Terbinafine and amphotericin B MICs for 90% of strains tested were 1.6 and 0.4 ,ug/ml against Aspergillus fumigatus (16 strains), 0.8 and 3.2 ,ug/mI against Aspergillusflavus (10 strains), and 0.4 and 1.6 jig/ml against Aspergillus niger (10 strains), respectively. For all species tested, the minimal inhibitory and fungicidal concentrations for 90% of strains of both drugs were identical and the inoculum size did not have a major effect on the results.Aspergillus species can cause life-threatening opportunistic infections among immunocompromised patients. Invasive aspergillosis is difficult to diagnose and often responds poorly to treatment. Amphotericin B remains the drug of choice despite its considerable toxicity.Terbinafine belongs to a new class of antifungal agents (8). It is a synthetic naphthalenemethanamine that inhibits squalene epoxidase, a key enzyme in ergosterol biosynthesis of fungi (3,5). Its mode of action is highly selective, i.e., it is much more inhibitory to fungal than to mammalian sterol biosynthesis (4). Terbinafine can be administered orally, and preliminary studies in humans indicate that it is well tolerated (7).To evaluate the potential use of terbinafine in human aspergillosis, we compared its in vitro inhibitory and lethal activities against 36 clinical isolates of Aspergillus species with the activities of amphotericin B and ketoconazole.(These data were presented in part at the 27th Source and preparation of antifungal agents. Amphotericin B (Fungizone) was obtained from E. R. Squibb & Sons, Princeton, N. J., and terbinafine was obtained from Sandoz Ltd., Basel, Switzerland. Amphotericin B was initially diluted in distilled water, and terbinafine was initially diluted in dimethyl sulfoxide. Ketoconazole was initially diluted in absolute ethanol. Further drug dilutions were made in distilled water at log 2 concentrations ranging from 0.025 to 51.2 ,ug/ml. Finally, 0.1-ml portions of each concentration were added to wells of microdilution plates.Source and preparation of fungal species. The Aspergillus strains tested were isolated from patients treated at Memorial Sloan-Kettering Cancer Center from January 1985 to April 1987. The control organisms for amphotericin B were Candida albicans B311 and C. tropicalis CT4. The organisms (16 strains of Aspergillus fumigatus, 10 strains of A. flavus, and 10 strains of A. niger) were subcultured on Sabouraud dextrose agar and incubated at 35°C for 4 days. Spores were harvested with 0.02% Tween 80, centrifuged (10 min, 1,500 x g), suspended in distilled water, and counted in * Corresponding author. a hemacytometer. The spores were diluted in 2 x yeast nitrogen glucose broth (Difco Laboratories; 5% glucose) to produce the high (105 spores per ml) and low (2 x 103 spores per ml) inocula.Susceptibility testing. Portions of 0.1 ml of each spore suspension at either the high or low inoculum were added to microdilution tray wells that contained no drug (control wells), terbinafine, amphotericin B, or ketoconazole. Plates were incubated at 35°C for 4...