Etiology and outcomes of neutropenic fever in Malawi: A prospective, observational cohort.

Abstract

e18831 Background: Neutropenic fever (NF) is a common, life-threatening complication of chemotherapy. Over 80% of high risk hematological malignancy cases will develop NF, with mortality rates reaching 10%. In high income countries (HIC), NF prevention guidelines including prophylaxis have decreased treatment-related mortality. In sub-Saharan Africa (SSA), little is known about the causes of NF as many deaths occur in facilities without blood culture availability. Pathogens in SSA may vary from those in HIC due to high burden of HIV infection as well as different antibiotic stewardship, endemic infections, and chemotherapy regimens. Unknown bacterial spectrum and resistance patterns prevent extrapolation of NF guidelines to SSA. Methods: Patients undergoing chemotherapy at Kamuzu Central Hospital (KCH) in Lilongwe, Malawi were eligible to participate in the study if they were starting a new chemotherapy regimen, lived within 200 kilometers, and (1) had a hematologic malignancy of any kind or (2) had a solid tumor and were HIV-infected. Enrolled patients were given a thermometer and if a temperature of 38º C or greater was recorded, they were instructed to return for management, including admission and antibiotics. Evaluation consisted of complete blood count with differential, urine and blood cultures, and other testing at the clinician’s discretion. All patients were followed through 90 days post fever event or completion of chemotherapy, whichever came later. Results: 106 patients were screened and 50 were enrolled. The primary reason for screen failure was household distance from KCH. Of the enrollees, 26 (52%) were men and 26 (52%) were HIV positive with an average ART treatment time of 7 years and CD4 count of 293 cells/mL. Diagnoses included: aggressive lymphoma (36%), Hodgkin lymphoma (22%), low-grade lymphoma (10%), multicentric Castleman disease (6%), non-Hodgkin lymphoma, NOS (4%), breast cancer (6%), cervical cancer (4%), and other solid tumors (12%). There were 23 febrile events recorded from 15 patients (10 HIV– and 5 HIV+), with no relationship between HIV status and febrile events (p = 0.08). Of the 23 events, a causative agent was isolated in 13 cases (E. coli (6), Malaria (3) and S. pneumonia (2), Pseudomonas (1) and C. freundii (1). Of the 6 E.coli isolates, 3 were found to be multidrug resistant (ie resistant to fluoroquinolones and cephalosporins) and treatment was escalated with a carbapenem. All patients but one survived; death was attributed to Pseudomonas bacteremia. Conclusions: This study demonstrates the initial laboratory trends of febrile events in cancer patients in Lilongwe, Malawi. GNRs, followed by malaria, were the most common source of infection and GNRs were associated with high rates of resistance. Endemic and resistant bacterial infections should be considered when treating NF. Larger studies are needed across the region to identify the pathogens most commonly associated with NF.