OBJECTIVES: To compare the efficacy and safety of indinavir 800 mg three times a day, ritonavir 600 mg twice a day, and a combination of ritonavir 400 mg twice a day and saquinavir 400 mg twice a day, when administered with two nucleoside analogues. DESIGN: A randomized, open-labelled, controlled trial. Two hundred and eighty-four patients started randomized treatment. The primary end-point was the proportion of patients with HIV RNA of 200 copies/ml or less (Roche Amplicor) and HIV RNA of 20 copies/ml or less (Roche ultradirect assay) at 6 months. Analysis was performed as intent-to-treat, and missing values were accounted for as failures. RESULTS: As of 1 May 1998, 269 patients should have completed 24 weeks of treatment. The proportion of patients with HIV RNA of 200 copies/ml or less was 71% (indinavir), 67% (ritonavir), and 82% (ritonavir + saquinavir), P = 0.07. In antiretroviral drug-naive patients (n = 119), the corresponding figures were 63, 57, and 89% (P < 0.01), whereas among drug-experienced patients (n = 165) 77, 74, and 77% had HIV RNA of 200 copies/ml or less (P = 0.90). The same pattern was observed in the ultradirect analysis. All three regimens were generally safe, but significantly more patients in the ritonavir group (37%) stopped treatment because of adverse drug reactions compared with the indinavir group (8%) and the ritonavir plus saquinavir group (16%) (P < 0.001). CONCLUSIONS: Treatment with saquinavir plus ritonavir in combination with two nucleoside analogues is generally safe, and has superior short-term antiviral efficacy compared with indinavir and ritonavir also combined with two nucleoside analogues in antiretroviral drug-naive patients. Further follow-up is needed to determine the durability of the viral response.