Shortening antibiotic therapy duration for hospital-acquired bloodstream infections in critically ill patients: a causal inference model from the international EUROBACT-2 database.

Summary

In a prospective international ICU cohort (EUROBACT-2), among 550 eligible HA-BSI patients, 7–10 day antibiotic courses were associated with lower 28-day treatment failure (OR 0.64) driven by fewer subsequent infectious complications (OR 0.58), with no difference in mortality or persistent infection. Most infections were catheter-related and caused by Enterobacterales; longer courses were more often used for S. aureus or difficult-to-treat organisms.

Key Findings

• Short-course (7–10 days) therapy reduced 28-day treatment failure (OR 0.64, 95% CI 0.44–0.93) versus long-course (14–21 days).
• Reduction driven by fewer subsequent infectious complications (OR 0.58, 95% CI 0.37–0.91); mortality and persistent infection were similar.
• Most HA-BSIs were catheter-related (33%) and due to Enterobacterales (39%); longer durations were used more often for S. aureus and difficult-to-treat organisms.

Clinical Implications

Consider 7–10 day courses for uncomplicated HA-BSI in ICU when source control is achieved and pathogens are susceptible, while individualizing for S. aureus or difficult-to-treat organisms.

Limitations

• Non-randomized design with residual confounding and indication bias despite IPTW
• Pathogen and source distributions differed between groups; findings apply to carefully selected uncomplicated HA-BSI

Future Directions

Conduct randomized trials and pragmatic multicenter studies to confirm optimal duration by pathogen and source, and integrate biomarker-guided stopping rules.