CROI 2019 Abstract eBook

Abstract eBook

Poster Abstracts

913 CHANGING CONTEXTUAL FACTORS POST-HIV DIAGNOSIS PREDICT 5-YR MORTALITY IN SOUTH AFRICA Ingrid V. Bassett 1 , Ai Xu 1 , Janet Giddy 2 , Laura M. Bogart 3 , Andrew Boulle 4 , Lucia Millham 1 , Elena Losina 5 , Robert A. Parker 1 1 Massachusetts General Hospital, Boston, MA, USA, 2 McCord Hospital, Durban, South Africa, 3 RAND Corporation, Santa Monica, CA, USA, 4 University of Cape Town, Cape Town, South Africa, 5 Brigham and Women’s Hospital, Boston, MA, USA Background: Changes in an individual’s contextual factors following HIV diagnosis may influence long-term outcomes. We evaluated how changes to contextual factors between HIV diagnosis and 9-month follow-up predict 5-year mortality among HIV-infected individuals in Durban, South Africa. Methods: We used baseline and 9-month survey data from the Sizanani Trial (NCT01188941) in which adults (≥18y) were enrolled prior to HIV testing in 4 outpatient sites between Aug 2010-Jan 2013. We assessed social support, mental health, and competing needs, meaning financial constraints that required deciding between meeting basic needs (food, clothing, or housing) or receiving healthcare. We used the South African National Population Register to ascertain vital status; median follow-up time was 5.8y (IQR 5.2-6.5). We used random survival forests to identify the most important 9-month variables predicting time to subsequent mortality. We incorporated these predictors into a Cox proportional hazards model that included age, sex, and starting ART by 9 months a priori; the full model included changes in social support, mental health, and competing needs between baseline and 9 months. Results: Among 1,154 HIV-infected participants with valid South African ID numbers, 905 (78%) had baseline and 9-month data available. Mean age was 36 years, 49%were female, and 109 (12%) participants died after 9-month follow up. Time-independent parameters that increased mortality risk included male sex (HR 1.41, 95% CI 0.96-2.08) and not starting ART (HR 1.48, 95% 0.97-2.26). Less social support at 9 months compared to baseline significantly increased mortality risk (HR 1.17, 95% CI 1.03-1.33). Going without basic needs or healthcare at both baseline and 9 months more than doubled mortality risk compared to not going without these at either time point (HR 2.45, 95% CI 1.03- 5.79). A change from not foregoing basic needs or healthcare to afford the other at baseline to needing to do so at 9-months increased mortality slightly more (HR 2.71, 95% CI 1.47-4.99) when also compared to not foregoing basic needs or healthcare at either time point. Conclusion: Less social support and changes in competing needs between time of HIV diagnosis and 9-month follow-up significantly increase long-term mortality risk. Reassessing contextual factors during follow-up and targeting interventions to increase social support and affordability of seeking care may reduce long-termmortality for HIV-infected individuals in South Africa. 914 MORTALITY RATE AND ASSOCIATED RISK FACTORS AMONG HIV- INFECTED ADULTS ON ART IN KENYA Jacques Muthusi 1 , Irene Mukui 2 , Evelyne Ngugi 1 , Tai Ho Chen 1 , Kenneth Masamaro 1 , Samuel M. Mwalili 1 , Peter W. Young 1 , Emily C. Zielinski-Gutierrez 1 1 US CDC Nairobi, Nairobi, Kenya, 2 Ministry of Health, Nairobi, Kenya Background: Since the early 2000s, Kenya has scaled up antiretroviral therapy (ART) for HIV-infected adults and children. However, significantly high HIV related deaths have been reported (approximately 36000 deaths per year as at 2016). We investigated mortality rate and associated risk factors among HIV- infected adults on ART in Kenya. Methods: We conducted a retrospective national survey of HIV infected patients aged 15 years and above, who initiated ART from October 2003 to September 2013 at a representative sample of health facilities in Kenya. We abstracted data from patient medical records, including documented deaths during the study period. Patients were censored at the end of the study if they were still active in care. We used Chi-square statistics to compare patient characteristics by outcome status. We used Cox regression model to identify factors associated with mortality. Survey design parameters including weights, clustering and stratification were considered in all analyses. Results: Of 2517 adult patients initiated on ART during the study period, 1850 (74%) had documented outcomes at the end of the study period. Sixty four percent (1178) were female, 1110 (60%) had been on ART for < 5 years, and 768 (60%) were enrolled with WHO stage 1–2 with median age at ART initiation of 35.1 years (inter-quartile range [IQR] 28.8–42.8). Median follow-up time was 4.1 person-years (PY) (IQR 2.1–6.6) and was significantly different among patients who died (0.6) versus those censored (4.4), p<.01. The total follow-up time was 8172 PY and resulted in 156 (8.4%) deaths and 1694 censored patients. The

overall mortality rate was 1.9/100 PY. Patients who died were more likely to be male than female (12% versus 7%, p<.01), with WHO stage 3–4 (9%) compared to stage 1–2 (2%), p<.01, and had been of ART for < 5 years (13%) versus those with >= 5 years (2%), p<.01. The main factors associated with mortality were male sex (adjusted hazard ratio [aHR] 2.0; 95% confidence interval [CI] 1.1-3.5, p=0.01), WHO stage 3–4 vs 1–2 (aHR 6.9; 95% CI 3.3–13.7, p<.01) and been on ART for < 5 years (aHR 9.1; 95% CI 4.0–20.6, p<.01) (Table 1). Conclusion: Despite accessibility of ART, HIV related deaths continue to be reported especially among men, adult patients initiating ART with advanced disease, and during early years of treatment. There is a need to improve strategies for HIV case identification and close monitoring of patients during early years of initiating ART. Male-targeted intervention are also needed.

Poster Abstracts

915 MORTALITY RATES AND CAUSES OF DEATH ACCORDING TO INCLUSION PERIOD IN HIV/HCV PATIENTS Mathieu Chalouni 1 , Dominique Salmon 2 , Marc-Antoine Valantin 2 , Firouze Bani-Sadr 3 , Eric Rosenthal 4 , Laure Esterle 5 , Philippe Sogni 6 , Linda Wittkop 5 , for the ANRS CO13 HEPAVIH study group 1 INSERM, Bordeaux, France, 2 Assistance Publique – Hôpitaux de Paris, Paris, France, 3 CHU de Reims, Reims, France, 4 L’Université Nice Sophia Antipolis, Nice, France, 5 L’Université de Bordeaux, Bordeaux, France, 6 Paris Descartes University, Paris, France Background: Availability of direct acting antivirals (DAA) against HCV has potentially changed mortality rates and underlying causes of death in HIV/HCV co-infected patients. We aimed to compare the three-year mortality rates and underlying causes in a cohort of HIV/HCV co-infected patients, according to inclusion periods reflecting anti-HCV treatment periods. Methods: The ANRS CO13 HEPAVIH cohort is a nationwide cohort of HIV/ HCV co-infected patients with prospective data collection. We defined three inclusion periods: before 2011 (no DAA), between 2011 and 2014 (1st generation DAA), after 2014 (DAA period). Mortality rates were estimated overall and by cause (hepatic, HIV, cardiovascular, non-HIV non-HCV cancer, others (cerebral hemorrhage, overdose, septic shock, suicide, unknown cause…)) using Aalen- Johansen method accounting for competitive risks. Impact of inclusion period on all-cause and cause-specific mortality was evaluated using Cox proportional

CROI 2019 357