CROI 2020 Abstract eBook

Abstract eBook

Poster Abstracts

cardiovascular (CVD), renal, non-AIDS-defining cancers (NADC), hypertension, diabetes, and chronic obstructive pulmonary disease. Marginal structural models estimated the risk of all-cause mortality among PLWH with 1, 2 and ≥3 comorbidities (versus none). Results: Overall, 51% of 8,405 PLWH, and 30% of 42,025 HIV-negative individuals developed ≥1 comorbidity by the end of follow-up. With the exception of the CVD-NADC combination, PLWH had higher all-cause mortality rates for all singular and combinations of diseases (see Figure). The largest disparity in mortality rate was related to renal disease (in isolation), where PLWH had a rate >30 times higher than that of HIV-negative controls. Among PLWH and the HIV-negative controls, a liver-NADC combination was associated with the highest mortality rate per 1000 person-years: 106.6 (95% confidence interval: 73.57-139.64) and 78.2 (46.24-110.16), respectively. After adjustment for demographic and time-dependent treatment-related confounders, PLWH with 1, 2 and ≥3 comorbidities were, respectively, 3.15 (2.57-3.86), 5.95 (4.65-7.61) and 12.96 (15.59-40.80) times more likely to die than PLWH without comorbidities. Conclusion: Compared to HIV-negative controls, after adjusting for similar morbidities, PLWH experienced substantial excess in mortality rates. Additionally, we observed a strong positive dose-response between the number of morbidities and the risk of mortality among PLWH. These results highlight the critical role that additional morbidities continue to pose as drivers of mortality among PLWH within a publicly funded province-wide ART program.

Santiago, Chile, 7 Weill Cornell Medicine, New York, NY, USA, 8 Instituto Nacional de Infectologia Evandro Chagas, Rio de Janeiro. Brazil Background: Increased survival among persons living with HIV (PLWH) receiving antiretroviral therapy (ART) has been documented in the United States, Canada, and Europe. However, sparse data exist on life expectancy in low- and middle-income settings where ART is increasingly available. We therefore calculated life expectancy gains among PLWH initiated ART within the Caribbean, Central and South America network for HIV epidemiology (CCASAnet). Methods: We included PLWH started on ART and ≥20 years old between 2003- 2017 from CCASAnet sites in Haiti, Mexico, Honduras, Peru, Argentina, Chile, and Brazil. PLWH contributed person-time until the first of death, last cohort contact, database closure, or December 2017. Due to differences in general population life expectancies and clinical sites, we stratified analyses by Haiti vs. all other sites. We used the Chiang method for abridged life tables to calculate life expectancy at age 20 for three eras (2003-2008, 2009-2012, and 2013-2017) overall and by demographic and clinical characteristics at ART initiation. As mortality ascertainment varies by country, mortality rates were weighted for probability of loss to follow-up (LTFU) using adjusted Poisson regression models. Results: Among 30,688 PLWH included, 17,491 (57%) were from Haiti, of whom 57%were female, 23% initiated ART in 2003-2008, 32% in 2009-2012, and 45% in 2013-2017. Of those from other sites, 23%were female, and 7% initiated ART before 2003, 29% in 2003-2008, 26% in 2009-2012, and 38% in 2013-2017. At ART initiation, 36% of PLWH from Haiti and 46% from all other sites had CD4+ count <200 cells/µL (17%missing). There were 1,470 deaths and 7,154 LTFU among PLWH from Haiti and 1,167 deaths and 3,174 LTFU at other sites. Crude and weighted mortality rates markedly decreased among all age groups over calendar eras. There were accompanying significant improvements in life expectancy, approaching that of the general population (61 years in Haiti and 69 years at other sites, in 2013-2017), though disparities by sex were significant in Haiti (Figure). While life expectancy improved over time, disparities by CD4+ count, education, and tuberculosis at or prior to ART persisted. Conclusion: Life expectancy among PLWH on ART has significantly improved in Latin America and approaches that of the general population. Persistent disparities in life expectancy by sex, CD4+ count, education, and history of tuberculosis highlight vulnerable populations in the region.

Poster Abstracts

870 EXCESS MORTALITY AFTER CANCER DIAGNOSIS IN PERSONS WITH HIV Michael J. Silverberg 1 , Varada Sarovar 1 , Derek Satre 2 , Stacey Alexeeff 1 , Wendy Leyden 1 , Jennifer O. Lam 1 , Alexandra N. Anderson 1 , Rulin Hechter 3 , Haihong Hu 4 , Julia L. Marcus 5 , William J. Towner 3 , Qing Yuan 3 , Michael A.Horberg 4 1 Kaiser Permanente Division of Research, Oakland, CA, USA, 2 University of California San Francisco, San Francisco, CA, USA, 3 Kaiser Permanente Southern California, Pasadena, CA, USA, 4 Kaiser Permanente Mid-Atlantic States, Rockville, MD, USA, 5 Harvard Pilgrim Health Care Institute, Boston, MA, USA Background: Persons with HIV (PWH) may have reduced survival after a cancer diagnosis compared with uninfected persons, yet prior studies are limited by insufficient control for confounding from access to care and other comorbidities, and do not often account for the increased background mortality rate of PWH. Methods: We conducted a cohort study during 2000-2016 of adult PWH who were members of Kaiser Permanente in Northern California, Southern California, or Mid-Atlantic States, which provide comprehensive cancer and HIV care. Uninfected adults were matched 10:1 to PWH by age, sex, race/ethnicity, medical center, and year. The primary outcome was all-cause mortality per 1,000 person-years (py), and the predictors were HIV status, and cancer grouped as: any cancer; AIDS-defining cancers (ADC); non-AIDS-defining cancers (NADC); virus-unrelated NADCs; virus-related NADCs; and HPV-related NADCs (see Table footnote). We first computed mortality rate differences (RD), separately by HIV status, to measure the increased mortality rates after cancer (RD>0 denotes higher mortality rates after cancer). Next, we modeled mortality rates using additive Poisson regression, including terms for HIV status, cancer, and an HIV*cancer interaction term. The interaction term represents the excess mortality rate associated with cancer among PWH as compared with uninfected persons. Adjusted models included terms for demographics, smoking, body

869 EXCESS MORTALITY AMONG PLWH WITH MULTIMORBIDITY COMPARED TO HIV-NEGATIVE CONTROLS Ni Gusti Ayu Nanditha 1 , Grace Zheng 2 , Hiwot M.Tafessu 1 , Taylor McLinden 1 , Andreea Bratu 1 , Robert S. Hogg 1 , Julio S. Montaner 1 , Viviane D. Lima 1 1 British Columbia Centre for Excellence in HIV/AIDS, Vancouver, BC, Canada, 2 University of British Columbia, Vancouver, BC, Canada Background: Antiretroviral therapy (ART) and gains in life expectancy have increased the likelihood of people living with HIV (PLWH) developing comorbidities. We examined which chronic comorbidities, experienced in isolation or in combination, led to higher mortality rates among PLWH compared to HIV-negative controls. Secondarily, we assessed the impact of multimorbidity on all-cause mortality among PLWH. Methods: This population-based cohort study used longitudinal individual- level data on all treated PLWH and 1:5 age-sex-matched HIV-negative controls in British Columbia (BC), Canada. Eligible participants were ≥19 years old and enrolled in the Comparative Outcomes and Service Utilization Trends Study between 2001 and 2012 for ≥1 year. Comorbidities were identified from provincial administrative health databases (i.e., hospitalizations, outpatient physician, and pharmacy records). Selected comorbidities included liver,

CROI 2020 325