In patients with moderate to severe histoplasmosis associated with AIDS, the preferred treatment has been the deoxycholate formulation of amphotericin B. However, serious side effects are associated with use of amphotericin B. To compare amphotericin B with liposomal amphotericin B for induction therapy of moderate to severe disseminated histoplasmosis in patients with AIDS. Randomized, double-blind, multicenter clinical trial. 21 sites of the U.S. National Institute of Allergy and Infectious Diseases Mycoses Study Group. 81 patients with AIDS and moderate to severe disseminated histoplasmosis. Clinical success, conversion of baseline blood cultures to negative, and acute toxicities that necessitated discontinuation of treatment. Clinical success was achieved in 14 of 22 patients (64%) treated with amphotericin B compared with 45 of 51 patients (88%) receiving liposomal amphotericin B (difference, 24 percentage points [95% CI, 1 to 52 percentage points]). Culture conversion rates were similar. Three patients treated with amphotericin B and one treated with liposomal amphotericin B died during induction (P = 0.04). Infusion-related side effects were greater with amphotericin B (63%) than with liposomal amphotericin B (25%) (P = 0.002). Nephrotoxicity occurred in 37% of patients treated with amphotericin B and 9% of patients treated with liposomal amphotericin B (P = 0.003). Liposomal amphotericin B seems to be a less toxic alternative to amphotericin B and is associated with improved survival.