Three-year immune reconstitution in PI-sparing and PI-containing antiretroviral regimens in advanced HIV-1 disease.

Background: The long-term immunological benefit of protease inhibitor (PI)-sparing antiretroviral therapy (ART) using non-nucleoside reverse transcriptase inhibitors (NNRTIs) remains poorly investigated.
Methods: A total of 120 ART-naive, HIV-1-infected participants were included in the immunology substudy of INITIO, an international randomized trial comparing two NRTIs (didanosine + stavudine) combined with either: one NNRTI (efavirenz; EFV), one non-boosted PI (nelfinavir; NFV), or one NNRTI + one PI (EFV/NFV). CD4+ T-cell counts, HIV-1 plasma RNA load (VL), T-cell phenotype, T-cell proliferation and IFN-gamma production against opportunistic/recall and HIV-1 antigens/peptides were compared at baseline and at week (W) 96 and W156.
Results: Participants (37 EFV, 44 NFV, 39 EFV/NFV) had similar baseline VL; median CD4+ T-cell counts/mm3 were: 144 (64-303) EFV, 212 (42-313) NFV and 257 (86-331) EFV/NFV. At W156, the proportion of patients with VL < or =50 copies/ml was not different between the arms (P=0.3). From baseline to W156 there was a significant increase in CD4+ T-cell counts (P<0.001) and in naive CD4+ T cells (P<0.001), with no difference between arms and percentages of total and activated CD8+ T cells decreased significantly (P<0.001) in all arms. The decrease in activated memory CD4+ T-cells was significantly greater in the EFV arm at W96 (P=0.03) and W156 (P=0.01), but did not persist after adjusting for baseline CD4+ T-cell counts. During follow-up, responses to opportunistic pathogens increased in all patients while specific T-cell responses to HIV-1-p24 and gp160 recombinant proteins or to Gag and Nef peptides were not restored.
Conclusion: Regimens using/sparing PIs provide similar levels of long-term immune reconstitution even in patients with low CD4+ T-cell counts.