Escherichia coli bloodstream infections in the western interior of British Columbia, Canada: a population-based cohort study.

Abstract

Our population-based study objectives were to describe characteristics and outcomes of Escherichia coli bloodstream infections (BSIs), and to evaluate factors associated with outcomes. We included incident E. coli BSIs from western interior residents (British Columbia, Canada; 04/2010-03/2020). We obtained data including patient demographics, location of onset, infection focus, Charlson comorbidity index (CCI), antimicrobial resistance, 30-day all-cause mortality and length of hospital stay (LOS). Using multivariable logistic regression models fitted with generalised estimating equations, we estimated factors associated with 30-day mortality and long post-infection LOS (>75th percentile). We identified 1080 incident E. coli BSIs in 1009 patients. The crude incidence and 30-day mortality rates were 59.1 BSIs and 6.8 deaths/100 000 person-years, respectively. The 30-day case fatality risk was 11.5%. Compared to community-acquired E. coli BSIs, either healthcare-associated or nosocomial cases had higher odds of 30-day mortality. Older cases, non-urogenital BSI foci and CCI ⩾ 3 had higher odds of 30-day mortality compared to younger cases, urogenital foci and CCI < 3. In patients that survived to discharge, those with extended-spectrum β-lactamase (ESBL)-producing E. coli BSIs, nosocomial BSIs, and CCI ⩾ 3 had higher odds of long post-infection LOS compared to those with non-ESBL-producing, community-acquired and healthcare-associated, and CCI < 3. There is a substantial disease burden from E. coli BSIs.