Evaluation of YeastIdent and Uni-Yeast-Tek yeast identification systems

The accuracy of the new API YeastIdent system and the Flow Laboratories Uni-Yeast-Tek identification kit with an expanded data base was evaluated in comparison to the API 20C yeast identification system by three laboratories. A total of 489 test isolates were used, biased toward yeasts commonly encountered in clinical specimens. Isolates not in a system's data base were not counted in the evaluation of that system. For isolates in their data base, YeastIdent was 55% accurate and Uni-Yeast-Tek was 40% accurate. By the manufacturer's criteria of reliable identification without additional tests, both systems failed to identify many common and uncommon species. The limited number of substrates and difficulties in assessing results obtained with 11 of the API YeastIdent substrates and apparent errors in the expanded Uni-Yeast-Tek data base appeared to be major factors limiting the accuracy of these systems.