The Effects of Maraviroc Versus Efavirenz in Combination With Zidovudine/Abacavir on the CD4/CD8 Ratio in Treatment-naïve HIV-infected Individuals

The Merit study was a randomized, double-blind, active-comparator multicenter, phase IIb/III study, treatment- naive patients (aged at least 16 years) with R5 HIV-1, and with plasma viral load (HIV-1 RNA) above 2000 copies/ml, who received MVC 300mg q.d., MVC 300mg b.i.d., or EFV 600mg q.d., each in combination with ZDV/3TC 300 mg/150mg b.i.d. Key exclusion criteria included prior treatment with EFV, ZDV, 3TC, or any antiretroviral for more than 14 days at any time, and evidence of resistance to EFV, ZDV, or 3TC, as indicated by the presence of at least one nucleoside-associated mutations conferring resistance to ZDV, or phenotypic resistance to ZDV, at least one mutation conferring resistance to 3TC or phenotypic resistance to 3TC, or at least one mutation responsible for EFV resistance or phenotypic resistance to EFV.

Following a planned analysis at week 16, the MVC q.d. arm was discontinued for not meeting prespecified efficacy criteria, and the study continued with two treatment arms. The sponsor was unblinded at the 48-week analysis time point, but the investigators and patients remained blinded until the 96-week analysis. The study was then fully unblinded following the last patient's 96-week visit, and patients were enrolled in a nominal 3-year open-label phase. Efficacy and safety data from the 240-week (nominal 5-year) study duration have been recently published (Cooper D. et al, AIDS 2014).

The longitudinal data of the Merit study (240-week follow-up) will be analysed. The current long-term follow-up of the MERIT study, the extensive registry of both AIDS and non-AIDS clinical events and the randomization to a therapeutic intervention including maraviroc, will allow to evaluate the effects of maraviroc vs. efavirenz on the CD4/CD8 ratio trajectories.

All randomized subjects included in MERIT will be included in this exploratory post hoc analysis. For the principal objective, longitudinal changes in CD4 and CD8 counts and in the CD4/CD8 ratio will be assessed using generalized estimating equations. Interaction terms will be created to assess whether these changes over time differed significantly between treatment arms. Kaplan-Meier methods will be used to calculate the rates of CD4/CD8 normalization at 0.4 and 1 cut-offs and cumulative probabilities. Cox proportional hazard models will be used to compare probabilities of CD4/CD8 normalization by treatment arm.