Objective: To determine whether hospitalization in an intensive care unit (ICU) with a routine multidrug resistant organisms (MDRO) screening policy increases appropriate initial antibiotic therapy (IAT) for treatment of patients with ICU-acquired bloodstream infections (BSIs).

Methods: A sub-analysis of the EUROBACT-2 prospective international cohort study, including adult patients with ICU-acquired BSI. Exposure was admitting ICU's MDRO screening policy, classified as universal on-admission ± weekly screening versus no screening. Primary outcome was appropriate, empiric IAT given within 24 hours of BSI onset, that was concordant with antimicrobial susceptibilities. Associations were evaluated using mixed-effects logistic regression models. Subgroup analyses by BSI source, causative pathogen, local epidemiology and time period were completed.

Results: Overall, 1,800 patients with ICU-acquired BSIs from 304 ICUs were included. Patients were significantly more likely to receive appropriate IAT in the first 24 hours of the BSI onset when hospitalised in MDRO screening versus no screening ICUs (51.8% versus 43.1%; p<0.001). Considering patient demographics, clinical severity, clustering effects, and ICU-level factors the protective effect of MDRO screening remained significant (adjusted odds ratio 1.47, 95% CI 1.09-1.99; p=0.012). This association was strongest in patients with intra-abdominal or respiratory BSI sources, monomicrobial Gram-negative BSIs, ICUs with lower antimicrobial resistance prevalence, in the pre-COVID period and in ICUs performing both on-admission and weekly screening.

Conclusions: More appropriate IAT for ICU-acquired BSIs was given when patients were hospitalised in ICUs practicing routine MDRO screening. These results support the role of regular screening in improving empiric antibiotic decision making for ICU-acquired BSIs.