INTRODUCTION

Top Abstract Introduction Materials & Methods Results & Discussion References

Substrate assimilation is based on the development of growth of an organism in the presence of chemically pure substrates, and it is the conventional method used for the identification of yeasts and yeast-like fungi. By using a better basal medium and more carbon compounds than were previously evaluated by other investigators, Wickerham and Burton (11) demonstrated the usefulness of assimilation tests for the classification of yeasts in 1948. These conventional assimilation methods, which were simplified in 1975 (4), remain tedious and time-consuming. The increased incidence of yeast infections among immunocompromised patients demanded even simpler methods of identification, which led to the development of several commercial kits during the mid-1970s. Among the early commercial methods (1, 5, 7, 8), the API 20C Yeast Identification system (bioMérieux Vitek, Inc., Hazelwood, Mo.) was modified to its current version and has been evaluated by several investigators (1, 5). The API 20C system permits the accurate use of 19 assimilation tests for the identification of most clinically important yeasts after 72 h of incubation. The prompt and accurate identification of yeasts is becoming more important because new antifungal agents with different activities against the various species are being developed and the association of common, emerging, and new yeast pathogens with severe infection continues to increase among patients with compromised, cell-mediated immunity and neutropenia.

A single-substrate, chromogen micromethod has been designed by Innovative Diagnostic Systems, L.P., Norcross, Ga. (RapID Yeast Plus System). The RapID Yeast Plus System is based upon enzymatic reactions of chromogenic substrates involving preformed enzymes and allows the differentiation of yeasts, yeast-like fungi, and similar organisms recovered from human clinical specimens after only 4 h of incubation. The purpose of the present study was to compare the ability of the RapID Yeast Plus System with that of the API 20C System to identify 447 clinical isolates of yeasts.

	MATERIALS AND METHODS

Top Abstract Introduction Materials & Methods Results & Discussion References

Cultures. A set of 447 isolates from two medical centers representing the genera and species listed in Tables 1, 3, and 4 were studied. Candida albicans ATCC 14053, Candida glabrata ATCC 2001, Candida (Yarrowia) lipolytica ATCC 9773, Candida kefyr (Candida pseudotropicalis) ATCC 2512, and Cryptococcus laurentii ATCC 66036 were tested as quality control isolates. When the expected enzymatic reactivity results were obtained with the control isolates as recommended by the manufacturer, the clinical isolates were tested.

RapID Yeast Plus System. The RapID Yeast Plus System uses a qualitative micromethod with 18 conventional and chromogenic substrates (2): 1% glucose, maltose, sucrose, trehalose, and raffinose (cavities 1 to 5, respectively); 1% fatty acid ester (cavity 6); 0.05% p-nitrophenyl-N-acetyl--D-galactosaminide, p-nitrophenyl--D-glucoside, p-nitrophenyl--D-glucoside, o-nitrophenyl--D-galactoside, p-nitrophenyl--D-galactoside, p-nitrophenyl--D-fucoside, p-nitrophenyl phosphate, and p-nitrophenyl phosphorylcholine (cavities 7 to 14, respectively); 0.3% urea (cavity 15); and 0.01% proline -naphthylamide, histidine -naphthylamide, and leucyl-glycine -naphthylamide (cavities 16, 17, and 18, respectively).

API 20C system. Molten (50°C) API basal medium ampoules were inoculated with yeast colonies, and the suspension was standardized to a density below 1+ (lines can be clearly distinguished) on a Wickerham card. Each cupule was inoculated, and the trays were incubated for 72 h at 30°C. Cupules showing turbidity significantly heavier than that of the negative control cupule (0 cupule) were considered positive. Identification was made by generating a microcode and using the API 20C Analytical Profile Index or the Voice Response System (for profiles not found in the index). Morphology on cornmeal was also evaluated as determined by the manufacturer.

Analysis of the data. For each isolate, the identifications obtained by the two methods were compared; and each method was evaluated for its ability to identify the isolates (i) to the species level, (ii) to the genus level or when additional tests (low-probability identifications) were required to distinguish between two or more possible species, (iii) for its discrepant identifications, and (iv) for its failure to provide an identification (no codes) (see Tables 1, 3, and 4). When the six-digit microcode provided an identification with a low percentage of probability (<94%), the recommended additional tests described below were performed by conventional methods (10). The percentages of agreement at the species level with additional tests (low-probability identifications) and without and disagreements between the two methods were obtained (see Tables 1, 3, and 4). Some isolates with discrepant identifications between the two methods or for which the RapID Yeast Plus System failed to provide an identification were sent to the Clinical Microbiology (Mycology) Laboratory of the New York University Medical Center (NYU) for confirmation of identification by noncommercial assimilation and fermentation methods, morphological studies, and other tests such as thermotolerance tests and tests for urease and KNO₃ assimilation (11).

	RESULTS AND DISCUSSION

Top Abstract Introduction Materials & Methods Results & Discussion References

In 1979, Land et al. (5) reported a 97% agreement between the identifications obtained by a conventional method and the API 20C yeast identification system when the latter test was used in conjunction with morphological characteristics. During the last two decades, the performance of new commercial methods for yeast identification has often been compared to that of the API 20C system. Because of this, although some (>200) of the isolates evaluated had previously been identified by conventional methods, we reidentified each isolate by the two methods.

Prior evaluations of the RapID Yeast Plus System by the manufacturer (2) have claimed a 95.5% correlation with the API 20C system for the identification of 378 isolates, but the species evaluated were not described. The manufacturer's differential chart (2), however, claims the possible identification of 42 species; we evaluated 36 of these 42 species. As indicated in Table 1, the two methods were in agreement to the species level, without additional tests, for the identification of 202 of the 211 (95.7%) yeasts grouped as common yeast pathogens. This group of species comprised the yeasts most frequently recovered from patients with severe yeast infections (10). Another four (1.9%) common yeasts were identified by the RapID Yeast Plus System with additional tests, e.g., morphological and thermotolerance tests. Among the discrepant identifications (Table 2) between the two methods for isolates in this group, the identification by the API 20C System of three Candida guilliermondii and two Candida tropicalis (sucrose-negative variety) by the RapID Yeast Plus System were confirmed by conventional methods in the NYU laboratory. Our results are similar to those obtained by other investigators in a prior evaluation of the RapID Yeast Plus System with the same species (3).

As indicated in Table 3, the overall level of agreement between the two methods for the identification of the yeasts that we grouped as emerging species of Candida was 79.1% (72 of the 91 isolates tested). An additional six yeasts (6.6%) were identified by supplementary conventional tests, e.g., thermotolerance tests and the presence of hyphae and of a pellicle in broth. Although the level of agreement between the two methods was low for these Candida spp. as a whole (Table 3), the two methods showed good agreement in the identification of Candida kefyr, Candida lambica, and Candida lipolytica strains. Representative strains of these species were reidentified by conventional methods. A previous study demonstrated that the RapID Yeast Plus System identified 94.1% of 304 yeast and yeast-like isolates (3), but those investigators evaluated mostly (264 of the 304) common yeast pathogens. Furthermore, only one to three isolates of each of the new and emerging species of pathogens were included in their study (3).

Table 4 indicates that the level of agreement between the two methods of yeast identification for the fungi grouped as emerging yeast and yeast-like pathogens was similar to the one for the emerging species of Candida: 75.2% without additional tests and 83.5% with the aid of additional tests. The additional tests needed were tests for ascospore production (Hansenula anomala); thermotolerance tests; nitrate, lactose, and raffinose assimilation tests; capsule formation; and hypha and pigmentation production (Cryptococcus spp. and Rhodotorula spp.) tests. The major discrepancies between the two methods were observed when evaluating isolates of Cryptococcus albidus, Cryptococcus laurentii, Rhodotorula minuta, and Trichosporon beigelii (Table 2). The RapID Yeast Plus method was also unable to identify some isolates of these species as well as isolates of Blastoschizomyces capitatus (1 of 9 isolates) and Hansenula anomala (2 of 17 isolates). One of the problems was the difficulty in interpreting the color change indicative of positive reactions, especially when reading wells 6 (lipase), 16, 17, and 18 (proline, histidine, and leucyl-glycine -naphthylamide, respectively). The NYU laboratory confirmed the identifications of two Rhodotorula rubra isolates obtained by the RapID Yeast Plus System. These isolates were identified as Rhodotorula minuta by the API 20C system. For seven isolates of Hansenula anomala, including the two isolates that the RapID Yeast Plus System failed to identify (listed as no code in Table 4), additional tests were performed for confirmation of the identifications. Although conventional methods (10) identified these isolates as probable Hansenula anomala, ascospore production was observed in only one of these strains. Nine isolates of Cryptococcus albidus, Cryptococcus laurentii, and Trichosporon beigelii had discrepant identifications by the two methods. The identifications of these isolates by the API 20C system were also confirmed by conventional tests. One site also tested direct inoculation into the RapID Yeast Plus System from the primary isolation medium, inhibitory mold agar. Only 48 of 97 isolates tested (data not shown) were correctly identified. This emphasizes the importance of using the incubation conditions and media specified by the manufacturer.

In conclusion, although rapid and easy methods are needed for clinical laboratories that have not been able to switch to automated procedures for rapid yeast identification, such as the Vitek and MicroScan systems (6, 9), the RapID Yeast Plus System should be used with caution when identifying the less common yeasts and yeast-like pathogens. However, some of the incorrectly identified isolates belong to the species less frequently isolated from clinical specimens. Our data also suggest that the RapID Yeast Plus method is an accurate, rapid, and cost-effective tool in the clinical laboratory for the identification of common and many of the new and emerging Candida spp., Cryptococcus neoformans, and certain other yeasts and yeast-like fungi.