Abstract
Candida auris is a multidrug-resistant yeast that causes a wide spectrum of infections, especially in intensive care settings. We investigated C. auris prevalence among 102 clinical isolates previously identified as Candida haemulonii or Candida famata by the Vitek 2 system. Internal transcribed spacer region (ITS) sequencing confirmed 88.2% of the isolates as C. auris, and matrix-assisted laser desorption ionization–time of flight mass spectrometry (MALDI-TOF MS) easily separated all related species, viz., C. auris (n = 90), C. haemulonii (n = 6), C. haemulonii var. vulnera (n = 1), and Candida duobushaemulonii (n = 5). The in vitro antifungal susceptibility was determined using CLSI broth microdilution (CLSI-BMD), the Vitek 2 antifungal susceptibility test, and the Etest method. C. auris isolates revealed uniformly elevated fluconazole MICs (MIC50, 64 μg/ml), and an alarming percentage of isolates (37%) exhibited elevated caspofungin MICs by CLSI-BMD. Notably, 34% of C. auris isolates had coexisting elevated MICs (≥2 μg/ml) for both fluconazole and voriconazole, and 10% of the isolates had elevated coexisting MICs (≥2 μg/ml) to two additional azoles, i.e., posaconazole and isavuconazole. In contrast to reduced amphotericin B MICs by CLSI-BMD (MIC50, 1 μg/ml) for C. auris, elevated MICs were noted by Vitek 2 (MIC50, 8 μg/ml), which were statistically significant. Candida auris remains an unnoticed pathogen in routine microbiology laboratories, as 90% of the isolates characterized by commercial identification systems are misidentified as C. haemulonii. MALDI-TOF MS proved to be a more robust diagnostic technique for rapid identification of C. auris. Considering that misleading elevated MICs of amphotericin B by the Vitek AST-YS07 card may lead to the selection of inappropriate therapy, a cautionary approach is recommended for laboratories relying on commercial systems for identification and antifungal susceptibility testing of rare yeasts.
INTRODUCTION
In recent years, two species, namely, Candida pseudohaemulonii and Candida auris, which are phylogenetically closely related to Candida haemulonii in the Metschnikowiaceae clade, have been described (1). The yeast C. auris, isolated from the external ear canal of a Japanese patient, was described as a new species in 2009 (2). This pathogen was recently recognized as an emerging multidrug-resistant (MDR) yeast that can cause a wide spectrum of infections, ranging from fungemia to deep-seated infections, especially in intensive care settings (3–8). Candida auris is reported to be misidentified as C. haemulonii, Candida famata, and Rhodotorula glutinis by commercial identification systems, such as Vitek 2 and API20C-AUX, and exhibits a unique susceptibility profile (5–8). Notably, the potential of clonal transmission of C. auris, highly elevated MICs to fluconazole, and reduced susceptibility to voriconazole, caspofungin, and flucytosine are matters of serious concern (7–9). Therefore, accurate identification is important, because treatment strategies are often directed by species characterization of Candida. Further, a recent report of misleadingly high MICs of amphotericin B and caspofungin using the commercially available Vitek 2 automated system in a patient with pericarditis due to C. auris highlights issues of inappropriate treatment strategies if accurate susceptibility testing is not available (4). In the present era, molecular techniques and a growing database of fungal genome sequences have facilitated the reliable identification of phylogenetically related and phenotypically identical species. Herein, we investigated the prevalence of C. auris in a collection of clinical isolates previously phenotypically identified as C. haemulonii/C. famata, using internal transcribed spacer region (ITS) sequencing and evaluated matrix-assisted laser desorption ionization–time of flight mass spectrometry (MALDI-TOF MS) for rapid and reliable identification of this yeast. Further, their in vitro antifungal susceptibilities to the 10 antifungals were determined using the CLSI broth microdilution (CLSI-BMD) method, and data obtained by CLSI were compared with those obtained by the commercial Vitek 2 system and the Etest method.
(Part of this study was presented at the Interscience Conference on Antimicrobial Agents and Chemotherapy [54th ICAAC], M-1203, slide session, Washington, DC, USA [10].)
MATERIALS AND METHODS
Isolates.
A total of 102 clinical isolates, from individual patients, previously identified as Candida haemulonii/C. famata by the Vitek 2 compact system (bioMérieux, Marcy I'Etoile, France) in 4 tertiary care hospitals in Delhi, North India, and a single center in Kochi, Kerala, South India, from 2010 to 2014 were included. Three centers were about 1,000-bed general hospitals, and the remaining 2 were a pediatric hospital and a transplant center. The pediatric hospital had neonatal intensive care units and surgical intensive care facilities. The isolates were mainly from patients with candidemia (blood; n = 78), and other specimens from invasive Candida infections included gangrenous tissue (n = 4), pleural fluid (n = 6), and peritoneal fluid (n = 7). Also, 7 isolates from urine (n = 4) and sputum (n = 3) specimens from immunocompromised patients were included. The control and type strains of three C. auris isolates from Korea (KCTC 17809, KCTC 17810) and Japan (DSM 21092T) and one isolate each of C. haemulonii strain CBS 7802 and C. duobushaemulonii strain CBS 7798T were also analyzed.
Phenotypic characterization.
The isolates were identified by standard mycological procedures, including colony color on CHROMagar Candida medium (Difco, Becton Dickinson & Company, Baltimore, MD, USA) and morphology on rice Tween 80 agar. Growth patterns at different temperatures, 37°C, 42°C, and 45°C, were also observed (1). Additionally, the assimilation profile of all yeast isolates was done by commercially available API strips (ID32C; bioMérieux, Marcy I'Etoile, France), which were read and interpreted at 48 h.
Sequencing of ITS region.
Genomic DNA was extracted from all test isolates along with reference strains as described by Xu et al. (11). DNA was amplified and sequenced using the ITS-1 (5′-TCCGTAGGTGAACCTTGCGG-3′) and ITS-4 (5′-TCCTCCGCTTATTGATATGC-3′) primers, which amplify the ITS region of the ribosomal subunit (8). Sequences were aligned, and GenBank Basic Local Alignment Search Tool (BLAST) searches were performed for species identification. For phylogenetic analyses, the ITS gene sequences of the C. auris, C. haemulonii, and C. duobushaemulonii isolates were aligned with the ClustalW program (version 1.82), and the final alignments were edited manually. A neighbor-joining (NJ) tree based on ITS gene sequences using 2,000 bootstrap replications was generated using MEGA version 5 (12). The sequences of the reference/type strains of C. auris from Japan (JCM 15448T) and Korea (KCTC 17809 and KCTC 17810), along with C. haemulonii (CBS 5150, Portugal; CBS 7801, United States), C. haemulonii var. vulnera (CNMCL-7462, Spain), and C. duobushaemulonii (CBS 7799, USA), were retrieved from GenBank and included for the analysis.
MALDI-TOF MS.
The ethanol-formic acid extraction procedure was followed according to the manufacturer's protocol for the identification of yeast isolates (13). The spectra were analyzed using the Flex Control 3.1 software (Bruker Daltonics, Inc., Billerica, MA, USA) and MALDI Biotyper OC version 3.1 (Bruker Daltonics, Bremen, Germany). Score values were analyzed as per manufacturer recommendations: a score of ≥2 indicated confidence to the species level, 1.7 to 1.99 indicated confidence to the genus level, and <1.7 indicated no identification.
MALDI data analysis.
The MALDI Biotyper version 3 database contains spectra of 3 strains of C. auris, two from Korea (KCTC 17809 and KCTC 17810) and a type strain from Japan (DSM 21092T). For phylogenetic analysis, spectra of 90 C. auris isolates were added manually to the library for the creation of a score-oriented dendrogram in Biotyper as described previously for Aspergillus species (14). The mass spectra of each quadruplicate of the respective isolates with a score value of >2 were considered for dendrogram preparation. Additionally, available spectra of reference strains of C. auris from Japan (DSM 21092T) and Korea (KCTC 17809 and KCTC 17810) and of C. haemulonii (CBS 5149T and CBS 5150), C. duobushaemulonii (CBS 7799 and CBS 7800), and C. pseudohaemulonii (CBS 10004 and CBS 12453T) in the database were imported in the software for the analysis of the dendrogram. The dendrogram was generated by using the respective functionality of the MALDI Biotyper 3.1 offline client. The spectra of all the isolates tested were analyzed by a score-oriented dendrogram using an arbitrary distance level of 1,000 as the cutoff.
AST. (i) CLSI-BMD method.
Antifungal susceptibility testing (AST) was carried out using the Clinical and Laboratory Standards Institute broth microdilution method (CLSI-BMD), following the M27-A3 guidelines (15). Antifungals tested were amphotericin B (AMB; Sigma, St. Louis, MO, USA), fluconazole (FLU; Pfizer, Groton, CT, USA), itraconazole (ITC; Lee Pharma, Hyderabad, India), voriconazole (VRC; Pfizer), posaconazole (POS; Merck, Whitehouse Station, NJ, USA), isavuconazole (ISA; Basilea Pharmaceutica, Basel, Switzerland), flucytosine (5-FC; Sigma), caspofungin (CAS; Merck), micafungin (MFG; Astellas, Toyama, Japan), and anidulafungin (AFG; Pfizer). RPMI 1640 medium with glutamine without bicarbonate (Sigma) buffered to pH 7 with 0.165 mol/liter 3-N-morpholinepropanesulfonic acid (MOPS; Sigma) was used. Drug-free and yeast-free controls were included, and microtiter plates were incubated at 35°C and read visually after 24 h, as validated recently by Pfaller et al. (16, 17). CLSI-recommended Candida krusei ATCC 6258 and Candida parapsilosis ATCC 22019 were used as quality control strains, and two reference strains of C. auris (KCTC 17809 and DSM 21092T) were also included. Excepting AMB, the MIC endpoints for all the antifungals were defined as the lowest drug concentration that caused 50% growth inhibition vis-à-vis the drug-free controls. The MIC for AMB was defined as the lowest concentration at which there was 100% inhibition of growth. The susceptibility for all the isolates was performed by two different personnel on two occasions, which revealed reproducible results.
(ii) Vitek 2 Compact system using an AST-YS07 card.
Susceptibility was determined using an AST-YS07 card, which tests the MIC of 6 antifungals, i.e., FLU, 5-FC, VRC, AMB, CAS, and MFG. All the C. auris isolates were tested as per the manufacturer's instructions. The time of incubation ranged from 18 to 27 h, based on the rate of growth in the drug-free control well, and the results were expressed as MICs in micrograms per milliliter.
(iii) Etest method.
Further, the isolates which revealed >2-fold discrepancies in the antifungal MICs by the above-described two methods were also tested for susceptibility by Etest using Etest Technical Guide 4: Antifungal Susceptibility of Yeasts (AB Biodisk, bioMérieux, Solna, Sweden), as described previously (19, 20). The antifungals tested were AMB, CAS, and VRC. Briefly, the inoculum density of 0.5 × 103 to 2.5 × 103 cells/ml prepared for the CLSI-BMD test was used, and the test medium included RPMI 1640 with 1.5% agar supplemented with glucose (2%) and was buffered to pH 7.0 with MOPS. In addition, AMB was also tested on antibiotic medium 3 (AM3) agar plates. The plates were inoculated by dipping a sterile cotton swab into the inoculum and streaking it across the entire surface of the agar in three directions. The plates were dried for 15 min before the Etest strips (AB Biodisk) were applied and kept at 35°C and read visually after 24 h. The Etest MIC was defined as the drug concentration at which the border of the elliptical zone of complete inhibition intersected the scale on the antifungal test strip.
Statistical analysis.
Statistical analyses were performed with SPSS version 20.0 (SPSS, Chicago, IL, USA). MIC values from the CLSI-BMD, Vitek 2, and Etest methods were assessed by using the Student t test (paired sample). The Etest MIC endpoints, which were in between the 2-fold dilution scale of the CLSI method, were rounded to the corresponding next upper 2-log dilution to simplify comparisons. The discrepancies among MIC endpoints of ±2 dilutions (two wells) were used to calculate the essential agreement (EA).
FKS gene sequencing.
Candida auris isolates with elevated CLSI-BMD MICs of CAS (MICs ≥ 1 μg/ml) were subjected to sequencing of the FKS1 and FKS2 genes. Considering that the genome sequence of C. auris is not yet available, the published mutations in the FKS gene of echinocandin-resistant Candida glabrata isolates were used to analyze the mutations in the FKS gene of C. auris isolates by sequence homology (21, 22). Genomic DNA was amplified and sequenced for hot spot regions of both the genes. The primers were designed based on the C. glabrata FKS1 and FKS2 gene sequences (GenBank accession no. XM_446406 and XM_448401, respectively). The primers for FKS1 (FKS1HSF, 5′-CATTGCTATTTTTCTCAGTCATGC-3′; FKS1HSR, 5′-CCAACGGAAAAGACAGTGTTGA-3′) and FKS2 (FKS2HSF, 5′-CTGTGACATTTTTCATTGCTG-3′; FKS2HSR, 5′-TCCAAGGAGTTAAGATGGAAATACC-3′) were designed using Primer3 software version 4 (http://primer3.ut.ee/). DNA sequences were analyzed with Sequencing Analysis software version 5.3.1 (Applied Biosystems). Consensus sequences were made using BioEdit software (version 7.0.5.3) and were aligned with hot spot FKS regions of reference C. glabrata (GenBank accession no. HM366439 for FKS1 and HM366442 for FKS2).
Nucleotide sequence accession numbers.
The sequences determined in this study were deposited in GenBank under accession no. KF689009 to KF689022, KC692039 to KC692052, and KP862745 to KP862818.
RESULTS
Vitek 2 identified 100 isolates as C. haemulonii (88 with 91 to 98% identity and 12 with a low discrimination profile), whereas the remaining 2 isolates were identified as C. famata (93% identity). Of the 102 isolates, 88.2% (n = 90) were confirmed as C. auris by ITS sequencing. The remaining 12 isolates were identified as C. haemulonii (n = 6), C. haemulonii var. vulnera (n = 1), and C. duobushaemulonii (n = 5). All C. auris (n = 90) isolates showed smooth, white to cream-colored colonies on Sabouraud dextrose agar (SDA), whereas they developed a pink color on CHROMagar Candida medium. Microscopic examination showed ovoid to elongated budding yeast cells occurring singly or in pairs. No pseudohyphae were formed on rice Tween 80 agar after 4 to 8 days of incubation at 28°C. They grew well at 37°C and 42°C. In contrast, C. haemulonii and C. duobushaemulonii isolates revealed pseudohyphae with blastoconidia and did not grow at 42°C. All the C. auris isolates were positive for assimilation of N-acetylglucosamine, succinate, and gluconate, whereas negative results were recorded for C. haemulonii and C. duobushaemulonii.
ITS sequences of the 90 isolates (GenBank accession no. KF689009 to KF689022, KC692039 to KC692052, KP862745 to KP862805, and KP862818) showed 99% homology (query coverage ranging from 98 to 100%) with C. auris isolates in GenBank (accession no. HE797773 and AB375772). However, the ITS sequences of 6 isolates (GenBank accession no. KP862806 to KP862811) showed 100% homology with C. haemulonii (GenBank accession no. KM014586 and JX459689), and 5 isolates (GenBank accession no. KP862813 to KP862817) showed 100% homology with C. duobushaemulonii (GenBank accession no. KM361511 and KJ476202). Also, a solitary isolate (VPCI 715/P/14; GenBank accession no. KP862812) showed 100% homology with C. haemulonii var. vulnera (GenBank accession no. JX459686). The ITS tree yielded 3 distinct clades with a good bootstrap value (99%) and enabled the differentiation of C. auris from C. haemulonii and C. duobushaemulonii strains (Fig. 1). All the Indian C. auris (n = 90) strains exhibited 99 to 100% sequence similarity among themselves. However, C. auris isolates from Japan and Korea formed a separate group away from Indian C. auris isolates in the same clade. The other species of the C. haemulonii complex were well differentiated in 2 separate clades. Candida duobushaemulonii (n = 5) formed clade 2, whereas Candida haemulonii (n = 6) and C. haemulonii var. vulnera (n = 1) formed two groups in clade 3.
FIG 1.
Phylogenetic tree based on partial ITS sequences of Indian C. auris (n = 90), C. duobushaemulonii (n = 5), C. haemulonii (n = 6), and C. haemulonii var. vulnera (n = 1) isolates using neighbor-joining analysis with 2,000 bootstrap replications. Sequences of reference strains of Candida haemulonii (CBS 5150, Portugal; CBS 7801, United States), C. duobushaemulonii (CBS 7799, United States), and C. haemulonii var. vulnera (CNMCL-7462, Spain) along with Japanese (JCM 15448T) and Korean (KCTC-17809 and KCTC-17810) C. auris isolates were retrieved from GenBank for the analysis. Bootstrap values are shown above the branches.
MALDI-TOF MS.
The MALDI-TOF mass spectra obtained for all C. auris isolates were evaluated against the original MALDI Biotyper OC version 3.1 database. The mean MALDI-TOF MS score of the tested C. auris isolates was 2.167 (range, 2.013 to 2.347). Of the 90 C. auris isolates, the majority (n = 77, 85.5%) were identified to the species level at the first attempt with a score value of >2. The remaining isolates (n = 13) were also identified as C. auris but with a score value of <2 and >1.7. These isolates were repeated and revealed high score values in the second attempt. All the C. haemulonii (n = 6) and C. duobushaemulonii (n = 5) strains were identified to the species level (score value, >2). Further, a solitary isolate of C. haemulonii was also identified to the variety level with a score value of >2 (C. haemulonii var. vulnera). The dendrogram clearly revealed separation of members of the Metschnikowiaceae clade in 4 phylogroups (Fig. 2). The mass spectra of the Indian C. auris isolates showed marked similarity, whereas the Japanese (n = 1) and Korean C. auris isolates (n = 2) exhibited variations in mass spectra among themselves and with those of Indian C. auris isolates, resulting in a separate cluster in C. auris (phylogroup 4). The dendrogram generated was in agreement with the phylogenetic NJ tree with ITS sequences.
FIG 2.
Score-oriented dendrogram of the main spectra (MALDI Biotyper 3.1; Bruker Daltonics) by using average linkages clustering the MALDI-TOF spectra of Indian Candida auris (n = 90) along with the C. haemulonii (n = 6), C. duobushaemulonii (n = 5), and C. haemulonii var. vulnera (n = 1) isolates. Candida auris (DSM 21092T, KCTC-17809, and KCTC-17810) and C. pseudohaemulonii (JCM 12453T and CBS 10004), C. haemulonii (CBS 5150 and CBS 5149T), C. haemulonii var. vulnera (CL-7073), and C. duobushaemulonii (CBS 7800 and CBS 7799) were added to make the clustering in MALDI-TOF MS more robust. The isolates were classified into 4 phylogroups. The inset depicts the variation among Indian, Japanese, and Korean C. auris isolates leading to two clusters in phylogroup 4.
In vitro susceptibility and FKS mutation analysis.
The in vitro susceptibility data and the MIC distribution of C. auris isolates using different methods along with essential agreements between the tested methods are presented in Tables 1 and 2.
TABLE 1.
In vitro antifungal susceptibility profile of C. auris, C. haemulonii, and C. duobushaemulonii strains by the CLSI M27-A3 broth microdilution method
| Species tested | MIC parameter | MIC (μg/ml)a |
|||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|
| AMB | ITC | VRC | ISA | POS | FLU | 5-FC | CAS | MFG | AFG | ||
| C. auris (n = 90 isolates) | GM | 0.8 | 0.15 | 1.01 | 0.18 | 0.06 | 36 | 0.5 | 0.58 | 0.11 | 0.23 |
| MIC50 | 1 | 0.125 | 1 | 0.25 | 0.06 | 64 | 0.25 | 0.5 | 0.125 | 0.125 | |
| MIC90 | 4 | 0.5 | 8 | 2 | 2 | 64 | 8 | 1 | 0.25 | 0.5 | |
| MIC range | 0.125–8 | <0.03–2 | <0.03–16 | <0.015–4 | <0.015–8 | 4–>64 | <0.125– >64 | 0.125–8 | <0.015–8 | <0.015–8 | |
| C. haemulonii (n = 7 isolates) | GM | 12.7 | 0.314 | 0.31 | 0.027 | 0.05 | 32 | 2.82 | 0.19 | 0.28 | 0.44 |
| MIC50 | 16 | 0.25 | 0.5 | 0.015 | 0.125 | 64 | 32 | 0.25 | 0.25 | 0.5 | |
| MIC90 | 16 | 0.5 | 2 | 0.25 | 0.25 | 64 | 64 | 1 | 1 | 1 | |
| MIC range | 4–16 | 0.25–0.5 | <0.03–4 | <0.015–0.5 | 0.015–0.25 | 2–>64 | 0.125–>64 | 0.06–1 | 0.125–1 | 0.25–1 | |
| C. duobushaemulonii (n = 5 isolates) | GM | 16 | 0.315 | 0.068 | 0.023 | 0.11 | 6.35 | 0.125 | 0.14 | 0.35 | 0.56 |
| MIC50 | 16 | 0.25 | 0.06 | 0.015 | 0.125 | 8 | 0.125 | 0.125 | 0.5 | 0.5 | |
| MIC90 | 16 | 0.5 | 0.125 | 0.06 | 0.25 | 16 | 0.125 | 0.25 | 1 | 1 | |
| MIC range | >16 | 0.25–0.5 | <0.03–0.125 | <0.015–0.06 | 0.03–0.125 | 1–16 | <0.125 | 0.06–0.25 | 0.125–1 | 0.5–1 | |
AMB, amphotericin B; ITC, itraconazole; VRC, voriconazole; ISA, isavuconazole; POS, posaconazole; FLU, fluconazole; 5-FC, 5-flucytosine; CAS, caspofungin; MFG, micafungin; AFG, anidulafungin.
TABLE 2.
Distribution of MICs of amphotericin B, caspofungin, and voriconazole obtained by 3 different methods for Candida auris (n = 90) strains
| Drug testeda | Test method | No. of isolates at MIC (μg/ml) |
MIC (μg/ml) |
||||||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| <0.03 | 0.03 | 0.06 | 0.125 | 0.25 | 0.5 | 1 | 2 | 4 | 8 | 16 | >16 | MIC50 | MIC90 | ||
| AMB | CLSI-BMD | 2 | 16 | 23 | 35 | 4 | 6 | 4 | 1 | 4 | |||||
| Vitek 2 | 1 | 48 | 41 | 8 | 16 | ||||||||||
| Etest | 5 | 1 | 4 | 25 | 54 | 1 | 0.5 | 1 | |||||||
| CAS | CLSI-BMD | 1 | 29 | 27 | 25 | 1 | 4 | 3 | 0.5 | 1 | |||||
| Vitek 2 | 21 | 34 | 28 | 7 | 0.5 | 4 | |||||||||
| Etest | 9 | 1 | 9 | 22 | 33 | 5 | 4 | 7 | 0.25 | 2 | |||||
| VRC | CLSI-BMD | 1 | 4 | 7 | 8 | 18 | 17 | 18 | 6 | 3 | 3 | 5 | 1 | 8 | |
| Vitek 2 | 3 | 5 | 12 | 28 | 16 | 14 | 10 | 2 | 1 | 4 | |||||
| Etest | 1 | 3 | 2 | 8 | 15 | 36 | 12 | 3 | 7 | 3 | 1 | 16 | |||
AMB, amphotericin B; CAS, caspofungin; VRC, voriconazole.
CLSI-BMD.
FLU exhibited no activity against 89% (n = 80) of C. auris isolates (MIC of 16 to >64 μg/ml), whereas the remaining 10 isolates revealed a MIC of 4 μg/ml. Similarly, an elevated MIC90, i.e., 8 μg/ml, was noted for VRC. Notably, 58% of C. auris isolates (n = 52) showed VRC MICs of ≥1 μg/ml. In contrast, MIC50 values of POS (0.06 μg/ml) and ISA (0.25 μg/ml) were relatively low compared to that of VRC (Table 1). Also, 11% of C. auris isolates revealed MICs of ≥1 μg/ml for both POS and ISA, and a solitary isolate showed a MIC of ≥1 μg/ml only to ISA. All C. auris isolates showed reduced MICs to ITC (geometric mean [GM] MIC, 0.15 μg/ml). Furthermore, C. auris isolates had AMB MIC50 values of 1 μg/ml; however, 15.5% (n = 14) of the isolates revealed MICs of ≥2 μg/ml for AMB. Moreover, elevated GM MICs were observed for CAS (0.58 μg/ml) in comparison to MFG (0.11 μg/ml) and AFG (0.23 μg/ml). Notably, 37% (n = 33) of the C. auris isolates revealed MICs of ≥1 μg/ml to CAS. Also, all the echinocandins had no activity in 8% (n = 7) of the isolates, with MICs ranging from 4 to >8 μg/ml (Table 2). Further, 88% of C. auris isolates had reduced MICs to 5-FC (GM MIC, 0.4 μg/ml), whereas 11 isolates showed highly elevated MICs (≥32 μg/ml). In contrast to C. auris, all the C. haemulonii (n = 7) and C. duobushaemulonii (n = 5) isolates had markedly elevated AMB MICs ranging from 4 to 16 μg/ml. Also, variable FLU MICs were observed for C. haemulonii (MIC range, 2 to >64 μg/ml) and C. duobushaemulonii (MIC range, 1 to 16 μg/ml). However, reduced MICs of VRC (MIC range, 0.03 to 0.5 μg/ml) were noted for both the C. haemulonii and C. duobushaemulonii isolates, except a solitary isolate of C. haemulonii, which showed a MIC of 4 μg/ml. Moreover, both the C. haemulonii and C. duobushaemulonii isolates exhibited reduced GM MICs to ISA (0.027 μg/ml and 0.023 μg/ml), followed by POS (0.05 μg/ml and 0.11 μg/ml) and ITC (0.31 μg/ml). Also, in contrast to C. auris (GM MIC, 0.58 μg/ml), both C. haemulonii (GM MIC, 0.19 μg/ml) and C. duobushaemulonii (GM MIC, 0.14 μg/ml) showed reduced MICs to CAS. However, a wide MIC range (<0.125 to 64 μg/ml) for 5-FC was observed for C. haemulonii, while reduced MICs (GM MIC, 0.125 μg/ml) were found for C. duobushaemulonii (Table 1).
FKS gene sequencing of C. auris isolates with elevated caspofungin MICs (≥1 μg/ml).
Amplification of FKS1 and FKS2 regions generated amplicons of 391 bp and 460 bp, respectively. Mutations reported for caspofungin-resistant C. glabrata were not observed in the FKS1 and FKS2 regions of any of the tested C. auris strains.
Vitek 2.
In contrast to low AMB MICs recorded by CLSI for C. auris, exceptionally elevated AMB MICs (CLSI MIC50 of 1 μg/ml compared to Vitek MIC50 of 8 μg/ml) were noted, which were statistically significant (P < 0.0001). Barring a solitary isolate of C. auris, all the other isolates had MICs of ≥8 μg/ml for AMB (Table 2). Also, the AMB MICs of C. haemulonii and C. duobushaemulonii ranged from 8 to 16 μg/ml, which was in concordance with CLSI. Vitek 2 MIC50 values of VRC (1 μg/ml), CAS (0.5 μg/ml), and MFG (0.125 μg/ml) of C. auris isolates were in 100% agreement with those by the CLSI method. Vitek 2 MIC50 values of FLU (32 μg/ml) and 5-FC (1 μg/ml) were within ±2 dilutions by CLSI-BMD.
Etest.
Similar to CLSI MICs, low MIC50 values of AMB for C. auris isolates were observed by Etest on AM3 medium (0.5 μg/ml) and on RPMI agar (1 μg/ml). Except a solitary isolate, all C. auris isolates showed MICs of ≤1 μg/ml for AMB. The MIC50 (1 μg/ml) of VRC was similar to that by CLSI-BMD. In contrast, Etest CAS MICs were better differentiated than CLSI-BMD MICs, and Vitek showed a wide range from 0.002 to 4 μg/ml. Specifically, 33 C. auris isolates which showed MICs of ≥1 μg/ml for CAS by CLSI revealed highly variable MICs ranging from 0.064 to 4 μg/ml by Etest. Interestingly, 26 of these 33 isolates revealed MICs of ≤1 μg/ml for CAS by Vitek. The remaining 7 isolates had MICs of 4 μg/ml, which were in agreement with the CLSI-BMD MICs. Further, Etest MICs of C. haemulonii and C. duobushaemulonii for AMB, CAS, and VRC were within ±2 dilutions of CLSI MICs.
Agreement between methods.
The essential agreement within ±2 dilutions for the comparison of 24-hour CLSI-BMD with Vitek 2 and Etest results showed 10% and 81% for AMB, 90% and 48% for CAS, and 91% and 79% for VRC, respectively.
DISCUSSION
The present study highlights that Candida auris remains an unnoticed pathogen in routine microbiology laboratories in India, as 90% of the isolates characterized by commercial identification systems misidentify this yeast as C. haemulonii. In the past 5 years, Candida auris has emerged as a significant pathogen in tertiary care general hospitals and a pediatric center in north and south India, representing 8.6% to 30% of cases of candidemia (7, 8). The actual prevalence of C. auris in varied clinical settings in India is unexplored, as the majority of centers do not perform molecular or MALDI-TOF MS-based identification. In this work, a large number of C. auris isolates were tested for antifungal susceptibility with three methods which showed uniform fluconazole resistance and an alarming percentage of isolates (37%) exhibiting elevated caspofungin MICs by CLSI-BMD. Taken together, 10% of isolates showed highly elevated MICs to 4 antifungals drugs (AMB, FLU, CAS, VRC) by the CLSI-BMD method. Notably, 34% of isolates had coexisting elevated MICs for two commonly used azoles, i.e., FLU and VRC (MICs of ≥2 μg/ml), and 10% of the isolates had elevated coexisting MICs (≥1 μg/ml) to two additional azoles, i.e., POS and ISA. Considering the frequent prevalence of MDR strains of C. auris in the intensive care units and other wards of 5 different hospitals in the present series, the accurate identification and antifungal susceptibility testing of this yeast is pertinent for guiding therapy and determining the prognosis in such settings. Also, accurate identification of the cryptic species C. auris is important in assessing the epidemiology and pathogenicity of the disease caused by this underreported pathogen in different geographic areas. In the past 5 years, C. auris fungemia has been reported from South Africa, South Korea, Japan, and India (3, 5–8). All of the reports from these countries confer the major issue of notable elevated MICs for azoles and caspofungin in C. auris and its misidentification by phenotypic methods. The present study employed MALDI-TOF MS, a more robust diagnostic technique, for rapid identification. The strength of the present study is that it developed a comprehensive reference database built with a large number of molecularly characterized C. auris strains from different geographical regions of India to supplement the Bruker Biotyper library, which has a database of only 3 strains from Korea and Japan. Not only was interspecies differentiation well characterized, but also the mass spectra variation at the intraspecies level separated C. auris isolates from India. It is pertinent to mention here that, previously, Indian C. auris isolates have been reported to exhibit differences in biochemical profiles compared to the Japanese and the Korean C. auris isolates (7, 8). Similarly, genotypic variation among C. auris isolates from different geographical regions has also been observed with M13 fingerprinting and amplified fragment length polymorphism analysis (7). It is evident from the present study that the high resolution and discriminatory power of MALDI-TOF MS facilitate differentiation of closely related cryptic species within the Metschnikowiaceae clade (23), which has also been documented previously for Mucorales, particularly the Lichtheimia species (24).
Another issue of concern is the misleading highly elevated MICs of AMB with Vitek automated readings in all C. auris isolates tested. The overall EA between Vitek automated readings and the CLSI-BMD method for AMB was very low (10%). Notably, the reference CLSI-BMD method in the present series showed reduced AMB MICs in 84% of C. auris isolates. Similarly, low AMB MICs (0.25 to 1 μg/ml) by CLSI-BMD were reported for 20 C. auris isolates from South Korea by Shin et al. (25). However, these authors observed a high EA (100%) between the CLSI-BMD and Vitek method for AMB, which is in contrast with the observations in the present study. This deviation could be attributed to the low number of isolates tested (25). Major errors of azole susceptibility in 218 isolates of 5 Candida species using another commercial automated reading system (ATB FUNGUS 3) have been reported recently from China, resulting in pseudohigh rates of antifungal resistance (26). In fact, the erroneously elevated MICs by the Vitek 2 automated reading method not only may lead to inappropriate selection of antifungal therapy but also depict false rates of high antifungal resistance in epidemiological studies. Further, the percentage of C. auris isolates that showed elevated CAS MICs (≥1 μg/ml) by CLSI-BMD in the present series (37%) declined to 12% using the Etest. The lower Etest MIC values than CLSI-BMD MIC values for CAS have also been reported earlier for other Candida species (27). Recently, the performance of the CAS Etest based on the recently revised CLSI breakpoints for Candida isolates showed that 13.1% were misclassified as intermediate or resistant (28). Also, marked interlaboratory variation has been observed with both CLSI-BMD and the EUCAST method for CAS susceptibility (29). In order to investigate the resistance mechanism with respect to elevated CAS MICs, in the present series we attempted to sequence FKS hot spot regions in C. auris isolates using known FKS C. glabrata primers due to a lack of published genomic data for C. auris. Although none of the isolates with elevated CAS MICs harbored mutations reported for echinocandin-resistant C. glabrata, the possibility of other mutations not reported so far could not be ruled out. Future studies on complete genomic analysis are warranted to detect true antifungal resistance in this significant pathogen.
Finally, C. auris is emerging as a serious multidrug nosocomial pathogen in many centers in India, which could be reliably and rapidly identified by MALDI-TOF MS. Notwithstanding the fact that routine laboratories heavily rely on commercial systems for identification and antifungal susceptibility testing for yeasts, a cautionary approach is recommended for isolates showing elevated MICs with these systems.
ACKNOWLEDGMENTS
J.F.M. received grants from Astellas, Basilea, and Merck. He has been a consultant to Astellas, Basilea, and Merck and received speaker's fees from Merck and Gilead. All other authors report no potential conflicts of interest.
The authors alone are responsible for the content and writing of the paper.
REFERENCES
- 1.Lachance MA, Boekhout T, Scorzetti G, Fell JW, Kurtzman CP. 2011. Candida Berkhout, p 987–1278. In Kurtzman CP, Fell JW, Boekhout T (ed), The yeasts, a taxonomic study, 5th ed Elsevier Science B.V., Amsterdam, The Netherlands. [Google Scholar]
- 2.Satoh K, Makimura K, Hasumi Y, Nishiyama Y, Uchida K, Yamaguchi H. 2009. Candida auris sp. nov., a novel ascomycetous yeast isolated from the external ear canal of an inpatient in a Japanese hospital. Microbiol Immunol 53:41–44. doi: 10.1111/j.1348-0421.2008.00083.x. [DOI] [PubMed] [Google Scholar]
- 3.Kim MN, Shin JH, Sung H, Lee K, Kim EC, Ryoo N, Lee JS, Jung SI, Park KH, Kee SJ, Kim SH, Shin MG, Suh SP, Ryang DW. 2009. Candida haemulonii and closely related species at 5 university hospitals in Korea: identification, antifungal susceptibility, and clinical features. Clin Infect Dis 48:e57–e61. doi: 10.1086/597108. [DOI] [PubMed] [Google Scholar]
- 4.Khillan V, Rathore N, Kathuria S, Chowdhary A. 2014. A rare case of breakthrough fungal pericarditis due to fluconazole-resistant Candida auris in a patient with chronic liver disease. JMM Case Rep doi: 10.1099/jmmcr.0.T00018. [DOI] [Google Scholar]
- 5.Lee WG, Shin JH, Uh Y, Kang MG, Kim SH, Park KH, Jang HC. 2011. First three reported cases of nosocomial fungemia caused by Candida auris. J Clin Microbiol 49:3139–3142. doi: 10.1128/JCM.00319-11. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 6.Magobo RE, Corcoran C, Seetharam S, Govender NP. 2014. Candida auris-associated candidemia, South Africa. Emerg Infect Dis 20:1250–1251. doi: 10.3201/eid2007.131765. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 7.Chowdhary A, Sharma C, Duggal S, Agarwal K, Prakash A, Singh PK, Jain S, Kathuria S, Randhawa HS, Hagen F, Meis JF. 2013. New clonal strain of Candida auris, Delhi, India. Emerg Infect Dis 19:1670–1673. doi: 10.3201/eid1910.130393. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 8.Chowdhary A, Anil Kumar V, Sharma C, Prakash A, Agarwal K, Babu R, Dinesh KR, Karim S, Singh SK, Hagen F, Meis JF. 2014. Multidrug-resistant endemic clonal strain of Candida auris in India. Eur J Clin Microbiol Infect Dis 33:919–926. doi: 10.1007/s10096-013-2027-1. [DOI] [PubMed] [Google Scholar]
- 9.Oh BJ, Shin JH, Kim MN, Sung H, Lee K, Joo MY, Shin MG, Suh SP, Ryang DW. 2011. Biofilm formation and genotyping of Candida haemulonii, Candida pseudohaemulonii, and a proposed new species (Candida auris) isolates from Korea. Med Mycol 49:98–102. doi: 10.3109/13693786.2010.493563. [DOI] [PubMed] [Google Scholar]
- 10.Chowdhary A, Kathuria S, Prakash A, Meis JF. 2014. Pitfalls in phenotypic identification of Candida auris and comparison of in-vitro susceptibilities using CLSI-BMD and VITEK2, abstr M-1203. Abstr 54th Intersci Conf Antimicrob Agents Chemother (ICAAC). American Society for Microbiology, Washington, DC. [Google Scholar]
- 11.Xu J, Ramos AR, Vilgalys R, Mitchell TG. 2000. Clonal and spontaneous origins of fluconazole resistance in Candida albicans. J Clin Microbiol 38:1214–1220. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 12.Tamura K, Peterson D, Peterson N, Stecher G, Nei M, Kumar S. 2011. MEGA5: molecular evolutionary genetics analysis using maximum likelihood, evolutionary distance, and maximum parsimony methods. Mol Biol Evol 28:2731–2739. doi: 10.1093/molbev/msr121. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 13.Jensen RH, Arendrup MC. 2011. Candida palmioleophila: characterization of a previously overlooked pathogen and its unique susceptibility profile in comparison with five related species. J Clin Microbiol 49:549–556. doi: 10.1128/JCM.02071-10. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 14.De Carolis E, Posteraro B, Lass-Flörl C, Vella A, Florio AR, Torelli R, Girmenia C, Colozza C, Tortorano AM, Sanguinetti M, Fadda G. 2012. Species identification of Aspergillus, Fusarium and Mucorales with direct surface analysis by matrix-assisted laser desorption ionization time-of-flight mass spectrometry. Clin Microbiol Infect 18:475–484. doi: 10.1111/j.1469-0691.2011.03599.x. [DOI] [PubMed] [Google Scholar]
- 15.Clinical and Laboratory Standards Institute (CLSI). 2008. Reference method for broth dilution antifungal susceptibility testing of yeasts; approved standard—3rd ed CLSI document M27-A3. CLSI, Wayne, PA. [Google Scholar]
- 16.Pfaller MA, Boyken LB, Hollis RJ, Kroeger J, Messer SA, Tendolkar S, Diekema DJ. 2008. Validation of 24-hour fluconazole MIC readings versus the CLSI 48-hour broth microdilution reference method: results from a global Candida antifungal surveillance program. J Clin Microbiol 46:3585–3590. doi: 10.1128/JCM.01391-08. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 17.Pfaller MA, Boyken LB, Hollis RJ, Kroeger J, Messer SA, Tendolkar S, Diekema DJ. 2011. Validation of 24-hour posaconazole and voriconazole MIC readings versus the CLSI 48-hour broth microdilution reference method: application of epidemiological cutoff values to results from a global Candida antifungal surveillance program. J Clin Microbiol 49:1274–1279. doi: 10.1128/JCM.02437-10. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 18.Reference deleted.
- 19.Pfaller MA, Messer SA, Bolmstrom A. 1998. Evaluation of Etest for determining in vitro susceptibility of yeasts isolates to amphotericin B. Diagn Microbiol Infect Dis 32:223–227. [DOI] [PubMed] [Google Scholar]
- 20.Wanger A, Mills K, Nelson PW, Rex JH. 1995. Comparison of Etest and National Committee for Clinical Laboratory Standards broth macrodilution method for antifungal susceptibility testing: enhanced ability to detect amphotericin-B resistant Candida isolates. Antimicrob Agents Chemother 39:2520. doi: 10.1128/AAC.39.11.2520. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 21.Garcia-Effron G, Lee S, Park S, Cleary JD, Perlin DS. 2009. Effect of Candida glabrata FKS1 and FKS2 mutations on echinocandin sensitivity and kinetics of 1,3-β-d-glucan synthase: implication for the existing susceptibility breakpoint. Antimicrob Agents Chemother 53:3690–3699. doi: 10.1128/AAC.00443-09. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 22.Pfaller MA, Diekema DJ, Jones RN, Castanheira M. 2014. Use of anidulafungin as a surrogate marker to predict susceptibility and resistance to caspofungin among 4,290 clinical isolates of Candida by using CLSI methods and interpretive criteria. J Clin Microbiol 52:3223–3229. doi: 10.1128/JCM.00782-14. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 23.Cendejas-Bueno E, Kolecka A, Alastruey-Izquierdo A, Theelen B, Groenewald M, Kostrzewa M, Cuenca-Estrella M, Gómez-López A, Boekhout T. 2012. Reclassification of the Candida haemulonii complex as Candida haemulonii (C. haemulonii group I), C. duobushaemulonii sp. nov. (C. haemulonii group II), and C. haemulonii var. vulnera var. nov.: three multiresistant human pathogenic yeasts. J Clin Microbiol 50:3641–3651. doi: 10.1128/JCM.02248-12. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 24.Schrödl W, Heydel T, Schwartze VU, Hoffmann K, Grosse-Herrenthey A, Walther G, Alastruey-Izquierdo A, Rodriguez-Tudela JL, Olias P, Jacobsen ID, de Hoog GS, Voigt K. 2012. Direct analysis and identification of pathogenic Lichtheimia species by matrix-assisted laser desorption ionization-time of flight analyzer-mediated mass spectrometry. J Clin Microbiol 50:419–427. doi: 10.1128/JCM.01070-11. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 25.Shin JH, Kim MN, Jang SJ, Ju MY, Kim SH, Shin MG, Suh SP, Ryang DW. 2012. Detection of amphotericin B resistance in Candida haemulonii and closely related species by use of the Etest, Vitek-2 yeast susceptibility system, and CLSI and EUCAST broth microdilution methods. J Clin Microbiol 50:1852–1855. doi: 10.1128/JCM.06440-11. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 26.Zhang L, Wang H, Xiao M, Kudinha T, Mao LL, Zhao HR, Kong F, Xu YC. 2014. The widely used ATB FUNGUS 3 automated readings in China and its misleading high MICs of Candida spp. to azoles: challenges for developing countries' clinical microbiology labs. PLoS One 9:e114004. doi: 10.1371/journal.pone.0114004. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 27.Pfaller MA, Messer SA, Mills K, Bolmström A, Jones RN. 2001. Evaluation of Etest method for determining caspofungin (MK-0991) susceptibilities of 726 clinical isolates of Candida species. J Clin Microbiol 39:4387–4389. doi: 10.1128/JCM.39.12.4387-4389.2001. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 28.Arendrup MC, Pfaller MA, Danish Fungaemia Study Group 2012. Caspofungin Etest susceptibility testing of Candida species: risk of misclassification of susceptible isolates of C. glabrata and C. krusei when adopting the revised CLSI caspofungin breakpoints. Antimicrob Agents Chemother 56:3965–3968. doi: 10.1128/AAC.00355-12. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 29.Espinel-Ingroff A, Arendrup MC, Pfaller MA, Bonfietti LX, Bustamante B, Canton E, Chryssanthou E, Cuenca-Estrella M, Dannaoui E, Fothergill A, Fuller J, Gaustad P, Gonzalez GM, Guarro J, Lass-Flörl C, Lockhart SR, Meis JF, Moore CB, Ostrosky-Zeichner L, Pelaez T, Pukinskas SR, St-Germain G, Szeszs MW, Turnidge J. 2013. Interlaboratory variability of caspofungin MICs for Candida spp. Using CLSI and EUCAST methods: should the clinical laboratory be testing this agent? Antimicrob Agents Chemother 57:5836–5842. doi: 10.1128/AAC.01519-13. [DOI] [PMC free article] [PubMed] [Google Scholar]