PHILADELPHIA, PA—No difference in clinical outcomes was found between subjects with HIV randomly assigned to a non-nucleoside reverse-transcriptase inhibitor (NNRTI) versus ritonavir-boosted protease inhibitor (PI/r)-based initial antiretroviral therapy (ART), a comprehensive meta-analysis of randomized controlled trials concluded at IDWeek 2014.

Results of randomized controlled trials have suggested that NNRTIs may cause faster virological suppression, which PI/rs may recover more CD4 cells, said Álvaro H Borges, MD, MSc, of the Centre for Health & Infectious Diseases Research, Department of Infectious Diseases, Rigshospitalet, University of Copenhagen, Denmark.

However, “It is unknown if differences in immunological and virological endpoints are clinically relevant, as individual randomized controlled trials have not been powered to compare clinical outcomes,” he added.

The investigators included 27 randomized controlled trials in their meta-analysis, which compared NNRTI- vs. PI/r-based initial ART. Primary end point was death or progression to AIDS. Risk ratios of virological suppression were calculated.

Of 9,515 patients in the trials, 4,848 were assigned to an NNRTI; 3,636 received efavirenz, and 1,212 received nevirapine. Of the 4,667 patients on PI/r, 2,661 received lopinavir, 1,498 atazanavir, and 508 received a different PI/r.

Death or progression to AIDS occurred in 305 patients in the NNRTI arm and 294 in the PI/r arm (RR 1.03; 95% CI 0.89–1.19) in 11 randomized controlled trials (n=4,239). Death occurred in 255 patients in the NNRTI arm vs. 249 in the PI/r arm (RR 1.02; 95% CI 0.87-1.28) in 21 randomized controlled trials (n=8671). Progression to AIDS occurred in 194 patients in the NNRTI arm vs. 185 in the PI/r arm (RR 1.06; 95% CI 0.87–1.29) in 10 randomized controlled trials (n=4809).

Overall treatment discontinuation (RR 1.15; 95% CI 0.96–1.38) in 22 randomized controlled trials (n=8,703) and toxicity (RR 1.25; 95% CI 0.87–1.79) in 19 randomized controlled trials (n=6,685), were similar between the NNRTI and PI/r arms. However, in 15 randomized controlled trials (n=5,830), risk of discontinuation due to randomized controlled trial-defined virological failure was higher with NNRTI (RR 1.89; 95% CI 1.02–3.51).

At Week 48, in 13 randomized controlled trials (n=4,010), no difference between NNRTI- and PI/r-based ART was observed for virological suppression (RR 1.04; 95% CI 0.96–1.12), or increase of CD4 cells (MD: -6.2 cells [-16.32–2.42]; NS -6.2 [-15.5, 3.1])

“An individual patient data meta-analysis is warranted to investigate differences in adverse event profiles and the dynamics of immune recovery and virological suppression with individual NNRTI- or PI/r-based ART regimens,” Dr. Borges concluded.