False amoxicillin/clavulanic acid susceptibility in Bacteroides fragilis using gradient strip tests

BACKGROUND: Repeatedly, too low MIC results were obtained in Bacteroides fragilis quality assessment strains, using gradient strip tests with a ratio of amoxicillin:clavulanic acid of 2:1. We aimed to find the most accurate available gradient strip tests for susceptibility testing of amoxicillin/clavulanic acid in B. fragilis in comparison with agar dilution with EUCAST methodology and breakpoints.

METHODS: Twenty-seven clinical B. fragilis isolates were investigated using gold standard EUCAST amoxicillin/clavulanic acid agar dilution (fixed clavulanic acid concentration at 2 mg/L, with increasing amoxicillin concentrations) as well as three commercial gradient strip tests: XL (ratio), AUG (ratio) or AMC (fixed concentration).

RESULTS: Using agar dilution (fixed concentration), 19 isolates were susceptible, 1 isolate was susceptible increased exposure (I) and 7 isolates were resistant. Categorical agreement of the gradient strip tests with agar dilution (fixed concentration) was 70% for XL (ratio), 71% for AUG (ratio) and 89% for AMC (fixed concentration). Very major error rates in comparison with agar dilution (fixed concentration) were 100%, 0%, and 0%, respectively.

CONCLUSIONS: EUCAST breakpoint usage in amoxicillin/clavulanic acid susceptibility tests for B. fragilis should be accompanied by EUCAST methodology. When using alternative methods such as gradient strip tests, a higher degree of alignment with EUCAST methodology, such as using fixed clavulanic acid concentrations, improves precision.