Approximately 700,000 cases of invasive candidiasis were estimated annually, with approximately >350,000 of them resulting in death¹. Candida albicans is the main culprit patho- gen of candidiasis. However, nonalbicans species such as C. glabrata (Nakaseomyces glabratus), C. parapsilosis, C. tropi- calis, C. guilliermondii (Meyerozyma guilliermondii), C. krusei (Pichia kudriavzevii), and C. auris (Candidozyma auris) have become more common over the last two decades. They were identified as the important causes of candidemia and are likely more resistant to antifungal drugs^2-6.

The effective management of candidiasis requires early diagnosis and timely, appropriate medical treatment. Con- sequently, antifungal susceptibility testing (AFST) has to be performed routinely. Broth microdilution (BMD) is the reference AFST method, and two BMD protocols have been developed by two groups—the Clinical and Laboratory Stand- ards Institute (CLSI) and the European Committee on Anti- microbial Susceptibility Testing (EUCAST). Proposed protocols include the CLSI M27 and EUCAST methods for pathogenic Candida species^7,8. These reference methods are expensive, labor-intensive, and time-consuming; therefore, they are not suitable for routine use in a laboratory. To overcome these shortcomings, the disk diffusion (DD) test or the Kirby-Bauer method⁹ has been developed as the CLSI M44-A docu- ment^10,11. The test was performed using Mueller-Hinton agar with 2% glucose and 0.5 μg/ml methylene blue dye (MHA/ GMB). A breakpoint is interpreted by the inhibition-zone diameter. The DD test, which was suggested as an alternative, is a simple, inexpensive, and accurate method for AFST^12-16.

Besides, some commercial kits have also been developed for AFST, such as the Thermo Scientific^TM Sensititre^TM YeastOne^TM YO10 AST Plate (SYO) (Thermo Scientific Inc., UK) and the Vitek 2 AST-YS09 (VT) (bio bioMérieux, FR). The SYO that uses BMD is a 96-well plate containing serial dilutions of nine antifungal agents. Resazurin (Alamar Blue) was added to each well to provide a colorimetric growth indicator. The minimal inhibitory concentration (MIC) is deter- mined by a change in color from blue to pink in the well after a 24-h incubation period. Many studies have demonstrated a high interlaboratory reproducibility using this kit^17-28. The VT is a card panel containing 64 wells with dried concentrations of six antifungal drugs and a medium. The MIC values were determined using spectrophotometry of the automated Vitek 2 system. The kit was used to detect the MIC of clinically relevant Candida species^28-38 , and many studies reported high agreement with the CLSI and EUCAST reference methods. The results of % agreement values between SYO, VT, and DD, with the CLSI, are reviewed and concluded in Table 1.

In this study, AFST for six Candida species was performed using SYO, VT, and DD tests. The susceptibility profiles were investigated and assessed in an attempt to find the most appropriate AFST for use in a routine laboratory.

1. Candida strains

From January 2021 to February 2023, 40 Candida strains had been isolated from the clinical specimens of patients with candidiasis, and those patients were admitted at the Srinagarind Hospital, Faculty of Medicine, Khon Kaen University, Thailand. These strains included 30 blood samples, two from bile fluid, peritoneal dialysis fluid, and Jackson drain fluid, and one from lymph node tissue, pleural fluid, percutaneous drainage, and thrombus. All strains were identified using a MALDI-TOF Autoflex maX mass spectrometer (Bruker, Bremen, DE). The identification results showed 15 strains of C. albicans, 16 of C. tropicalis, 3 of C. glabrata, 3 of C. parapsilosis, 2 of C. guilliermondii, and 1 of C. krusei. All of them were grown on sheep blood agar (SBA) and Sabouraud dextrose agar and stored in 20% glycerol at -70℃ for long-term preservation.

C. krusei ATCC 6258 and C. albicans ATCC 90028 were used in the study as controls for the AFST according to the CLSI recommendation.

2. Antifungal susceptibility testing

Each Candida strain was cultured on SBA at 37℃ for 48 h. The inoculum was prepared by suspending yeast colonies in sterile 0.85% saline solution to no. 0.5 McFarland for the SYO and DD methods, and no. 1.8-2.2 McFarland for VT.

The SYO kit contains the YeastOne YO10 AST plate filled with nine antifungal drugs at different concentrations. The drug concentrations were as following; 0.125-8 mg/mL of amphotericin B (AMB), 0.125-256 mg/mL of fluconazole (FLU), 0.015-16 mg/mL of itraconazole (ITR), 0.008-8 mg/mL of posaconazole (POS), voriconazole (VOR), caspofungin (CAS) and micafungin (MICA), 0.015-8 mg/mL of anidulafungin (ANI), and 0.06-64 mg/mL of flucytosine (5FC). The procedure was performed according to the manufacturer's guidelines. In brief, the 0.5 McFarland yeast suspension was diluted in the YeastOne medium and automatically filled into the YeastOne YO10 AST plate by the YeastOne machine. The plate was incubated at 37℃ for 24 h, and the MIC results were visually read. The breakpoints were interpreted by the CLSI^11,39-44 criteria as shown in Table 2.

The VT kit included the Vitek AST-YS09 card containing six antifungal agents, such as 0.25-8 mg/mL of AMB, 0.5-64 mg/mL of FLU, 0.125-8 mg/mL of VOR, 0.125-8 mg/mL of CAS, 0.06-8 mg/mL of MICA, and 1-64 mg/mL of 5FC. The procedure was performed according to the manufacturer's guidelines. The 1.8-2.2 McFarland yeast inoculum was diluted  in normal saline and automatically filled into the Vitek AST-YS09 card by the Vitek 2 system, where the MIC results were also automatically measured and interpreted for the break- points as shown in Table 2.

The DD test was performed according to the CLSIM44-A¹⁰ by using a sterile swab soaked with a yeast inoculum suspension to smear and spread the yeast cells over the entire surface of an MHA/GMB agar plate. After letting the plate dry for 5-10 min, five antifungal disks (Liofichem S.r.l, IT), including 10-μg AMB, 5-μg CAS, 25-μg FLU, 8-μg ITR, and 1-μg VOR, were placed on the surface of the medium. The plate was incubated at 37℃ for 20-24 h. The diameters of the inhibition zones were measured for breakpoint inter- pretation^10,11,44 as shown in Table 2.

3. Discrepancy analysis among each AFST

Each method was assigned as a reference value for com- paring its breakpoints with the others. The following par- ameters, such as very major error (VME), major error (ME), minor error (mE), and categorical agreement (CA), were investigated, counted, and calculated. "VME" is an error in  which a strain is resistant according to the reference but sensitive using the test method (false sensitive). "ME" refers to an error in which a strain is sensitive according to the reference but resistant according to the test method (false resistance), and "mE" indicates an error in which a strain is resistant or sensitive according to the reference but inter- mediate by the test method, or vice-versa. "CA" is the pro- portion of test strains showing the same breakpoint when using the reference and the test method²⁸. According to recommendations, the acceptable VME, ME, and mE rates should be <3%, <3%, and <10%, respectively^10,43,44.

1. Antifungal susceptibility testing

Using all three methods, the reference strains, C. krusei ATCC 6258 and C. albicans ATCC 90028, revealed acceptable MIC values of the antifungal drugs tested.

To be confirmed by the three tests, almost all 15 strains of C. albicans were sensitive (S)/wild-type (WT) strains against nine antifungal drugs. All 16 strains of C. tropicalis and two strains of C. guilliermondii were S/WT to AMB, 5FC, and all echinocandin. All strains of C. glabrata and C. krusei and two strains of C. parapsilosis were S/WT to AMB and 5FC. The resistant (R)/nonwild-type (NWT) phenotype was found in two strains of C. tropicalis strains. They were resistant to FLU and VOR.

The SYO found that all C. tropicalis strains were NWT against POS and one NWT-strain against ITC. DD could detect the greatest number of resistant strains in other drugs, viz. 10 of FLU and 9 of VOR. All the details of the susceptibility profiles are shown in Table 3.

2. Discrepancy analysis among each AFST

No discrepancy was found between SYO and VT in 5FC and MICA. The discrepancy among SYO, VT, and DD was found in AMB, CAS, FLU, VOR, and ITR. The errors, viz. VME, ME, mE, and %CA, were counted and calculated as shown in Table 4. The greatest number of errors from the three tests was found in FLU, such as 9 VME, 9 ME, and 24 mE. The errors of DD as the reference were the most numerous; i.e., 22 VME, 1 ME, and 30 mE, but the CA was 84.38% to VT and 86.00% to SYO. When assigned SYO and VT as a reference, only 3 mE with 98.75% CA were found between them. Candida tropicalis strains showed discrepancy break- points in three drugs (FLU, VOR, and ITR), whereas C. albicans strains showed discrepancy breakpoints only in CAS.