Background:  NNRTI- and NRTI-resistance are increasingly reported in SA after the rollout of NNRTI-based cART. PI-based regimens are seldom used as first line therapy in this region. Phidisa II was a 2x2 factorial randomized, controlled trial comparing LPV/r vs EFV + ZDV/ddI vs. d4T/3TC in ART-naïve subjects. This substudy sought to examine the incidence and type of genotypic resistance in subjects who had viral suppression (viral load (VL) <400 copies/mL) at 6-months, with subsequent VL rebound.

Methods:  Phidisa II subjects who had VL <400 copies/mL at 6-month, and had 2 consecutive VL >1000 copies/mL, were identified. Genotypic testing (GT) was performed on stored samples from the date of the 2^nd consecutive viral rebound. Pre-cART samples of those subjects with major genotypic mutations at rebound, were then sent for GT to identify pre-cART mutations.

Results:  Seventy-one of 1,771 Phidisa II subjects fulfilled the criteria of suppression at 6-month with subsequent rebound. The median age was 35.7 years, 38% were female, median baseline CD4 = 107 cells/mm³, VL = 5.2 log₁₀ copies/mL, 94 % were HIV-1 subtype C. Median time to rebound was 13.9 months, median VL at rebound = 3.6 log₁₀ copies/mL. Major PI mutations were detected in 1/36 (2.8%) LPV/r recipients, whereas NNRTI mutations were detected in 17 of 33 (51.5%) EFV recipients. K103N and V106M were the most common NNRTI mutations, detected in 9 and 5 samples, respectively. No Y181C mutation was seen. NRTI mutations were detected in 17/61 (27.9%) samples. At rebound, 11/34 (32.4%) 3TC recipients had M184V mutations. D67N+K70R mutations were found in 2/28 (7.1%) ZDV/ddI recipients (both were on EFV, with NNRTI mutations), other thymidine-associated mutations (TAM) were not seen. Both K103N + M184V mutations were reported in 7/13 (53.8%) subjects, who received both EFV and 3TC. No resistance mutations were detected at baseline. Subjects with RT mutations at failure had significantly lower median CD4 count (218 vs 278 cells/mm³, P =0.05) and significantly higher VL (4.1 vs 3.6 log₁₀ copies/mL, P =0.04) at rebound.

Conclusion:  In this randomized trial, despite failure on LPV/r, major PI mutations were rarely detected, whereas NNRTI- and M184V mutations are most frequently seen at rebound in those who received EFV +/- 3TC. Patients on EFV+d4T+3TC and had these mutations were essentially on d4T monotherapy at rebound. LPV/r-based regimen maybe reasonable first line option in SA and has less likelihood for resistance at failure.