CROI 2016 Abstract eBook

Abstract Listing

Poster Abstracts

transmissions from HIV-positive males to their female partners in 2009 stratifying by the attributes of HIV-positive men, including their primary risk factor for acquiring HIV: male- to-male sex (MSM), heterosexual sex, and injection drug use (IDU). Results: An estimated 864,301 HIV-positive men had 950,178 female sex partners, of whom 873,508 were of HIV-negative or unknown sero-status (discordant). We estimated that, on average, HIV-positive MSM had 0.3 HIV-discordant female sex partners, heterosexual men had 3.0 female partners, and persons who inject drugs (PWID) had 2.3 partners. Most transmissions to females were frommain partners (82%). Of estimated HIV transmissions to females, 27% of the males had MSM as their primary risk factor for acquiring HIV, 35% heterosexual sex, and 31% IDU. Transmissions to females varied by male partner’s race and primary risk factor. White males transmitting HIV to females mostly acquired HIV through heterosexual sex (73%); among Hispanic males, IDU was the most common primary risk factor (50%); the primary risk factor of black males varied: for an estimated 34% it was MSM, 38% heterosexual sex, 23% IDU, and 5% both MSM and IDU. Conclusions: Reducing HIV among high-prevalence populations, especially MSM and IDU, can have important benefits for other populations. Most males transmitting HIV to female had MSM or IDU risk factors. Although a small percentage of HIV-positive MSM reported female partners, MSMmake up a large portion of the prevalent HIV infections and consequently comprise a substantial proportion of transmissions to females.

1046 Utility of the 1% HIV Prevalence Threshold in Defining Concentrated HIV Epidemics Michael Pickles 1 ; Marie-Claude Boily 1 ; Sharmistha Mishra 2 1 Imperial Coll London, London, UK; 2 Univ of Toronto, Toronto, ON, Canada

Background: Regions with HIV prevalence < 1% are defined as experiencing ‘concentrated’ HIV epidemics, with larger epidemics labeled as ‘generalized’. The surveillance and prevention implications include prioritizing resources on key populations (e.g. persons engaged in sex work [SW]) in ‘concentrated’ epidemics or on the general population otherwise. We hypothesize that the 1% threshold does not distinguish HIV epidemics by the underlying risk factors that sustain HIV spread, such as unprotected SW. Methods: We developed synthetic HIV epidemics using deterministic mathematical models of heterosexual HIV transmission, using sub-national demographic, behavior, and biological data in Sub-Saharan Africa synthesized from a systematic review of these data across subpopulations. We generated 9000 plausible HIV epidemics of three types, according to risk factors that sustain HIV spread: concentrated (which we define as where unprotected SWwas necessary and sufficient for HIV spread), generalized (defined as unprotected SW insufficient on its own for HIV spread), and mixed (defined as sustained epidemics that were neither concentrated nor generalized). We estimated the diagnostic performance of the general population HIV prevalence to distinguish a concentrated HIV epidemic from other types. Results: The sensitivity, specificity, positive predictive value, and negative predictive value of the 1% prevalence threshold was 65%, 92%, 80%, and 84% respectively early in the HIV epidemics (circa 1990), indicating that 35% of the time the 1% threshold would fail to identify an epidemic exclusively sustained by sex work. If epidemic types were equally likely in real-life, then there is an 80% probability that an epidemic with <1% HIV prevalence is driven solely by sex work, and a 16% probability that an epidemic called ‘generalized’ because HIV prevalence is ≥1%, is in fact driven only by sex work. Sensitivity increased and specificity decreased later in the epidemics because rising condom use during sex work had reduced HIV prevalence in concentrated HIV epidemics. The findings were similar for differentiating between concentrated/mixed from generalized epidemics. Conclusions: The 1% HIV prevalence threshold is neither sensitive nor specific enough to adequately define HIV epidemics, and hence should not form the basis of how surveillance and interventions should be prioritized. The 1% threshold should be replaced with approaches which distinguish epidemic types based on risk factors that sustain HIV spread.

Poster Abstracts

1047 The Role of YoungWomen in the HIV Epidemic in Benin, West Africa John R. Williams 1 ; Eugene L. Geidelberg 1 ; Michel Alary 2 ; Sharmistha Mishra 3 ; Marie-Claude Boily 1 1 Imperial Coll London, London, UK; 2 Univ Laval, Québec, QC, Canada; 3 Univ of Toronto, Toronto, ON, Canada

Background: Young women are particularly vulnerable to acquiring HIV in Sub-Saharan Africa. But their contribution to onward HIV transmission has not been explored. In Cotonou, Benin, HIV prevalence declined from 3% to 0.5% (1998 to 2008) among men but remained stable at 3.6% among women. The prevalence decline in men followed a

448

CROI 2016