Comparative evaluation of four diagnostic methods for Candida dubliniensis identification: diagnostic accuracy and statistical concordance with PCR
摘要
This study evaluated the diagnostic performance of three phenotypic and biochemical methods - the germ tube test, the Bichro-Dubli Fumouze assay, and the VITEK 2 YEAST system - against polymerase chain reaction as the reference method for identifying Candida dubliniensis in clinical oral yeast isolates. A total of 140 yeast isolates from oral and subgingival samples of immunocompetent patients were examined. Each isolate was identified by the three conventional methods and by PCR amplification of a species-specific 119 bp fragment. The diagnostic parameters calculated included sensitivity, specificity, predictive values, and agreement coefficients. PCR confirmed 37 isolates (26.4%) as Candida dubliniensis. The germ tube test showed limited accuracy, with moderate sensitivity and specificity. The Bichro-Dubli Fumouze assay exhibited high sensitivity (89.2%) and specificity (97.1%), while the VITEK 2 system achieved the best performance, with sensitivity of 91.9% and specificity of 99.0%. Both tests demonstrated strong agreement with PCR results. The Bichro-Dubli Fumouze and VITEK 2 YEAST systems provided rapid and highly accurate identification of Candida dubliniensis, whereas the germ tube test alone was unreliable. A stepwise diagnostic approach combining a rapid screening test followed by confirmatory testing is recommended, reserving PCR for ambiguous or epidemiological cases.