Using an isolate of Aspergillus fumigatus that is less susceptible in vivo to amphotericin B than most other isolates, we compared different doses of liposomal nystatin (L-nystatin), liposomal amphotericin B (L-amphotericin), and amphotericin B lipid complex (ABLC) with amphotericin B deoxycholate. Four experiments with intravenously infected neutropenic mice were conducted. A dose of L-nystatin at 10 mg/kg of body weight was toxic (the mice had fits or respiratory arrest). The optimal dosage of L-nystatin was 5 mg/kg daily on days 1, 2, 4, and 7 (90% survival). This was superior to L-amphotericin (5 mg/kg [P = 0.24] and 1 mg/kg [P < 0.0001]), ABLC (5 mg/kg [P = 0.014] and 1 mg/kg [P < 0.0001]), and amphotericin B deoxycholate (5 mg/kg [P = 0.008]). In terms of liver and kidney cultures, L-nystatin (5 mg/kg) was superior to all other regimens (P = 0.0032 and <0.0001, respectively). Higher doses of L-amphotericin (25 and 50 mg/kg) in one earlier experiment were more effective (100% survival) than 1 mg of L-amphotericin per kg and amphotericin deoxycholate (5 mg/kg) in terms of mortality and both liver and kidney culture results and to L-amphotericin (5 mg/kg) in terms of liver and kidney culture results only. ABLC (25 mg/kg) given daily for 7 days was superior to ABLC (50 mg/kg [P = 0.03]) but not to ABLC at 5 mg/kg or amphotericin B deoxycholate in terms of mortality, although it was in terms of liver and kidney culture results. No dose-response for amphotericin B (5 and 1 mg/kg) was demonstrable. In conclusion, in this stringent model, high doses of L-amphotericin and ABLC could overcome reduced susceptibility to amphotericin B deoxycholate, but all were inferior to 5- to 10-fold lower doses of L-nystatin.