Follow-up blood cultures in Gram-negative bloodstream infections: predictors of collection and association with mortality-a two-centre cohort study.

INTRODUCTION: Given the high morbidity and mortality of Gram-negative bloodstream infections (GN-BSI), follow-up blood cultures (FUBCs) have received increasing attention in clinical practice. However, evidence on their impact remains conflicting. This study assessed the association of FUBCs with 90-day all-cause mortality, accounting for immortal time bias. METHODS: This retrospective cohort study included 637 hospitalized adults with GN-BSI from two Swiss tertiary-care hospitals (April 2022-April 2024). FUBCs were defined as blood cultures obtained after availability of the index culture result up to 168 h after collection. Endpoints included predictors of FUBC collection and association with 90-day mortality. Sensitivity analyses evaluated FUBCs collected 48-120 h after the index culture. RESULTS: Mean age was 71.3 years and 46% were female. Most infections were community-acquired (81%), predominantly of genitourinary origin (53%). Overall, 90-day mortality was 12.6%. In 83.5% of FUBCs, no pathogen was isolated, and persistent GN-BSI was identified in 26/33 positive FUBC (13%). FUBC collection was associated with longer antibiotic therapy (median 14 vs. 10 days; p < 0.001), hospital stay (median 10 vs. 7 days; p < 0.001) and higher 90-day mortality (18 vs. 10%). Predictors of FUBC collection included intensive-care unit admission, cardiac devices or intravascular catheters, advanced chronic kidney disease, leukaemia, paralysis, hospital-acquired infection, COPD, and non-urinary/non-biliary sources. Regarding the primary outcome, FUBC collection was not associated with mortality after adjustment for cofounding variables (HR 1.41 [95% CI 0.89-2.23]. However, positive FUBCs were found to be an independent predictor of increased 90-day mortality. CONCLUSION: FUBCs were markers of disease severity but were not independently associated with mortality, while being linked to prolonged antibiotic use and hospitalization.