This is the first completed prospective randomized clinical efficacy trial of antifungals in the treatment of invasive aspergillosis (IA) and the first to compare the clinical efficacy of two dosages of liposomal amphotericin B (L-AmB) for IA in neutropenic patients with cancer or those undergoing bone marrow transplantation. Eighty-seven of 120 patients were eligible and evaluable. Clinical responses were documented for 26 (64%) of 41 patients receiving 1 mg/(kg.d) (L-AmB-1) and 22 (48%) of 46 receiving 4 mg/(kg.d) (L-AmB-4). Radiologic response rates were similar: 24 (58%) of the L-AmB-1 recipients and 24(52%) of the L-AmB-4 recipients. The six-month survival rates were 43% (L-AmB-1) and 37% (L-AmB-4). These differences were not significant. The numbers of deaths directly due to IA at 6 months were similar: 9 (22%) of 41 L-AmB-1 recipients and 9 (20%) of 46 L-AmB-4 recipients. No other variable independently influenced survival, apart from central nervous system IA. L-AmB is effective in treating approximately 50%-60% of patients who have IA. A 1-mg/(kg.d) dosage is as effective as a 4-mg/(kg.d) dosage, and no advantages to use of the higher, more expensive, dosage has been observed.