CROI 2019 Abstract eBook

Abstract eBook

Poster Abstracts

resources and to evaluate the impact of changes in demographic trends and policies on the Mexico HIV epidemic.

epidemiology, these observations could illustrate the role of migrant workers leaving their home towns for work. Conclusion: This analysis of HIV CRF01_AE dynamics within China emphasizes the key roles of highly populated regions and human mobility in the dispersal of the HIV epidemic. Such information could be important in developing prevention strategies.

Poster Abstracts

878 WITHDRAWN / INTENTIONALLY UNASSIGNED 879 LOCAL HUMAN MOBILITY AND TRANSMISSION ROUTES OF HIV CRF01_ AE ACROSS CHINA Simon Dellicour 1 , Bin Zhao 2 , Xiaoxu Han 2 , Minghui An 2 , Hong Shang 2 , Jin Zhao 3 , Ping Zhong 4 , Davey M. Smith 5 , Bram Vrancken 1 , Antoine Chaillon 5 1 Katholieke University Leuven, Leuven, Belgium, 2 China Medical University, Shenyang, China, 3 Shenzhen Center for Disease Control and Prevention, Shenzhen, China, 4 Shanghai Municipal Center for Disease Control and Prevention, Shanghai, China, 5 University of California San Diego, La Jolla, CA, USA Background: China is witnessing a rapid growth of its HIV epidemic and now accounts for 3% of the global HIV incidence. This upsurge is driven by new HIV infections among men who have sex with men (MSM) often with CRF01_AE infection. Here, we characterized the dynamics and determinants of diffusion of HIV-1 CRF01_AE across China. Methods: We applied a multistep phylogenetic approach on a large dataset of CRF01 AE pol sequences sampled across China. We first performed an overall maximum likelihood phylogenetic inference to identify well-supported monophyletic clades. All clades of size ≥ 3 identified were used to perform a discrete phylogeographic inference to evaluate the dispersal history across Chinese provinces. We then applied a generalized linear model (GLM) to test the association of epidemiologic factors and connectivity (i.e. geographic distances between each location and the intensity of air traffic passenger flow) with lineage dispersal frequencies among the provinces (Fig 1A). Results: A total of 6,800 unique CRF01_AE HIV partial pol sequences from participants originating from China were collected between 2004-2017. After combining these with 6,423 publicly available HIV-1 CRF01_AE pol sequences with known sampling country, we identified 59 clades of size ≥ 3 that represent 458 sequences from 17 provinces. The discrete phylogeographic analysis based on these clades revealed varying levels of virus exchange between the sampled provinces (Fig. 1B). The GLM analysis also suggested that viral migration was strongly associated with the population density in both source and recipient provinces (Bayes Factor, BF>25.103 and 105.9 respectively), but not with the intensity of air passenger flows associated with a negative GLM coefficient (BF=59.5). Finally, the geographical distance between provinces was also negatively associated with viral movement (i.e. increased distance associated with less transition, negative BF>25.103), Fig. 1C. The negative correlations with both the air traffic and geographical distances underline the importance of local mobility (e.g. between adjacent provinces) in spreading the virus. Based on

880 SUBSTANCE USE DISORDERS ASSOCIATED WITH MORTALITY AMONG HIV+ IN WASHINGTON, DC Rupali K. Doshi 1 , Morgan Byrne 1 , Matthew E. Levy 1 , Leah Varga 2 , Irene Kuo 1 , Michael A. Horberg 3 , Amanda D. Castel 1 , Anne K. Monroe 1 , for the DC Cohort Executive Committee 1 George Washington University, Washington, DC, USA, 2 District of Columbia Department of Health, Washington, DC, USA, 3 Kaiser Permanente Mid-Atlantic States, Rockville, MD, USA Background: Substance use disorders (SUDs) are common among people living with HIV (PLWH) and may make achievement of optimal health outcomes challenging. We described the prevalence of alcohol, opioid and stimulant use disorders among PLWH in the DC Cohort and assessed the association of SUDs with viral suppression and death. Methods: We analyzed diagnosis and treatment data for participants enrolled in the DC Cohort (2011-2017), a longitudinal study of PLWH receiving care at 14 clinical sites in Washington, DC, and reported on the prevalence of overall SUD, alcohol, opioid, stimulant, and polysubstance (all 3) use disorders at enrollment or during follow-up, and prevalence of hepatitis B (HBV) and hepatitis C (HCV) among the SUD groups. We used multivariable Cox proportional hazard models to evaluate the association of SUD with all-cause mortality, adjusting for demographics, CD4, and viremia copy-years. We calculated adjusted prevalence ratios (aPR) to assess the association of age at HIV diagnosis, gender, and race/ ethnicity with SUDs, and association of SUDs with viral suppression (VS; <200 copies/ml at most recent measurement), adjusting for current age, gender, race/ ethnicity, and mode of HIV transmission. Results: Of 8,507 adults, 2,929 (34.4%) had history of any SUD. The most prevalent SUDs were: 73.6% only alcohol, 9.0% alcohol/stimulant, 7.0% only stimulants, 3.4% only opioids, 3.4% alcohol/opioids and 2.8% polysubstance use disorders. Chronic HCV was highest among those with alcohol/opioids (62%), polysubstance (60%), and opioid use disorders (49%). After adjustment, those with any SUD were older at HIV diagnosis (median 49.9 vs. 45.6 years), more likely to be Black (vs. White; aPR 1.37, 95% CI 1.23, 1.52) or Latino (aPR 1.29, 95% CI 1.10, 1.52), less likely to be female (aPR 0.75, 95% CI 0.69, 0.81), and more likely to have any mode of HIV transmission other than men who have sex with men (MSM; all p<.0001). Based upon 388 deaths, SUD was independently associated with all-cause mortality (aHR 1.49, 95% CI 1.18, 1.89). Overall VS was 87%, and SUD was not significantly associated with VS (aPR 0.99, 95% CI 0.97, 1.01).

CROI 2019 342