Résumé
BACKGROUND: Poor and delayed microbiological documentation of episodes of febrile neutropenia (EFN) deserves improvement. We assessed the impact of a new blood culture (BC) sampling protocol to optimize the diagnosis of bloodstream infection during EFN, compared with standard of care protocol.METHODS: This pre/post intervention included patients who presented an EFN in a pediatric hematology-oncology center. Data were compared between 1-year periods P1 (110 EFN, 53 patients) and P2 (124 EFN, 53 patients). Pre-intervention settings were 1−2 mL of blood cultured per BC set and several samplings over days (multisampling strategy) during period P1 vs. one unique early sampling of a large volume of blood (0.5–60 mL) depending on patient weight during period P2 (single-sampling weight-adapted strategy). Microbial detection and time-to-diagnosis were evaluated.RESULTS: Seventeen EFNs were microbiologically documented in P1 (15.5%) and 26 in P2 (21%). The rate of positive BC sets increased during P2 (10.4% vs. 5.8%). All cases of bacteremia were documented by BC drawn during the first 4 days of fever, and during P2 by samples obtained on the first day of fever.CONCLUSIONS: Bacteremia detection was improved. This proof-of-concept study shows benefits of combining the single-sampling strategy with large weight-adapted blood sampling strategy during EFN.