CROI 2019 Abstract eBook

Abstract eBook

Poster Abstracts

adjusting for age, gender, VL, CD4 count, and mid-upper arm circumference (MUAC). Subjects with VL ≥500 cpmwithout DRM and those with DRM pre-ART were excluded from this analysis. Results: Among 621 individuals with VL data at 6 and/or 12 months after starting ART (17.7%with TB), 101 had VL ≥500 cpm (16.3%); DRM was detected in 64/98 (65.3%; sequencing failure in 3 subjects). All 64 had NNRTI resistance and 35 (54.7%) had both NRTI and NNRTI resistance. Pre-ART resistance (according to WHO SDRM 2009) was detected in 7/56 (12.5%; pre-ART samples missing from 8 subjects). Acquisition of DRMwas associated with pre-ART VL, CD4 count, and MUAC in univariate analysis (Table 1). In multivariate analysis, VL and MUAC remained significantly associated with acquired DRM. TB was not associated with acquisition of DRM; 12/64 (18.8%) patients with DRM had TB. Conclusion: DRM was detected in a majority of patients with VL ≥500 cpm during ART at Ethiopian health centers, and was mostly due to acquired drug resistance. The risk of DRM acquisition was associated with high pre-ART VL and lowMUAC, but was not associated with concomitant TB.

sustainability of a national ART programme. Nationally representative HIVDR surveys should be routinely implemented to assess predicted efficacy of current and possible future ARV regimens as the programme expands.

538 MODELING THE IMPACT OF DOLUTEGRAVIR INTRODUCTION ON NNRTI RESISTANCE IN SOUTH AFRICA Anthony Hauser 1 , Katharina Kusejko 2 , Leigh F. Johnson 3 , Huldrych F. Günthard 2 , Julien Riou 1 , Gilles Wandeler 4 , Matthias Egger 1 , Roger Kouyos 2 1 Institute of Social and Preventive Medicine, Bern, Switzerland, 2 University Hospital Zurich, Zurich, Switzerland, 3 University of Cape Town, Cape Town, South Africa, 4 University Hospital of Bern, Bern, Switzerland Background: The success of the global scale-up of antiretroviral therapy (ART) is threatened by rising HIV resistance to non-nucleoside reverse transcriptase inhibitors (NNRTI). First-line regimens including dolutegravir (DTG) could help mitigate this problem due to its higher genetic barrier to resistance. We extended the previously-developed MARISA model to simulate the impact of DTG introduction on NNRTI pre-treatment drug resistance (PDR) in South Africa in 2018-2038. Methods: MARISA (Modelling Antiretroviral drug Resistance In South Africa) is an epidemiological model describing the emergence and transmission of NNRTI resistance from 2005. It is parameterized with data from IeDEA Southern African cohorts, and with Thembisa/UNAIDS and literature estimates. We extended MARISA to account for the introduction of DTG in 2018 under two scenarios: DTG as first-line regimen for ART-initiators, or DTG for all patients. Considering potential safety issues of DTG during pregnancy, we assessed the impact of prescribing DTG to a) all eligible men, b) all men and all women out of reproductive age (older than 49, that is 17.5% of adult women on ART in 12 IeDEA South African cohorts), c) all men and all women out of reproductive age or using modern contraception methods (62% of adult women according to World Bank data), and d) all men and all women. For DTG, we assumed a similar efficacy compared to NNRTI but no resistance. Results: The model projections show that introducing DTG would lead to a noticeable reduction of NNRTI PDR in all scenarios compared to the continuation of the current situation (Fig 1). DTG could lead to the virtual disappearance of NNRTI PDR (1% by 2038) if given to all adult patients regardless of treatment status and gender. Limiting DTG to ART-initiators would allow for the stabilization then decrease of NNRTI PDR to 13% by 2038. In both scenarios, NNRTI PDR would continue to rise if DTG is restricted to men, but may be reduced to 12% by 2038 if also provided to women aged >49, or even to 1% if also provided to women using contraception. Conclusion: Our model shows the potential benefit of the introduction of DTG for attenuating the rise of NNRTI PDR. As safety issues related to neural tube defects in newborns may limit the use of DTG in women with child-bearing potential, the model shows that the effect of introducing DTG would be largely reduced if its use is limited to men only. However, this can be almost completely overcome if DTG is used in women with low risk of pregnancy.

537 BASELINE HIV DRUG RESISTANCE AMONG TREATMENT-NAIVE PATIENTS IN ESWATINI, MAXART TRIAL Shaukat Khan 1 , Anita Hettema 1 , Fiona Walsh 2 , Sikhathele Mazibuko 3 , Velephi Okello 3 , Charlotte Lejeune 1 , Jilian Sacks 2 , Michael R. Jordan 4 , John Koo 5 , P. Richard Harrigan 6 , Chanson J. Brumme 5 , Hope R. Lapointe 6 1 Clinton Health Access Initiative, Mbabane, Swaziland, 2 Clinton Health Access Initiative, Boston, MA, USA, 3 Ministry of Health, Mbabane, Swaziland, 4 Tufts University, Boston, MA, USA, 5 British Columbia Centre for Excellence in HIV/AIDS, Vancouver, BC, Canada, 6 University of British Columbia, Vancouver, BC, Canada Background: The Kingdom of Eswatini (formerly known as Swaziland) has the highest global adult prevalence of HIV at 27.2%. The country has expanded access to HIV testing services and increased antiretroviral treatment (ART) coverage in recent years from 49% (2013) to 85% (2017). The MaxART Early Access to ART for All Implementation Trial launched in 2014 to assess the scalability and clinical outcomes of offering ART to all people living with HIV (PLHIV) in Eswatini, regardless of CD4 cell count and WHO clinical stage. As a secondary endpoint, we sought to determine the extent of HIV drug resistance (HIVDR) in all treatment-naïve individuals initiating ART in the Hhohho region of Eswatini. Methods: The trial was a 3-year randomized stepped-wedge design open to enrolment for PLHIV attending 14 rural health facilities in the Hhohho region. Exclusion criteria included age (<18yo), pregnancy, breastfeeding, and previous antiretroviral (ARV) drug exposure except for prevention-of-mother-to-child- transmission interventions. Pre-ART plasma samples were genotyped at the BC Centre for Excellence in HIV/AIDS, Canada. Sanger sequences were generated targeting the protease and reverse transcriptase genes. HIVDR was predicted using the Stanford HIVdb algorithm (v.8.6.1). Results: 3485 PLHIV were enrolled, with pre-ART samples and HIV sequences available for 2626 (75.4%) and 2585 (74.2%) participants, respectively. HIVDR was detected in 658 (25.5%) sequences, with 289 sequences (11.2%) containing mutations conferring HIVDR to first-line drugs efavirenz/nevirapine (EFV/ NVP; Table 1). E138A alone was detected in over 13% of sequences, accounting for over half of the inferred HIVDR to non-nucleoside reverse transcriptase inhibitors (NNRTI). HIVDR to EFV/NVP was associated with being female, younger age, and CD4 cell count ≥ 200 cells/mm3 (p<0.05). Female sex was also associated with longer time from diagnosis to ART initiation, and less advanced clinical status at enrolment (p<0.05). Dual-class HIVDR to nucleoside reverse transcriptase inhibitor and NNRTI drugs was rare (n=15/2585). Conclusion: Moderate levels of HIVDR to EFV and NVP warrant re-assessment of first-line ARV regimens in Eswatini. These findings are critical to evaluate the national progress towards 90-90-90 targets and assess the long-term

Poster Abstracts

CROI 2019 202