OBJECTIVE To evaluate and describe the parameters and characteristics of different drug regimens in children infected with human immunodeficiency virus (HIV). METHODS Randomized, open-label, multicenter study. METHODS Pediatric HIV research clinics in the United States and Puerto Rico. METHODS Twenty-one HIV-infected children, aged 3-14 years, who were clinically stable and treated with the same antiretroviral therapy for 16 weeks or longer. METHODS In step 1, children were randomized to receive one of three treatment regimens: zidovudine plus lamivudine, ritonavir plus zidovudine and lamivudine, or ritonavir plus stavudine. Patients originally assigned to the zidovudine plus lamivudine group in step 1 were eligible to progress to step 2 if their HIV RNA values at week 12, 24, or 36 were 10,000 copies/ml or greater but 100,000 copies/ml or less. In step 2 they received a regimen of ritonavir plus stavudine and nevirapine. RESULTS Seven children were randomized to each of the three treatment regimens. Concentrations of the agents were quantitated at steady state after observed doses, and the pharmacokinetic parameters were determined. Nevirapine concentrations were not determined. One child was excluded from analysis because pharmacokinetic parameters could not be estimated. Ritonavir oral clearance was slower in the pooled cohort of children who received stavudine compared with zidovudine and lamivudine. Stavudine oral clearance was marginally faster when combined with ritonavir and nevirapine compared with only ritonavir. CONCLUSIONS Therapy for HIV is complex, and pharmacodynamic data indicate that relationships exist between systemic concentrations of antiretroviral drugs and virologic response. Careful drug interaction studies have not been conducted for all treatment regimens, and it will not be surprising if unexpected interactions are found. Pharmacokinetic studies to address these considerations should be viewed as a fundamental component of antiretroviral drug development, as they represent a tool to improve pharmacotherapy for HIV-infected children.