1090. Practicing Antimicrobial Stewardship: De-escalating Empiric Antibiotics in Patients with Acute Myelogenous Leukemia and Neutropenic Fever

BACKGROUND: In the era of increased antibiotic resistance, minimizing the use of broad-spectrum antibiotics is essential. We sought to determine whether there was a difference in risk of recurrent fever in patients with acute myelogenous leukemia (AML) and neutropenic fever without an identifiable source in which antibacterials were de-escalated prior to neutrophil recovery compared with those that continued until recovery. METHODS: We performed a retrospective chart review of adult patients with AML undergoing induction chemotherapy at Mount Sinai Hospital in New York, NY from 2009–2017. Neutropenic fever was defined as a temperature of 100.4°F for 1 hour or single temperature of 101°F in a patient with an absolute neutrophil count (ANC) of less than 500 cells/μL. Febrile patients were treated with cefepime, piperacillin–tazobactam, or a carbapenem. De-escalation was defined as changing from one of these antibiotics to antibacterial prophylaxis such as levofloxacin, or discontinuing antibiotics. The primary outcome was recurrent neutropenic fever. Secondary outcomes were adverse events related to antibiotics, intensive care unit (ICU) transfer, and all-cause mortality. RESULTS: Of 390 AML patients undergoing induction chemotherapy, 135 had a neutropenic fever; of whom, 77 had no identifiable infectious source. Of those 77, 38 had antibiotics de-escalated prior to ANC recovery (“short”) and 39 had antibiotics continued until ANC recovery or discharge (“long”). Demographics were similar (Table 1). The median number of antibiotic days for the first fever was 9 in the short group and 15 in the long group (P = 0.0008) (Table 2). Risk of recurrent fever was 46% lower in the short group compared with the long group (hazard ratio 0.54, 95% CI: 0.34–0.88; P = 0.01). There was no significant difference in ICU transfer (P = 0.11) and in-hospital mortality (P = 0.36) between the short and long groups (Table 2). There were 7 adverse drug outcomes, 2 in the short group and 5 in the long group (Table 3). CONCLUSION: Antibiotic de-escalation in AML patients with neutropenic fever with no identifiable infectious source was associated with a lower rate of recurrent fever without affecting ICU transfer, adverse drug events, and death. Physicians should consider de-escalation prior to ANC recovery in the appropriate setting. [Image: see text] [Image: see text] [Image: see text] DISCLOSURES: All authors: No reported disclosures.