Evaluation of the New Micronaut-Candida System Compared to the API ID32C Method for Yeast Identification

Abstract

A new system, Micronaut-Candida, was compared to API ID32C to identify 264 yeast (Candida albicans, C. parapsilosis, C. tropicalis, C. krusei, C. inconspicua, C. norvegensis, C. lusitaniae, C. guilliermondii, C. dubliniensis, C. pulcherrima, C. famata, C. rugosa, C. glabrata, C. kefyr, C. lipolytica, C. catenulata, C. neoformans, Geotrichum and Trichosporon species, Rhodotorula glutinis, and Saccharomyces cerevisiae) clinical isolates. Results were in concordance in 244 cases. Eighteen out of the 20 of discordant results were correctly identified by Micronaut-Candida but not by API ID32C, as confirmed by PCR ribotyping.

The frequency of invasive fungal infections has increased dramatically during the past 3 decades (10). Correct identification is important for accurate therapy, as some fungi possess intrinsic resistance to certain antifungals (4). Conventional methods are time-consuming, while DNA-based methods are not available to all laboratories. Relatively simple tests based on assimilation reactions (API ID32C, API 20C, API Candida) are commercially available, though they may sometimes yield equivocal results or even lead to misidentification (1, 3, 5, 11).

A new microplate-based system with four tests on a single plate, Micronaut-Candida (Merlin Diagnostika GmbH, Bornheim, Germany), has been developed to identify medically important yeasts. Its database contains 31 species of six genera. A test contains 21 biochemical reactions and three controls (8 chromogenic substrates [N-acetyl-β-d-galactosaminidase, α-galactosidase, l-prolinaminopeptidase, p-nitrophenyl-β-glucoronidase, l-phenylalaninaminopeptidase, α-glucosidase, and β-glucosidase plus control], 14 carbohydrate assimilation tests [melibiose, d-xylose, l-rhamnose, gentibiose, d-glucose, inositol, cellobiose, saccharose, trehalose, galactose, maltose, lactose, raffinose, and assimilation control], and urease test with its control). After 24 h, results are read and interpreted automatically using the Micronaut Skan device and the Micronaut software.

(This work was presented in part at the 16th Congress of the International Society for Human and Animal Mycology, Paris, France, 2006 [13a].)

The aim of our study was to identify 264 yeast clinical isolates with Micronaut-Candida and with API ID32C (BioMerieux, Marcy l'Etoile, France). We also tested 13 ATCC strains (Candida albicans 14053, C. albicans 10231, C. parapsilosis 22019, C. tropicalis 750, C. krusei 6258, C. inconspicua 16783, C. norvegensis 22977, C. lusitaniae 38533, C. guilliermondii 6260, C. dubliniensis CD36, C. pulcherrima 18406, C. famata 36239, and C. rugosa 2142). The majority of isolates were isolated from throat, sputum, blood, wound, urine, and vagina from in- and outpatients between November 2005 and April 2006. All C. parapsilosis isolates belonged to C. parapsilosis sensu stricto, determined as described earlier (13). Candida dubliniensis isolates were derived from a previous study (12). Some less common yeast species (Geotrichum and Trichosporon species) were from our collection.

Yeasts grown on Sabouraud dextrose agar were tested simultaneously with API ID32C and Micronaut-Candida according to the manufacturers’ instructions. API ID32C and Micronaut-Candida were read after 48 and 24 h, respectively. In case of discrepancy, both methods were repeated. If these were in accord, we accepted them as final. Otherwise or when additional tests (microscopic morphology, esculin hydrolysis, growth at 45°C, nitrate assimilation) yielded equivocal results, PCR ribotyping was performed in each case and ribotyping results were accepted as valid identification (5).

The proportion of misidentifications was determined for each test and the relative accuracies of results were compared using Fisher's exact test.

API ID32C correctly identified all tested ATCC strains. Micronaut-Candida misidentified C. norvegensis and C. rugosa ATCC strains, both as C. valida according to the additional test proposed (microscopic morphology). As C. pulcherrima is not included in the Micronaut database, Micronaut could not identify the C. pulcherrima ATCC strain. Other ATCC strains were correctly identified.

Results obtained by both methods without additional tests were in concordance for C. albicans (n = 40), C. glabrata (n = 40), C. parapsilosis (n = 35), C. tropicalis (n = 35), C. krusei (n = 35), C. kefyr (n = 10), C. dubliniensis (n = 9), C. guilliermondii (n = 7), C. lusitaniae (n = 7), C. lipolytica (n = 4), S. cerevisiae, (n = 8), C. famata (n = 2), C. catenulata (n = 2), C. rugosa (n = 1), C. neoformans (n = 2), and Geotrichum (n = 2) and Trichosporon (n = 2) species isolates. In the cases of one of two C. rugosa and two of two Rhodotorula glutinis isolates, Micronaut-Candida needed additional tests (microscopic morphology and nitrate assimilation, respectively) to accord with API ID32C.

API ID32C using an extra test (esculin hydrolysis) misidentified all tested C. inconspicua strains (n = 12) as C. norvegensis strains, as confirmed by PCR ribotyping (6). These strains were correctly identified with Micronaut-Candida using an additional test (microscopic morphology). Five C. lusitaniae strains were misidentified as C. famata and a single C. tropicalis strain was misidentified as C. humicolus with API ID32C but not with Micronaut-Candida. Two C. pulcherrima clinical isolates were correctly identified with API ID32C; Micronaut-Candida did not identify these strains, as C. pulcherrima is not included in the Micronaut database. These identities were also confirmed by PCR ribotyping.

Micronaut-Candida correctly identified Candida species occurring frequently in clinical practice (C. albicans, C. glabrata, C. parapsilosis, C. tropicalis, and C. krusei) within 24 h. Additionally, Candida species possessing primary (C. krusei) or secondary (C. glabrata) resistance to fluconazole (6, 9) were also identified accurately within 1 workday. Both methods correctly identified all tested C. dubliniensis isolates without additional tests. As C. dubliniensis, similarly to C. glabrata, may acquire fluconazole resistance readily during treatment (7), correct differentiation of this species from C. albicans within 24 h can be important.

API ID32C but not Micronaut-Candida misidentified 5 of 12 C. lusitaniae strains as C. famata strains. This may be problematic in clinical situations, as C. famata, in contrast to C. lusitaniae, often shows high MICs to triazoles and caspofungin (2, 8), while C. lusitaniae may be resistant to amphotericin B (4).

Micronaut-Candida never misidentified C. inconspicua, but an extra test (microscopic morphology) and extra days were needed to properly identify this inherently fluconazole-resistant species (5). Thus, fast and correct identification of C. inconspicua remains unresolved.

In summary, only 2 clinical isolates were unidentified/misidentified by Micronaut-Candida, while 18 were unidentified/misidentified by API ID32C (0.76% versus 6.82%; P < 0.001).

Micronaut-Candida is an easy-to-perform method to identify the most common Candida species within 24 h. Micronaut-Candida without extra tests identified 241 (91.3%) yeast isolates concordantly to API ID32C. In case of the 18 discordant isolates, ribotyping confirmed the Micronaut-Candida results.

Rare species with proven decreased susceptibility to certain antifungals (2, 5, 8, 10) were identified correctly by Micronaut-Candida with (C. inconspicua) or without extra tests (C. kefyr, C. guilliermondii, and C. lusitaniae). Thus, this new system seems to be a reliable and useful method for identification of medically relevant yeasts in routine mycology laboratories.