Risk factors for treatment failure in patients receiving β-lactam/β-lactamase inhibitor combinations for Enterobacteriaceae bloodstream infection: A retrospective, single-centre, cohort study.

• BL/BLI combination was the most common antibiotic regimen used for Enterobacteriaceae bloodstream infection (E-BSI). • All-cause 30-day mortality rate was low (6.8%) in patients receiving appropriate empirical/definitive BL/BLI therapy. • Non-susceptibility to ESCs was not associated with increased 30-day mortality or higher 90-day relapse rate. • In multivariate analysis, the only predictor of relapse was immunosuppression, mainly in patients with higher SOFA score. The aim of this study was to investigate risk factors for treatment failure in patients receiving in vitro-active therapy with β-lactam/β-lactamase inhibitor (BL/BLI) for Enterobacteriaceae bloodstream infection (E-BSI). This was a retrospective, single-centre study of patients diagnosed with E-BSI at an Italian centre over a 4-year period. Exclusion criteria were age <18 years, clinical data unavailable, polymicrobial BSI, failure to receive in vitro-active therapy and death within 72 h from drawing the index blood culture. Patients who received BL/BLI as appropriate empirical and/or definitive therapy for ≥50% of the total treatment duration were selected. The primary endpoint was all-cause 30-day mortality. The secondary endpoint was 90-day relapse. Of 1319 eligible patients, 835 were selected. A total of 714 received BL/BLI as appropriate empirical therapy, of whom 522 remained on BL/BLI as definitive therapy and 192 shifted to another antibiotic for <50% of the treatment duration; 121 received BL/BLI as definitive therapy only. Non-susceptibility to extended-spectrum cephalosporins (NS-ESCs) was detected in 207 episodes (24.8%). All-cause 30-day mortality was 6.8%. In multivariate analysis adjusted for NS-ESC, independent predictors of mortality were Charlson comorbidity index, septic shock, Proteus spp. and CVC-related BSI, whilst urinary source was a protective factor. The 90-day relapse rate was 4.2%. Immunosuppression was the main independent predictor for relapse. BL/BLI was the most common antibiotic administered to patients with E-BSI in this cohort. Among patients appropriately treated with BL/BLI, failure rates were low and were primarily associated with underlying diseases, clinical severity at BSI onset and infection source. [ABSTRACT FROM AUTHOR]