BACKGROUND: Although standardized broth dilution methods for antifungal susceptibility testing are available, easier testing procedures are desirable. We evaluated the E-test (AB disk, Sweden) as a possible alternative instead of NCCLS (National Committee for Clinical Laboratory Standards) broth macrodilution method. METHODS: Fifty-two bloodstream isolates of Candida spp. (including 11 C. albicans, 13 C. tropicalis, 18 C. parapsilosis, 1 C. glabrata, 4 C. krusei, 2 C. pelliculosa, 2 C. lipolytica, and 1 C. guilliermondii) were tested. Amphotericin B and fluconazole MICs for each isolate were determined by both NCCLS broth macrodilution method and E-test. The results of E-test for Candida spp. were compared with those of NCCLS macrodilution method. For selecting plating media for E-test, we compared E-test results in two different media (RPMI and Casiton medium) using five ATCC Candida strains. RESULTS: As E-test media, we selected RPMI medium for amphotericin B and Casitone medium for fluconazole because of higher agreement with NCCLS method. The E-test and NCCLS method of 52 Candida spp. yielded a very narrow range of MICs (0.064-2.0 microgram/mL) for amphotericin B and a broad range of MICs (0.5-64 microgram/mL) for fluconazole. The agreements of E-test within one doubling dilutions of the macrodilution reference were 90.4% (24h and 48h) for amphotericin B, and 90.4% (24h) and 96.2% (48h) for fluconazole. CONCLUSION: The E-test is a valuable alternative to the NCCLS macrodilution method for amphotericin B and fluconazole susceptibility testing of Candida species.