CROI 2019 Abstract eBook

Abstract eBook

Poster Abstracts

relationship durations) and age-specific herd immunity (ASHI) that protects adolescents entering the sexually active population. In sensitivity analyses, we found ASHI was the biggest driver of the success of age-based TasP. Over time, ASHI gives rise to an ever-expanding “AIDS-free generation” that drives HIV to extinction. Conclusion: As testing rates increase in response to UNAIDS 90-90-90 goals, we suggest that efforts to link all young people to care and treatment could be an effective long-term strategy for ending the HIV epidemic. Youth focused treatment will be particularly important in low and middle income countries with demographic ‘youth bulges’ that are increasing the number of young people at risk for infection.

estimated the relative difference in cumulative HIV incidence between 2020 and 2030. Results: By 2017, the modeled HIV prevalence was 17% overall (total population, including FSW), and 43% among FSW. Under Scenario A, HIV incidence declines to 0.59 per 1000 person-years by 2030. Scenario B (90-90-90 reached in the overall population) is actually achieved if the 2017 rates of testing and treatment are maintained; however the cascade among FSW only reaches 81-60-83 by 2020. As a result, incidence only declines to 1.22 per 1000 person-years by 2030, and the model projects a 60% increase in cumulative new infections in the total population between 2020 and 2030 versus Scenario A. Conclusion: Heterogeneity in HIV transmission risks across the 10-10-10 could undermine the projected impact of achieving 90-90-90 across the Southern African region. Efforts to meet and surpass UNAIDS targets among key populations such as FSW and their clients should be prioritized to maximize incidence reductions and achieve pandemic control by 2030.

Poster Abstracts

1092 TasP COVERAGE MAY INCREASE WITHOUT SELECTING FOR MORE VIRULENT HIV: A MODELING STUDY Sarah E. Stansfield , Joshua T. Herbeck, Geoffrey S. Gottlieb, Neil Abernethy, James Murphy, John E. Mittler, Steven M. Goodreau University of Washington, Seattle, WA, USA Background: HIV-1 set point viral load (SPVL) influences both transmission potential and disease progression and is a proxy for HIV virulence. Multiple test-and-treat models have found that increasing the proportion of people who receive treatment selects for viruses with higher SPVL, i.e. higher virulence. Here we extend these modeling studies to evaluate the potential impact of different risk, treatment, and transmission scenarios on the evolution of HIV virulence. Methods: We extend a stochastic, dynamic network model (EvoNetHIV) in which sexual network structure and behavioral parameters are derived from modeling studies of HIV among US men who have sex with men. Key agent attributes include SPVL and current viral load; SPVL was partially heritable so virulence could evolve over time. Our main input is treatment coverage and main output is mean population SPVL (MPSPVL). We vary the transmission model (increasing vs. plateauing transmission at very high viral loads), relationship patterns (relationships averaging 2.45 years and 1 sex act/day vs 100 days and 0.2 sex acts/day), and ART schemes (ART beginning at a fixed interval after infection vs. stochastic interval driven by testing) in isolation and in combination to determine those most integral to observed outcomes. In each case we explore mean times to ART initiation of 1-6 years. Results: In scenarios most similar to those previously published, we confirmed that higher ART coverage led to higher MPSPVL. In contrast, in scenarios in which ART occurs immediately after individuals test positive, with shorter relationships and less frequent sex acts, and with the increasing transmission function, increasing levels of ART instead either led to no significant MPSPVL change or selected for viruses with lower MPSPVLs. Further analyses showed that changing any of these factors was enough to eliminate the relationship between high treatment levels and low MPSPVL and in some cases to reverse this pattern. Conclusion: Under a set of realistic, data-derived modeling assumptions, we found that MPSPVL remains unchanged and/or decreases with higher ART coverage. These findings emphasize the impact of epidemiological conditions and model design on predicted evolutionary outcomes. Our results suggest that, under some realistic conditions, vigorous test-and-treat strategies may not need to face a previously-reported tradeoff in which increasing coverage fuels evolution of greater virulence.

1091 WHO IS LEFT IN 10-10-10? IMPORTANCE OF REACHING KEY POPULATIONS WITH THE HIV CASCADE

Jesse Knight 1 , Huiting Ma 1 , Stefan Baral 2 , Sosthenes Ketende 2 , Oluwasolape M. Olawore 2 , Amrita Rao 2 , Carly Comins 2 , Sheree Schwartz 2 , Harry Hausler 3 , Katherine Young 3 , Mfezi Mcingana 3 , Vijayanand Guddera 3 , Sindy Matse 4 , Zandile Mnisi 4 , Sharmistha Mishra 5 1 St. Michael’s Hospital, Toronto, ON, Canada, 2 Johns Hopkins University, Baltimore, MD, USA, 3 TB/HIV Care Association, Cape Town, South Africa, 4 Ministry of Health, Mbabane, Swaziland, 5 University of Toronto, Toronto, ON, Canada Background: Achievement of the UNAIDS 90-90-90 targets for HIV cascade of care (90% of infected are diagnosed, 90% of diagnosed are on treatment, 90% of those on treatment are virally suppressed) by 2020 is predicted to end the AIDS pandemic by 2030. We sought to determine the influence of heterogeneity across the remaining 10-10-10 on the epidemic features after the UNAIDS targets are achieved in a high-prevalence HIV epidemic in Southern Africa. Methods: We built a deterministic mathematical model of heterosexual HIV transmission to simulate a high-prevalence epidemic in a Southern African context (using demographic health survey and female sex worker (FSW) survey data from eSwatini and South Africa). The model includes 6 different populations at risk for HIV, including FSW and clients; 4 sexual partnership types; and the HIV cascade (undiagnosed, diagnosed, on ART, and virally suppressed), Figure 1 (a-b). The model simulates observed HIV prevalence ratios by risk group, and trends in cascade of care to 2017. We then compared two scenarios where A) 90-90-90 is achieved in all populations, including FSW; B) 90-90-90 is achieved in the overall population, but not among FSW - and

CROI 2019 430