2666. De-escalation of Broad-Spectrum Antibiotics in Hematopoietic Stem Cell Transplant Patients During Initial Episode of Febrile Neutropenia

BACKGROUND: Febrile neutropenia (FN) is a common and serious complication in patients undergoing hematopoietic stem cell transplant (HSCT). Typically, broad-spectrum antibiotics (BSA) are promptly initiated with controversy on timing of de-escalation. ECIL 2013 guidelines suggest de-escalation after 72 hours if the patient is infection free and afebrile for at least 48 hours. Conversely, the 2011 IDSA recommends continuing BSA in patients who defervesce until absolute neutrophil count (ANC) recovery. In 2014, our center’s practice changed to early de-escalation and we sought to compare outcomes between the two practices. METHODS: We retrospectively analyzed patients who underwent a HSCT in 2013 and 2017 with an episode of FN and negative infectious work up. The standard care group (SCG) were continued on BSA until ANC recovery. The early de-escalation group (EDG) de-escalated to fluoroquinolone prophylaxis at least 24 hours prior to ANC recovery after the patient was fever free for 48 hours. The primary end-point was duration of BSA. Secondary endpoints included 30-day mortality, re-hospitalization and length of stay (LOS) from FN. Median values were compared with the Mann–Whitney test. RESULTS: Among 229 HSCT patients, 155 (68%) developed FN post-transplant and of those 97 (63%) were without infection (13 EDG and 84 SCG). Initial FN duration of BSA was less in the EDG (3.09 days vs. 4.69 days, P = 0.069). Total antibiotic free days to 30 day follow-up were similar (EDG 24.08 vs SCG 25.19, P = 0.81). Duration of neutropenia was less in the SCG with 7.99 days compared with 11.69 days in the EDG (P = 0.007), but duration of initial fever was less in the EDG (2.55 days vs. 3.33 days, P = 0.023). 30 day mortality was 0% in both groups. Rates of re-hospitalization within 30 days were approximately the same (7.1% vs. 7.6%). LOS from FN was not significantly different with 6.68 days in SCG and 7.75 days in EDG (P = 0.140). More new bacterial infections were identified within 30 days of FN in the SCG than the EDG (10.7% vs. 7.6%). CONCLUSION: Early BSA de-escalation resulted in no significant difference in LOS from FN and fewer days of BSA with 30-day mortality and re-hospitalization rates similar. This data suggest de-escalating BSA prior to ANC recovery is likely safe and leads to less BSA exposure, but more multi-center data are needed. DISCLOSURES: All authors: No reported disclosures.