Case 10-2004 — A 58-Year-Old Man with Acute Myeloid Leukemia and Fever after Chemotherapy
http://www.100md.com
《新英格兰医药杂志》
Presentation of Case
Dr. Bimalangshu R. Dey: A 58-year-old man with a large sacral ulcer was admitted to the hospital for chemotherapy to treat acute myeloid leukemia.
The patient had had recurrent infections of a pilonidal cyst for 20 years; the infections had required surgical débridement on multiple occasions. Six weeks before admission to this hospital, he was admitted to another hospital with a temperature as high as 39.4°C with chills and rigors. The pilonidal cyst had a malodorous, purulent discharge and was surrounded by fluctuant, swollen tissue. Surgical incision and drainage were performed. The white-cell count was 7500 per cubic millimeter, with 6 percent circulating blasts. The patient was treated with intravenous antibiotics, and the wound developed some granulation. A bone marrow biopsy and aspiration were performed, and the aspirate was reported to show a hypercellular marrow with 15 percent myeloblasts and multilineage dysplasia — features suggestive of a myelodysplastic syndrome with excess blasts. The patient's fever resolved, and he was discharged after two weeks.
Eighteen days before admission to this hospital, he was seen as an outpatient at this hospital's Cancer Center for consultation regarding the diagnosis and treatment of his myelodysplastic syndrome. He was asymptomatic and afebrile. Laboratory tests were repeated (Table 1 and Table 2), and a repeated bone marrow biopsy was scheduled. Four days later, chills, fever, and night sweats developed in association with a recurrent malodorous discharge from the presacral wound, with mild local pain, and the patient was admitted the next day.
Table 1. Hematologic Laboratory Data.
Table 2. Blood Chemical Values.
The temperature was 38.3°C. There was an open wound, 10.0 by 7.5 cm, that was over the sacrum; the wound was 4.0 cm deep, and there was granulation tissue at its base and surrounding erythema. The results of laboratory tests are shown in Table 1 and Table 2. Vancomycin, levofloxacin, and metronidazole were administered intravenously. The peripheral-blood smear contained 19 percent blasts and promonocytes (Table 1 and Figure 1A). The bone marrow aspirate and biopsy specimen had features characteristic of acute myeloid leukemia (Figure 1B). A magnetic resonance imaging (MRI) study of the pelvis showed no evidence of osteomyelitis or a rectosacral fistula. The fever resolved. Because of the infected ulcer, the decision was made to defer induction chemotherapy for the acute leukemia, and he was discharged on the seventh hospital day, with instructions to take oral antibiotics and sitz baths and to use topical papain ointment.
Figure 1. Specimens of Peripheral Blood and Bone Marrow Obtained on the First Day of the First Admission.
A smear of peripheral blood (Panel A) shows an immature monocytoid cell with a folded nucleus, multiple prominent nucleoli, and moderately abundant pale cytoplasm (left) and a blast containing an Auer rod (right) (Wright's stain, x1200). The bone marrow–biopsy specimen contains large blast cells with folded nuclei (Panel B; hematoxylin and eosin, x500). Megakaryocytic dysplasia, identified by the small cell size and simplified nuclei with separated nuclear lobes, is evident (arrows).
Six days after he was discharged, the patient's white-cell count had risen to 88,000 per cubic millimeter, with 46 percent blasts. The next day he was readmitted for chemotherapy to treat his rapidly evolving acute leukemia.
The patient was married and worked as a manager. His wife and children were well, and he had had no infectious exposures. There was no family history of leukemia. He had smoked one pack of cigarettes daily until one month before admission, and he drank alcohol occasionally. He had no other medical problems, and he stated that he had not recently had headache, nausea, vomiting, chest pain, shortness of breath, altered bowel habits, loss of appetite, or weight loss.
On physical examination, the patient was afebrile. The blood pressure was 126/70 mm Hg, the pulse 84 beats per minute, and the respiratory rate 16 breaths per minute. Examination of the skin, chest, abdomen, arms, and legs disclosed no abnormalities. There was an ulcerated area, 6.0 by 4.0 cm, over the sacrum; it was 3.0 cm deep, with granulation tissue at the base. The edges were clean and tender, with slight erythema and induration in the surrounding area; there was no fluctuance or purulent discharge. The results of laboratory tests performed on the day of the patient's readmission to this hospital are shown in Table 1 and Table 2.
Pathological Discussion
Dr. Robert P. Hasserjian: The peripheral-blood smear obtained at the time of the first admission to this hospital contained immature monocytoid cells and a few blasts with Auer rods (Figure 1A). The specimen obtained from the bone marrow biopsy was markedly hypercellular and contained predominantly large cells with folded nuclei and moderate amounts of pale cytoplasm, features characteristic of monoblasts (Figure 1B). Blasts constituted 45 percent of the cells in the bone marrow aspirate. Dysplasia was present in all three cell lines, as shown by hypogranular neutrophils, small megakaryocytes with hypolobulated nuclei, and nuclear irregularity in the few erythroid elements identified. Immunophenotyping of a bone marrow aspirate by flow cytometry revealed CD45+, dimly CD34+, HLA-DR+ blasts expressing the myeloid markers CD33, myeloperoxidase, and CD117 and the monocytic markers CD64 and CD14. Cytogenetic analysis revealed a normal karyotype (46,XY). These findings confirmed the diagnosis of acute myeloid leukemia.
According to the current World Health Organization classification of myeloid neoplasms, this case is classified as acute myeloid leukemia with multilineage dysplasia because of dysplastic features in more than one cell line.1 This type of acute leukemia may arise as a primary disorder or may evolve from a preexisting myelodysplastic syndrome. Although the bone marrow specimen from the other hospital was not reviewed here, the fact that it showed multilineage dysplasia with fewer than 20 percent myeloblasts supports the interpretation that this is an acute leukemia evolving from a myelodysplastic syndrome.
Pathological Diagnosis
Acute myeloid leukemia with multilineage dysplasia after a myelodysplastic syndrome.
Presentation of Case (Continued)
Dr. Dey: Remission-induction chemotherapy for acute myeloid leukemia was begun on the second hospital day with idarubicin at a dose of 18 mg per square meter of body-surface area given intravenously daily for three days and cytarabine at a dose of 200 mg per square meter by continuous infusion daily for seven days. Oral levofloxacin, acyclovir, and fluconazole were given as prophylaxis. On the fifth hospital day, four days after the initiation of chemotherapy, fever to 38.3°C developed, and oral mucosal lesions were noted. The levofloxacin was discontinued, and vancomycin and ceftazidime were administered, but on the sixth day, the patient's temperature rose to 40.0°C. The results of stool cultures and assays were negative for Clostridium difficile. On the eighth hospital day, the temperature again rose to 40.0°C, and diarrhea developed. A renal ultrasound image showed no obstruction. On the ninth day, the patient became hypotensive, with a systolic blood pressure of 70 mm Hg and a decline in urine output. Methylprednisolone and dopamine were administered.
He was transferred to the medical intensive care unit. Voriconazole was substituted for the fluconazole. The next day the trachea was intubated because of hypoxemia and progressive acidosis; the arterial pH was 7.23, the arterial carbon dioxide tension was 51 mm Hg, and the arterial oxygen tension was 67 mm Hg. The presacral wound appeared unchanged. Multiple blood, urine, and stool cultures were sterile, and examination of urine and stool specimens showed no abnormalities. Computed tomographic (CT) scanning of the chest without the use of contrast material showed bilateral perihilar air-space disease, with scattered peripheral nodular opacities in both lungs, particularly in the upper lobes — findings most consistent with an infectious process. CT scanning of the abdomen and pelvis revealed no abscesses. Granulocyte colony-stimulating factor and metronidazole were administered.
Over the next three days, the patient was febrile and hypotensive, his creatinine level continued to rise, and his urine output continued to decrease. Mechanical ventilation was continued, and norepinephrine was administered. Bronchoscopy with bronchoalveolar lavage showed no evidence of infection; Gram's stains, fungal smears, and cultures of bronchoalveolar lavage specimens revealed no abnormalities.
On the 14th day, the temperature rose to 38.8°C. The norepinephrine was discontinued. The next day the patient was unresponsive, despite the discontinuation of treatment with sedative agents for 12 hours. On the 16th day, anuria with increasing fluid retention developed, and hemodialysis was initiated. The temperature rose to 38.3°C. He continued to be unresponsive. CT scanning of the head showed no abnormalities. Examination of a bone marrow specimen on the 14th day after chemotherapy was begun showed hypocellularity, without evidence of residual leukemia. On the 18th hospital day, the patient was more alert, but the temperature rose to 39.4°C. A CT scan of the chest disclosed improvement in the air-space disease in both lungs, except that a focus in the left upper lobe had become more consolidated. Quinupristin–dalfopristin was added to the patient's medications on the 19th day because vancomycin-resistant Enterococcus faecalis was detected in cultures of the sacral ulcer and on a rectal swab. A sputum culture grew vancomycin-sensitive enterococcus. Cultures of the blood and urine were sterile. On the 20th day, continuous venovenous hemodialysis was begun. Between the 20th and 26th days, despite a gradual rise in absolute neutrophil count to normal by the 25th day (Table 1), a fever with temperatures ranging from 38.3 to 39.4°C persisted, and the patient's trachea remained intubated; dialysis and treatment with vasoactive drugs were necessary to maintain an adequate systemic arterial pressure.
Differential Diagnosis
Dr. Dey: As the hematologist treating this patient, I thought it likely that the patient's clinical deterioration after the fifth hospital day was due to multiple factors. The diarrhea was probably due to the antimicrobial therapy and the treatment with cytarabine. The high fevers and hypotension could have indicated infection, tumor lysis with cytokine release, or both, although drug-induced fever caused by the cytarabine and mucositis with bacterial superinfection could not be ruled out. The inability to obtain diagnostic microbiologic cultures could have been due to the use of broad-spectrum antimicrobial agents, or it could have indicated that the fever had a noninfectious cause. The acute renal failure could have been the result of the tumor lysis syndrome or severe hypotension and subsequent acute tubular necrosis. The patient's stuporous mental state was probably mainly metabolic in origin and may have included components of uremic encephalopathy, hepatic encephalopathy, and both sedation and fever. Hypotension, infection, hyperalimentation, or drugs may have caused hepatocellular injury, which resulted in elevated values on liver-function testing.
Because of the persistence of fever despite the hematologic recovery during the fourth week of hospitalization, I was concerned that an undiagnosed infection was present and requested a consultation with the Infectious Disease Division.
Discussion of Management
Dr. Jay A. Fishman: Infection in immunocompromised persons is often rapidly progressive and life threatening. Thus, the appropriate design of initial empirical antimicrobial regimens for such patients is critically important. Although progress has been made in the prevention and treatment of many infections common to such hosts, the case under discussion poses two common problems for the clinician. First, if fever in a patient receiving broad-spectrum antimicrobial agents is due to infection, which organisms are likely to be involved? Second, which antimicrobial agents should be used? If the spectrum of coverage is too narrow, the patient may die of sepsis, but at the same time the negative consequences of excess antimicrobials (e.g., resistance, toxic effects, drug interactions, and allergic reactions) are not trivial.
Guidelines for Treating Infections in Immunocompromised Patients
Progress in the care of patients with neutropenia is reflected in a series of excellent guidelines developed for categories of infection that are common to this group, particularly patients with fever and neutropenia.2,3,4,5,6,7,8,9 These guidelines provide information about the pathogens most often associated with infections in each patient group, as well as approaches to the selection of antimicrobial agents. Such guidelines are very useful as a starting point for therapy.
Unfortunately, the application of general guidelines to specific patients is often difficult. A patient may have allergies or other conditions, such as impairment of renal or hepatic function, that limit the use of certain antimicrobial agents. The epidemiologic features of infections vary according to institution and practice for prescribing antimicrobial agents. Guidelines become outdated as new diagnostic tools and therapies become available. In addition, opinion is not uniform with regard to the optimal use of antimicrobial agents in febrile, immunocompromised patients. The evaluation and comparison of data from clinical trials are often complicated by variations in study design and end-point analysis. The best use of vancomycin, aminoglycosides, antifungal therapies, immune globulins, and other agents remains uncertain, as does their timing.
Finally, many fevers in immunocompromised patients are noninfectious in origin; processes that include pulmonary emboli, graft-versus-host disease, drug toxicity, or engraftment syndromes may cause fever. Even in patients with clinical evidence of infection, a specific pathogen is not identified in more than half the cases. Infections due to uncommon viruses (e.g., human herpesvirus 6, parvovirus B19, and metapneumovirus) or to organisms that are difficult to isolate in microbiologic culture (i.e., slow-growing organisms, those requiring special mediums, or cell-wall-deficient bacteria) may go unrecognized. Patients often have multiple, simultaneous conditions, such as infections, adverse drug reactions, and coexisting medical conditions that influence both the susceptibility to infection and the ability of physicians to establish a diagnosis.
There are few guidelines for the treatment of patients, such as this one, who have fever that develops or persists after recovery from neutropenia.3,10 Some localized infectious processes, such as pneumonia, perineal abscess, and fungal infections, may become clinically apparent only after leukocytes return to mount an inflammatory response. The degree of residual impairment of immunity in such patients is unknown, although they are generally less prone to rapidly progressive sepsis than they are when neutropenia is present. Unfortunately, it is often impossible to obtain a microbiologic diagnosis on the basis of cultures obtained from patients receiving broad-spectrum antimicrobial agents.
General Principles
Four principles may be used to guide the selection of antimicrobial agents. First, the treatment of any documented infection must be completed and resolution of the infection documented. Second, after the resolution of neutropenia, physical examination and radiologic studies often reveal sites of infection that were previously unrecognized. Samples should be taken from these sites for histologic or microbiologic evaluation. Third, additional antimicrobial agents should be selected on the basis of the patient's medical history, exposure history, and aspects specific to his or her clinical condition. Fourth, the physician needs to remember that the fever may not be of infectious origin; other causes of fever, such as drugs, may need to be considered. The strategy must be flexible enough to be modified as new data emerge and clinical conditions evolve.
In immunocompromised patients, there is a limited window of opportunity during which antimicrobial therapy can effectively prevent death from infection. This observation has two implications: most patients will receive empirical therapy (at least until microbiologic data are available), and procedures that define the nature and extent of infection are essential. These procedures include radiographic examinations (particularly MRI and CT scans), biopsies to obtain specimens for histologic examination and culture, and other microbiologic assays (based on detection of antigens or specific nucleic acids).11 The use of aggressive diagnostic approaches is especially important because of the relative absence of physical signs or symptoms of infection in many immunocompromised patients. Infection is often advanced or disseminated by the time of diagnosis. Thus, the suspicion that infection is present is critical. Minor abnormalities may be signs of impending sepsis or of tissue-invasive infection (Table 3).
Table 3. Common Signs of Unrecognized Infection in Immunocompromised Patients.
Risk of Infection
For each patient, I ask, "How great is this person's risk of infection? And for which infections are the risks greatest?" For any patient, this risk is determined by the relationship between his or her epidemiologic exposures and a measure of his or her susceptibility to infection, which is termed the net state of immune suppression. The net state of immune suppression is a function of all the factors contributing to the person's risk of infection. Most patients have several types of immune defects.
Pathogens that may be present (so-called epidemiologic exposures) may be intrinsic or extrinsic, and they may have been acquired in the hospital or community; moreover, exposures may be recent or distant in time. The potential pathogens vary according to the nature of the patient's immune deficits. An organism is likely to become invasive or to "set up shop" if the host lacks functional components of the immune system that are needed to control or kill that pathogen. Thus, extracellular bacteria are common pathogens in the absence of neutrophils or other phagocytes, and the majority of infections in patients with neutropenia are derived from endogenous flora.12 In contrast, viruses have greater importance as a cause of opportunistic infection in the absence of cell-mediated cytotoxic (T-cell) responses.
The risk of infection is also determined by the burden of organisms (i.e., the intensity of exposure), the native virulence of the organism, and the nature of the exposure (e.g., whether it is gastrointestinal, is pulmonary, originates at a catheter, or arises from old granulomas). In a person who is not immunocompromised (Table 4), the number of organisms at various sites varies greatly. Thus, in patients with neutropenia, the frequency of infection originating in the gastrointestinal tract may reflect the burden of organisms in the gastrointestinal tract as well as the frequency of injury to the gastrointestinal mucosa. Shifts in patterns of colonization are commonly associated with hospitalization and with the use of antimicrobial agents. Each episode of neutropenia, chemotherapy, infection, catheter insertion, skin breakdown, mucositis, or colitis increases the likelihood that infection or colonization of tissues will persist and that colonizing organisms will be resistant to various antimicrobial agents. Access to information about the patient's prior infections and exposures to antimicrobial agents is essential.
Table 4. Approximate Concentration of Bacteria in Tissues from the Normal Human Host.
Among immunocompromised patients, over time exposure of organisms to selection pressure from antimicrobial agents has shifted the distribution of common bacterial flora from predominantly gram-positive bacteria (e.g., staphylococci) to gram-negative organisms (e.g., Escherichia coli) to resistant gram-negative organisms (e.g., pseudomonas species) and back to resistant gram-positive organisms (e.g., vancomycin-resistant enterococci, fluoroquinolone-resistant streptococci). With prolonged neutropenia, along with improved treatment of bacterial infections and the selection pressure of prophylaxis, resistance to azole antifungal agents has emerged. Resistance of cytomegalovirus to ganciclovir has been observed. As a result, sicker patients tend to become progressively more difficult to treat (Table 5).
Table 5. Factors Contributing to Colonization with Nosocomial Flora.
Care of This Patient
Discussions about this patient between myself and Dr. Dey began by telephone on the 26th hospital day, as follows: "I have a patient with acute myeloid leukemia who has received chemotherapy with prolonged neutropenia, which is resolving. He now has fever, is critically ill, and has some unique risks for infection." These features of the case, including the large skin wound in the sacral area and the lack of a clinical response to multiple antimicrobial agents, became the essential clues to management for this person (Table 6). The goal of consultation was not to stop the use of various antimicrobial agents (although a laudatory goal), nor was it to treat persistent fever, but rather, the goal was to determine whether infection was present and to select the optimal therapy on the basis of pathogens known to be or likely to be present.
Table 6. Clinical Features of This Patient.
On physical examination, the patient appeared to be critically ill; his trachea was intubated, he was sedated, and he was receiving continuous venovenous hemodialysis in the intensive care unit. His temperature was 38.3°C, his blood pressure 80/60 mm Hg while he was receiving pressor agents, his heart rate 82 beats per minute, and his respiratory rate 14 breaths per minute with mechanical ventilation. There was a deep sacral ulcer, 5 by 8 cm. There was granulation tissue and no exudate, and no bone was exposed. The remainder of the findings on examination were normal. Radiographic tests (CT and MRI) revealed no deep infection in the abdomen, chest, or perirectal area. His white-cell count had climbed to 4600 per cubic millimeter, with an absolute neutrophil count of 3174 per cubic millimeter while he was receiving granulocyte colony-stimulating factor.
Potential pathogens in this patient fell into two groups: documented infectious organisms and likely pathogens. Multiple blood cultures had been sterile; a single isolate of coagulase-negative staphylococcus from blood drawn from a Hickman catheter was considered a contaminant. A sputum culture had grown a few yeast (Candida albicans) and moderate numbers of vancomycin-susceptible E. faecalis. A Gram's-stained sputum specimen contained few neutrophils and no organisms. Cultures of both a rectal swab and the pilonidal-cyst ulcer grew E. faecalis (resistant to vancomycin, doxycycline, erythromycin, and tetracycline and sensitive to ampicillin, quinupristin–dalfopristin, chloramphenicol, linezolid, and rifampin).
There was no clinical response to the addition of multiple antimicrobial agents; in fact, the patient's fever was higher. No focus of infection had become manifest with the return of neutrophils. We suspected that if he had sepsis, it was derived from his deep skin ulcer, although this area did not appear to be grossly infected. It also seemed possible that some of the antimicrobial agents might have been contributing to his fever and hypotension. The period of greatest risk (three weeks of absolute neutropenia) had passed. He had completed courses of guideline-specified antimicrobial therapy during the weeks he had neutropenia and had received, in addition, a variety of broad-spectrum antimicrobial agents. Therefore, I thought that his antibiotic coverage could be safely narrowed. I suggested that we treat the obvious skin ulcer with intravenous ampicillin (for susceptible enterococcus from the wound and rectum) and metronidazole (for presumed anaerobic organisms from the same areas) and that we eliminate the other drugs.
Dr. Dey: The granulocyte colony-stimulating factor was discontinued on the 27th day. The quinupristin–dalfopristin, ceftazidime, and acyclovir were discontinued on the 28th day. Within 48 hours, the patient became afebrile, became fully awake, and no longer required pressor agents or ventilatory support. Abnormalities in the results of liver-function tests improved markedly after discontinuation of the quinupristin–dalfopristin. Cultures remained sterile.
A repeated bone marrow biopsy was performed the day before discharge and showed hypercellularity with trilineage hematopoiesis, without evidence of leukemia. The patient gradually recovered and was discharged after 44 days of hospitalization. He is currently undergoing consolidation chemotherapy as an outpatient.
Dr. Philip Amrein (Hematology and Oncology): Is it possible that some patients who are receiving the appropriate antibiotics and remain febrile even while their cultures remain sterile do indeed have sepsis? Is it possible that the cultures do not grow pathogens because of the presence of the antibiotics?
Dr. Fishman: In some cases, microbiologic cultures are sterilized by antimicrobial agents. This is particularly true of patients with low-grade bacteremic infections and those with loculated infections, such as biliary obstruction, small-bowel perforations, or other small abscesses, as was suspected in this patient.
Dr. Nancy Lee Harris (Pathology): In this patient, do you think the fever was caused by drugs or by infection?
Dr. Fishman: I think that drugs clearly had a role in causing his fever. When we stopped the quinupristin–dalfopristin and voriconazole, the patient's condition improved rapidly without the addition of new antimicrobial agents. The chemotherapeutic agents may also have contributed to the fever. However, it is equally likely that he had intermittent bacteremia from his deep pilonidal cyst. So I suspect that he had a combination of infections and adverse drug reactions.
Dr. Howard Weinstein (Pediatric Hematology and Oncology): Is there any role for granulocyte transfusions in patients with neutropenia? And what is the role of prophylactic oral antibiotics in patients who have neutropenia but no fever? Most of the algorithms for empirical antibiotic therapy in pediatric oncology are for patients with neutropenia and fever; we do not give antibiotics to afebrile patients.
Dr. Fishman: Granulocyte transfusions remain controversial, with advocates and detractors in equal numbers. Transfusion reactions are common, and in many patients the granulocyte counts increase only slightly after transfusion. There is a substantial risk of transmission of cytomegalovirus and other viruses. However, some controlled clinical trials have shown a benefit in patients with documented infections. Many of us would try granulocyte transfusions if other approaches failed. Clinical trials are needed in this area.
The routine use of antibiotics in neutropenia in the absence of fever should be limited to specific high-risk patients: those likely to have prolonged and severe neutropenia, those who have had prior episodes of fever or infection during neutropenia, and those who have a likely source of infection, such as this patient who had an open ulcer. In patients with a history of herpes simplex virus or varicella–zoster infections or of thrush, prophylaxis is also worthwhile. Those receiving high-dose corticosteroids may benefit from trimethoprim–sulfamethoxazole prophylaxis against pneumocystis infection. As much as possible, I prefer to individualize prophylaxis for each patient, with risk stratification for specific infections.
Dr. Fishman reports serving as a consultant to Gilead, Roche, and Imerge BT; receiving research support from Fujisawa Healthcare; receiving educational program support from Fujisawa Healthcare and Roche; and serving on the speakers' bureaus of Pfizer, Roche, Fujisawa Healthcare, and Aventis.
Source Information
From the Transplant Infectious Disease and Compromised Host Program, Infectious Disease Division (J.A.F.), the Bone Marrow Transplant Unit, Hematology Oncology Division (B.R.D.), and the Department of Pathology (R.P.H.), Massachusetts General Hospital; and the Departments of Medicine (J.A.F., B.R.D.), and Pathology (R.P.H.), Harvard Medical School.
References
Brunning R, Matutes E, Harris NL, Flandrin G. Acute myeloid leukemia with multilineage dysplasia. In: Jaffe ES, Harris NL, Stein H, Vardiman J, eds. Pathology and genetics of tumours of haematopoietic and lymphoid tissues. Vol. 3 of World Health Organization classification of tumours. Lyons, France: IARC Press, 2001:88-9.
Hughes WT, Armstrong D, Bodey GP, et al. 2002 Guidelines for the use of antimicrobial agents in neutropenic patients with cancer. Clin Infect Dis 2002;34:730-751.
Pizzo PA, Robichaud KJ, Gill FA, Witebsky FG. Empiric antibiotic and antifungal therapy for cancer patients with prolonged fever and granulocytopenia. Am J Med 1982;72:101-111.
Schimpff SC. Empiric antibiotic therapy for granulocytopenic cancer patients. Am J Med 1986;80:13-20.
Klastersky J, Zinner SH, Calandra T, et al. Empiric antimicrobial therapy for febrile granulocytopenic cancer patients: lessons from four EORTC trials. Eur J Cancer Clin Oncol 1988;24:Suppl 1:S35-S45.
Viscoli C. The evolution of the empirical management of fever and neutropenia in cancer patients. J Antimicrob Chemother 1998;41:Suppl D:65-80.
Rolston KV, Mullen CA, Wade JC. NCCN: fever and neutropenia. Support Care Cancer 2002;10:58-64.
European Organization for Research and Treatment of Cancer (EORTC) International Antimicrobial Therapy Cooperative Group, National Cancer Institute of Canada-Clinical Trials Group. Vancomycin added to empirical combination antibiotic therapy for fever in granulocytopenic cancer patients. J Infect Dis 1991;163:951-958.
National Comprehensive Cancer Network home page. (Accessed February 5, 2004, at http://www.nccn.org.)
Barton TD, Schuster MG. The cause of fever following resolution of neutropenia in patients with acute leukemia. Clin Infect Dis 1996;22:1064-1068.
Donowitz GR, Harman C, Pope T, Stewart FM. The role of the chest roentgenogram in febrile neutropenic patients. Arch Intern Med 1991;151:701-704.
Bodey GP, Buckley M, Sathe YS, Freireich EJ. Quantitative relationships between circulating leukocytes and infection in patients with acute leukemia. Ann Intern Med 1966;64:328-340.(Jay A. Fishman, M.D., Bim)
Dr. Bimalangshu R. Dey: A 58-year-old man with a large sacral ulcer was admitted to the hospital for chemotherapy to treat acute myeloid leukemia.
The patient had had recurrent infections of a pilonidal cyst for 20 years; the infections had required surgical débridement on multiple occasions. Six weeks before admission to this hospital, he was admitted to another hospital with a temperature as high as 39.4°C with chills and rigors. The pilonidal cyst had a malodorous, purulent discharge and was surrounded by fluctuant, swollen tissue. Surgical incision and drainage were performed. The white-cell count was 7500 per cubic millimeter, with 6 percent circulating blasts. The patient was treated with intravenous antibiotics, and the wound developed some granulation. A bone marrow biopsy and aspiration were performed, and the aspirate was reported to show a hypercellular marrow with 15 percent myeloblasts and multilineage dysplasia — features suggestive of a myelodysplastic syndrome with excess blasts. The patient's fever resolved, and he was discharged after two weeks.
Eighteen days before admission to this hospital, he was seen as an outpatient at this hospital's Cancer Center for consultation regarding the diagnosis and treatment of his myelodysplastic syndrome. He was asymptomatic and afebrile. Laboratory tests were repeated (Table 1 and Table 2), and a repeated bone marrow biopsy was scheduled. Four days later, chills, fever, and night sweats developed in association with a recurrent malodorous discharge from the presacral wound, with mild local pain, and the patient was admitted the next day.
Table 1. Hematologic Laboratory Data.
Table 2. Blood Chemical Values.
The temperature was 38.3°C. There was an open wound, 10.0 by 7.5 cm, that was over the sacrum; the wound was 4.0 cm deep, and there was granulation tissue at its base and surrounding erythema. The results of laboratory tests are shown in Table 1 and Table 2. Vancomycin, levofloxacin, and metronidazole were administered intravenously. The peripheral-blood smear contained 19 percent blasts and promonocytes (Table 1 and Figure 1A). The bone marrow aspirate and biopsy specimen had features characteristic of acute myeloid leukemia (Figure 1B). A magnetic resonance imaging (MRI) study of the pelvis showed no evidence of osteomyelitis or a rectosacral fistula. The fever resolved. Because of the infected ulcer, the decision was made to defer induction chemotherapy for the acute leukemia, and he was discharged on the seventh hospital day, with instructions to take oral antibiotics and sitz baths and to use topical papain ointment.
Figure 1. Specimens of Peripheral Blood and Bone Marrow Obtained on the First Day of the First Admission.
A smear of peripheral blood (Panel A) shows an immature monocytoid cell with a folded nucleus, multiple prominent nucleoli, and moderately abundant pale cytoplasm (left) and a blast containing an Auer rod (right) (Wright's stain, x1200). The bone marrow–biopsy specimen contains large blast cells with folded nuclei (Panel B; hematoxylin and eosin, x500). Megakaryocytic dysplasia, identified by the small cell size and simplified nuclei with separated nuclear lobes, is evident (arrows).
Six days after he was discharged, the patient's white-cell count had risen to 88,000 per cubic millimeter, with 46 percent blasts. The next day he was readmitted for chemotherapy to treat his rapidly evolving acute leukemia.
The patient was married and worked as a manager. His wife and children were well, and he had had no infectious exposures. There was no family history of leukemia. He had smoked one pack of cigarettes daily until one month before admission, and he drank alcohol occasionally. He had no other medical problems, and he stated that he had not recently had headache, nausea, vomiting, chest pain, shortness of breath, altered bowel habits, loss of appetite, or weight loss.
On physical examination, the patient was afebrile. The blood pressure was 126/70 mm Hg, the pulse 84 beats per minute, and the respiratory rate 16 breaths per minute. Examination of the skin, chest, abdomen, arms, and legs disclosed no abnormalities. There was an ulcerated area, 6.0 by 4.0 cm, over the sacrum; it was 3.0 cm deep, with granulation tissue at the base. The edges were clean and tender, with slight erythema and induration in the surrounding area; there was no fluctuance or purulent discharge. The results of laboratory tests performed on the day of the patient's readmission to this hospital are shown in Table 1 and Table 2.
Pathological Discussion
Dr. Robert P. Hasserjian: The peripheral-blood smear obtained at the time of the first admission to this hospital contained immature monocytoid cells and a few blasts with Auer rods (Figure 1A). The specimen obtained from the bone marrow biopsy was markedly hypercellular and contained predominantly large cells with folded nuclei and moderate amounts of pale cytoplasm, features characteristic of monoblasts (Figure 1B). Blasts constituted 45 percent of the cells in the bone marrow aspirate. Dysplasia was present in all three cell lines, as shown by hypogranular neutrophils, small megakaryocytes with hypolobulated nuclei, and nuclear irregularity in the few erythroid elements identified. Immunophenotyping of a bone marrow aspirate by flow cytometry revealed CD45+, dimly CD34+, HLA-DR+ blasts expressing the myeloid markers CD33, myeloperoxidase, and CD117 and the monocytic markers CD64 and CD14. Cytogenetic analysis revealed a normal karyotype (46,XY). These findings confirmed the diagnosis of acute myeloid leukemia.
According to the current World Health Organization classification of myeloid neoplasms, this case is classified as acute myeloid leukemia with multilineage dysplasia because of dysplastic features in more than one cell line.1 This type of acute leukemia may arise as a primary disorder or may evolve from a preexisting myelodysplastic syndrome. Although the bone marrow specimen from the other hospital was not reviewed here, the fact that it showed multilineage dysplasia with fewer than 20 percent myeloblasts supports the interpretation that this is an acute leukemia evolving from a myelodysplastic syndrome.
Pathological Diagnosis
Acute myeloid leukemia with multilineage dysplasia after a myelodysplastic syndrome.
Presentation of Case (Continued)
Dr. Dey: Remission-induction chemotherapy for acute myeloid leukemia was begun on the second hospital day with idarubicin at a dose of 18 mg per square meter of body-surface area given intravenously daily for three days and cytarabine at a dose of 200 mg per square meter by continuous infusion daily for seven days. Oral levofloxacin, acyclovir, and fluconazole were given as prophylaxis. On the fifth hospital day, four days after the initiation of chemotherapy, fever to 38.3°C developed, and oral mucosal lesions were noted. The levofloxacin was discontinued, and vancomycin and ceftazidime were administered, but on the sixth day, the patient's temperature rose to 40.0°C. The results of stool cultures and assays were negative for Clostridium difficile. On the eighth hospital day, the temperature again rose to 40.0°C, and diarrhea developed. A renal ultrasound image showed no obstruction. On the ninth day, the patient became hypotensive, with a systolic blood pressure of 70 mm Hg and a decline in urine output. Methylprednisolone and dopamine were administered.
He was transferred to the medical intensive care unit. Voriconazole was substituted for the fluconazole. The next day the trachea was intubated because of hypoxemia and progressive acidosis; the arterial pH was 7.23, the arterial carbon dioxide tension was 51 mm Hg, and the arterial oxygen tension was 67 mm Hg. The presacral wound appeared unchanged. Multiple blood, urine, and stool cultures were sterile, and examination of urine and stool specimens showed no abnormalities. Computed tomographic (CT) scanning of the chest without the use of contrast material showed bilateral perihilar air-space disease, with scattered peripheral nodular opacities in both lungs, particularly in the upper lobes — findings most consistent with an infectious process. CT scanning of the abdomen and pelvis revealed no abscesses. Granulocyte colony-stimulating factor and metronidazole were administered.
Over the next three days, the patient was febrile and hypotensive, his creatinine level continued to rise, and his urine output continued to decrease. Mechanical ventilation was continued, and norepinephrine was administered. Bronchoscopy with bronchoalveolar lavage showed no evidence of infection; Gram's stains, fungal smears, and cultures of bronchoalveolar lavage specimens revealed no abnormalities.
On the 14th day, the temperature rose to 38.8°C. The norepinephrine was discontinued. The next day the patient was unresponsive, despite the discontinuation of treatment with sedative agents for 12 hours. On the 16th day, anuria with increasing fluid retention developed, and hemodialysis was initiated. The temperature rose to 38.3°C. He continued to be unresponsive. CT scanning of the head showed no abnormalities. Examination of a bone marrow specimen on the 14th day after chemotherapy was begun showed hypocellularity, without evidence of residual leukemia. On the 18th hospital day, the patient was more alert, but the temperature rose to 39.4°C. A CT scan of the chest disclosed improvement in the air-space disease in both lungs, except that a focus in the left upper lobe had become more consolidated. Quinupristin–dalfopristin was added to the patient's medications on the 19th day because vancomycin-resistant Enterococcus faecalis was detected in cultures of the sacral ulcer and on a rectal swab. A sputum culture grew vancomycin-sensitive enterococcus. Cultures of the blood and urine were sterile. On the 20th day, continuous venovenous hemodialysis was begun. Between the 20th and 26th days, despite a gradual rise in absolute neutrophil count to normal by the 25th day (Table 1), a fever with temperatures ranging from 38.3 to 39.4°C persisted, and the patient's trachea remained intubated; dialysis and treatment with vasoactive drugs were necessary to maintain an adequate systemic arterial pressure.
Differential Diagnosis
Dr. Dey: As the hematologist treating this patient, I thought it likely that the patient's clinical deterioration after the fifth hospital day was due to multiple factors. The diarrhea was probably due to the antimicrobial therapy and the treatment with cytarabine. The high fevers and hypotension could have indicated infection, tumor lysis with cytokine release, or both, although drug-induced fever caused by the cytarabine and mucositis with bacterial superinfection could not be ruled out. The inability to obtain diagnostic microbiologic cultures could have been due to the use of broad-spectrum antimicrobial agents, or it could have indicated that the fever had a noninfectious cause. The acute renal failure could have been the result of the tumor lysis syndrome or severe hypotension and subsequent acute tubular necrosis. The patient's stuporous mental state was probably mainly metabolic in origin and may have included components of uremic encephalopathy, hepatic encephalopathy, and both sedation and fever. Hypotension, infection, hyperalimentation, or drugs may have caused hepatocellular injury, which resulted in elevated values on liver-function testing.
Because of the persistence of fever despite the hematologic recovery during the fourth week of hospitalization, I was concerned that an undiagnosed infection was present and requested a consultation with the Infectious Disease Division.
Discussion of Management
Dr. Jay A. Fishman: Infection in immunocompromised persons is often rapidly progressive and life threatening. Thus, the appropriate design of initial empirical antimicrobial regimens for such patients is critically important. Although progress has been made in the prevention and treatment of many infections common to such hosts, the case under discussion poses two common problems for the clinician. First, if fever in a patient receiving broad-spectrum antimicrobial agents is due to infection, which organisms are likely to be involved? Second, which antimicrobial agents should be used? If the spectrum of coverage is too narrow, the patient may die of sepsis, but at the same time the negative consequences of excess antimicrobials (e.g., resistance, toxic effects, drug interactions, and allergic reactions) are not trivial.
Guidelines for Treating Infections in Immunocompromised Patients
Progress in the care of patients with neutropenia is reflected in a series of excellent guidelines developed for categories of infection that are common to this group, particularly patients with fever and neutropenia.2,3,4,5,6,7,8,9 These guidelines provide information about the pathogens most often associated with infections in each patient group, as well as approaches to the selection of antimicrobial agents. Such guidelines are very useful as a starting point for therapy.
Unfortunately, the application of general guidelines to specific patients is often difficult. A patient may have allergies or other conditions, such as impairment of renal or hepatic function, that limit the use of certain antimicrobial agents. The epidemiologic features of infections vary according to institution and practice for prescribing antimicrobial agents. Guidelines become outdated as new diagnostic tools and therapies become available. In addition, opinion is not uniform with regard to the optimal use of antimicrobial agents in febrile, immunocompromised patients. The evaluation and comparison of data from clinical trials are often complicated by variations in study design and end-point analysis. The best use of vancomycin, aminoglycosides, antifungal therapies, immune globulins, and other agents remains uncertain, as does their timing.
Finally, many fevers in immunocompromised patients are noninfectious in origin; processes that include pulmonary emboli, graft-versus-host disease, drug toxicity, or engraftment syndromes may cause fever. Even in patients with clinical evidence of infection, a specific pathogen is not identified in more than half the cases. Infections due to uncommon viruses (e.g., human herpesvirus 6, parvovirus B19, and metapneumovirus) or to organisms that are difficult to isolate in microbiologic culture (i.e., slow-growing organisms, those requiring special mediums, or cell-wall-deficient bacteria) may go unrecognized. Patients often have multiple, simultaneous conditions, such as infections, adverse drug reactions, and coexisting medical conditions that influence both the susceptibility to infection and the ability of physicians to establish a diagnosis.
There are few guidelines for the treatment of patients, such as this one, who have fever that develops or persists after recovery from neutropenia.3,10 Some localized infectious processes, such as pneumonia, perineal abscess, and fungal infections, may become clinically apparent only after leukocytes return to mount an inflammatory response. The degree of residual impairment of immunity in such patients is unknown, although they are generally less prone to rapidly progressive sepsis than they are when neutropenia is present. Unfortunately, it is often impossible to obtain a microbiologic diagnosis on the basis of cultures obtained from patients receiving broad-spectrum antimicrobial agents.
General Principles
Four principles may be used to guide the selection of antimicrobial agents. First, the treatment of any documented infection must be completed and resolution of the infection documented. Second, after the resolution of neutropenia, physical examination and radiologic studies often reveal sites of infection that were previously unrecognized. Samples should be taken from these sites for histologic or microbiologic evaluation. Third, additional antimicrobial agents should be selected on the basis of the patient's medical history, exposure history, and aspects specific to his or her clinical condition. Fourth, the physician needs to remember that the fever may not be of infectious origin; other causes of fever, such as drugs, may need to be considered. The strategy must be flexible enough to be modified as new data emerge and clinical conditions evolve.
In immunocompromised patients, there is a limited window of opportunity during which antimicrobial therapy can effectively prevent death from infection. This observation has two implications: most patients will receive empirical therapy (at least until microbiologic data are available), and procedures that define the nature and extent of infection are essential. These procedures include radiographic examinations (particularly MRI and CT scans), biopsies to obtain specimens for histologic examination and culture, and other microbiologic assays (based on detection of antigens or specific nucleic acids).11 The use of aggressive diagnostic approaches is especially important because of the relative absence of physical signs or symptoms of infection in many immunocompromised patients. Infection is often advanced or disseminated by the time of diagnosis. Thus, the suspicion that infection is present is critical. Minor abnormalities may be signs of impending sepsis or of tissue-invasive infection (Table 3).
Table 3. Common Signs of Unrecognized Infection in Immunocompromised Patients.
Risk of Infection
For each patient, I ask, "How great is this person's risk of infection? And for which infections are the risks greatest?" For any patient, this risk is determined by the relationship between his or her epidemiologic exposures and a measure of his or her susceptibility to infection, which is termed the net state of immune suppression. The net state of immune suppression is a function of all the factors contributing to the person's risk of infection. Most patients have several types of immune defects.
Pathogens that may be present (so-called epidemiologic exposures) may be intrinsic or extrinsic, and they may have been acquired in the hospital or community; moreover, exposures may be recent or distant in time. The potential pathogens vary according to the nature of the patient's immune deficits. An organism is likely to become invasive or to "set up shop" if the host lacks functional components of the immune system that are needed to control or kill that pathogen. Thus, extracellular bacteria are common pathogens in the absence of neutrophils or other phagocytes, and the majority of infections in patients with neutropenia are derived from endogenous flora.12 In contrast, viruses have greater importance as a cause of opportunistic infection in the absence of cell-mediated cytotoxic (T-cell) responses.
The risk of infection is also determined by the burden of organisms (i.e., the intensity of exposure), the native virulence of the organism, and the nature of the exposure (e.g., whether it is gastrointestinal, is pulmonary, originates at a catheter, or arises from old granulomas). In a person who is not immunocompromised (Table 4), the number of organisms at various sites varies greatly. Thus, in patients with neutropenia, the frequency of infection originating in the gastrointestinal tract may reflect the burden of organisms in the gastrointestinal tract as well as the frequency of injury to the gastrointestinal mucosa. Shifts in patterns of colonization are commonly associated with hospitalization and with the use of antimicrobial agents. Each episode of neutropenia, chemotherapy, infection, catheter insertion, skin breakdown, mucositis, or colitis increases the likelihood that infection or colonization of tissues will persist and that colonizing organisms will be resistant to various antimicrobial agents. Access to information about the patient's prior infections and exposures to antimicrobial agents is essential.
Table 4. Approximate Concentration of Bacteria in Tissues from the Normal Human Host.
Among immunocompromised patients, over time exposure of organisms to selection pressure from antimicrobial agents has shifted the distribution of common bacterial flora from predominantly gram-positive bacteria (e.g., staphylococci) to gram-negative organisms (e.g., Escherichia coli) to resistant gram-negative organisms (e.g., pseudomonas species) and back to resistant gram-positive organisms (e.g., vancomycin-resistant enterococci, fluoroquinolone-resistant streptococci). With prolonged neutropenia, along with improved treatment of bacterial infections and the selection pressure of prophylaxis, resistance to azole antifungal agents has emerged. Resistance of cytomegalovirus to ganciclovir has been observed. As a result, sicker patients tend to become progressively more difficult to treat (Table 5).
Table 5. Factors Contributing to Colonization with Nosocomial Flora.
Care of This Patient
Discussions about this patient between myself and Dr. Dey began by telephone on the 26th hospital day, as follows: "I have a patient with acute myeloid leukemia who has received chemotherapy with prolonged neutropenia, which is resolving. He now has fever, is critically ill, and has some unique risks for infection." These features of the case, including the large skin wound in the sacral area and the lack of a clinical response to multiple antimicrobial agents, became the essential clues to management for this person (Table 6). The goal of consultation was not to stop the use of various antimicrobial agents (although a laudatory goal), nor was it to treat persistent fever, but rather, the goal was to determine whether infection was present and to select the optimal therapy on the basis of pathogens known to be or likely to be present.
Table 6. Clinical Features of This Patient.
On physical examination, the patient appeared to be critically ill; his trachea was intubated, he was sedated, and he was receiving continuous venovenous hemodialysis in the intensive care unit. His temperature was 38.3°C, his blood pressure 80/60 mm Hg while he was receiving pressor agents, his heart rate 82 beats per minute, and his respiratory rate 14 breaths per minute with mechanical ventilation. There was a deep sacral ulcer, 5 by 8 cm. There was granulation tissue and no exudate, and no bone was exposed. The remainder of the findings on examination were normal. Radiographic tests (CT and MRI) revealed no deep infection in the abdomen, chest, or perirectal area. His white-cell count had climbed to 4600 per cubic millimeter, with an absolute neutrophil count of 3174 per cubic millimeter while he was receiving granulocyte colony-stimulating factor.
Potential pathogens in this patient fell into two groups: documented infectious organisms and likely pathogens. Multiple blood cultures had been sterile; a single isolate of coagulase-negative staphylococcus from blood drawn from a Hickman catheter was considered a contaminant. A sputum culture had grown a few yeast (Candida albicans) and moderate numbers of vancomycin-susceptible E. faecalis. A Gram's-stained sputum specimen contained few neutrophils and no organisms. Cultures of both a rectal swab and the pilonidal-cyst ulcer grew E. faecalis (resistant to vancomycin, doxycycline, erythromycin, and tetracycline and sensitive to ampicillin, quinupristin–dalfopristin, chloramphenicol, linezolid, and rifampin).
There was no clinical response to the addition of multiple antimicrobial agents; in fact, the patient's fever was higher. No focus of infection had become manifest with the return of neutrophils. We suspected that if he had sepsis, it was derived from his deep skin ulcer, although this area did not appear to be grossly infected. It also seemed possible that some of the antimicrobial agents might have been contributing to his fever and hypotension. The period of greatest risk (three weeks of absolute neutropenia) had passed. He had completed courses of guideline-specified antimicrobial therapy during the weeks he had neutropenia and had received, in addition, a variety of broad-spectrum antimicrobial agents. Therefore, I thought that his antibiotic coverage could be safely narrowed. I suggested that we treat the obvious skin ulcer with intravenous ampicillin (for susceptible enterococcus from the wound and rectum) and metronidazole (for presumed anaerobic organisms from the same areas) and that we eliminate the other drugs.
Dr. Dey: The granulocyte colony-stimulating factor was discontinued on the 27th day. The quinupristin–dalfopristin, ceftazidime, and acyclovir were discontinued on the 28th day. Within 48 hours, the patient became afebrile, became fully awake, and no longer required pressor agents or ventilatory support. Abnormalities in the results of liver-function tests improved markedly after discontinuation of the quinupristin–dalfopristin. Cultures remained sterile.
A repeated bone marrow biopsy was performed the day before discharge and showed hypercellularity with trilineage hematopoiesis, without evidence of leukemia. The patient gradually recovered and was discharged after 44 days of hospitalization. He is currently undergoing consolidation chemotherapy as an outpatient.
Dr. Philip Amrein (Hematology and Oncology): Is it possible that some patients who are receiving the appropriate antibiotics and remain febrile even while their cultures remain sterile do indeed have sepsis? Is it possible that the cultures do not grow pathogens because of the presence of the antibiotics?
Dr. Fishman: In some cases, microbiologic cultures are sterilized by antimicrobial agents. This is particularly true of patients with low-grade bacteremic infections and those with loculated infections, such as biliary obstruction, small-bowel perforations, or other small abscesses, as was suspected in this patient.
Dr. Nancy Lee Harris (Pathology): In this patient, do you think the fever was caused by drugs or by infection?
Dr. Fishman: I think that drugs clearly had a role in causing his fever. When we stopped the quinupristin–dalfopristin and voriconazole, the patient's condition improved rapidly without the addition of new antimicrobial agents. The chemotherapeutic agents may also have contributed to the fever. However, it is equally likely that he had intermittent bacteremia from his deep pilonidal cyst. So I suspect that he had a combination of infections and adverse drug reactions.
Dr. Howard Weinstein (Pediatric Hematology and Oncology): Is there any role for granulocyte transfusions in patients with neutropenia? And what is the role of prophylactic oral antibiotics in patients who have neutropenia but no fever? Most of the algorithms for empirical antibiotic therapy in pediatric oncology are for patients with neutropenia and fever; we do not give antibiotics to afebrile patients.
Dr. Fishman: Granulocyte transfusions remain controversial, with advocates and detractors in equal numbers. Transfusion reactions are common, and in many patients the granulocyte counts increase only slightly after transfusion. There is a substantial risk of transmission of cytomegalovirus and other viruses. However, some controlled clinical trials have shown a benefit in patients with documented infections. Many of us would try granulocyte transfusions if other approaches failed. Clinical trials are needed in this area.
The routine use of antibiotics in neutropenia in the absence of fever should be limited to specific high-risk patients: those likely to have prolonged and severe neutropenia, those who have had prior episodes of fever or infection during neutropenia, and those who have a likely source of infection, such as this patient who had an open ulcer. In patients with a history of herpes simplex virus or varicella–zoster infections or of thrush, prophylaxis is also worthwhile. Those receiving high-dose corticosteroids may benefit from trimethoprim–sulfamethoxazole prophylaxis against pneumocystis infection. As much as possible, I prefer to individualize prophylaxis for each patient, with risk stratification for specific infections.
Dr. Fishman reports serving as a consultant to Gilead, Roche, and Imerge BT; receiving research support from Fujisawa Healthcare; receiving educational program support from Fujisawa Healthcare and Roche; and serving on the speakers' bureaus of Pfizer, Roche, Fujisawa Healthcare, and Aventis.
Source Information
From the Transplant Infectious Disease and Compromised Host Program, Infectious Disease Division (J.A.F.), the Bone Marrow Transplant Unit, Hematology Oncology Division (B.R.D.), and the Department of Pathology (R.P.H.), Massachusetts General Hospital; and the Departments of Medicine (J.A.F., B.R.D.), and Pathology (R.P.H.), Harvard Medical School.
References
Brunning R, Matutes E, Harris NL, Flandrin G. Acute myeloid leukemia with multilineage dysplasia. In: Jaffe ES, Harris NL, Stein H, Vardiman J, eds. Pathology and genetics of tumours of haematopoietic and lymphoid tissues. Vol. 3 of World Health Organization classification of tumours. Lyons, France: IARC Press, 2001:88-9.
Hughes WT, Armstrong D, Bodey GP, et al. 2002 Guidelines for the use of antimicrobial agents in neutropenic patients with cancer. Clin Infect Dis 2002;34:730-751.
Pizzo PA, Robichaud KJ, Gill FA, Witebsky FG. Empiric antibiotic and antifungal therapy for cancer patients with prolonged fever and granulocytopenia. Am J Med 1982;72:101-111.
Schimpff SC. Empiric antibiotic therapy for granulocytopenic cancer patients. Am J Med 1986;80:13-20.
Klastersky J, Zinner SH, Calandra T, et al. Empiric antimicrobial therapy for febrile granulocytopenic cancer patients: lessons from four EORTC trials. Eur J Cancer Clin Oncol 1988;24:Suppl 1:S35-S45.
Viscoli C. The evolution of the empirical management of fever and neutropenia in cancer patients. J Antimicrob Chemother 1998;41:Suppl D:65-80.
Rolston KV, Mullen CA, Wade JC. NCCN: fever and neutropenia. Support Care Cancer 2002;10:58-64.
European Organization for Research and Treatment of Cancer (EORTC) International Antimicrobial Therapy Cooperative Group, National Cancer Institute of Canada-Clinical Trials Group. Vancomycin added to empirical combination antibiotic therapy for fever in granulocytopenic cancer patients. J Infect Dis 1991;163:951-958.
National Comprehensive Cancer Network home page. (Accessed February 5, 2004, at http://www.nccn.org.)
Barton TD, Schuster MG. The cause of fever following resolution of neutropenia in patients with acute leukemia. Clin Infect Dis 1996;22:1064-1068.
Donowitz GR, Harman C, Pope T, Stewart FM. The role of the chest roentgenogram in febrile neutropenic patients. Arch Intern Med 1991;151:701-704.
Bodey GP, Buckley M, Sathe YS, Freireich EJ. Quantitative relationships between circulating leukocytes and infection in patients with acute leukemia. Ann Intern Med 1966;64:328-340.(Jay A. Fishman, M.D., Bim)