Reporting of antibiotic resistance in randomized controlled trials of new antibiotics: a systematic review.

BACKGROUND: Planning and reporting of antibiotic resistance at baseline and as an outcome in randomized controlled trials (RCTs) of new antibiotics is important for trial interpretation. OBJECTIVES: To assess the planning and reporting of antibiotic susceptibility testing (AST) in recent randomized controlled trials of new antibiotics. METHODS: Data sources: MEDLINE, EMBASE, Web of Science, Cochrane Central, and Ai2 Paper Finder searched from January 2017 to July 2025, with no language restrictions; complemented by backward and forward citation screening for secondary publications and registry results, and the WHO antibacterial pipeline. STUDY ELIGIBILITY CRITERIA: RCTs of new antibiotics listed in phase 2/3 of the 2025 WHO antibacterial pipeline. PARTICIPANTS: Adults and children with bacterial infections. INTERVENTIONS: New antibiotics versus any comparator.Assessment of risk of bias: Cochrane RoB 2.Methods of data synthesis: Four pre-specified resistance domains were extracted - real-time AST, central-laboratory baseline AST, AST-stratified primary efficacy outcome assessment, and assessment of resistance acquisition during follow-up. We calculated proportions of trials and arms with planning and reporting of each domain. RESULTS: Fifty-seven RCTs evaluating 23 new antibiotics across 119 arms were included. Real-time baseline AST was reported in 17/57 (29.8%) trials, including 6/62 (9.7%) intervention and 17/57 (29.8%) comparator arms. Central-laboratory baseline AST was reported in 48/57 (84.2%). The primary efficacy outcome was stratified by baseline AST in 36/48 (75%) with available baseline AST data. Resistance acquisition was reported in 22/57 (38.6%) based on clinical infections; none of the trials conducted surveillance for resistance development. Only 2/57 trials planned, and 9/57 reported, all four resistance domains. CONCLUSIONS: In RCTs of antibiotics approved or in late-stage development since 2017, resistance assessment remained inconsistently planned and reported. Future trials should provide local laboratories with real-time AST tools, pre-specify resistance-stratified analyses, and assess resistance acquisition through routine surveillance samples.