BACKGROUND: Limited viral load (VL) testing in human immunodeficiency virus (HIV) treatment programs in low-income countries often delays detection of treatment failure. The impact of remaining on failing protease inhibitor (PI)-containing regimens is unclear. METHODS: We retrospectively tested VL in 2164 stored plasma samples from 386 patients randomized to receive lopinavir monotherapy (after initial raltegravir induction) in the Europe-Africa Research Network for Evaluation of Second-line Therapy (EARNEST) trial. Protease genotypic resistance testing was performed when VL >1000 copies/mL. We assessed evolution of PI resistance mutations from virological failure (confirmed VL >1000 copies/mL) until PI monotherapy discontinuation and examined associations using mixed-effects models. RESULTS: Median post-failure follow-up (in 118 patients) was 68 (interquartile range, 48-88) weeks. At failure, 20% had intermediate/high-level resistance to lopinavir. At 40-48 weeks post-failure, 68% and 51% had intermediate/high-level resistance to lopinavir and atazanavir; 17% had intermediate-level resistance (none high) to darunavir. Common PI mutations were M46I, I54V, and V82A. On average, 1.7 (95% confidence interval 1.5-2.0) PI mutations developed per year; increasing after the first mutation; decreasing with subsequent mutations (P < .0001). VL changes were modest, mainly driven by nonadherence (P = .006) and PI mutation development (P = .0002); I47A was associated with a larger increase in VL than other mutations (P = .05). CONCLUSIONS: Most patients develop intermediate/high-level lopinavir resistance within 1 year of ongoing viral replication on monotherapy but retain susceptibility to darunavir. Viral load increased slowly after failure, driven by non-adherence and PI mutation development. CLINICAL TRIALS REGISTRATION: NCT00988039.