Misidentification of clinical yeast isolates by using the updated Vitek Yeast Biochemical Card

Abstract

The Vitek Yeast Biochemical Card (YBC) is widely used as a rapid identification (RI) (within 48 h) system for clinical yeast isolates. We compared the RI results obtained by the YBC technique with matched results obtained with the API 20C system. The RI of germ tube-negative yeasts isolated from 222 clinical specimens was performed with the YBC system, and the results were compared with those of standard identifications obtained by using the API 20C system and morphology, with additional biochemical reactions performed as required. Commonly isolated yeasts (Candida albicans [n = 29], Candida tropicalis [n = 40], Torulopsis [Candida] glabrata [n = 28], Candida parapsilosis [n = 12], and Cryptococcus neoformans [n = 14]) were generally well identified (115 of 123 [93%] identified correctly, with only C. albicans, C. tropicalis, and C. neoformans mis- or unidentified more than once). The RI of less commonly isolated yeasts included in the YBC database, however, was less successful (54 of 99 [55%] correct). The YBC card failed to identify 42% (10 of 24) of Candida krusei isolates, 80% (4 of 5) of Candida lambica isolates, 88% (7 of 8) of Trichosporon beigelii isolates, and 83% (10 of 12) of Cryptococcus isolates (non-C. neoformans species). For most identification failures (79%; 42 of 53) there was no identification by the end of 48 h; the other identification failures (21%; 11 of 53) gave definite but incorrect identifications. Of eight rare clinical yeast isolates not included in the Vitek database, six were correctly, not identified, while two (25%) were falsely assigned a definite RI (one Hansenula fabianii isolate was identified as Rhodotorula glutinis, and one Hansenula isolate [non-Hansenula anomala] was identified as Hansenula anomala). While the Vitek YBC rapidly and adequately identifies common yeast isolates, it fails in the RI of more unusual organisms.