The in vitro activity of fluconazole against 143 Candida spp. obtained from the bloodstreams of 143 hospitalized patients from 1995 to 1997 was studied. Susceptibility tests were carried out by two macrodilution methods, the M27-A and a modified M27-A method (0. 165 M, pH 7/morpholinepropanesulfonic acid-buffered RPMI 1640 medium supplemented with 20 g of D-dextrose per liter), and by the agar diffusion method (with 15-microg fluconazole [Neo-Sensitab] tablets). With 2 microg of fluconazole per ml, 96.92% of 65 C. albicans isolates, 86.2% of 58 C. parapsilosis isolates 7 of 8 C. tropicalis isolates, and 1 of 6 C. glabrata isolates were inhibited. Only one strain of C. albicans and one strain of C. tropicalis were resistant. The agreement between the two macrodilution methods was greater than 90% within +/-2 log2 dilutions for all strains except C. glabrata (83.3%) and C. tropicalis (87.5%). Generally, MICs were 1 log2 dilution lower in glucose-supplemented RPMI 1640 medium. No correlation between zone sizes and MICs was found. All strains susceptible by the diffusion test were susceptible by the dilution method, but the converse was not necessarily true. Interestingly, inhibition zones were smaller for C. albicans, for which the geometric mean MIC was 0.29 microg/ml and the mean inhibition zone diameter was 25.7 mm, while for C. parapsilosis the geometric mean MIC was 0.96 microg/ml and the mean inhibition zone diameter was 31. 52 mm. In conclusion, the two macrodilution methods give similar results. The modified M27-A method with 2% dextrose has the advantage of shortening the incubation time and simplifying the endpoint determination.