Objective: To assess the durability of the undetectability of HIV plasma viraemia (pV) and to determine the factors associated with virological rebound (VR) in HIV-infected adults on protease inhibitor (PI)-sparing highly active antiretroviral therapy (HAART). The development of resistance mutations during virologically successful therapy and VR was also analysed.
Materials and methods: One hundred and twenty-six HIV-infected adults on PI-sparing HAART were prospectively followed from April 1998 to December 2002: Group 1, naive for antiretroviral drugs (n = 26); Group 2, previously PI-HAART-exposed patients (n = 19); Group 3, previously exposed to suboptimal therapy (n = 81). Genotypic resistance tests on peripheral blood mononuclear cells or on plasma RNA (when feasible) were carried out when undetectable HIV pV was demonstrated for at least 48 weeks. Additionally, patients showing a therapy adherence >95% developing VR were also tested at rebound, at simplification and during previous suboptimal therapy exposure.
Results: The median follow-up time was 630 [329-903] days. VR was considered as two consecutive pV levels >50 copies/mL. Twenty-two (17.5%) patients developed VR. Only therapy adherence <95% was independently associated with VR (adjusted hazard ratio: 8.42; 95% CI: 3.33-21.27). Twenty (40%) of the 50 patients with pV < 50 copies/mL for at least 48 weeks showed at least one thymidine-associated mutation (TAM) but none had NNRTI-resistance mutations. Ten (83.3%) of 12 available adherent patients showing VR harboured NNRTI-resistance-associated mutations; 50% of them were considered as wild-type strains at simplification time. However, the TAM number and resistance mutations profile found on suboptimal exposure were very similar to those found at VR on simplification therapy.
Conclusions: PI-sparing HAART allows maintenance of successful long-term control of HIV replication, adherence to therapy being the main factor associated with VR. However, a small proportion of patients on simplification regimen may develop VR regardless of therapy compliance. VR on PI-sparing HAART is characterized by the emergence of NNRTI cross-resistance mutations. Finally, TAMs 'archived' during previous suboptimal exposures are partially involved in subsequent VR on simplification HAART.