Determinants of residual viraemia during combination HIV treatment: Impacts of baseline HIV RNA levels and treatment choice.

Abstract

Effective HIV therapy reflects suppression of plasma HIV RNA levels below assay detection thresholds, although lower levels of "residual viraemia" have also been demonstrated over extended periods of effective antiretroviral treatment. Here we examine the determinants of HIV RNA suppression below the standard assay threshold (40 HIV-1 RNA copies/mL) as well as factors associated with detectable HIV RNA below this reported detection limit. Between 2007 and 2010, 11 575 consecutive viral load (VL) tests were obtained from 1540 patients, including 356 on effective antiretroviral therapy followed since initiation (1996-2001: n = 165; 2002-2009: n = 191). Analyses modelled the probability of an undetectable VL given successful suppression to < 200 copies/mL, and the probability of residual viraemia given an undetectable result. Detectable HIV RNA amplification was demonstrated in 20% of samples with a VL result < 40 copies/mL. Longitudinal analyses from 356 patients revealed that the likelihood of achieving results < 40 copies/mL was increased with current nonnucleoside reverse transcriptase inhibitor (NNRTI) therapy [odds ratio (OR) 2.0; P < 0.05] and reduced with prior virological rebound (OR 0.5; P < 0.05). In contrast, the presence of detectable HIV RNA < 40 copies/mL was strongly associated with pretreatment HIV RNA levels among those on current protease inhibitor (PI) treatment (OR 1.5 per log10 copies/mL increase; P = 0.02) as well as those on NNRTIs (OR 1.7; P = 0.002). While HIV treatment history was associated with plasma HIV RNA levels below the detection limit, residual viraemia results were dominantly determined by pretreatment VL. These findings support the concept of a stable, long-lived reservoir of latently infected cells as a source of residual viraemia despite effective HIV treatment.