Background. IMPAACT-P1060 demonstrated short-term superiority of lopinavir/ritonavir (LPV/r) over nevirapine (NVP) in antiretroviral therapy (ART), regardless of prior NVP exposure (PrNVP). However, NVP-based ART had a marginal benefit in CD4% and growth. We compared five-year outcomes from this clinical trial.

Methods. HIV-infected ART-eligible children were enrolled into two cohorts based on PrNVP and randomized to NVP- or LPV/r-based ART. The Data Safety Monitoring Board recommended unblinding results in both cohorts due to superiority of LPV/r for the primary endpoint: stopping randomized treatment, virologic failure (VF), or death by 6 months. Participants were offered a switch in regimens (if on NVP) and continued observational follow-up. We compared time to VF or death, death, and CD4% and growth changes using intent-to-treat analyses. Additionally, inverse probability weights were used to account for treatment switching and censoring.

Results. As of September 2014, 329 (73%) of the 451 enrolled participants were still in follow-up (median: 5.3 years, IQR: 4.3–6.4), with 52% on NVP and 88% on LPV/r as originally randomized. NVP arm participants had significantly higher risk of VF or death (adjusted hazard ratio [aHR]:1.90, 95% CI, 1.37–2.65) but not death alone (aHR:1.65, 95% CI, .72–3.76) compared to participants randomized to LPV/r. Mean CD4% was significantly higher in the NVP arm up to one year after ART initiation, but not beyond. Mean weight z-scores were marginally higher with NVP arm, but height z-scores did not differ. Similar trends were observed in sensitivity analyses.

Conclusions. These findings support the current WHO recommendation of LPV/r in first-line ART regimens for HIV-infected children.