CROI 2020 Abstract eBook

Abstract eBook

Poster Abstracts

(DRD) is the leading cause of death amongst PWID. DRD in Scotland, a high income country, were reported at the highest on record in 2018. Amongst PWID in Scotland, the HIV prevalence is estimated at 2.3%, a recent increase explained by an outbreak in Glasgow. We sought to retrospectively evaluate the non-AIDS crude mortality rate (CMR), amongst PWID diagnosed with HIV in Scotland, and examine causes of death, over time. Methods: A retrospective cohort review of all patients diagnosed with HIV, with injecting drug use identified as the transmission risk factor, between 1st January 2000 and 31st December 2018, with follow up until 30th June 2019, identifying those who died. Data was collated on basic demographics, virological markers, cause of death and time from HIV diagnosis to death. The main outcome measure was the all-cause crude mortality rate (CMR) and primary cause of death over time. Drug-related mortality was defined as per the National Records of Scotland definition. Results: 413 PWID were diagnosed with HIV in the study period. 32% (133/413) were female. Mean age at diagnosis increased over time from 33 years (diagnosed 2000-04) to 39 years (diagnosed 2015-18). Until 30th June 2019, 22% (92/413) of these had died. Mean age of death was 42 years with no change over time. The non-AIDS CMR (per 100 PYFU) was 6.96 in those diagnosed from 2015-2018. The cause of death by year of HIV diagnosis are shown in figure 1; death from HIV/AIDs decreased from 61% (14/23) in those diagnosed 2000-2004 to 0% in those diagnosed from 2015-2018. DRD increased from 43% (10/23) to 63% (17/27) for the same groups. Conclusion: Mortality amongst PWID living with HIV in Scotland is increasing over time and non-AIDS CMR is now higher than previously reported PWID living with HIV. HIV/AIDS is no longer a cause of death in this cohort and DRD are increasing, despite extensive, free to access, addiction and recovery services. To tackle the rising mortality rate, alongside high quality HIV care this cohort require novel interventions including heroin assisted treatment and highlight the case for safer drug consumption facilities.

for baseline CD4 count, age, gender and transmission risk group with mortality analyses additionally adjusted for baseline viral load. Results: Of 33206 PLHIV, 4056 died during 183,683 person-years follow-up. 9,623 (28.7%) of PLHIV were non-drinkers, whilst 19,738 (58.9%), 3,320 (9.9%), and 857 (2.6%) had low, medium, and high alcohol intake, respectively. PLHIV with medium and high alcohol intake had higher odds of detectable viral load at baseline (aORs 1.14 [95%CI 1.05, 1.24] and 1.57 [1.36, 1.83], respectively) compared with low intake (figure). Medium- and high-drinkers had faster time to detectable viral load than those with low intake, aHRs 1.13 (1.02, 1.26) and 1.60 (1.35, 1.90), respectively. For mortality, aHRs compared with low intake were 1.33 (1.24, 1.42) for non-drinkers, 1.20 (1.06, 1.36) for medium intake, and 1.99 (1.70, 2.33) for high intake. Conclusion: Among PLHIV, high or medium alcohol intake is associated with higher mortality than low intake. Higher mortality risk for non-drinkers is likely due to a “sick-quitter” effect. PLHIV with medium or high alcohol intake were more likely than those with low intake to have a detectable viral load at baseline. Interventions to reduce unhealthy alcohol use among PLHIV should be considered.

Poster Abstracts

896 DEPRESSION AND VIROLOGIC REBOUND AMONG PATIENTS WITH HIV IN THE UNITED STATES Quan M. Vu 1 , R. L. Shouse 1 , John Weiser 1 1 CDC, Atlanta, GA, USA Background: The relationship between depression and HIV virologic rebound among persons with HIV has not been characterized. We analyzed nationally representative data from the Medical Monitoring Project (MMP) to examine the association between depression and virologic rebound among adults with diagnosed HIV in the United States. Methods: We used data collected during 6/2015-5/2018 fromMMP, a surveillance system that produces nationally representative estimates of behavioral and clinical characteristics among adults with diagnosed HIV. Demographic characteristics were collected through interview. Data on viral loads and diagnoses including clinical depression and substance use disorder were abstracted frommedical records during the two years prior to interview. A total of 7133 patients who were prescribed antiretroviral therapy (ART), had an initial HIV RNA viral load [VL] level <50 copies/mL, and had at least 1 subsequent VL measure during 2 years of follow-up were included in the analysis (9232 person-years). We estimated weighted incidence rates of virologic rebound (defined as having a VL ≥200 copies/mL following viral suppression during follow-up) and used Cox proportional hazards modeling to estimate the association between depression and the time to first virologic rebound, adjusting for selected covariates and accounting for sample weights and design. Results: Overall, 27% of patients had depression. The weighted incidence rate of virologic rebound was 9.2 per 100 person-years (95% confidence interval [CI] = 9.1-9.4) among patients with depression, and was 6.8 per 100 person-years (95% CI = 6.7-6.9) among patients without depression. In a multivariable Cox proportional hazards model that accounted for sample weights and design, and controlled for age group, gender, race/ethnicity, and diagnosis of substance use disorder, factors known to be associated with viral rebound, patients with depression had a significantly higher hazard of virologic rebound during follow- up, compared with patients without depression (adjusted hazard ratio = 1.35, 95% CI = 1.13-1.61, p <0.001; Figure).

895 ASSOCIATIONS OF ALCOHOL CONSUMPTION WITH VIRAL SUPPRESSION AND ALL-CAUSE MORTALITY Adam Trickey 1 , Suzanne Ingle 1 , Lei Zhang 1 , Mario Sarcletti 2 , Matthias Cavassini 3 , Michael Saag 4 , Heidi M. Crane 5 , Derek Satre 6 , Michael J. Silverberg 6 , Amy C. Justice 7 , Jonathan Sterne 1 , for the ART-CC and the COMpAAS consortium 1 University of Bristol, Bristol, UK, 2 Innsbruck Medical University, Innsbrusk, Austria, 3 Lausanne University Hospital, Lausanne, Switzerland, 4 University of Alabama at Birmingham, Birmingham, AL, USA, 5 University of Washington, Seattle, WA, USA, 6 Kaiser Permanente Division of Research, Oakland, CA, USA, 7 Yale University, New Haven, CT, USA Background: Unhealthy alcohol use may lead to higher morbidity and mortality among people living with HIV (PLHIV), either directly or by influencing the success of antiretroviral therapy (ART). We investigated associations of alcohol intake with viral suppression and all-cause mortality. Methods: We assembled data from 5 cohorts participating in the Antiretroviral Therapy Cohort Collaboration that provided AUDIT-C alcohol measures (categorised as no drinking and low [reference group], medium and high intake). Eligible PLHIV were aged ≥16 years and started ART 1996-2018. The date of AUDIT-C measure after ART start was taken as baseline, with follow up censored at the first of loss to follow-up or death. We used logistic regression to estimate adjusted odds ratios (aOR) for detectable viral load at baseline and Cox models (stratified by cohort) to estimate adjusted hazard ratios (aHR) for virological failure among those with undetectable baseline viral load (censoring at 3 years after ART start) and for all-cause mortality. Models were adjusted

CROI 2020 335