CROI 2017 Abstract e-Book

Abstract eBook

Poster and Themed Discussion Abstracts

Methods: The time profile of efficacy of an ALVAC/gp120 regimen with 12-month boost was fit to RV144 results or increased to 50% cumulative efficacy at 24 months. This time- varying efficacy profile was implemented in EMOD-HIV v2.5, a microsimulation model that has been fit to the HIV epidemic, and includes demographics, risk stratification, and HIV testing/cascade of care. In future projections, we varied the scale-up of treatment to examine its interplay with the 20-year impact of a vaccine to be started in 2027. Coverage was assumed to be 50% or 80%with boosters to continue two-yearly for ten years with up to 20% attrition per dose. Results: A partially effective vaccine could reduce HIV incidence in South Africa by up to 21%with 80% coverage. The most efficient ages to start immunization are 15 in women and 20 in men, with 16-29 HIV infections averted per 1000 individuals vaccinated. If gender differences in vaccination age were impractical to implement, then ages 15 and 18 would be equally optimal for vaccination, with 10-23 HIV infections averted per 1000 vaccinated, a result highly sensitive to concurrent scale-up of ART. In contrast, combining HIV vaccination with the current HPV vaccine program among 9-11 year olds would result in less than one HIV infection averted per 1000 vaccinated due to waning immunity. Maximum cost-effective prices of a vaccine, calculated from the ratio of the net budget impact to DALYs averted, varied widely depending on cost-effectiveness thresholds and ART scale-up, and to a lesser extent on efficacy, age at vaccination, and attrition. Conclusion: Partially effective HIV vaccines with rapid waning of immunity could substantially reduce HIV incidence if vaccination schedules were aligned with the ages of highest HIV incidence and high coverage levels were achieved. Reaching a new target population with a complex immunization schedule not aligned with other schedules may pose an implementation challenge in South Africa.

Poster and Themed Discussion Abstracts

1037 COST-EFFECTIVENESS OF PREEXPOSURE PROPHYLAXIS ACROSS COUNTIES IN WESTERN KENYA Anna Bershteyn 1 , Daniel J. Klein 1 , Adam N. Akullian 1 , Zindoga Mukandavire 2 , Graham F. Medley 2 , Wanjiru Mukoma 3 , Michael K. Kiragu 3 , Kennedy Mutai 4 , Katharine Kripke 5 , for the OPTIONS Consortium 1 Inst for Disease Modeling, Bellevue, WA, USA, 2 London Sch of Hygiene and Trop Med, London, United Kingdom, 3 LVCT Hlth, Nairobi, Kenya, 4 National AIDS Control Council, Ministry of Hlth, Nairobi, Kenya, 5 Avenir Hlth, Washington, DC, USA Background: Pre-exposure prophylaxis (PrEP) is recommended when risk of acquiring HIV is high. Implementation planning for PrEP is now underway in Kenya. This analysis compares the estimated cost-effectiveness of PrEP in the six counties comprising the former Nyanza Province in Western Kenya, which exhibit a range of epidemic characteristics, including highly generalized, highly concentrated, and mixed epidemics. Methods: The microsimulation model EMOD-HIV v2.5, fitted to demographic, programmatic, and epidemic data for six counties in Western Kenya, was used to assess the cost- effectiveness of PrEP. Female sex workers (FSW) were included based on a recent FSW enumeration; male clients of FSWwere included to balance the number of clients reported by FSW. Projections with PrEP provision to FSW, “medium”-risk adolescent girls and young women (AGYW), who are not identified as FSW but still at elevated risk of HIV infection, and all medium-risk young adults were compared to projections without PrEP. The person-years of PrEP provided per HIV infection averted over a twenty-year time horizon was used as a proxy for PrEP cost-effectiveness. Upper bounds for PrEP impact assumed that present-day trends in antiretroviral therapy (ART) coverage continue; lower bounds for PrEP impact assumed achievement of UNAIDS 90-90-90 ART targets. Results: Providing PrEP to FSWwas more cost-effective compared to providing PrEP to “medium” risk adolescents and young adults in the general population. Regardless of PrEP coverage, the populations for which PrEP was most cost-effective were FSW in generalized or mixed epidemic contexts. In contrast, the cost-effectiveness of PrEP in a concentrated epidemic setting such as Kisii County, which had the second-highest proportion of adult women participating in FSW but an overall low HIV prevalence, was similar to that of providing PrEP to medium-risk AGYW in generalized or mixed epidemics. PrEP for medium-risk AGYW in mixed epidemics such as Kisumu was more cost-effective than PrEP for FSW in concentrated epidemics such as Kisii. Conclusion: Transmission modeling suggests that the most cost-effective population for providing PrEP is FSW in mixed or generalized epidemic. FSW in concentrated epidemic contexts and high-risk AGYW in mixed or generalized epidemic contexts are both important populations to consider for PrEP.

CROI 2017 448