In a prospective, randomized clinical trial, the toxicity of 1 mg of amphotericin B (AmB) per kg of body weight per day infused in 5% dextrose was compared with that of AmB infused in lipid emulsion in children with malignant disease. In an analysis of 82 children who received a full course of 6 days or more of AmB (117 courses), it was shown that there were significant increases in plasma urea and creatinine concentrations and in potassium requirement after 6 days of therapy with both AmB infused in dextrose and AmB infused in lipid emulsion, with there being no difference between the two methods of AmB administration. An intent-to-treat comparison of the numbers of courses affected by acute toxicity (fever, rigors) and chronic toxicity (nephrotoxicity) also indicated that there was no significant difference between AmB infused in dextrose (78 courses) and AmB infused in lipid emulsion (84 courses). The pharmacokinetics of AmB were investigated in 20 children who received AmB in dextrose and 15 children who received AmB in lipid emulsion. Blood samples were collected up to 24 h after administration of the first dose, and the concentration of AmB in plasma was analyzed by a high-performance liquid chromatography assay. The clearance (CL) of AmB in dextrose (0.039 ± 0.016 liter · h¹ · kg¹) was significantly lower (P < 0.005) than the CL of AmB in lipid emulsion (0.062 ± 0.024 liter · h¹ · kg¹). The steady-state volume of distribution for AmB in dextrose (0.83 ± 0.33 liter · kg¹) was also significantly lower (P < 0.005) than that for AmB in lipid emulsion (1.47 ± 0.77 liter · kg¹). Although AmB in lipid emulsion is apparently cleared faster and distributes more widely than AmB in dextrose, this study did not reveal any significant advantage with respect to safety and tolerance in the administration of AmB in lipid emulsion compared to its administration in dextrose in children with malignant disease.