CROI 2016 Abstract eBook

Abstract Listing

Poster Abstracts

916 Risk Factors for HIV in an Outbreak Among Persons Who Inject Drugs, Indiana 2015 Monita Patel 1 ; Jesse Shields 2 ; MichaelT. Spiller 1 ; Caitlin Conrad 3 ; Jeremy Roseberry 3 ; DorothyWaterhouse 2 ; Pamela Pontones 3 ; Philip Peters 1 ; Joan Duwve 3 ; JohnT. Brooks 1 1 CDC, Atlanta, GA, USA; 2 Clark County Hlth Dept, Jeffersonville, IN, USA; 3 Indiana State Dept of Hlth, Indianapolis, IN, USA Background: In early 2015, an outbreak of HIV infections among persons who inject drugs (PWID) was identified in Scott County in rural southeastern Indiana. Investigation indicated that by July 31, 2015, a single strain of HIV had infected over 170 persons in a matter of months. To identify risk factors for HIV infection and inform control efforts for this unprecedented event, we conducted a cohort study of HIV-positive and HIV-negative PWID in this community. Methods: We included residents of Scott County who had been named syringe-sharing contacts by an HIV-infected person diagnosed from Oct 1, 2014 to July 31, 2015 and were investigated and HIV tested as part of contact tracing efforts through August 6, 2015. We conducted bivariate analysis of routinely collected self-reported sociodemographic and behavioral data by HIV status. Chi-square and Kruskal-Wallis tests were used to test for differences between categorical and continuous variables, respectively. Variables that differed significantly by HIV test status (p<0.05) were examined as potential risk factors. Log-binomial regression modeling with stepwise backwards elimination was used to estimate adjusted risk ratios (adj-RR) and 95% confidence intervals (95% CIs) for potential risk factors. Results: Of 315 investigated contacts named as syringe-sharing partners, 196 (62%) were both HIV tested and had demographic and behavioral data available: 51% tested HIV- positive. The cohort was 58%male, 98% non-Hispanic white, and had a median age of 33 years. Among cohort members, the annual income was <$10,000 for 92%, and 54% had been incarcerated in the past 12 months. HIV risk behaviors were highly prevalent (Table 1). Sharing injection equipment, sex with a person known to be HIV-positive, and sex with a PWID were substantially more frequent among persons who tested HIV-positive than among those who did not, but these differences were not statistically significant. Only the number of times that a person had been named by an HIV-infected person as a syringe-sharing contact was significantly associated with testing HIV-positive: adj-RR (per each time named as syringe-sharing partner): 1.90, 95% CI: 1.49-2.41. Conclusions: In HIV outbreaks linked to injection drug use, investigation and control efforts should prioritize persons for investigation and HIV testing according to their extent of syringe sharing as indicated by the frequency with which they are named syringe-sharing contacts by HIV-infected persons.

Poster Abstracts

917 Cross-Sectional vs Longitudinal HIV Incidence Estimates in PeopleWho Inject Drugs Andrew F. Longosz 1 ; Shruti H. Mehta 2 ; Gregory D. Kirk 2 ; Jacquie Astemborski 1 ;Thomas C. Quinn 3 ; Susan H. Eshleman 3 ; Oliver Laeyendecker 4 ; for the AIDS Link to Intravenous Experiences (ALIVE) cohort 1 Johns Hopkins Bloomberg Sch of PH, Baltimore, MD, USA; 2 Johns Hopkins Univ, Baltimore, MD, USA; 3 Johns Hopkins Univ Sch of Med, Baltimore, MD, USA; 4 NIH, Bethesda, MD, USA Background: HIV incidence can be estimated in cross-sectional surveys and longitudinal cohorts. We compared HIV incidence at enrollment (baseline) and during follow-up in a cohort of people who inject drugs (PWID) in Baltimore, MD. We explored potential explanations for the any difference incidence (differential loss to follow-up or behavioral modification). Methods: Samples were obtained from the AIDS Link to Intravenous Experiences (ALIVE) cohort: 2,938 PWID at baseline (1988-89; 89% Black, 81%male) and 1,772 individuals at the first follow-up visit (4-12 months post enrollment, 92% Black, 79%male). All participants were asked questions about demographics, injection drug use and sexual practices. Blood was drawn at each time point for HIV screening. HIV incidence was determined cross-sectionally at baseline using a multi assay algorithm and longitudinally at follow-up. Differential lost to follow-up and behavioral characteristics were analyzed between baseline and the 1 st follow-up visit (occurred 4-12 months past baseline).

387

CROI 2016