Abstract
BACKGROUND AND OBJECTIVES: There are few Australian data regarding the burden of hospital-onset bloodstream infections (HO-BSIs). To quantify the impact of antimicrobial-susceptible and -resistant HO-BSIs on patient outcomes by augmenting laboratory-based surveillance data. METHODS: We performed a retrospective cohort study at a tertiary referral hospital in Melbourne, Australia, from 2015 to 2020. We linked administrative data with bloodstream infection surveillance data from the Australian Group on Antimicrobial Resistance. We performed cause-specific Cox proportional hazards regression to quantify the impact of HO-BSI on inpatient mortality and discharge alive, with separate models for Enterobacterales, Staphylococcus aureus, Enterococcus species and the non-fermenting Gram-negative bacilli (NFGNB), Pseudomonas aeruginosa and Acinetobacter species, compared to admissions without HO-BSI. Excess length of stay (LOS) was estimated using multistate models. RESULTS: The cohort of 278 984 admissions included 814 (0.3%) HO-BSIs. Enterobacterales were the most frequent pathogens, followed by enterococci, S. aureus and NFGNB (incidence 3.62, 2.34, 1.11 and 0.80 events per 10 000 patient-days, respectively). Both antimicrobial-resistant and -susceptible HO-BSI increased risk of death and LOS compared with admissions without HO-BSI. Antimicrobial-resistant and -susceptible HO-BSIs, respectively, increased LOS by 5.7 days (95% CI: 4.9-6.5) and 4.1 days (95% CI: 3.8-4.5) for Enterobacterales, 4.9 days (95% CI: 4.5-5.4) and 3.1 days (95% CI: 2.6-3.6) for enterococci, and 6.3 days (95% CI: 5.3-7.3) and 9.8 days (95% CI: 9.1-10.5) for S. aureus. CONCLUSIONS: Antimicrobial-susceptible and -resistant HO-BSIs have a substantial impact on patient outcomes. We demonstrated the feasibility of leveraging a national laboratory-based surveillance system to quantify the impact of HO-BSI.