Abstract
Fifteen years after the introduction and expanded access to antiretroviral therapy (ART) in resource limited countries (RLC), viral suppression, a key objective of its effectiveness is not achieved in many patients infected with HIV. Unlike developed countries, where often more powerful and robust ART coupled with a routine virological monitoring limit the emergence of resistance, the current situation of RLC is more worrying. Operational research conducted here aims to improve care and monitoring of patients living in these countries, and receiving ART according to the WHO public health approach. While recommended routinely since 2013, HIV viral load (VL) is rarely available and monitoring remains primarily clinical or immunological.In several countries in Central Africa and West, we documented high rates of HIV drug resistance (HIVDR) (99%) and multi-drug resistance (97%) in patients with virologic failure after long-term first-line ART, consisting of 2 nucleoside reverse transcriptase inhibitors (NRTI, 3TC+AZT/d4T) associated with 1 non-NRTI (NNRTI, EFV/NVP). This multi-drug resistance also reduces significantly the effectiveness of second-generation NNRTIs and of the NRTI, TDF, which is now currently recommended in 2nd line by the WHO. More unusually, we also reported that high VLs (> 5 log 10 copies / ml) are associated with multi-resistance rather than an insufficient adherence in these patients who were largely asymptomatic. Without improvement of patient monitoring, through expanded access to VL testing to enable early diagnosis of virological failure, a potential new epidemic caused by HIVDR strains could emerge in sub-Saharan Africa. This could jeopardize efforts to achieve the "90-90-90" objective (90% of diagnosed infections, 90% of diagnosed patients on ART, 90% of patients treated durable virologic success) of UNAIDS and WHO in 2020. Moreover, our second study in a district hospital in Cameroon illustrates that these ambitious targets will probably be even more difficult to achieve in decentralized areas, often having inadequate laboratory infrastructure to acquire complex VL platforms.We then successfully transferred an open and polyvalent universal VL assay in a national reference laboratory in Cameroon. This assay can detect HIV-1 group M, N, O and P and their simian ancestors. With this RT-qPCR HIV-1/SIVcpz/SIVgor assay, HIV-1 M variants were better quantified (+0.47 log10 copies / mL) compared to Abbott RealTime HIV-1 assay which is approved by the FDA. At the VL threshold of virological failure defined by the WHO (1000 copies/mL), the correlation of the results between the two tests was 95%, with a higher sensitivity for the new test.We finally explored how to adapt this open test at optimal large-scale application for dried blood spot (DBS), to make VL testing available in decentralized areas, given the current lack of point of care test for VL. We compared different extraction methods of nucleic acids that could potentially reduce the contribution of the pro-viral DNA present in DBS, with the aim to improve their specificity without success. Instead, we have shown that with most of the RT-qPCR VL assays, many patients would be inappropriately eligible to perform a repeated VL testing (or switched to a costly 2nd line ART), which increase significantly costs of national programs.In conclusion, the studies from this thesis have contributed to improve current knowledge about the effectiveness of ART programs and services and also on the use of virological monitoring tools adapted to patients living with HIV in RLCs.