One hundred eighty-one antiretroviral-experienced, protease inhibitor-naive, clinically stable HIV-infected children between 4 months and 17 years of age were randomly assigned to receive one of four combination regimens to evaluate the change in plasma HIV RNA, safety, and tolerance when changing antiretroviral therapy to a protease inhibitor-containing combination regimen. All four regimens contained stavudine; in addition children received nevirapine plus ritonavir, lamivudine plus nelfinavir, nevirapine plus nelfinavir, or lamivudine plus nevirapine plus nelfinavir. Twelve additional children chose to receive stavudine plus lamivudine plus nelfinavir, with nelfinavir given bid, rather than tid as for the main regimens. Overall, 51% (89/176; 95% CI 43-58%) of the children on the randomized portion of the study had an HIV RNA response (≤400 copies/ml) on at least two of the three HIV RNA determinations taken at Weeks 8, 12, and 16. At Week 24 the proportion of children with an HIV RNA response still on initial therapy was 47% (83/176; 95% CI 40-55%) and ranged from 41 to 61% for the four randomized treatment arms. Rash was frequently seen (27%) on the treatment arms containing nevirapine. At Week 24 64% (7/11, 95% CI 31-89%) of the children on the bid nelfinavir combination regimen were still on initial therapy with an HIV RNA response as compared with 46% (23/50; 95% CI 32-61%) on the corresponding tid nelfinavir combination regimen. A change in antiretroviral therapy to a protease inhibitor-containing regimen was associated with a virological response rate of approximately 50% for this patient population.