Agreement of Direct Antifungal Susceptibility Testing from Positive Blood Culture Bottles with the Conventional Method for Candida Species

Early availability of antifungal susceptibilities can ensure timely institution of targeted therapy in candidemia, which can improve patient outcomes. This study prospectively determines the agreement between the results of direct testing of antifungal susceptibilities from blood culture bottles by disk diffusion and Etest and the results of standardized susceptibility testing methods; direct testing would allow susceptibility results to be available 1 to 2 days earlier. A total of 104 blood cultures with different Candida species (28% C. albicans, 27% C. parapsilosis, 26% C. tropicalis, etc.) were evaluated between January 2012 and May 2013 for agreement of fluconazole, voriconazole, and amphotericin B susceptibility results by disk diffusion. Agreement in MICs obtained by Etest was determined for fluconazole (21 isolates), voriconazole (28 isolates), amphotericin (29 isolates), and caspofungin (29 isolates). The kappa scores for categorical agreement were highest for fluconazole by disk diffusion (0.902, standard error [SE] -0.076) and Etest (1.00, SE = 0.218) and for amphotericin B by disk diffusion (1.00, SE = 0.098). The Pearson correlation (r) of zone diameters was strongest for fluconazole (0.69) and amphotericin (0.70) and moderate for voriconazole (0.60), and the Pearson correlation of MICs was strongest for fluconazole (0.94) and caspofungin (0.88). However, the moderate correlation of amphotericin MICs with zone diameters (-0.42) precludes the use of amphotericin B disk diffusion for susceptibility testing. There were no very major errors; however, there were 1 (1%) major and 5 (4.8%) minor errors with disk diffusion and 4 (13.3%) minor errors with Etest. Thus, antifungal disk diffusion directly from blood culture bottles is a rapid and easy method for fluconazole and voriconazole susceptibility testing for timely tailoring of candidemia therapy.