Background: IMPAACT P1060 demonstrated short-term superiority of lopinavir/ritonavir (LPV/r)-based ART over nevirapine (NVP) for the primary endpoint (stopping randomized treatment, virologic failure [VF] or death by 24 weeks) in HIV-infected children regardless of NVP exposure at birth. In contrast, those on NVP had marginally superior improvements in CD4% and growth (weight and height z-scores). Longer-term outcomes are presented.
Methods: HIV-infected ART-eligible infants and children (2m – 3y) from 6 African countries and India enrolled into 2 cohorts based on prior NVP exposure (PrNVP) and were randomized to initiate NVP or LPV/r (with zidovudine and lamivudine). The study DSMB recommended closing enrollment and unblinding cohorts with PrNVP (2009) and no PrNVP (2010) due to superiority of the LPV/r arm for the primary endpoint. Participants could switch regimens and continue in observational follow-up. Randomized and observational data were combined in intent-to-treat (ITT) analyses to investigate long-term trends in VF and death (Cox proportional hazards models) and CD4% and growth (generalized estimating equations). Additional analyses were performed using marginal structural models (MSM) to account for treatment switching and censoring. Models were adjusted for demographics and HIV disease status at entry.
Results: 229 participants were randomized to NVP and 222 to LPV/r based ART. As of January 2014, 75% were still in follow-up (median follow-up 4.6 years [IQR: 3.7-5.7]). From their original randomization, 48% were still on NVP while 81% were still on LPV/r. Participants in the NVP arm had significantly shorter time to VF (adjusted hazard ratio [aHR]: 1.91, 95%CI: 1.37-2.65) but not death (aHR: 1.64, 95%CI: 0.72-3.75). Mean CD4% and weight z-scores were higher in the NVP arm at 1 year, by 1.5% and 0.23 respectively (p<0.05), while height z-scores were not significantly higher by 0.15 (p=0.10). By the second year of follow-up, differences were no longer statistically significant. Similar trends were observed in MSM models for all outcomes.
Conclusions: Long-term virologic suppression was superior in children on LPV/r-based ART compared to NVP-based regimens. Early modest gains in CD4% and growth associated with NVP were no longer statistically significant beyond 1 year after ART initiation. These findings further support the current WHO recommendation for LPV/r-based ART as first line therapy for HIV-infected children aged < 3 years.