We used a broth microdilution method to study the susceptibility to fluconazole of yeasts isolated from oropharyngeal samples (108 cases) or blood cultures (56 cases), and compared the susceptibility-patterns of strains obtained in patients receiving or not receiving fluconazole prophylaxis. The concentration at which the optical density was reduced to 50 % or more of drug-free controls (inhibitory concentration 50 % or IC 50) was taken as the end-point. The mean IC 50 were : Candida albicans, 1.26 mu g/ml; C. parapsilosis, 2.38 mu g/ml; C. glabrata, 4.54 mu g/ml; C. krusei, 12.58 mu g/ml. Nine of 108 C. albicans isolates had a reduced susceptibility to fluaconazole with IC 50 ranging from 3 to 16 mu g/ml; when comparing yeasts obtained from patient receiving or not receiving fluconazole at the time of sampling the mean IC 50 of C. albicans isolates were 3.11 or 0.93 mu g/ml respectively (p < 0.05). Thus despite an overall susceptibility superior to that of other Candida species tested (p = 0, 0001), some C. albicans isolates may be resistant to fluconazole in vitro. Moreover our data enhance the likelyhood of a selection of significantly reduced susceptibility C. albicans isolates during fluconazole treatment. On the other hand fluconazole treatment did not significantly alter the susceptibility of C. krusei, and in all C. krusei and C. glabrata isolates the IC 50 ranged from 2 to 64 mu g/ml. Given these findings we consider unreasonable to treat with fluconazole an infection due to a Candida spp. isolate with an IC 50 of 2 mu g/ml or above.