Fludarabine

Background: Little is known about lung injury caused by fludarabine therapy.

Objectives: To establish a case definition, to describe management, and to identify risk factors for fludarabine-related pulmonary toxicity.

Design: Case-control study.

Setting: Tertiary-care US Army teaching hospital.

Patients: Individuals treated with fludarabine at our institution between January 1989 and June 2000.

Measurements: Cases of fludarabine-related pulmonary toxicity were defined as follows: dyspnea, fever, hypoxemia, and radiographic infiltrates seen in a patient treated with fludarabine; cases were excluded if there was evidence of pulmonary infection or progression of underlying lymphoproliferative disease affecting the lungs. For each ease, demographic data, medical history, radiographic information, available bronchoscopy and pathology data, and details of treatment were reviewed. Cases were compared with fludarabine-treated control subjects to identify potential risk factors. Comparisons were made with regard to age, gender, history of underlying lung disease, lymphoproliferative diagnosis, prior chemotherapy, fludarabine treatment regimen, and pretreatment chest radiograph.

Results: During the study period, 105 patients were treated with fludarabine. The incidence of fludarabine-related pulmonary toxicity using our definition was 8.6% (95% confidence interval [CI], 3.2 to 13.9%). One patient died before this entity was suspected; the remainder of the patients underwent bronchoscopy to exclude infection. Patients were treated with corticosteroids with subjective and objective benefits. One patient later died of apparent infection during steroid therapy. One patient was retreated with fludarabine and symptoms of lung toxicity developed again. Patients (n = 9) were similar to control subjects (n = 96) with respect to age, gender, history of underlying lung disease, previous chemotherapy, and fludarabine regimen. Patients with chronic lymphocytic leukemia were 13.3 (95% CI, 1.6 to 300.6) times more likely to have toxicity develop than patients treated with fludarabine for other diagnoses. There was a trend toward an increased incidence in patients with interstitial infiltrates apparent on prefludarabine chest radiographs.

Conclusions: A variety of lung conditions arise in patients treated with fludarabine; however, this agent seems to cause direct pulmonary toxicity. After performing an appropriate evaluation to exclude infection, corticosteroids are an effective therapy. The relative frequency of this condition and potential for mortality underscore the need for increased clinician awareness of fludarabine-related pulmonary toxicity and its risk factors.

Key words: chronic lymphocytic leukemia; drug toxicity; fludarabine

Abbreviations: CI = confidence interval; CLL = chronic lymphocytic leukemia; PCP = Pneumocystis carinii pneumonia

**********

Fludarabine monophosphate (Berlex Laboratories; Montville, NJ) is a nucleoside analog that is widely used in the treatment of low-grade lympho-proliferative malignancies including chronic lymphocytic leukemia (CLL) and low-grade non-Hodgkin's lymphoma. (1) As with other nucleoside analogs, fludarabine produces bone marrow suppression and can lead to profound leukopenia. Rare cases of autoimmune hemolytic anemia have been reported. (2-4) Neurotoxicity, ranging from weakness and confusion to coma, has also been reported in patients treated with high doses of fludarabine. (5) Patients treated with fludarabine are at increased risk for opportunistic pulmonary infections. (6,7) Specifically, these patients are susceptible to respiratory tract infections with unusual pathogens such as Listeria monocytogenes, cytomegalovirus, respiratory syncytial virus, and Pneumocystis carinii pneumonia (PCP). (8-12) The increased incidence of PCP in patients treated with fludarabine, particularly those who have been heavily pretreated with alkylating agents, has led to recommendations for PCP prophylaxis in many instances. (13) In cases in which patients who are treated with fludarabine have symptoms develop that are consistent with lower respiratory tract infection, prompt evaluation is indicated.

Chemotherapeutic agents such as bleomycin, cyclophosphamide, and methotrexate are associated with pulmonary toxicity. (14) Despite a vast experience with fludarabine, little clinical data are available regarding a possible association between this agent and lung toxicity. To date, there have been eight published reports outlining a total of 10 patients who had a syndrome develop consistent with possible lung toxicity secondary to fludarabine. (15-22) These are isolated case reports, and there remains no established case definition for this poorly characterized entity.

We hypothesized that pulmonary toxicity from fludarabine is a discrete entity, and we developed a disease definition. We reviewed our experience with noninfectious pulmonary complications observed in patients treated with fludarabine at our institution and examined our diagnostic approach, management, and outcome in suspected cases.

MATERIALS AND METHODS

Case Definition and Patient Selection

We reviewed the records of patients treated with fludarabine at our institution between January 1989 and June 2000 to determine whether fludarabine-related pulmonary toxicity was present. The case definition used was as follows: dyspnea, fever, hypoxemia, and radiographic infiltrates in a patient treated with fludarabine. Cases were excluded if there was evidence of either confirmed or suspected pulmonary infection or progression of the patient's underlying lymphoproliferative syndrome involving the lung.

Fever was defined by a temperature > 38[degrees]C on any single reading taken within 24 h of development of a radiographic infiltrate. Hypoxemia was defined by either pulse oximetry or arterial blood gas analysis demonstrating hemoglobin saturation [less than or equal to] 92% while breathing room air at rest. Radiographic infiltrates could be alveolar, interstitial, or mixed. Cases in which infiltrates were present before treatment with fludarabine were also considered. In suspected cases of fludarabine-related pulmonary toxicity, radiographs were reviewed by a radiologist and a pulmonologist. Infection was excluded either by bronchoscopy with BAL or at autopsy. BAL was performed in standard fashion in the segment or segments of the lung that appeared most involved radiographically. The National Cancer Institute criteria for progression of disease were applied to patients with CLL. (23) In addition, subjects with other low-grade lymphoproliferative conditions were assessed clinically by a hematologic oncologist (J.C.B.) for evidence of progression of disease.

For each case, patient demographic variables (gender, age, cancer diagnosis, and any cancer pretreatment), treatment regimen, and interval between treatment and development of lung toxicity were obtained. Pretreatment laboratory results and chest radiographs as well as laboratory and radiographic data during the episode of fludarabine-related pulmonary toxicity were reviewed. Bronchoscopy findings including BAL data (cellularity, microbiology, and flow cytometry) were collected. Biopsy results obtained from transbronchial biopsy, surgical lung biopsy, or autopsy were also reviewed. Information regarding the treatment of suspected cases of pulmonary toxicity as well as outcomes were also recorded. Control subjects were all patients who were treated with fludarabine but did not have toxicity develop. The two groups were compared for age, gender, history of underlying lung disease (treatment for asthma or COPD), diagnosis for which fludarabine was prescribed, prior and concurrent chemotherapy, and pretreatment chest radiograph.

Statistical Analysis

Comparisons were made between the groups of patients with and without fludarabine-related pulmonary toxicity. Student's t test was used to analyze means for continuous variables with parametric distributions. For categorical data, the Fisher's exact test was used. We calculated the odds ratios and 95% confidence intervals (CIs) associated with proportions for factors associated with development of toxicity. All tests of significance were two tailed. Significance was defined by a p [less than or equal to] 0.05.

RESULTS

Patients

A total of 105 patients were treated with fludarabine at our institution between January 1989 and June 2000. From this group, using our case definition, a cohort of nine patients with fludarabine-related pulmonary toxicity were identified. There included seven men and two women (mean [+ or -] SD age, 66.3 [+ or -] 10.4 years; range, 54 to 83 years). Eight patients received fludarabine for B-cell CLL, and one patient received fludarabine for mantle cell lymphoma. Only two patients received chemotherapy before treatment with fludarabine: one patient had been treated with chlorambucil and another with the combination of cyclophosphamide, adriamycin, vincristine, and prednisone. No patients received external beam radiation. Five patients were treated with fludarabine alone, two patients were treated with fludarabine in combination with cyclophosphamide, and two patients were treated with fludarabine in combination with rituximab immunotherapy. The baseline chest radiographic finding was normal in five of the patients. Two patients had increased interstitial markings before treatment with fludarabine, and two patients had evidence of hilar or mediastinal lymphadenopathy.

Comparison of Subjects Developing Toxicity and Control Subjects

The incidence of toxicity was 8.6% (95% CI, 3.2 to 13.9%). Subjects who had pulmonary toxicity develop were compared to those without pulmonary toxicity (Table 1). Eight of 44 patients (18.2%) with CLL treated with fludarabine had toxicity develop, compared with 1 of 61 patients (0.2%) patients with other malignancies; this difference was statistically significant (p = 0.004). CLL patients treated with fludarabine were more likely to have lung toxicity develop, with an odds ratio of 13.3 (95% CI, 1.6 to 300.6). As shown in Table 1, there were no differences between the two groups in terms of gender, previous chemotherapy for lymphoproliferative disorder, prior lung disease, or the use of fludarabine in combination with chemotherapy or other immunotherapies. There was a trend toward an increased likelihood for toxicity in patients with chest radiographs that demonstrated radiographic infiltrates before therapy with fludarabine (p = 0.11).

Clinical Features of Cases

The onset of symptoms after fludarabine treatment ranged from 3 days after a first cycle of fludarabine treatment to 6 days after the seventh cycle of treatment. Three patients had toxicity develop after the first cycle, three patients after the second cycle, one patient after the third cycle, one patient after the fourth cycle, and one patient after the seventh cycle of fludarabine. The WBC count obtained before the cycle of chemotherapy that was implicated in each case of fludarabine-related pulmonary toxicity varied widely (median, 4,400/mL; range, 1,000 to 115,200/mL). Seven patients had new interstitial or mixed interstitial and alveolar infiltrates evident on plain chest radiography. Another patient had an unremarkable chest radiograph but had evidence of interstitial infiltrates on chest CT. One patient had preexisting interstitial infiltrates that were not significantly changed with the onset of pulmonary toxicity. Infiltrates were located in the upper lobes in one case, in the lower lobes in two cases, and diffusely in the remaining six cases. The pattern was interstitial in five cases and mixed interstitial-alveolar in four cases. The most common chest radiograph findings were diffuse interstitial infiltrates (n = 4) and diffuse interstitial-alveolar infiltrates (n = 2). In two cases there were small pleural effusions. There were no cases of cavitation, pneumothorax, or cysts.

Bronchoscopy was performed in eight patients. As shown in Table 2, BAL cell count and differential was performed in six of the eight procedures. BAL cellularity was increased in all cases; however, there was no consistent pattern to the type of predominant cell. Gram's stain, Giemsa stain, acid-fast stain, direct fluorescent antibodies for PCP, bacterial culture, mycobacterial culture, and fungal culture as well as specific assays for Legionella, herpes virus, and cytomegalovirus were performed in each case. All patients were treated with broad-spectrum antibiotics, including trimethoprim/sulfamethoxazole, pending microbiology results.

Flow cytometry was performed on BAL fluid in seven cases. No monoclonal cell populations were identified. Transbronchial biopsy (n = 3) and surgical lung biopsy (n = 3) were performed in selected cases. Biopsy specimens, in general, revealed diffuse chronic interstitial inflammation and fibrosis. In two specimens, there seemed to be reactive pneumocyte atypia. In one case, there were findings consistent with diffuse alveolar damage. One open-lung biopsy revealed poorly formed caseating granulomatous inflammation, and one transbronchial biopsy was noted to have a single noncaseating granuloma; in each case stain and culture findings for microorganisms were negative. In no biopsy specimen was there evidence of malignancy. One patient died of progressive ventilatory failure; autopsy was performed and revealed evidence of interstitial fibrosis with organizing pneumonia consistent with a drug-related pneumonitis. There was no evidence of malignancy on the autopsy specimen.

Treatment of Presumed Toxicity

One patient died before the diagnosis of fludarabine-related pulmonary toxicity was suspected. The remaining eight patients were treated with corticosteroids. Four patients were initially treated with IV methylprednisolone, and four patients were treated with oral prednisone (1 mg/kg/d). All patients were initially treated as inpatients under careful observation, and all were discharged with steroid therapy. One patient had gradual improvement in his symptoms over 12 months. Six patients were noted to have a more rapid subjective improvement, as well as evidence of either an increase in hemoglobin saturation on pulse oximetry or a decreased need for supplemental oxygen after initiation of corticosteroid therapy. The median time to response in these patients was 4 days, with a range of 1 to 14 days. The final patient had a prompt initial response, but recrudescence of symptoms developed during the steroid taper. She required reevaluation for infection with bronchoscopy and BAL, and ultimately required surgical lung biopsy to exclude other etiologies for increasing interstitial infiltrates. One month into her second steroid taper, she was admitted to another hospital in respiratory distress and required intubation. Her intubation was complicated by cardiac arrest and she was not successfully resuscitated. Her family declined a postmortem examination. Review of records and radiographs from that hospitalization suggested that the patient, while receiving corticosteroids, had developed a lobar pneumonia in the setting of progressive interstitial infiltrates. Steroid taper was eventually possible in all other cases. Only one subject was retreated with fludarabine; this patient had signs and symptoms redevelop consistent with pulmonary toxicity that again responded to steroid therapy.

DISCUSSION

In a recent randomized trial, fludarabine was compared with chlorambucil as initial therapy in patients with CLL, and was shown to provide higher response rates and longer duration of remission and progression-free survival. (1) Fludarabine is likely to be used with increasing frequency in the treatment of CLL and other low-grade lymphoproliferative disorders. We are the first to describe a series of patients with chronic lymphoproliferative disorders who developed a syndrome consistent with fludarabine-related pulmonary toxicity. The incidence of toxicity in our patient population was 8.6%. A variety of lung conditions arise in the population of patients treated with fludarabine. These symptoms may mimic infection; therefore, pulmonary specialists, oncologists, and physicians in hospital-based practice need to be familiar with lung injury as a potential complication of fludarabine therapy.

Identification of new drug toxicities is difficult. When the drug is administered to patients with complicated medical conditions, distinguishing drug effect from underlying disease can be even more complicated. Our case definition--dyspnea, fever, hypoxia, and radiographic infiltrates with the exclusion of infection or malignancy--is similar to definitions used in the description of other drug-related pulmonary toxicities. (24-25) Our definition is also consistent with many of the previously published case reports of fludarabine-related pulmonary toxicity. Most of the previously published cases involved subjects who had been pretreated with alkylating chemotherapeutic agents known to cause pulmonary toxicity. In our series, however, the majority of patients had not received prior chemotherapy. Although some were simultaneously treated with other chemotherapeutic agents, more than one half were treated with fludarabine alone. Previous chemotherapy and combination therapy with fludarabine and another agents were not risk factors for development of pulmonary toxicity in our study. This strengthens the potential relationship between fludarabine and the observed pulmonary findings. The number of cycles of fludarabine before development of toxicity was broad, although two thirds of the cases followed either the first or second cycle of fludarabine. There does not appear to be a consistent radiographic finding in these cases. BAL cellularity is increased in each case; however, BAL lymphocytosis does not appear to be a requisite of the diagnosis. Histologically, this appears to be a fibrotic interstitial process with variable chronicity. The finding of granulomatous inflammation on two specimens suggests that hypersensitivity may play a role in some cases. Importantly, no biopsy specimen demonstrated evidence of malignancy; stains and cultures, when performed, did not reveal infection. Although there is no histologic appearance that confirms the diagnosis of a drug-related effect, we advocate biopsy, when possible, to exclude other etiologies for the pulmonary findings.

Patients who receive fludarabine for CLL are more likely to have pulmonary toxicity develop. Whether this is related to the immunobiology of CLL or to other factors associated with differences in patients with CLL compared with the other lymphoproliferative disorders is uncertain. Patients with CLL have a high proportion of circulating tumor cells that constituently overexpress cytokines such as tumor necrosis factor-[alpha] and interleukin-6. (26,27) This may promote a milieu for lung injury when rapid cell lysis is noted. The hypothesis of tumor lysis is supported by the relatively high frequency of previously untreated patients in our series. One would expect a more marked cytokine release in those individuals with the highest levels of circulating leukemia cells; however, only two patients had WBC counts > 100,000/mL before fludarabine treatment. Five of the remaining six patients all had WBC counts < 5,000/mL. Autopsy studies of CLL patients have commonly demonstrated pulmonary infiltration with CLL lymphocytes; fludarabine-related pulmonary toxicity may, therefore, represent a localized intrapulmonary tumor lysis syndrome or a cytokine-mediated inflammatory process directly within the lungs. There was a trend toward a significant difference between patients who had pretreatment chest radiographs demonstrating interstitial infiltrates. Our inability to identify other risk factors may be related to sample size and underrecognition of this entity.

Recognizing that fludarabine is associated with profound immunosuppression, particularly altered cell-mediated immunity, significant effort must be made to exclude infectious processes as the etiology for the pulmonary findings. Bronchoscopy with BAL is a moderately invasive tool widely available to aid in the diagnosis of respiratory infections. The sensitivity of BAL in the diagnosis of pneumonia has been studied in populations similar to individuals treated with fludarabine. In patients who undergo bone marrow transplant, the sensitivity of BAL in the diagnosis of lower respiratory tract infections ranges from 60 to 95%. (28) In patients with AIDS, sensitivity ranges from 64 to 91%. The sensitivity of BAL with direct fluorescent antibodies for PCP in this population is 82 to 90%. (29) Similar to previous case reports, broad-spectrum antimicrobial therapy was continued in our subjects while awaiting initial BAL microbiology results.

Eight of these patients were noted to have initial improvement after initiation of steroid therapy. We also report two cases of fludarabine-related pulmonary toxicity that proved fatal: one patient died before the diagnosis was suspected, and another patient died, possibly from an infection, during steroid therapy. The potential for mortality highlights the need for expedient evaluation to exclude other potential causes for the respiratory finding, prompt initiation of corticosteroid therapy, and careful monitoring while receiving steroid treatment. We also confirm a prior finding of recrudescence of fludarabine-related pulmonary toxicity in a patient who was rechallenged with the drug. (18)

This study has several limitations. First, our study, although the largest reported series, has a relatively small sample size. The patient characteristics were varied, and a larger sample size would better characterize the disease. Second, this is a retrospective study. Because clinical testing such as pulse oximetry, pulmonary function testing, and chest radiography was not routinely performed in asymptomatic patients treated with fludarabine, subclinical lung toxicity may have gone unnoticed. Retrospective case series are limited by the criteria used in defining the cases. It may be possible for fludarabine-related pulmonary toxicity to coexist in the presence of either a lower respiratory tract infection or with progression of underlying malignant disease; such patients would be excluded from our review. For these reasons, the actual incidence of this disease may be underestimated. In creating our case definition, we attempted to choose clear and objective end points that would minimize the introduction of a sampling bias. By establishing this case definition, it will be possible to determine more accurately the incidence of fludarabine-related pulmonary toxicity, because this condition often may go unrecognized. Prospective research is needed to further our knowledge in this area.

Fludarabine, although generally safe and well tolerated, can cause pulmonary toxicity. Clinicians should consider this diagnosis in patients treated with fludarabine who have dyspnea, fever, hypoxemia, and interstitial infiltrates develop after infection and underlying malignant disease are ruled out. Patients with CLL are far more likely to have this syndrome develop than patients treated with fludarabine for other malignancies. The clinical presentation of this syndrome can be quite severe and even life threatening. Fludarabine-related pulmonary toxicity appears to respond to corticosteroid therapy and, in the majority of our cases, this response was noted within several weeks of initiation of therapy. Development of toxicity should preclude retreatment with fludarabine or other nucleoside analogs unless no other viable therapeutic options exist.

REFERENCES

(1) Rai K, Peterson B, Appelbaum F, et al. Fludarabine compared with chlorambucil as primary therapy for chronic lymphocytic leukemia. N Engl J Med 2000; 343:1750-1757

(2) Weiss RB, Freiman J, Kweder SL, et al. Hemolytic anemia after fludarabine therapy for chronic lymphocytic leukemia. J Clin Oncol 1998; 16:1885-1889

(3) Tosti S, Caruso R, D'Adamo F, et al. Severe autoimmune hemolytic anemia in a patient with chronic lymphocytic leukemia responsive to fludarabine-based treatment. Ann Hematol 1992; 65:238-239

(4) Di Raimondo F, Giustolisi R, Cacciola E, et al. Autoimmune hemolytic anemia in chronic lymphocytic leukemia patients treated with fludarabine. Leuk Lymphoma 1993; 11:63-68

(5) Chun HG, Leyland-Jones BR, Caryk SM, et al. Central nervous system toxicity of fludarabine phosphate. Cancer Treat Rep 1986; 70:1225-1228

(6) Sanders C, Perez EA, Lawrence HJ. Opportunistic infections in patients with chronic lymphocytic leukemia following treatment with fludarabine. Am J Hematol 1992; 39:314-315

(7) Anaissie EJ, Kontoyiannis DP, O'Brien S, et al. Infections in patients with chronic lymphocytic leukemia treated with fludarabine. Ann Intern Med 1998; 129:559-566

(8) Hequet O, de Jaureguiberry JP, Jaubert D, et al. Listeriosis after fludarabine treatment for chronic lymphocytic leukemia. Hematol Cell Ther 1997; 39:89-91

(9) Anaissie E, Kontoyiannis DP, Kantarjian H, et al. Listeriosis in patients with chronic lymphocytic leukemia who were treated with fludarabine and prednisone. Ann Intern Med 1992; 117:466-469

(10) Grigg A, Chapman R, Szer J. Fatal CMV pneumonia associated with steroid therapy after autologous transplantation in patients previously treated with fludarabine. Bone Marrow Transplant 1998; 21:619-621

(11) Eftekhari P, Lassoued K, Oksenheudler E, et al. Severe respiratory syncytial virus pulmonary infection in a patient treated with fludarabine for chronic lymphocytic leukemia. Ann Hematol 1998; 76:225-226

(12) Bastion Y, Coiffier B, Tigaud JD, et al. Pneumocystis pneumonia in a patient treated with fludarabine for chronic lymphocytic leukemia. Eur J Cancer 1991; 27:671

(13) Byrd JC, Hargis JB, Kester KE, et al. Opportunistic pulmonary infections with fludarabine in previously treated patients with low-grade lymphoid malignancies: a role for Pneumocystis carinii pneumonia prophylaxis. Am J Hematol 1995; 49:135-142

(14) Rosenow EC, Limper AH. Drug-induced pulmonary disease. Semin Respir Infect 1995; 10:86-95

(15) Hurst PG, Habib MP, Garewal H, et al. Pulmonary toxicity associated with fludarabine phosphate. Invest New Drug 1987; 5:207-210

(16) Kane GC, McMichael AJ, Patrick H, et al. Pulmonary toxicity and acute respiratory failure associated with fludarabine monophosphate. Respir Med 1992; 86:261-263

(17) Cervantes F, Salgado C, Montserrat E, et al. Fludarabine for prolymphocytic leukaemia and risk of interstitial pneumonitis [letter]. Lancet 1990; 336:1130

(18) Levin M, Aziz M, Opitz L. Steroid-responsive interstitial pneumonitis after fludarabine therapy. Chest 1997; 111: 1472-1473

(19) Foss FM, Ihde DC, Linnoila IR, et al. Phase II trial of fludarabine phosphate and interferon [alpha]-2a in advanced mycosis fungoides/Sesary syndrome. J Clin Oncol 1994; 12: 2051-2059

(20) Grever MR. Fludarabine monophosphate in chronic lymphocytic leukemia [letter]. Blood 1986; 68:223

(21) Montillo M, Tedeschi A, Centurioni R, et al. Treatment of relapsed adult acute lymphoblastic leukemia with fludarabine and cytosine arabinoside followed by granulocyte colony-stimulating factor (FLAG-GCSF). Leuk Lymphoma 1997; 25:579-582

(22) Puccio CA, Mittelman A, Lichtman SM, et al. A loading dose/continuous infusion schedule of fludarabine phosphate in chronic lymphocytic leukemia. J Clin Oncol 1991; 9:1562-1569

(23) Cheson BD, Bennett JM, Grever M, et al. National Cancer Institute Sponsored Working Group guidelines for chronic lymphocytic leukemia: revised guidelines for diagnosis and treatment. Blood 1996; 87:4990-4997

(24) Alarcon G, Kremer J, Macaluso M, et al. Risk factors for methotrexate-induced lung injury in patients with rheumatoid arthritis. Ann Intern Med 1997; 127:356-364

(25) Malik S, Myers J, DeRemee R, et al. Lung toxicity associated with cyclophosphamide use. Am J Respir Crit Care Med 1996; 154:1851-1856

(26) Mainou-Fowler T, Miller S, Proctor SJ, et al. The levels of TNF [alpha], IL4 and IL10 production by T-cells and B-cells in chronic lymphocytic leukaemia (B-CLL). Leuk Res 2001; 25:157-163

(27) Fayad L, Keating MJ, Reuben JM, et al. Interleukin-6 and interleukin-10 levels in chronic lymphocytic leukemia: correlation with phenotypic characteristics and outcome. Blood 2001; 97:256-263

(28) Mares DC, Wilkes DS. Bronchoscopy in the diagnosis of respiratory infections. Curr Opin Pulm Med 1998; 4:123-129

(29) Baselski VS, Wunderink PG. Bronchoscopic diagnosis of pneumonia. Clin Microbiol Rev 1994; 70:533-558

* From the Pulmonary and Critical Care Medicine Service (Drs. Helman and Shorr), the Hematology and Oncology Service (Dr. Byrd), and the Allergy and Immunology Service (Dr. Ales), Department of Internal Medicine, Walter Reed Army Medical Center, Washington, DC; and the Division of Hematology and Oncology (Dr. Byrd), The Ohio State University, Columbus, OH. The opinions or assertions contained herein are the private views of the authors and are not to be construed as official or as reflecting the view of the Department of the Army or the Department of Defense.

Manuscript received September 25, 2001; revision accepted December 24, 2001.

Correspondence to: Donald L. Helman, Jr., MD, Pulmonary and Critical Care Medicine Service, Walter Reed Army Medical Center, 6900 Georgia Ave, NW, Washington, DC 20307-5001; e-mail: donald.helman@na.amedd.army.mil

COPYRIGHT 2002 American College of Chest Physicians
COPYRIGHT 2003 Gale Group