasal discharge or eschar. Diagnosis depends on demonstration of tissue invasion on biopsy and is facilitated by the same imaging studies described earlier for aspergillosis. Effective therapy involves a combination of surgical debridement and high-dose amphotericin B, which is minimally effective alone. Even with optimal treatment, the response in neutropenic patients is relatively poor. The azoles have no activity against these molds and therefore no role in treating mucormycosis. 5. Fungi of emerging importance

There now exists nearly 300 species of fungi that are known to cause human disease and the list is rapidly expanding [65]. Oncology patients are particularly vulnerable to these new pathogens because host defense mechanisms are increasingly compromised by aggressive chemotherapy and radiation. The majority of these fungi are ubiquitous in the environment and consist of both yeasts and molds.

5.1 Yeasts In addition to Candida, emerging yeasts in neutropenic hosts include Trichosporon spp., Blastoschizornyces capitatus (previously Trichosporon cutaneurn), Malassezia spp., Rhodotorula spp., and Hansenula an.ornala. Fungemia or catheter-associated infection is the primary manifestation of disease in many of these new yeast pathogens; however, most are also capable of deep tissue invasion in severely immunocompromised patients. Risk factors for infection with these yeasts are similar to those associated, with candidiasis with neutropenia, previous antibiotic use, indwelling intravascular catheters, and gastrointestinal disruption being among the most important. As with Candida, nosocomial infection from organisms carried on the hands of healthcare workers has also been reported with several of these emerging yeast pathogens [66,67].

5.1.1 Trichosporon. The term trichosporonosis encompasses infection from either Trichosporon species or Blastoschizornyces capitatus. The major predisposing factors for invasive infection with these organisms are severe neutropenia and corticosteroid therapy [65,68]. As with candidiasis, trichosporonosis in the neutropenic patient may result in a wide range of clinical presentations, from acute sepsis to a more indolent process. Acute disseminated infection may yield positive blood or urine cultures and, in addition to fever, manifest with skin lesions, pulmonary infiltrates, or ocular involvement [69,70]. Alternatively, persistent fever in a patient recovering from myelosuppression may be the only clue to a more chronic infection, often characterized by abdominal organ abscesses. The latter presentation is more commonly seen with infection from B. capitatus [65,71]. Of note, the serum latex agglutination test for Cryptococcus neoformans is frequently positive in patients infected with Trichosporon species but not B. capitatus [72]. Microbiologically, B. capitatus can be distinguished from Trichosporon spp. by the production of anelloconidia rather than arthroconidia, both in vitro and in vivo [71]. Despite variable activity against Trichosporon, high-dose amphotericin B remains the first-line therapy for invasive disease, although the response is generally poor in the absence of immune recovery. There have been some reports that fluconazole may be more efficacious than amphotericin B and that the two agents should be used in combination for serious infections 1681. With the frequent use of fluconazole as a prophylactic agent, however, fluconazole-resistant strains of Trichosporon spp. and B. capitatus are increasingly being identified [66]. Both Trichosporon spp. and B. capitatus are resistant to 5-FC as well [73]. 5.1.2 Malassezia, Rhodotorula, and Hansenula. The majority of infections with these emerging yeasts occur as isolated fungemias in cancer patients with indwelling intravascular catheters. Despite their relatively low virulence com-

pared with Candida, each is capable of a variety of deep tissue infections in the compromised host [65,68,73]. Malassezia furf~uand M. pachydernzatis cause fungemia and other deep tissue infections in immunocompromised or debilitated patients, and have been associated with the administration of intralipid infusions [65,68]. Malassezia can be distinguished from Candida and other yeasts by its ovoid shape and its need for lipid-enriched media for growth. Rhodotorula is a common commensal that can be isolated from human skin, sputum, urine, and feces [74]. Rhodotorula, along with C. parapsilosis, were the two most common yeasts isoIated on the hands of healthcare workers in a study that documented this to be a common finding (greater than 70% of personal tested) [67]. Finally, Hansenula is another yeast being increasingly reported as cause of bloodstream infection in both cancer patients and other immunocompromised individuals [73]. Amphotericin B, with or without 5-FC, is generally more effective in treating Malnssezia, Rhodoforula, and Hansenula compared with Trichosporon; however, immune recovery is still critical in cases of advanced infection. In vitro susceptibility testing has indicated that Rhodotor~rlnand Hansenula are generally resistant to fluconazole; however, it remains a viable alternative in the treatment of Malassezia [68,73]. The issue of catheter removal in hematoginous infections with these low-virulence yeasts remains controversial because insufficient experience exists with these organisms. One study demonstrated equal resolution of Rhodotorula fungemia in 23 patients, half of whom received therapy with amphotericin B alone and half in conjunction with catheter removal [74]. However, a retrospective evaluation of fungemia from unusual yeasts in general (including atypical Cnndidn species) indicated that failure to remove an indwelling intravascular catheter was associated with higher mortality [73]. At present, it is recommended that all fungemias in immunocompromised hosts be treated with appropriate antifungal therapy and that indwelling catheters be removed whenever possible.

5.2 Molds Fungal infection due to uncommon molds has increased in the past few decades, especially among bone marrow transplant and leukemic patients who endure prolonged episodes of neutropenia. Acquisition of these organisms is usually via the respiratory route or through breaks in the skin. hyalohyphornycosis refers to infection from molds whose basic tissue form is that of hyphal elements consisting of branched or unbranched hyphae without pigment in their cell walls. Important molds in this group include F~lsari~~rn, Scedosporium, and Acrenzonium species. The dematiaceous fungi produce disease referred to as phaeohyphornycosis and comprise a group of molds with melanin in their cell wall imparting a darkly pigmented color. Bipolaris, Exserohil~~m, Exop hialn, Curvularia, Alternaria, and Phialophora are the most frequent genera to cause human disease in this group [68].

5.2.1 Hyalohyphomycosis Fusarium spp. F~isariumspp. are common plant pathogens that are ubiquitous in the soil. In humans they had long been associated with superficial infections such as onychomycosis, keratomycosis, and colonization of wounds and burns. Disseminated fusariosis was first documented in the early 1970s in a patient with acute leukemia and has since been recognized as an increasingly common is now pathogen in severely immunocompromised individuals [75]. F~lsari~lm the second most common pathogenic mold (following Aspergill~ls)in patients receiving cytotoxic chemotherapy and BMT for hematologic malignancies [76,77]. Severe neutropenia is the greatest risk factor for disseminated infection, which occurs most often in patients with hematologic malignancies [77,78].In most cases, the portal of entry is unknown; however, the respiratory tract and disruption of the skin have been the most often reported. Disseminated F~isarium infection is clinically characterized by fever, myalgias, and a high frequency of both skin lesions and positive blood cultures. The skin lesions, which occur in 60-80% of individuals, are usually multiple and either resemble ecthyma gangrenosum or form subcutaneous nodules [65,68,75]. F~isariunz,like Aspergillus, is vasotropic; however, the frequency of im (-60%) far exceeds that of positive blood cultures in F ~ ~ s n r i ~infections Aspergillus (38.2"C) and neutropenia (neutrophil count ,JS . c/5 D. c 3 g CJ O O C S V

203

c "5 O

N'3

S " E § S S^ o

aj

0 ^

>^ C

i2 E o 3 O -7

O

^

"J

id

> S

tC ao

o —