Background: Treatment management after repeated failure of antiretroviral therapy (ART) is difcult due to resistance and adherence challenges. For people who have failed non-nucleoside reverse transcriptase inhibitor-(NNRTI-) and protease inhibitor-(PI-) based regimens with no or limited resistance, remaining on PI-based ART is an option. Using data from an ART strategy trial (A5288) in low/middle-income countries which included this option, we explored whether predictors can be identifed distinguishing those who experienced further virologic failure from those who achieved and maintained virologic suppression. Methods: A5288 enrolled people with confrmed HIV-1 RNA≥1000 copies/mL after ≥24 weeks of PI-based ART and prior failure on NNRTI-based ART. This analysis focused on the 278 participants with no resistance to the PI being taken and no or limited nucleoside reverse transcriptase inhibitor (NRTI) resistance, who continued their PI with fexibility to change NRTIs. Proportional hazards models were used to evaluate predictors of virologic failure during follow-up (VF: confrmed HIV-1 RNA≥1000 copies/mL at ≥24 weeks of follow-up). Results: 56% of participants were female. At study entry, median age was 40 years, time on ART 7.8 years, CD4 count 169 cells/mm3, HIV-1 RNA 20,444 copies/mL; and 37% had NRTI resistance. The estimated proportion experiencing VF increased from 39% at week 24 to 60% at week 96. In multivariable analysis, signifcant predictors at study entry of VF were higher HIV-1 RNA (adjusted hazard ratio: 2.20 for ≥10,000 versus <10,000 copies/mL), lower age (1.96 for <30 versus ≥30 years), NRTI resistance (1.74 for present versus absent), lower CD4 count (1.73 for <200 versus ≥200 cells/mm3), and shorter ART duration (1.62 for <10 versus ≥10 years). There was a strong trend in proportion with VF at week 96 with the number of these fve risk factors that a participant had, varying from 8% for zero, to 31%, 40%, 73%, and 100% for one, two, three, and four/fve. Only 13% of participants developed new NRTI or PI resistance mutations.