The Evaluation of Bronchoscopy in Acute Leukemia: A Retrospective Chart Review

Pulmonary infiltrates are a well-documented complication in patients with acute leukemia and are treated according to guidelines established for the febrile and neutropenic patient. Obtaining specific identification of the causative organism, and therefore identifying the optimal treatment for the infection, has posed a significant challenge. Bronchoscopy has remained an option for a more definitive diagnosis, but the yield from this invasive method is limited and not without risk. Although some diagnostic yield has been shown, it remains unclear whether the data obtained from bronchoscopy affects clinical management. Specifically, this has become a more relevant question today in the advent of a newer generation of antifungal therapies. This information is of great clinical importance should the data reveal that bronchoscopy does not significantly affect medical management, thus sparing our patients from an unnecessary, invasive, and risky procedure. The objective of this study is to characterize the data collected from bronchoscopy and to identify whether information yielded from patients with acute leukemia alters the clinical management of antimicrobial therapy in this population at our institution. Patients included in this study were greater than 18 years of age with radiographic findings of pulmonary infiltrates and a diagnosis of acute lymphoblastic leukemia (ALL), acute myelogenous leukemia (AML), or Burkitt’s lymphoma/leukemia who underwent a bronchoscopy at our institution between March 1, 2001 and June 3, 2007. In this review of 96 patients, bronchoscopy failed to identify any pathogenic organism in 77.1% (99% CI, 64.3%–87.1%) of the cases. Nine of the 22 patients (40.9%) with identified pathogenic organisms had directed changes in their antimicrobial therapy within 10 days after bronchoscopy (99% CI: 16.2%–69.5%). Only 9.4% (99% CI, 3.3%–19.7%) of all patients benefitted from a directed change in antimicrobial therapy within 10 days after bronchoscopy. On the day of bronchoscopy, a total of 91.7% (99% CI, 81.7%–97.3%) of patients were currently receiving at least one broad spectrum antibiotic and 83.3% (99% CI, 71.4%–91.8%) of patients were receiving at least one antifungal agent. Seventy percent of our patients were on broad spectrum antibiotics for greater than five days. Comparison of the culture-positive to culture-negative groups indicated more positive cultures associated with a longer period of neutropenia [20.5 vs. 9.0; p = 0.03] but not length of hospitalization [24.5 vs. 16.0; p = 0.09] prior to bronchoscopy. These data emphasize that bronchoscopy performed in leukemia patients who have previously received extensive antimicrobial therapy rarely yields a specific pathogen that changes therapy. Important limitations of our study include the lack of a control group that did not undergo bronchoscopy, the aspergillus galactomannan assay for bronchoalveolar lavage fluid was not yet in use, and we collected limited data on the impact of negative culture data on clinician decisions. In the interim, clinicians should carefully consider how the information obtained from bronchoscopic evaluation, positive or negative, will influence overall management in the era of effective broad spectrum therapeutic agents.