CROI 2019 Abstract eBook

Abstract eBook

Poster Abstracts

y.o. men and in 20-24 y.o. women. Over the trial period these heterogeneities in contributions to the epidemic were increased (fig 1). Beyond the trial, PAF direct of 30-34 yo men was >15% in 2023-2028, twice that of women of the same age. Conclusion: This work illustrates the significant contribution of 25-34 y.o. men to HIV transmission in generalised HIV epidemics in SSA, and of young people when considering indirect transmissions. Future interventions that target sub- populations in SSA need to address such discrepancies. Estimates of PAF direct may be complemented from phylogenetic studies.

was 0.1% (range: -4.55, 12.4). In MSM clusters, median HIV incidence was 1.44% (range: 0.47, 2.96) and median absolute change was 0.04% (range: -2.08, 1.94). In adjusted cluster-level analysis, a one percentage point increase in PV was associated with a 0.38 percentage point increase in HIV incidence (95% CI: 0.2, 0.6; Figure). Individual risk factors of incident HIV in PWID included age, gender, marriage, education, recent needle sharing, sex work and HCV, and among MSM included age, income, recent male and female sexual partners and injection drug use. After controlling for individual risk, PV at baseline (incidence rate ratio [IRR] per 10% increase in PV: 2.51; 95% CI: 1.53, 4.13) and follow-up (IRR per 10% increase: 2.80; 95% CI: 1.79, 4.38) were significantly associated with individual incident infections at follow-up. Conclusion: Prior studies have demonstrated a strong cross-sectional association between community PV and HIV incidence. We provide further evidence of the role of PV as a surrogate for incidence by demonstrating that change in community PV predicts change in incidence and PV predicts future incident HIV infections independent of individual risk.

Poster Abstracts

835 HIV EPIDEMIC IN A BRIDGE POPULATION IN UKRAINE CALLS FOR NEW PREVENTION STRATEGIES

834 COMMUNITY PREVALENCE OF VIREMIA: A NEW STANDARD FOR BIO- BEHAVIORAL SURVEYS Sunil S. Solomon 1 , Gregory M. Lucas 1 , Allison M. McFall 2 , Aylur K. Srikrishnan 3 , Oliver Laeyendecker 1 , Muniratnam S. Kumar 3 , Syed Iqbal 3 , Shanmugam Saravanan 3 , Nandagopal Paneerselvam 3 , Pachamuthu Balakrishnan 3 , David D. Celentano 2 , Shruti H. Mehta 2 1 Johns Hopkins University School of Medicine, Baltimore, MD, USA, 2 Johns Hopkins Bloomberg School of Public Health, Baltimore, MD, USA, 3 YR Gaitonde Center for AIDS Research and Education, Chennai, India Background: Monitoring HIV programs requires estimates of HIV incidence which can be challenging to obtain. We characterize the change in HIV incidence over 4 years among people who inject drugs (PWID) and men who have sex with men (MSM) across 22 cities in India and evaluate associations with community prevalence of viremia (PV). Methods: These data are from a cluster-randomized trial among PWID (12 clusters) and MSM (10 clusters) in India that included baseline (2012-13) and follow-up (2016-17) respondent-driven sampling surveys of ~1000/cluster. Cross-sectional HIV incidence was estimated using a validated multi-assay algorithm incorporating LAg Avidity EIA, BioRad Avidity assay, CD4+ count and HIV RNA (limit of detection>150 copies/ml). PV was estimated as the percentage of persons with detectable HIV RNA in the cluster. Cluster-level linear regression assessed the association between change in PV and change in HIV incidence over 4 years controlling for study arm and baseline PV. Multi-level Poisson regression assessed the association between baseline and follow-up PV and individual-level incident HIV risk at follow-up accounting for individual-level correlates using a risk score Results: The median HIV incidence in PWID clusters at follow-up was 5.16% (range: 0, 18.5); the median absolute change (baseline to follow-up) in incidence

Marina Kornilova , Yana Sazonova Alliance for Public Health, Kyiv, Ukraine

Background: Ukraine is experiencing one of the severest HIV epidemic in the World. From its onset in 90th it was primarily an IDU driven. The tremendous efforts that community-based organizations investing in HIV prevention slow down the disease spread among key populations (KP) including people who inject drugs (PWID). The HIV prevalence has stabilized at the level of 22% (21. 9% in 2015 and 21.6% in 2017 sentinel surveys). But when the prevalence is high there is always a risk that the disease may spread to the bridging population. To estimate the HIV burden in one of the bridge groups we conducted the survey among sexual partners of PWID. Methods: The study population included 769 respondents from 10 cities of Ukraine recruited through a linked RDS in a period of May-August 2015. The sample included those who had sexual intercourse with the PWID in the last 90 days and haven’t inject drugs for the last 30 days. All participants were confidentially interviewed to assess their HIV risk behavior. Blood samples were tested for HIV, hepatitis B, C and syphilis by rapid combo tests. The analysis were done in SPSS 21.0 and RDS-Analyst. Results: Among all the participants 87,3%were females. Medium age – 32,2. 75,5% declared as never injected (72.4-78.5%). 97,6%were regular sex partners of their PWID-recruiters. Only 48.5% used condom during the last intercourse and 35.9% never used condom in the last 90 days. The HIV prevalence in the group of non-injecting sexual partners of PWID as high as 15,0% (12,4 – 17,5%). Among those who declared themselves as never injected – 9,2% (6,8 – 11,6%) and among those who never injected and were HCV-negative - 5,6% (4.2 – 6.8%). The prevalence of syphilis was 4,7% (3.2 – 6.2%). Only 4.6% of the respondents were clients of the HIV prevention programs; 3.6% have received

CROI 2019 325