Pulmonary infiltrates during chemotherapy-induced febrile neutropenia: incidence, patterns and outcomes

OBJECTIVE: To analyze the incidence, etiologies, radiographic patterns, and clinical outcomes of adult leukemics with prolonged febrile neutropenia and pneumonia. METHODS: A retrospective study was conducted at a tertiary care hospital. The medical records of adult patients with acute myeloid leukemia diagnosed between January 1989 and June 2000 and undergoing induction chemotherapy were included. Only the patients who presented with a pulmonary infiltrate, secondary leukemia (e.g., transformed chronic myeloid leukemia underlying myelodysplastic syndrome, or disease following alkylating agent therapy) were included and those developing infiltrates following consolidation chemotherapy were excluded. RESULTS: A total of 124 patients were admitted to the hospital with a diagnosis of AML during the study period. Thirty-one patients were excluded; 93 patients received induction chemotherapy and were included in the study analysis. The median age was 36 years (15 - 70 years); 58 males and 35 females. Sixty two percent patients received Cytosine Arabinoside (Ara-C), 17% received Etoposide, 11% received Ara-C and Mitoxantrone, and 6% received All-trans-retinoic Acid. The mean onset and duration of neutropenia were 5 and 15 days, respectively. Pulmonary infiltrates were identified during 45% of neutropenic episodes. A presumptive causative organism was isolated from 50% of patients with an infiltrate: Gram-positive bacteria were most common (47%) followed by Gram-negative bacilli (33%) and fungi (20%). Survival data were available for 88 patients; median disease free survival for the entire cohort was 7 months. Male sex (p=0.015), onset of neutropenia (p=0.02) and bilateral distribution of an infiltrate (p=0.03) were statistically significant predictors of early mortality. For patients with and without pneumonia, the median disease-free interval and overall survival were 2.5 and 4.6 months and 9 and 13 months (p=0.038 and p=0.095) respectively. CONCLUSION: Neutropenia occurred at a mean of 5.0 after initiation of induction chemotherapy. The majority of patients had bilateral pulmonary infiltrates. Male sex, onset of neutropenia and bilateral distribution of an infiltrate were found to be statistically significant predictors of early mortality