Antimicrobial susceptibility testing of Bacteroides species by disk diffusion: The NordicAST Bacteroides study.

OBJECTIVES

Antimicrobial susceptibility testing (AST) of anaerobic bacteria has until recently been done by MIC methods. We have carried out a multi-centre evaluation of the newly validated EUCAST disk diffusion method for AST of Bacteroides spp.

METHODS

A panel of 30 Bacteroides strains was assembled based on reference agar dilution MICs, resistance gene detection and quantification of cfiA carbapenemase gene expression. Nordic clinical microbiology laboratories (n = 45) performed disk diffusion on Fastidious Anaerobe Agar with 5% mechanically defibrinated horse blood (FAA-HB) for piperacillin-tazobactam, meropenem and metronidazole.

RESULTS

A total of 43/45 (95.6%) laboratories carried out disk diffusion per protocol. Intraclass correlation coefficients were 0.87 (0.80-0.93) for piperacillin-tazobactam, 0.95 (0.91-0.97) for meropenem and 0.89 (0.83-0.94) for metronidazole. For metronidazole, one media lot yielded smaller zones and higher variability than another. Piperacillin-tazobactam and meropenem zone diameters correlated negatively with cfiA expression. A meropenem zone diameter of <28 mm in B. fragilis indicated presence of cfiA. Piperacillin-tazobactam had the most false susceptible results. Categorical errors for this antimicrobial were particularly prevalent in cfiA-positive strains, and piperacillin-tazobactam had the highest number of comments describing zone reading difficulties.

CONCLUSIONS

Inter-laboratory agreement by disk diffusion was good or very good. The main challenges were media-related variability for metronidazole and categorical disagreement with the reference method for piperacillin-tazobactam in some cfiA-positive strains. An area of technical uncertainty specific for such strains may be warranted.