The susceptibilities of 25 clinical isolates of various Aspergillus species (Aspergillus fumigatus, A. flavus, A. terreus, A. ustus, and A. nidulans) to itraconazole (ITC) and amphotericin B (AMB) were determined using the standard proposed by NCCLS for antifungal susceptibility testing of[filamentous fungi, a modification of this method using spectrophotometric readings, and a colorimetric method using the tetrazolium salt 2,3-bis [2-methoxy-4-nitro-5-[(sulfenylamino) carbonyl]-2H-tetrazolium-hydroxide] (XTT). Five MIC end points for ITC (MIC-0, no visible growth or <or=5% the growth control value [GC]; MIC-1, slight growth or 6 to 25% the GC; MIC-2, prominent reduction in growth or 26 to 50% the GC; MIC-3, slight reduction in growth or 51 to 75% the GC; and MIC-4, no reduction in growth or 76 to 100% the GC) and one for AMB (MIC-0) were determined visually by four observers and spectrophotometrically. The intraexperimental (between the observers) and interexperimental (between the experiments) levels of agreement of the NCCLS and XTT methods exceeded 95% for MIC-0 of AMB and MIC-0 and MIC-1 of ITC. The MIC-2 of ITC showed lower reproducibility, although spectrophotometric reading and/or incubation for 48 h increased the interexperimental reproducibility from 85 to >93%. Between visual and spectrophotometric readings, high levels of agreement were found for AMB (approximately 97%) and MIC-1 (approximately 92%) and MIC-2 (approximately 88%) of ITC. Poor agreement was found for MIC-0 of ITC (51% after 24 h), since the spectrophotometric readings resulted in higher MIC-0 values than the visual readings. The agreement was increased to 98% by shifting the threshold level of MIC-0 from 5 to 10% relative optical density and by establishing an optical density of greater than 0.1 for the GC as the validation criterion. No statistically significant differences were found between the NCCLS method and the XTT method, with the levels of agreement exceeding 97% for MIC-0 of AMB and 83% for MIC-0, MIC-1, and MIC-2 of ITC. The XTT method and spectrophotometric readings can increase the sensitivity and the precision, respectively, of in vitro susceptibility testing of Aspergillus species.