CROI 2018 Abstract eBook

Abstract eBook

Poster Abstracts

Conclusion: Although effective viral suppression has led to significant increases in longevity and quality of life, ART is not yet able to fully restore life expectancy to a level comparable to that found in HIV-negative persons even when PWH are successfully treated in our Swedish context. If we were used the cut off at ≤ 50/mL the result might have been different. The risk of mortality decreases the longer an HIV patient is able to suppress their VL. Even when PLHIV are successfully treated there are several other important areas related to death, such as smoking and social factors, where data are still missing. 894 RATES OF MORTALITY AMONG SCHIZOPHRENIC PEOPLE LIVING WITH AND WITHOUT HIV Kalysha Closson 1 , Kate Salters 1 , Thomas L. Patterson 2 , Oghenowede Eyawo 1 , William Chau 1 , Monica Ye 1 , Mark Hull 1 , Viviane D. Lima 1 , Julio S. Montaner 1 , Robert S. Hogg 1 1 British Columbia Centre for Excellence in HIV/AIDS, Vancouver, BC, Canada, 2 University of California San Diego, San Diego, CA, USA Background: Schizophrenia (SZO) is a mental health condition that has important implications for morbidity and mortality outcomes, particularly for people living with HIV (PLHIV). As of yet, few studies have explored the impact of HIV and SZO on mortality. Methods: Using the Comparative Outcomes and Service Utilization Trends cohort study, a population-based retrospective cohort study examining health outcomes and service use of PLHIV and a 10% random sample of individuals in British Columbia (BC), SZO prevalence and mortality outcomes were estimated from 1998-2013. Prevalence of SZO was assessed using physician and hospital-based administrative data and International Classification of Disease 9/10 codes. Survival time by HIV-status was accessed by a Kaplan-Meier (KM) plot, with log-rank test for comparison. Age and sex–adjusted mortality rates were calculated by using 2016 Canada population as reference. The association between HIV and all-cause mortality among SZO+ individuals were examined using logistic regression. Results: Of 515,913 BC residents accessing medical services from 1998-2013 in our study sample, 2.6% (n=13,412) were PLHIV and were significantly more likely to be SZO+ compared to HIV- individuals (6.3% vs. 1.1%, p<0.001). Compared to SZO+/HIV-, SZO+/PLHIV were significantly (all p<0.001) more likely to be male (75% vs. 56%), live in an urban setting (91% vs. 88%), have a history of injection drug use (IDU) (75% vs. 20%), and ever be on anti- psychotic medication (49% vs. 39%). Age and sex standardized all-cause mortality rates (ASMR) were highest among PLHIV/SZO+ (66.9/1,000 person years [PY], 95%CI=50.6-83.1), compared to PLHIV/SZO- (SMR=39.5/1,000PY, 95%CI=36.6-42.3) and SZO+/HIV- (ASMR=28.2/1,000PY, 95%CI=26.5-30.0). The KM plot (Figure 1) indicate that time from SZO diagnosis to death was significantly shorter among PLHIV compared to HIV- individuals (p<0.001). In a confounding logistic regression model of all SZO+ individuals, HIV-status remained significantly associated with mortality (aOR=2.31, 95%CI=1.84-2.89), controlling for sex, baseline age, and IDU. Conclusion: PLHIV experience a six-times higher SZO prevalence compared to HIV- individuals, and among SZO+ individuals HIV is a risk factor for mortality. Moreover, PLHIV/SZO+ have higher mortality rates than PLHIV/SZO-. Physicians working with PLHIV/SZO+ that have high levels of IDU, should closely monitor treatment for SZO and HIV, so as to reduce mortality for this under-served, high-risk population.

895 CAUSES OF DEATH AND EARLY MORTALITY IN PEOPLE WITH HIV IN MEXICO (2004-2015) Yanink Caro-Vega 1 , Pablo F. Belaunzaran-Zamudio 1 , Jesús Alegre-Díaz 2 , Brenda Crabtree-Ramírez 1 , Raul Ramirez-Reyes 2 , Pablo Kuri-Morales 2 , Malaquias Lopez-Cervantes 2 , Juan Sierra-Madero 1 1 Instituto Nacional de Ciencias Médicas y Nutrición Salvador Zubirán, Mexico City, Mexico, 2 Secretaría de Salud, Mexico City, Mexico Background: Antiretroviral therapy (ART) use was followed by mortality reductions and a shift in the causes of death from AIDS to non-AIDS deaths in most settings. In Mexico, mortality due to HIV/AIDS has remained constant even after the expansion of ART use in 2004. Information regarding time from HIV diagnosis to death and causes of death may provide important information on the contribution of late diagnosis on mortality and therefore help to identify gaps in the early phases of the continuum of care. We aimed to estimate the proportion of people dying in the first year after HIV diagnosis, and to describe their causes of death in recent years. Methods: Using national death registry data, we identified all registered deaths among HIV-infected people between 2004-2015. We define early mortality (EM) as deaths occurring within the first year after HIV-diagnosis and very early mortality (VEM) as deaths in the first month after HIV–diagnosis. We describe changes in the proportion of EM by calendar year and stratified by gender using logistic models. We classified the main cause of death according to ICD-10, and described the most frequent causes of death by gender, calendar year and EM condition. Results: Between 2004 and 2015 there were 10,872 deaths in HIV-infected women and 48,824 in men. AIDS-related deaths occurred in 74% of all subjects. The most frequent AIDS-related causes of death were: pneumonia (n=14070, 23%), tuberculosis (n=4646, 8%), pneumocystis (n=2197, 4%), and AIDS malignancy (n=1281, 2%); non-AIDS-related causes were: sepsis (n=7352, n=12%) and non-AIDS malignancy (n=1281, 2%) (Fig 1A). When stratified by sex there were non-significant differences in causes of death over time. Fifty three percent (n=3,677) and 57% (n=17,824) were classified as EM in women and men, respectively and did not change along time in both groups. VEM increased from 15% to 22% p<0.001, with no differences by sex (Fig 1B). AIDS explained 75% of EM. Overall, men had a higher risk of EM (OR: 1.14 [95%CI: 1.09-1.18], p<0.01). Conclusion: The stable EM, increasing VEM and the high percentage of deaths related to AIDS are markers of late HIV diagnosis, which persists along time in Mexico. Non-AIDS causes of death are a small proportion, with modest increase in later years in acute myocardial infarction and non-AIDS malignancies. These results support that policy efforts should be directed to expand HIV diagnosis and early linkage to care.

Poster Abstracts

CROI 2018 340