Abstract
BACKGROUND: Nakaseomyces glabrata, formerly Candida glabrata, is an opportunistic yeast and emerging cause of human infections. The use of broth microdilution (BMD) methodologies for caspofungin (CSP) antifungal susceptibility testing (AFST) against N. glabrata is reported to be prone to high inter-laboratory variation. We aimed to compare CSP MICs of N. glabrata isolates from our institution with those obtained by the Reference Laboratory for the same isolates. METHODS: All clinically significant N. glabrata isolates from 2019 to 2021 inclusive were reviewed. AFST was performed locally using the VITEK2 system with the AST-YS08 card, while E-tests were performed at the Mycology Reference Laboratory (MRL), and agreement between these two methods was evaluated - categorical and essential. RESULTS: Forty-one isolates were reviewed during the study period - 30 from blood cultures, seven from intra-operative theatre specimens and four from sterile site drain fluids. Despite an essential agreement of 100 % within ±2 log(2) dilutions, marked discrepancies were noted in interpretative breakpoints between assays with 17 Minor and 16 Major category errors. Categorical agreement was 19.5 %, with the VITEK2 over-estimating resistance. A Mann-Whitney U-test assessed the relationship of MICs across the AFST modalities, and a statistically significant difference was noted, P<0.01, with a higher mean rank for VITKEK2 outputs. CONCLUSION: While the VITEK2 system is highly applicable, its performance for CSP AFST is unreliable and potentially results in the mis-classification of susceptible isolates as highlighted in our study. The use of VITEK2 AST-YS08 micafungin as a sentinel echinocandin should be explored and/or the evaluation of CSP-specific E-tests as utilized by the MRL. These methods appear more consistent and less prone to the variation seen with BMD for CSP.