Does time-to-antibiotics predict outcome of febrile neutropenia in pediatric cancer?

Abstract

9076 Background: Prolonged time-to-antibiotic administration (TTA) has been associated with poor outcomes in adults with bacterial meningitis, pneumonia, and sepsis. For pediatric oncology patients with febrile neutropenia (FN), TTA has been suggested as a quality-of-care (QOC) measure, but no prior reports have correlated TTA with outcome in pediatric FN. We sought to determine whether TTA is associated with adverse events in a cohort of pediatric FN patients. Methods: We designed a retrospective cohort study of all FN admissions at our institution from 2001-2009. Inclusion criteria were cancer treatment at our institution, fever, and neutropenia (ANC<500). Exclusion criteria were prior hematopoietic stem cell transplantation, severe sepsis on presentation, and transfer from an outside facility. The primary outcome measure was a composite of any in-hospital mortality, transfer to the pediatric intensive care unit (PICU), and receipt of ≥ 40ml/kg of fluid resuscitation within 24 hours of presentation. Age, gender, oncologic diagnosis, temperature, total white blood cell (tWBC) count, bacteremia, viral infection, presentation location, empiric antibiotics, and weekend admission were analyzed as covariates with TTA. TTA was log transformed. Random effects mixed logistic regression was used for both bivariable and multivariable analyses for the composite outcome. Results: We identified 653 subjects with 1,628 admissions for FN from 2001-2009. Median TTA was 114 minutes (range 10-536). 181 (11.1%) admissions met the composite outcome criteria (mortality-0.6%, PICU transfer – 4.7%, fluid resuscitation – 10.1%). In bivariable analysis, prolonged TTA was associated with poor outcome (Odds Ratio (OR) 1.26 [95%Confidence Interval (CI), 1.02,1.49], p=.037). In a multivariable regression model, age, temperature, bacteremia, tWBC count, viral infection, empiric antibiotics, oncologic diagnosis and presentation location were independently associated with the composite outcome, but TTA was not (OR 0.99 [95%CI, 0.73,1.34], p=.96). Conclusions: As a QOC measure, TTA may adequately measure the process of care but is not independently associated with outcome in pediatric cancer patients with FN.