Objective: Since it changes treatment alternatives, identification of Candida yeast which is the most frequent infectious agent in human beingson species level. Many laboratories deficient in the use of classical diagnostic methods resort to commercially available automated identification methods. But, diagnostic “accuracy” of commercial identification systems have demonstrated variable results in various comparative studies. The aim of the present work is to compare different identification methods used for Candida species.

Material and Methods: Distribution of 276 Candida species was analyzed in Gazi University Hospital Microbiology Laboratory during the period between June 2015 and June 2016, and 72 isolates of Candida spp. were selected according to ‘stratified random sampling’ analysis method so as to accurately represent distribution related to this one year period Candida species Identification results obtained from morphological evaluation on corn meal-tween 80 agar as a conventional method and ID32C yeast identification system in addition to MALDI-TOF identification patterns were compared. A second comparative analysis was performed for 27 isolates in 72 selected samples using rRNA region ITS gene sequencing and Phoenix™ system.

Results: When ID32C method was considered as the gold standard in species identification in this study, a total of 41 isolates (56%) were identifed accurately by all three methods. The remaining 31 isolates (44%) were misidentified by at least one of the three methods. Among 27 isolates that were reidentified by DNA sequencing method, PhoenixTM identification system (n=14: 52%), morphology on corn meal-tween 80 agar (n=16: 59%), MALDITOF (n=22: 81%), and ID32C identification method (n=21: 78%) accurately identified indicated number of Candida isolates.

Conclusion: Among the techniques analysed, we estimated that the MALDI-TOF system presents valuable and fast results; however if we consider the costs of the methods, the most reliable method is ID32C method based on carbohydrate assimilation in identifying Candida species. Other methods presented accuracy between 52% and 79%.