A Per-Protocol Analysis using Inverse Probability of Censoring Weights in a Randomized Trial of Initial PI versus NNRTI Regimens in Children.

Protocol adherence may influence measured treatment effectiveness in randomized controlled trials. Using data from a multicenter trial from Europe and North and South America in 2002-2009 of children with HIV-1 randomized to initial protease inhibitor (PI) versus non-nucleoside reverse transcriptase inhibitor (NNRTI) regimens, we generated time-to-event intention-to-treat (ITT) estimates of treatment effectiveness; applied inverse probability of censoring weights (IPCW) to generate per-protocol efficacy estimates; and compared shifts from ITT to per-protocol estimates across and within arms. In ITT analyses, 263 participants experienced 4-year treatment failure probabilities of 41.3% for PIs versus 39.5% for NNRTIs with risk difference 1.8% (95% confidence interval: -10.1, 13.7) and hazard ratio 1.09 (0.74, 1.60). In per-protocol analyses, failure probabilities were 35.6% for PIs versus 29.2% for NNRTIs with risk difference 6.4% (-6.7, 19.4) and hazard ratio 1.30 (0.80, 2.12). Within-arm shifts in failure probabilities from ITT to per-protocol analyses were 5.7% in PIs versus 10.3% in NNRTIs. Protocol non-adherence was non-differential across arms, suggesting that possibly better NNRTI efficacy may have been masked by differences in within-arm shifts deriving from differential regimen forgiveness, residual confounding, or chance. An IPCW per-protocol approach facilitated evaluation of relationships among adherence, efficacy, and forgiveness applicable to pediatric oral antiretroviral regimens.