Evaluation of early de-escalation of empiric antimicrobial therapy in acute leukemia patients with febrile neutropenia at a large academic medical center.

ABSTRACT

INTRODUCTION: Recent trials have shown early de-escalation of empiric antimicrobial therapy (EAT) in febrile neutropenia has led to less adverse effects with no difference in patient mortality. In 2019, our institution adjusted internal guidelines to de-escalate EAT after 7 days of intravenous anti-pseudomonal therapy in patients with signs of clinical recovery from febrile neutropenia and no evidence of infection. METHODS: This was a retrospective, single-center, observational, cohort study. Eligible patients were adults with acute myeloid leukemia (AML) or acute lymphoblastic leukemia (ALL) who received induction chemotherapy and developed febrile neutropenia without documented infection. Patients were separated based on EAT duration ≤ 9 days and > 9 days. Empiric antimicrobial therapy was defined as the initiation of an anti-pseudomonal beta-lactam. The primary outcome was the difference in number of EAT-free days. Secondary outcomes included fever recurrence, ICU admissions, fever duration, infections post de-escalation, and Clostridioides difficile infection (CDI). RESULTS: Forty-four encounters met inclusion. The EAT ≤ 9 days group had 7 more EAT-free days compared to the EAT > 9 days group (p < 0.001). No between-group differences were identified in terms of fever after EAT discontinuation (p = 0.335), ICU admission (p = 0.498), or CDI (p = 0.498). The EAT > 9 days group experienced longer initial fever (p < 0.001) and received addition of resistant Gram-positive coverage (p = 0.014). More patients receiving EAT > 9 days had a diagnosis of AML (p = 0.001). CONCLUSIONS: Shorter EAT duration did not lead to worse outcomes in patients with AML or ALL who received induction chemotherapy and developed febrile neutropenia without a documented infection source.