Evaluating Antibiotic De-escalation for Autologous Stem Cell Transplant Patients With Febrile Neutropenia in a Real-World Clinical Setting

Febrile neutropenia (FN) is a complication in approximately 90% of autologous stem cell transplant (SCT) patients. Guidelines support early broad-spectrum antibiotics (BSA) to prevent morbidity and mortality. However, in patients who are clinically stable and deemed to have a fever of unknown origin, the optimal duration of BSA is unknown. Accumulating evidence suggests that de-escalation of BSA in select patients may decrease duration of BSA exposure without compromising clinical outcomes such as infection, recurrent fever, and readmission. With this, a de-escalation protocol was implemented at Vanderbilt University Medical Center (VUMC) to identify autologous SCT patients who may benefit from early de-escalation of BSA. The objectives of this study were to analyze the impact of early empiric antibiotic de-escalation on the duration of BSA as well as its impact on the incidence of recurrent fever and documented infection in autologous SCT patients. This was a single-center, retrospective study evaluating patients older than 18 years of age who underwent autologous SCT and experienced an episode of FN from January 2018 to December 2022 at VUMC (N = 195). The protocol was initiated on January 1, 2020, to de-escalate BSA back to prophylaxis in stable neutropenic patients determined to have a fever of unknown origin. The primary outcome was the number of BSA days within 30 days. Secondary clinical outcomes included recurrent fever, documented infection, readmission, 30-day mortality, and 90-day non-relapsed mortality (NRM). Outcomes were compared across pre- and postprotocol groups with a Wilcoxon rank sum test, Pearson chi-square test, or regression analysis as appropriate. The median BSA duration was 4.7 and 2.7 days in the pre- and postprotocol groups, respectively (P < .001). Recurrent fever (14.2% versus 16.0%, P = .726), documented infection (1.7% versus 6.7%, P = .068), and readmission (13.3% versus 22.7%, P = .091) within 30 days were not significantly different between the two groups. Neither 30-day mortality (0.8% versus 1.3%, P = .736) nor 90-day NRM (0.8% versus 1.3%, P = .736) differed. The implementation of an early de-escalation protocol for autologous SCT patients who develop FN was associated with a reduction in duration of BSA compared to the preprotocol group without a significant difference in readmission, recurrent fevers, and documented infections. This study adds to existing evidence that early de-escalation of BSA in FN patients with a fever of unknown origin who are afebrile and clinically stable is safe and reduces unnecessary antibiotic use.