The epidemiology and impact of pretreatment HIV drug resistance in adults in South Africa.
MetadataShow full item record
HIV drug resistance (HIVDR) present prior to initiating or re-initiating antiretroviral therapy (ART), is known as pretreatment drug resistance (PDR). Conventionally, PDR is detected by Sanger sequencing. Drug resistant minority variants (DRMVs) that are not reliably detected by Sanger sequencing can be detected by next generation sequencing. The aims of this research were to assess levels of PDR in HIV hyper-endemic areas (with high HIV incidence and prevalence) in KwaZulu-Natal (KZN) province, trends of PDR in South Africa, and the impact of DRMVs on ART. To assess PDR in adults from KZN hyper-endemic areas, 1845 sequences were analyzed from two population-based HIV surveillance studies; a longitudinal HIV surveillance programme in northern KZN (2013-2014), and the HIV Incidence Provincial Surveillance System (HIPSS) in central KZN (2014-2015). Overall, 182/1845 (10.0%) had NNRTI-PDR mutations, and when analyzed by study year, NNRTI-PDR was 10.2% (CI:7.5-12.9) for the HIPSS study in 2014. To assess PDR trends in South Africa, 6880 HIV-1 sequences were collated from 38 datasets of ART-naïve adults (2000-2016). Increasing levels of PDR were observed, most marked from 2010. Crude pooled prevalence of NNRTI-PDR reached 10% in 2014, with a 1.18-fold (CI:1.13- 1.23) annual increase (p<0.001), consistent with findings from the HIPSS data. This provided the first evidence of high-level NNRTI-PDR in KZN and South Africa, supporting the transition to dolutegravir in standard first-line ART, as recommended by the World Health Organization when NNRTI-PDR reaches ≥10%. A case-control (2:1) study in HIV/TB co-infected adult patients was done to assess the impact of DRMVs at different thresholds. Cases were patients that initiated ART and had viral loads ≥1000 copies/mL after ≥6 months on ART, and controls were those that initiated ART and achieved virologic suppression through 24 months. Pre-ART NNRTI-resistance was associated with ART failure. NGS improved detection of HIVDR at lower thresholds, but reduced the specificity of identifying patients at risk of virologic failure, with the specificity reducing from 97% (CI:92-99) at 20% threshold, to 79% (CI:71-86) at 2% threshold. In all, the findings presented in this thesis provide a broad message about the need to improve quality in HIV prevention and treatment services.