CROI 2018 Abstract eBook

Abstract eBook

Poster Abstracts

970 HIV DIAGNOSES AMONG PEOPLE WHO INJECT DRUGS—UNITED STATES, 2010-2016 Sheryl Lyss 1 , Tianchi Zhang 2 , Alexandra M. Oster 1 1 CDC, Atlanta, GA, USA, 2 ICF, Atlanta, GA, USA Background: The number of HIV diagnoses among people who inject drugs (PWID) in the United States declined substantially over the past decade, then increased slightly in 2015, the year of a large HIV outbreak among PWID in Scott County, Indiana. We describe trends in HIV diagnoses among PWID to elucidate the observed increase. Methods: We used National HIV Surveillance System data reported through June 2017 for diagnoses occurring during 2010-2015 and preliminary data for 2016; missing data were imputed with standard methods. We included persons aged ≥13 years with HIV attributed specifically to injection drug use (IDU) (i.e., those attributed to both IDU and male-to-male sexual contact were not included). To assess the influence of the Scott County outbreak, we examined 2015 diagnoses including and excluding those reported from Scott County. Results: HIV diagnoses among PWID decreased 35% nationally during 2010-2014 from 3421 to 2229, with annual declines of 14% (2010-2011), 10% (2011-2012), 9% (2012-2013) and 7% (2013-2014) (Table). During 2010-2014, diagnoses declined 35% among both males and females; declines varied across racial/ethnic groups, age groups and U.S. Census regions and were <20% among whites (-10%), those aged 13-34 years (-19%), and in the West (-18%). During 2014-2015, diagnoses among PWID increased by 117 (5%) to 2347 nationally; increases among whites (+233; 32%), those aged 13-34 years (+161; 22%), and in the Midwest (+186; 85%) are consistent with the distribution of the 148 Scott County cases in 2015. Excluding Scott County cases, diagnoses among PWID decreased by 31 (1%) to 2199, but increased among whites (+89; 12%), those aged 13-34 years (+81; 11%), and in the Midwest (+38; 17%). The preliminary number of 2016 diagnoses among PWID is similar to that for 2014. Conclusion: The national increase in HIV diagnoses among PWID during 2014-2015 followed a slowing decline during 2010-2014. Although the Scott County outbreak was largely responsible for the increase in 2015 cases, without Scott County cases, during 2014-2015 declines stalled nationally, and diagnoses increased in some demographic groups with slower declines during 2010-2014. The extent to which trends in the number and shifting characteristics of HIV diagnoses among PWID reflect a growing number of PWID in the United States, increased testing, increased reporting, or other causes cannot be determined from these data alone. Vigilance is needed so that longstanding gains in preventing HIV among PWID are not reversed.

Methods: PWID aged ≥18 years were interviewed for 2012 National HIV Behavioral Surveillance using respondent-driven sampling. We tested participants’ dried blood spot specimens with the CDC-developed Bio-Rad Avidity Index incidence assay (BRAI, mean duration of recency of 240 days). A recent HIV infection was defined as having a reactive HIV screening result, no reported HIV diagnosis ≥12 months before the interview date, a BRAI avidity index ≤30%, and no viral suppression (viral suppression defined as HIV viral load <1,000 copies/mL). In bivariate analyses, we compared recently infected to HIV-negative PWID to evaluate factors associated with recent HIV acquisition. We also assessed differences between PWID with recent infections and PWID with non-recent HIV infections. Results: Of 9,901 eligible PWID, 50 (0.5%) were recently HIV infected. Compared to HIV-negative PWID, those recently HIV infected were more likely to inject stimulants (p<0.01), have a greater number of sex partners (p=0.02), and have had male-male sex in the past year (p<0.01). PWID who have sex with men had the highest proportion of recent infections (2.0%). Compared to those with non-recent infections, recently infected PWID were more likely to be young (p<0.01), be white (p=0.03), have a high school diploma (p<0.01), not have health insurance (p<0.01), share syringes (p=0.01), have a greater number of sex partners (p=0.01), and have condomless sex in the past year (p<0.01). Conclusion: Our findings suggest that sexual behavior is associated with recent HIV acquisition among PWID. Promoting not only safe injection practices but also safe sex practices will be key to preventing new HIV infections. Prevention efforts should prioritize PWID who have sex with men, who experienced a higher proportion of recent infections and may be harder to retain on biomedical interventions. Our finding that recently HIV-infected PWID in our study population were more likely to be white and high school graduates, compared to PWID who acquired HIV earlier in the epidemic who were mostly black and who did not graduate high school, suggests a demographic shift in PWID who recently acquired HIV. 972 DISTRIBUTIVE SYRINGE SHARING AND USE OF SYRINGE SERVICES PROGRAMS AMONG PWID Monica Adams, Qian An, Dita Broz, Janet C. Burnett , Cyprian Wejnert, Gabriela Paz-Bailey CDC, Atlanta, GA, USA Background: Although syringe sharing is a well-documented risk for HIV transmission among persons who inject drugs (PWID), few studies specifically examined distributive syringe sharing (DSS, i.e., passing on a used syringe to another PWID). Syringe services programs (SSPs) are effective at reducing injection risk behaviors and are important prevention interventions for both HIV-negative and HIV-positive PWID. We examine demographic and behavioral factors that may be associated with DSS and how the effect of SSP use on DSS differs by HIV sero-status. Methods: PWID aged ≥18 years were recruited in 20 U.S. cities for the 2015 National HIV Behavioral Surveillance using respondent-driven sampling, interviewed and offered HIV testing. Bivariate and multivariable analyses via log-linked Poisson regression with generalized estimating equations were conducted to examine associations between demographic and behavioral variables and DSS. The effect of SSP utilization on DSS by HIV sero-status was assessed by including an interaction between SSP and HIV sero-status in the final model. Multivariable analyses were adjusted for sampling design covariates. Prevalence ratios (PR) and 95% confidence intervals (CI) are presented. Results: Among 10,402 PWID, 41% reported DSS in the past 12 months. DSS was less likely to be reported among HIV-positive compared to HIV-negative PWID (21% vs. 43%, PR=0.52, CI: 0.45-0.60), and among those who primarily obtained syringes from SSPs versus those who did not (34% vs. 46%, PR=0.82, 95% CI: 0.76-0.88). After adjusting for gender, age, race/ethnicity, homelessness and arrest in the 12 months prior to interview, those who primarily used SSPs were less likely to report DSS than those who did not among both HIV-negative PWID (36% vs. 47%, adjusted PR=0.83, 95% CI: 0.78-0.89) and HIV-positive PWID (12% vs. 27%, adjusted PR=0.53, 95% CI: 0.38-0.73; see Table). Conclusion: Primary use of SSPs was associated with less DSS among PWID. This effect was greater among HIV-positive PWID, who may unintentionally transmit infection when sharing their used syringes with others. These findings support expansion of SSPs and referrals to SSP programs by service providers working with PWID. Specifically, clinicians providing care and treatment to HIV- positive PWID should consider linkage of these patients to SSPs.

Poster Abstracts

971 CHARACTERISTICS OF HIV INCIDENT INFECTIONS AMONG PERSONS WHO INJECT DRUGS IN THE US Johanna Chapin-Bardales , Silvina Masciotra, Amanda Smith, Brooke Hoots, WilliamM. Switzer, Wei Luo, S. Michele Owen, Gabriela Paz-Bailey CDC, Atlanta, GA, USA Background: Detecting and characterizing recent HIV infections can better describe subpopulations at increased risk of HIV and identify factors that contribute to ongoing HIV transmission. We characterized recent HIV infections among a sample of persons who inject drugs (PWID) from 19 US cities to inform monitoring of the HIV epidemic among PWID.

CROI 2018 371