We conducted a prospective, randomized, double-blind study comparing amphotericin B colloidal dispersion (ABCD) with amphotericin B in the empirical treatment of fever and neutropenia. Patients with neutropenia and unresolved fever after > or = 3 days of empirical antibiotic therapy were stratified by age and concomitant use of cyclosporine or tacrolimus. Patients were then randomized to receive therapy with ABCD (4 mg/[kg.d]) or amphotericin B (0.8 mg/[kg.d]) for < or = 14 days. A total of 213 patients were enrolled, of whom 196 were evaluable for efficacy. Fifty percent of ABCD-treated patients and 43.2% of amphotericin B-treated patients had a therapeutic response (P = .31). Renal dysfunction was less likely to develop and occurred later in ABCD recipients than in amphotericin B recipients (P < .001 for both parameters). Infusion-related hypoxia and chills were more common in ABCD recipients than in amphotericin B recipients (P = .013 and P = .018, respectively). ABCD appeared comparable in efficacy with amphotericin B, and renal dysfunction associated with ABCD was significantly less than that associated with amphotericin B. However, infusion-related events were more common with ABCD treatment than with amphotericin B treatment.