Don't mess with the machine - evaluation of fluconazole susceptibility testing for Candida glabrata using the new VITEK2 AST-YS08 card following species modification

Evaluation of the Vitek 2 system for antifungal susceptibility testing of Candida auris using a representative international panel of clinical isolates: overestimation of amphotericin B resistance and underestimation of fluconazole resistance

Although the Vitek 2 system is broadly used for antifungal susceptibility testing of Candida spp., its performance against Candida auris has been assessed using limited number of isolates recovered from restricted geographic areas. We therefore compared Vitek 2 system with the reference Clinical and Laboratory Standards Institute (CLSI) broth microdilution method using an international collection of 100 C. auris isolates belonging to different clades. The agreement ±1 twofold dilution between the two methods and the categorical agreement (CA) based on the Centers for Disease Control and Prevention's (CDC's) tentative resistance breakpoints and Vitek 2-specific wild-type upper limit values (WT-ULVs) were determined. The CLSI-Vitek 2 agreement was poor for 5-flucytosine (0%), fluconazole (16%), and amphotericin B (29%), and moderate for voriconazole (61%), micafungin (67%), and caspofungin (81%). Significant interpretation errors were recorded using the CDC breakpoints for amphotericin B (31% CA, 69% major errors; MaEs) and fluconazole (69% CA, 31% very major errors; VmEs), but not for echinocandins (99% CA, 1% MaEs for both micafungin and caspofungin) for which the Vitek 2 allowed correct categorization of echinocandin-resistant FKS1 mutant isolates. Discrepancies were reduced when the Vitek 2 WT-ULV of 16 mg/L for amphotericin B (98% CA, 2% MaEs) and of 4 mg/L for fluconazole (96% CA, 1% MaEs, 3% VmEs) were used. In conclusion, the Vitek 2 system performed well for echinocandin susceptibility testing of C .auris. Resistance to fluconazole was underestimated whereas resistance to amphotericin B was overestimated using the CDC breakpoints of ≥32 and ≥2 mg/L, respectively. Vitek 2 minimun inhibitory concentrations (MICs) >4 mg/L indicated resistance to fluconazole and Vitek 2 MICs ≤16 mg/L indicated non-resistance to amphotericin B.

Comparison of Six Antifungal Susceptibilities of 11 Candida Species Using the VITEK2 AST-YS08 Card and Broth Microdilution Method

We used a Vitek 2 AST-YS08 (YS08) system and the broth microdilution method (BMD) adopted by the Clinical and Laboratory Standards Institute (CLSI) to compare the susceptibility of 184 isolates of 11 Candida species to fluconazole, voriconazole, micafungin, caspofungin, amphotericin B, and flucytosine. In Candida albicans, the categorical agreement (CA) was 79.2%, 91.7%, 95.8%, and 95.8% for fluconazole, voriconazole, micafungin, and caspofungin, respectively. About 12.5% and 4.2% of very major errors were detected for fluconazole and voriconazole, respectively. C. glabrata showed excellent essential agreements (EAs) (>90%) for azoles but different MIC distributions for fluconazole and caspofungin. The CA between BMD fluconazole MICs and YS08 voriconazole MICs by the method-specific clinical breakpoint (CBP) was 90% in C. glabrata. Over 80% of C. glabrata and C. krusei isolates identified as micafungin-susceptible were labeled intermediate or resistant to caspofungin in YS08. In C. parapsilosis, 5.3% of very major errors and 10.5% of minor errors were found, whereas 33.3% of minor errors were observed in C. tropicalis for fluconazole. For C. tropicalis, 13 (61.9%) non-wild type (WT) isolates of fluconazole and 7 (33.3%) non-WTs of voriconazole were classified in YS08 as WT. For C. auris, the EAs were 93.3%, 100%, 82.2%, 97.8%, and 97.8% for fluconazole, voriconazole, micafungin, caspofungin, and amphotericin B, respectively. YS08 showed comparable results to the BMD. However, considering the lower YS08 fluconazole MIC results compared with BMD in Candida species and YS08 caspofungin results in C. glabrata and C. krusei, improvements are needed. IMPORTANCE The new Vitek 2 AST-YS08 (YS08) card has been updated to reflect the recently revised Clinical and Laboratory Standards Institute (CLSI) guideline. In this study, antifungal drug susceptibility tests were performed using the YS08 card and compared with the CLSI broth microdilution (BMD) method. In conclusion, YS08 showed similar results to BMD, including with C. auris. However, about 12.5% and 4.2% of major errors were detected for fluconazole and voriconazole, respectively, in C. albicans. More than 80% of C. glabrata and C. krusei isolates identified as susceptible to micafungin were labeled moderate or resistant to caspofungin in YS08. The categorical agreement between BMD fluconazole MICs and YS08 voriconazole MICs was 90% by the method-specific CBP of voriconazole, 80% by the current epidemiological cutoff value (ECV) (0.25 μg/mL) of voriconazole, and 85% by the previous ECV (0.5 μg/mL) of voriconazole. Further improvements in YS08 for the detection of fluconazole and echinocandin resistance are thus needed.

Evaluation of Two Commercial Broth Microdilution Methods Using Different Interpretive Criteria for the Detection of Molecular Mechanisms of Acquired Azole and Echinocandin Resistance in Four Common Candida Species

The abilities of the new Vitek 2 AST-YS08 (YS08) and Sensititre YeastOne (SYO) systems to detect the resistances of Candida isolates to azoles and echinocandins were evaluated. In total, 292 isolates, including 28 Candida albicans (6 Erg11 and 2 Fks mutants), 57 Candida parapsilosis (26 Erg11 mutants), 24 Candida tropicalis (10 Erg11 and 1 Fks mutants), and 183 Candida glabrata (39 Pdr1 and 13 Fks mutants) isolates, were tested. The categorical agreements (CAs) between the Clinical and Laboratory Standards Institute (CLSI) method and YS08 fluconazole MICs obtained using clinical breakpoints were 92.4% (C. albicans), 96.5% (C. parapsilosis), and 87.0% (C. tropicalis), and the CAs between the CLSI and SYO MICs were 92.3% (C. albicans), 77.2% (C. parapsilosis), 100% (C. tropicalis), and 98.9% (C. glabrata). For C. glabrata, the CAs with the CLSI micafungin MICs were 92.4% and 55.5% for the YS08 micafungin and caspofungin MICs, respectively; they were 100%, 95.6%, and 98.9% for the SYO micafungin, caspofungin, and anidulafungin MICs, respectively. YS08 does not provide fluconazole data for C. glabrata; the CA with the CLSI fluconazole MIC was 97.8% for the YS08 voriconazole MIC, using an epidemiological cutoff value (ECV) of 0.5 μg/ml. Increased CAs with the CLSI MIC were observed for the YS08 MIC using CLSI ECVs (for fluconazole and C. tropicalis, 100%; for micafungin and C. glabrata, 98.9%) and for the SYO MIC using method-specific ECVs (for fluconazole and C. parapsilosis, 91.2%; for caspofungin and C. glabrata, 98.9%). Therefore, the YS08 and SYO systems may have different abilities to detect mechanisms of azole and echinocandin resistance in four Candida species; the use of method-specific ECVs may improve the performance of both systems.