CROI 2017 Abstract e-Book

Abstract eBook

Poster and Themed Discussion Abstracts

928 ASSOCIATION OF SEX AND RACE WITH HIV-/NON-HIV–RELATED DEATH IN TREATED HIV INFECTION David B. Hanna 1 , Donald R. Hoover 2 , Qiuhu Shi 3 , Uriel Felsen 1 , Mark H. Kuniholm 4 , Mindy S. Ginsberg 1 , Howard D. Strickler 1 , Kathryn Anastos 1 1 Albert Einstein Coll of Med, Bronx, NY, USA, 2 Rutgers, the State Univ of New Jersey, New Brunswick, NJ, USA, 3 New York Med Coll, Valhalla, NY, USA, 4 State Univ of New York at Albany, Rensselaer, NY, USA Background: Whether there are systematic survival differences by sex and/or race among HIV+ individuals treated with antiretroviral therapy (ART) remains unclear. We assessed associations of sex and race with mortality among HIV+ patients initiating ART in a large US health system. Methods: The Einstein/Rockefeller/CUNY Center for AIDS Research’s HIV Clinical Cohort Database contains electronic medical records of HIV+ patients receiving care in the Montefiore Health System, the largest provider of HIV care in Bronx, New York. Inclusion criteria included intake at one of >15 outpatient clinics and ART initiation between 2006 and 2012. Self-reported race was categorized as black (Hispanic and non-Hispanic), white (Hispanic and non-Hispanic), and Hispanic of other, unspecified, or >1 race; all others were excluded. Deaths were ascertained by linkage to the National Death Index. Cox regression assessed associations of sex and race with time to death, adjusted for age, HIV transmission risk group, peak HIV-1 RNA and nadir CD4+ count as of ART initiation, neighborhood socioeconomic status, insurance type, and hepatitis C virus infection. Results: Of 2,108 eligible patients, 41%were female, 52% black, 8%white and 40% Hispanic of other or unspecified race. Median CD4+ count at ART initiation was 240 cells/uL (interquartile range 99-365). Patients were followed for a median 3.6 years, 7,044 person-years in total. Of 227 reported deaths (11% of the cohort), 58%were HIV-related based on underlying cause ICD-10 codes. Men had a greater overall hazard of death than women (adjusted hazard ratio [aHR] 1.4, 95% CI 1.1-1.9), and black patients had a greater hazard of death (aHR 1.8, 95% CI 1.04-3.2) than white patients. For HIV deaths, blacks had a higher hazard of death than whites (aHR 2.8, 95% CI 1.2-6.4). However, there was no difference in HIV-related mortality in men versus women (aHR 1.1, 95% CI 0.7-1.5). For non-HIV deaths, there was no association with race, but there was an association of male versus female sex with mortality (aHR 2.2, 95% CI 1.4-3.4). Conclusion: Among patients initiating ART, greater HIV-related mortality was observed among blacks than whites, whereas there was no difference by sex. Further understanding of the causes of this disparity is warranted to eliminate preventable HIV-related deaths. Our finding of higher mortality due to non-HIV causes in men versus women is comparable to findings in HIV-uninfected populations. 929 IMPACT OF INJECTION DRUG USE AND SELECTED COMORBIDITIES ON HEALTHY LIFE EXPECTANCY Robert S. Hogg , Oghenowede Eyawo, Wendy Zhang, Paul Serada, Mark Hull, Jean A. Shoveller, Karyn Gabler, Viviane D. Lima, Julio S. Montaner, for the Comparative Outcomes And Service UtilizationTrends (COAST) Study BC Cntr for Excellence in HIV/AIDS, Vancouver, British Columbia, Canada Background: We sought to understand the impact of select comorbidities [cancers, diabetes, heart, liver, lung and renal diseases, hepatitis B and C (HCV, HBV)] and history of injection drug use on life expectancy among persons living with HIV (PLHIV). We hypothesized that persons who inject drugs (PWID) would be more impacted by these select comorbidities, especially HCV, HBV and liver diseases, and spend less time in a healthy state than other PLHIV in British Columbia (BC), Canada. Methods: The Comparative Outcomes And Service Utilization Trends (COAST) study follows a retrospective cohort study design and includes individuals aged 19 years or older at baseline or during study follow-up. A cohort of PLHIV was constructed from all adults (≥20 years) known to be HIV-positive in BC who had a record of at least one detectable HIV plasma viral load, AIDS defining illness, or CD4 cell count; and who initiated highly active antiretroviral therapy (HAART) in BC between 25 June 1996 and 31 December 2012. Prevalence of select comorbidities was determined using case-finding algorithms based on a set of validated International Classification of Diseases, version 9 and 10 codes. All deaths were obtained from the vital event registry in BC. A healthy state was defined as the proportion of life expectancy comorbid free. We estimated comorbid-specific healthy life expectancy from 20 years of age and by history of injection drug use (IDU). Results: Our study consisted of 5,636 PLHIV with known IDU status aged ≥20 years on HAART. Compared to other PLHIV, PWID were more likely to be women, to be Indigenous and to have initiated HAART at CD4 counts below 200 (all p-values <0.001). PWIDs had significantly higher crude and adjusted mortality rates and shorter overall life expectancies without any comorbidities than other PLHIV [29.5 (1.1, standard error) vs. 53.2 (1.1) years, p<0.001]. Differences length of the healthy state between the two groups were largest for: having one or more of eight select comorbidities (53.6 vs. 70.3%), liver disease [76.6 vs. 92.2%], and Hepatitis B (76.7 vs 97.0%) (see Figure). Conclusion: While HAART has substantially improved the life expectancy for many PLHIV, PWID have not benefitted to the same degree and spend significantly less time in a healthy state due to the fact that they are more impacted by liver-related conditions. 930 LONGITUDINAL ASSOCIATION BETWEEN HIV STATUS AND MEDICALLY SIGNIFICANT FALLS Julie A. Womack 1 , Terrence E. Murphy 2 , Harini Bathulapalli 1 , Jonathan Bates 2 , Cynthia L. Gibert 3 , Maria Rodriguez-Barradas 4 , Phyllis Tien 5 , Michael T. Yin 6 , Cynthia A. Brandt 1 , Amy Justice 1

Poster and Themed Discussion Abstracts

CROI 2017 402