BACKGROUND Protease inhibitors are an effective component of combination antiretroviral treatment for children infected with human immunodeficiency virus 1 (HIV-1), but tolerance or toxicity issues sometimes require an alternative therapeutic strategy. METHODS HIV-1-infected children aged 2-17 years received combination therapy with either stavudine plus ritonavir or with zidovudine, lamivudine, and ritonavir as part of a randomized clinical trial. Twenty-one months after the start of the trial, ritonavir in capsule formulation became unavailable. The treatment regimen for 25 children was switched from ritonavir capsules to indinavir capsules (500 mg/m(2) every 8 h). The other study drugs remained unchanged. A matched-pairs analysis was performed to compare the results for these 25 children with the results for 25 matched children whose treatment regimen continued to include ritonavir (in liquid formulation). RESULTS There were no significant differences in the percentage of children with an HIV-1 RNA load of <or=200 copies/mL between the group receiving indinavir and the matched group receiving ritonavir (52% vs. 68%, at the start of indinavir treatment; 42% vs. 52%, at week 12; and 50% vs. 56%, at week 24). Similarly, there were no significant differences between the treatment groups with respect to median CD4 cell counts over time. Toxicities observed in the children treated with indinavir were most often flank pain or headache (16%), renal dysfunction (16%), hematuria (12%), and skin rash (12%) and were consistent with toxicities reported elsewhere. There was no evidence that the switch from ritonavir to indinavir therapy altered the pharmacokinetics of indinavir as a result of residual cytochrome P450 induction or inhibition caused by ritonavir. CONCLUSIONS The switch from one protease inhibitor (ritonavir) to another (indinavir) as a component of combination antiretroviral treatment in this patient population was a practical therapeutic strategy.