CROI 2017 Abstract e-Book

Abstract eBook

Oral Abstracts

112 COST-EFFECTIVENESS OF POLICY OPTIONS WHEN PRETREATMENT NNRTI DRUG RESISTANCE IS HIGH Andrew Phillips 1 , Valentina Cambiano 1 , Michael Jordan 2 , Fumiyo Nakagawa 1 , Marco Vitoria 3 , Meg Doherty 3 , Paul Revill 4 , Elliot Raizes 5 , Silvia Bertagnolio 4 , for the Working Group on Modelling Drug Resistance in Sub-Saharan Africa 1 Univ Coll London, London, UK, 2 Tufts Univ, Boston, MA, USA, 3 WHO, Geneva, Switzerland, 4 Univ of York, York, UK, 5 CDC, Atlanta, GA, USA Background: The World Health Organization recommends drug resistance surveillance in people initiating antiretroviral therapy (ART). If high levels of pre-ART non-nucleoside reverse transcriptase inhibitor (NNRTI) resistance are identified, options for ART program response must be considered. One possible option is to transition from efavirenz to dolutegravir, which is associated with a lower rate of resistance acquisition and greater tolerability. Methods: We used an individual-based model of HIV transmission and the effect of ART which considers specific drugs and resistance mutations (HIV Synthesis Model). Multiple potential epidemics/ART programs (setting scenarios) were generated through simulation, informed by data from sub-Saharan Africa. Parameters relating to ART adherence and interruption rate, ART monitoring strategy, and switch rate to a 2nd-line regimen after 1st-line failure were varied randomly within bounds reflecting data from the region. For each setting scenario, if the pre-ART resistance in 2016 was >10%we compared outcomes of potential policy options (Table) over 2016-2036 (20 yr time horizon). Costs of all aspects of HIV testing and care included, dolutegravir cost $44, efavirenz $38, cost effectiveness threshold $500, health systems perspective, 3% discount rate. Results: In over 2000 setting scenarios median HIV prevalence was 8% (5%-95% 5-17), 62% of HIV+ people were on ART (90% range 44-76), and 20% of ART initiators had prior drug exposure (90% range 9-34). Of ART initiators without / with prior ARV exposure the %with NNRTI resistance in majority virus was 9% (2-20) / 16% (5-34). As shown in the Table, a policy of transitioning from efavirenz to dolutegravir for all on 1st-line ART was predicted to lead to improved health outcomes and was cost effective (incremental cost effectiveness ratio (ICER) $80 per disability adjusted life year (DALY) averted). Conclusions were consistent in sensitivity analyses including a dolutegravir cost of $80/year. Updated results will be presented considering a wider range of policy options and extended sensitivity analyses. Conclusion: A future transition from efavirenz to dolutegravir may be cost effective in low income settings in sub-Saharan Africa. The level of pre-ART NNRTI drug resistance will be just one factor to consider when estimating the potential impact of this transition. Further studies, such as stepped-wedge trials, should be conducted to understand the real- life impact of such a transition.

Oral Abstracts

113 INCREASED RISK OF CART FAILURE AFTER LOW-LEVEL VIREMIA UNDER WHO GUIDELINES Lucas E. Hermans 1 , Michelle A. Moorhouse 2 , Sergio Carmona 2 , Diederick Grobbee 3 , Laura Marije Hofstra 3 , Douglas D. Richman 4 , Hugo Tempelman 5 , Willem D. Venter 2 , Annemarie Wensing 3 , for the Ndlovu Research Consortium 1 Univ Med Cntr Utrecht, Utrecht, Netherlands, 2 Univ of Witwatersrand, Johannesburg, South Africa, 3 Univ Med Cntr Utrecht, Utrecht, Netherlands, 4 Univ of California San Diego, San Diego, California, USA, 5 Ndlovu Care Group, Groblersdal, South Africa Background: Current WHO guidelines for cART in HIV-1 infected patients define failure of cART as viremia above 1000 copies/mL during therapy. Detectable viral load (VL) below 1000 copies/mL during cART (low-level viremia; LLV) has been linked to subsequent failure of cART in studies performed in high-income settings, where more stringent VL cut-off values are used. We report the prevalence of LLV and its impact on subsequent failure of cART in a large South African cohort managed according to WHO guidelines. Methods: HIV-positive patients from 19 urban and 38 rural South African HIV treatment sites were studied. Adult patients on cART for ≥20 weeks and with ≥1 VL performed ≥20 weeks after start of cART were included. LLV was defined as viremia between 50-1000 copies/mL and stratified according to level (51-199, 200-399, and 400-999 copies/mL) and duration. Studied outcomes were failure of cART (VL ≥1000 copies/mL) and switch to second line cART. In the subset of patients with ≥100 weeks of first line cART without failure and ≥3 VLs risk of failure after LLV was estimated with survival analysis using Cox proportional hazard models corrected for sex, age and nadir CD4. Results: 69,615 patients met inclusion criteria. Virological suppression <1000 copies/mL during cART was maintained in 80.9% of patients. LLV occurred in 23.3% of patients. A single measurement of LLV (sLLV) was more common than persistent LLV (pLLV) (78.6% vs 21.4%). LLV between 51 and 199 copies/mL (LLV51-199) was most commonly encountered (59.1%). In survival analysis (26,305 patients) LLV was associated with increased hazard of failure of cART (HR 2.8; CI-95% 2.7-2.9) and of ever switching to second line (HR 3.2; CI-95% 3.0-3.3) when compared to patients with <50 copies/mL. Risk of failure increased proportionally to level of LLV: HR 1.7 (CI-95% 1.5-1.9) for sLLV51-199, 2.7 (CI-95%

43

CROI 2017