Evaluation of PNA-FISH method for direct identification of Candida species in blood culture samples and its potential impact on guidance of antifungal therapy

Early antifungal therapy has a major influence on survival in candidemia. Rapid identification of the species has importance for the treatment, prediction of the species-specific primary resistance and variable antifungal susceptibility. Recently, molecular-based methods attempt to reduce the time between the positive signal of a blood culture and identification of the fungus. PNA-FISH (Peptide nucleic acid fluorescence in situ hybridization) assay distinguishes a number of frequently isolated Candida species in groups following the growth in blood culture. The aim of this study was to investigate the correlation of the species identified by PNA-FISH with conventional identification methods in yeast positive blood cultures and its influence on the selection of antifungal therapy. Specimens of adult patients diagnosed as yeast with Gram stain in signal-positive blood cultures between August to December 2013, were included in the study. The strains were concomitantly cultivated by subculturing from the blood culture bottles onto solid media and identified by conventional methods (germ tube test, ID32C and morphology on cornmeal Tween 80 agar). Rapid species identification was performed by Yeast Traffic Light PNA-FISH, which generates green flourescence for Candida albicans and Candida parapsilosis, yellow for Candida tropicalis, and red for Candida krusei and Candida glabrata. C.tropicalis was identified as a single species whereas the others were identified in pairs. The time points when the yeast positive blood culture bottle was received by the mycology laboratory and reporting of the species identification results by PNA-FISH and the conventional methods were recorded. Seven C.albicans, six C.glabrata, three C.parapsilosis, one C.tropicalis, one C.krusei, one Cryptococcus neoformans, one Saprochaete capitata (Blastoschizomyces capitatus), one C.albicans and Candida dubliniensis, one C.krusei and C.dubliniensis, and one C.glabrata and C.parapsilosis were identified by conventional methods in 23 specimens. Results of PNA-FISH and conventional methods were in full agreement in 19 of the 23 specimens (82.6%). Two specimens were negative by PNA-FISH and yielded S.capitata and C.neoformans which were not included in the test panel. In three specimens that were infected with multiple species, PNA-FISH detected only one of the species. On the other hand and in one specimen, PNA-FISH detected a second species (C.glabrata or C.krusei) that could not be isolated an