Temporal dynamics and risk factors for bloodstream infection with extended-spectrum β-lactamase-producing bacteria in previously-colonized individuals: National population-based cohort study

Abstract

Background. Little is known of the long-term risks of bloodstream infection (BSI) with extended spectrum β-lactamase-producing Enterobacteriaceae (EPE) in previously-colonized individuals. We investigated EPE-BSI risks and associated risk factors during 6 years following EPE colonization. Methods. We performed a population-based cohort study in Sweden using national health registers. Subjects were followed from their first EPE finding in feces (n = 5513) or urine (n = 17 189). The effects of co-morbidity, sociodemography, and outpatient antibiotic dispensation on EPE-BSI risks were assessed. The EPE-BSI risks were compared to those of 45 161 matched population- based reference subjects. Results. The cumulative 6-year EPE-BSI incidences were 3.8%, 1.6%, and 0.02% in the urine, feces, and reference cohorts, respectively. The incidences decreased exponentially during the first 6-12 months. Among EPE-exposed subjects, urological disorders were associated with the highest adjusted cause-specific hazard ratio (aCSHR) for subsequent EPE-BSIs (3.40, 95% confidence interval 2.47-4.69). The aCSHRs were between 1.62-2.20 for male sex, immunosuppression, diabetes, malignancy, lung disease, baseline urine source, and Klebsiella pneumoniae, compared to the Escherichia coli baseline sample. Antibiotics with selective activity against gram-negative bacilli-but mostly not EPE (trimethoprim-sulfamethoxazole, fluoroquinolones, oral cephalosporins, and penicillins with extended spectrums)-and pivmecillinam were associated with doubled EPE BSI risk during the 3 months after antibiotic dispensation in EPE-colonized subjects. Conclusions. EPE in urine or feces is a substantial risk factor for subsequent EPE-BSIs, but the risk declines rapidly during the first year after detection. In EPE-colonized individuals, specific risk factors can be used to identify subgroups for targeted interventions, such as eradication therapy.