OBJECTIVES: The aim of the study was to test the antiviral efficacy of a triple nucleoside reverse transcriptase inhibitor (NRTI) regimen, with potential beneficial metabolic effects, as maintenance therapy after induction with dual NRTIs and a boosted protease inhibitor (PI). METHODS: An open-label, noninferiority study was carried out. Antiretroviral therapy (ART)-naïve patients with CD4 count ≤ 350 cells/µL and HIV-1 RNA >30000 copies/mL (n=207) were treated with zidovudine/lamivudine and lopinavir/ritonavir. After achieving HIV-1 RNA <50 copies/mL on two consecutive occasions between weeks 12 and 24 after baseline, 120 patients (baseline: median HIV-1 RNA 5.19 log10 copies/mL; median CD4 count 180 cells/µL) were randomized to receive abacavir/lamivudine/zidovudine (ABC/3TC/ZDV) (n=61) or to continue the PI-based ART (n=59). RESULTS: For the proportions of patients (intention-to-treat; missing=failure) with HIV-1 RNA <400 copies/mL (PI group, 66%; ABC/3TC/ZDV group, 71%) and <50 copies/mL (PI group, 63%; ABC/3TC/ZDV group, 62%) at 96 weeks, switching to ABC/3TC/ZDV was noninferior compared with continuing the PI regimen; the difference in failure rate (ABC/3TC/ZDV minus PI) was -4.4 percentage points [95% confidence interval (CI) -21.0 to +12.3 percentage points] and +0.4 percentage points (95% CI -16.9 to +17.7 percentage points), respectively. In the per protocol analysis, the difference in virological failure for HIV-1 RNA >400 copies/mL (0 of 39 patients in the PI group and two of 45 patients in the NRTI group) and for HIV-1 RNA >50 copies/mL (two of 39 and three of 45 patients, respectively) was +4.4 percentage points (95% CI -2.1 to +11.0 percentage points) and +1.5 percentage points (95% CI -8.6 to +11.7 percentage points), respectively, also showing noninferiority. Serum lipids significantly improved in the NRTI group, but not in the PI arm. CONCLUSIONS: A single-class NRTI regimen after successful induction with standard ART had similar antiviral efficacy compared to continuation of a PI-based regimen at 96 weeks after baseline, with improved serum lipids.