Delays in time to antibiotics in inpatient oncologic neutropenic fever do not predict decreased overall survival.

Abstract

e18765 Background: Neutropenic fever (NF) is relatively common oncologic emergency. Present expert consensus is that anti-pseudomonas gram-negative antibiotics should be administered within 60 minutes of detecting NF. To date studies investigating this relationship in neutropenic fever patients have been either limited in size, or have failed to reliably establish a relationship between time to antibiotics (TTA) and clinical outcomes.While some studies have shown an association between TTA and outcomes in NF patients admitted from the Emergency Department, such studies do not control for the time that patients may have been febrile in the community. To address these factors, we conducted a retrospective study on the effect of TTA on mortality in oncologic patients who developed NF as inpatients. Methods: We performed retrospective chart review of all cases of NF at an NCI designated Cancer Center between 7/1/2016 and 3/27/2019. NF was defined as temperature of 101˚ F on one occasion, or 100.4˚ F sustained over 60 minutes, with an absolute neutrophil count (ANC) less than 500. TTA and survival were calculated via chart abstraction; patients lost to follow up within 180 days were censored to the 180-day mortality group. Relationship between TTA and overall survival (OS) was analyzed via multivariable Cox regression. We excluded patients that had non-cancer related NF, were transferred from another institution with NF, were admitted from the ED with NF, or transitioned to hospice. Only the first instance of NF in any admission was analyzed. Results: A total of 187 eligible cases were identified during the study period, mean age was 57.6 +/- 13.6, 100 (53.5%) cases were in males, 114 (61.0%) cases in Caucasians, 53 (28.3%) in Black People. The 3 most common disease subtypes were acute leukemia (42.8%), plasma cell dyscrasias (27.8%), and lymphoma (16.6%). TTA showed no significant correlation with OS at any timeframe studied. Low Charlson Comorbidity Index ( < 3) correlated with increased survival through ̃360 days, however the effect was non-significant at longer timeframes. Immediate antibiotic treatment ( < 40 mins) correlated with poorer patient prognosis and significantly decreased OS (HR 3.08;CI: 1.30-7.28; p 0.010). Conclusions: TTA was not associated with OS in our study. For inpatients with NF, even hours long TTA may not be long enough to result in adverse clinical outcomes. Unlike NF patients presenting to the ED, where true TTA may often be many hours or even days prior to arrival, a few hours-long TTA in the hospital may not be sufficiently long enough to cause significant patient harm. Interestingly, in our cohort, those who received antibiotics quickly had adverse outcomes. It may be that in patients who were clinically unstable, TTA was shorter given the urgency of the situation. Ultimately, this study’s findings question the applicability of the 60-minute guideline when used in the inpatient setting.