OBJECTIVE: To determine initial 24-week outcomes among prospectively enrolled patients with failure of initial antiretroviral therapy (ART). METHODS: Baseline virologic failure was defined as HIV-1 viral load greater than 1000 copies/ml. Second-line ART was informed by results of genotype testing and selected from agents in the South-African public sector. Twenty-four week endpoints included virologic suppression and mortality. RESULTS: The cohort consisted of 141 patients (median CD4 cell count and viral load at failure of 173 cells/μl and 17 500 copies/ml). The median prior duration of initial ART was 12.0 months. At least one major resistance mutation was found in 87% of patients.After 24 weeks of follow-up, intent-to-treat virologic suppression (<50 copies/ml) was 65%, as-treated virologic suppression was 78%, the median CD4 cell count improvement was 88 cells/μl and the mortality was 6%. The median CD4 cell count at initial virologic failure among those who died was 70 cells/μl, compared to 182 cells/μl among patients who survived (P = 0.01). Patients with wild-type virus at initial failure (N = 19) had inferior outcomes after switch. The presence of nucleoside analogue resistance mutations at failure did not affect early efficacy of boosted-protease inhibitor regimens. CONCLUSIONS: Virologic monitoring linked to resistance testing helped demonstrate the efficacy of lopinavir/ritonavir-containing second-line regimens in South Africa. A switch to second-line regimens in patients with virologic failure and drug resistance has substantial and rapid immunological and clinical benefits. Resistance testing identified a high-risk group without resistance who might benefit from increased medication access and/or adherence support. Copyright © 2010 Lippincott Williams & Wilkins.