EVALUATION OF THE MICROBIAL IDENTIFICATION SYSTEM FOR IDENTIFICATION OF CLINICALLY ISOLATED YEASTS

The Microbial Identification System (MIS; Microbial ID, Inc., Newark, Del.) was evaluated for the identification of 550 clinically isolated yeasts. The organisms evaluated were fresh clinical isolates identifie d by methods routinely used in our laboratory (API 20C and conventiona l methods) and included Candida albicans (n = 294), C. glabrata (n = 1 45), C. tropicalis (n = 58), C. parapsilosis (n = 33), and other yeast s (n = 20). In preparation for fatty acid analysis, yeasts were inocul ated onto Sabouraud dextrose agar and incubated at 28 degrees C for 24 h. Yeasts were harvested, saponified, derivatized, and extracted, and fatty acid analysis was performed according to the manufacturer's ins tructions. Fatty acid profiles were analyzed, and computer identificat ions were made with the Yeast Clinical Library (database version 3.8). Of the 550 isolates tested, 374 (68.0%) were correctly identified to the species level, with 87 (15.8%) being incorrectly identified and 89 (16.2%) giving no identification. Repeat testing of isolates giving n o identification resulted in an additional 18 isolates being correctly identified. This gave the MIS an overall identification rate of 71.3% . The most frequently misidentified yeast was C. glabrata, which was i dentified as Saccharomyces cerevisiae 32.4% of the time. On the basis of these results, the MIS, with its current database, does not appear suitable for the routine identification of clinically important yeasts .