[Activity of voriconazole against yeasts isolated from blood culture determined by two methods].

The in vitro activity of voriconazole has been determined by two methods: the reference M27-A2 and the marketed Sensititre YeastOne microdilution colorimetric method. The agreement (+/-2 dilutions) and correlation between methods as well as the percentage of errors has been determined. A total of 144 yeasts (47 Candida albicans, 52 C. parapsilosis, 13 C. tropicalis, 10 C. krusei, 9 C. glabrata, 2 C. guilliermondii, 1 C. colliculosa, 1 C. dubliniensis, 2 Trichosporum asahii, 1 T. mucoide, 1 Trichosporum spp., 1 Kloakera apis, 2 Pichia ohmeri, and 2 Rhodotorula glutinis) isolated from blood culture between October 2002 and May 2005 were assayed. Voriconazole has shown good in vitro activity. The rate of voriconazole-susceptible (MIC < or =1 mg/l) strains was 97% and the MIC90 0.25 mg/l by the two methods. The overall percentage of agreement between methods was 86% (range 44.23-100%) and the Pearson's coefficient of correlation was 0.961. Categorical agreement was strain dependent and ranged from 84.6% for emergent yeasts to 100% for the other species tested except for C. glabrata (66.6%). No major or very major errors were found, the percentage of minor errors being 1.38%. Only one C. tropicalis and one C. glabrata strain were resistant (MIC > or =4 mg/l) to voriconazole (1.38%) by the reference method. The colorimetric method identified the voriconazole-resistant C. tropicalis strain, and classified the C. glabrata as susceptible-dose dependent. The colorimetric method is a potential alternative method for testing the susceptibility of yeast in a clinical laboratory and identifies the susceptible strains (100% agreement) very well. Nevertheless, further studies including more voriconazole-resistant strains are required to determine the ability of the method to identify resistance, which is the goal of susceptibility tests.