We conducted a randomized, open-labeled clinical trial to compare the tolerability and efficacy of amphotericin B deoxycholate, prepared in 5% dextrose or Intralipid (Kabi Pharmacia, Saint-Quentin-en-Yvelines, France), in the treatment of AIDS-associated cryptococcal meningitis in Burundi. Forty-four patients were assigned to receive amphotericin B/dextrose (0.7 mg/[kg.d]) for 14 days; the dose was then reduced to 1 mg/kg every other day for 28 days (infused over 6 hours). Forty-six patients were assigned to receive Intralipid/amphotericin B at a 50% higher dosage (1 mg/[kg.d]) for 14 days; the dose was then reduced to 1.5 mg/kg every other day for 28 days (infused over 2 hours). Intralipid significantly decreased the incidence of fever (P = .02) and chills (P = .0001) related to the infusion of amphotericin B deoxycholate. Analysis of the time to the onset of increased levels of serum creatinine (creatinine level, > 150 mumol/L) showed that Intralipid/amphotericin B was more nephrotoxic (P = .03). The percentage of patients who were clinically cured or had improvement in their conditions and successful mycological outcome was similar in both therapeutic groups, but analysis of the time to the first negative cerebrospinal fluid culture showed a nearly significant difference that favored Intralipid/amphotericin B (P = .07). Intralipid reduced the infusion-related toxicity of amphotericin B deoxycholate without altering its antifungal efficacy but did not confer substantial benefit against renal toxicity that would allow the unitary dosage of amphotericin B deoxycholate to be increased safely.