Candida auris MIC testing by EUCAST and CLSI broth microdilution, and gradient diffusion strips; to be or not to be amphotericin B resistant?

Objectives: Reported amphotericin B resistance rates for Candida auris vary considerably. This may reflect clinically relevant differences in susceptibility, technical issues with testing or adoption of a clinical breakpoint (BP) that bisects the wild-type population. We compared reference methods and two gradient diffusion strips using a shared C. auris strain collection.

Methods: Forty C. auris strains from nine US states and ≥3 clades were included. Fourteen MIC data sets were generated using EUCAST E.Def 7.4, CLSI M27Ed4, Etest and MTS (Liofilchem) strip MICs. MICs ≤1 mg/L were classified as susceptible.

Results: EUCAST and CLSI amphotericin B MIC testing were robust across included method variables. The modal MIC was 1 mg/L, distributions unimodal and narrow with similar GM-MICs (0.745-1.072); however, susceptibility classification varied (0-28% resistance). Gradient diffusion strip testing resulted in wider and bimodal distributions for 8/9 data sets. If adopting, per manufacturer's protocol, double inoculation for the Etest method, the modal MIC increased to 2-4 mg/L and resistance rates to 45-63% versus 25-30% with the single inoculation. The EUCAST, CLSI, Etest and MTS strip MICs correlated to the OD of drug-free control EUCAST wells suggesting that some isolates grew better than others and that this was associated with MIC.

Conclusions: The EUCAST and CLSI MIC results were in close agreement, whereas the strip test showed wider and bimodal distributions with reader to reader and centre to centre variation. Our study adds to the concern for commercial MIC testing of amphotericin B against C. auris and suggest the current breakpoint leads to random susceptibility classification.