Antimicrobial resistance in the Bacteroides fragilis group: National multicenter survey and Bayesian modelling study, France, 2022–2023

Résumé fourni par la source

OBJECTIVES: Epidemiological studies on Bacteroides fragilis group (BFG) species and their antimicrobial resistance (AMR) profiles remain scarce worldwide and are lacking in France. The aim of this study was to provide comprehensive national data on the distribution of BFG species, isolation sites, AMR and its potential regional disparity. METHODS: A retrospective multicenter study was conducted from January 2022 to December 2023 using routine data from 45 French hospitals. A total of 9458 BFG isolates identified to the species level by MALDI-ToF were included. Antimicrobial susceptibility testing (AST) was performed according to French guidelines using either disc diffusion or gradient diffusion methods. AST results were collected for ten antibiotics: amoxicillin-clavulanate (AMC), piperacillin-tazobactam (PIT), imipenem (IMI), clindamycin (CLI), metronidazole (MET), moxifloxacin, tigecycline, linezolid, rifampicin, and chloramphenicol. A Bayesian statistical model was applied to estimate regional resistance frequencies with corresponding credible intervals. RESULTS: Bacteroides fragilis sensu stricto (B. fragilis) was the most prevalent species (62.4%), followed by Bacteroides thetaiotaomicron (14.2%). Resistance to IMI and MET remained low, while AMC, PIT, and CLI resistance varied markedly across species: 5.2%, 3.6%, and 34.4% for B. fragilis; 18.0%, 50.1%, and 72.9% for B. thetaiotaomicron; and 32.1%, 19.5%, and 75.2% for Parabacteroides distasonis, respectively. Bayesian modelling produced robust regional estimates for B. fragilis, revealing noticeable regional disparities, particularly for AMC, PIT, and MET. CONCLUSION: Bayesian modeling revealed substantial interspecies and interregional differences in BFG resistance, highlighting the importance of species-level identification and localized surveillance to inform empirical antibiotic therapy.

Ce résumé expose les affirmations des auteurs. BNTIC ne l’interprète pas comme une validation indépendante des résultats.