CROI 2025 Abstract eBook

Abstract eBook

Poster Abstracts

1164 Projecting Demographics and Causes of Death Among People With HIV in Western Europe to 2050 Julie Ambia 1 , Ard van Sighem 2 , Nikos Pantazis 3 , M. John Gill 4 , Sophie Abgrall 5 , Antonella d'Arminio Monforte 6 , Robert Zangerle 7 , Andreu Bruguera 8 , Janne Vehreschild 9 , Matthias Cavassini 10 , Inma Jarrin 11 , Fabrice Bonnet 12 , Jonathan A. C. Sterne 1 , Suzanne M. Ingle 1 , Adam Trickey 1 1 University of Bristol, Bristol, UK, 2 Stichting HIV Monitoring, Amsterdam, Netherlands, 3 National and Kapodistrian University of Athens, Athens, Greece, 4 University of Calgary, Calgary, Canada, 5 Institut Pierre Louis d'Epidémiologie et de Santé Publique, Paris, France, 6 Icona Foundation, Milan, Italy, 7 Medical University of Innsbruck, Innsbruck, Austria, 8 Centre d'Estudis Epidemiològics Sobre les ITS i Sida de Catalunya, Barcelona, Spain, 9 University of Cologne, Cologne, Germany, 10 Lausanne University Hospital, Lausanne, Switzerland, 11 Instituto de Salud Carlos III, Madrid, Spain, 12 Bordeaux University Hospital, Bordeaux, France Background: Antiretroviral therapy (ART) scale-up in Western Europe has reduced the incidence of AIDS, increased life expectancies, and led to an aging population of people with HIV (PWH). We aimed to predict cause-specific mortality patterns in Western Europe in the context of changes in demographics and ART use. Methods: A deterministic, compartmental model was parameterised and calibrated using epidemiological data for 2010-2022 on Western Europe from the European Centre for Disease Prevention and Control and UNAIDS, including population sizes, new HIV cases, ART coverage, demographics, and migrants with HIV. Data on 147,698 PWH on ART in 8 European countries were taken from the Antiretroviral Therapy Cohort Collaboration, including cause-specific mortality rates by demographics and CD4 cell counts. The model accounted for age, sex at birth/mode of HIV acquisition, ART status, CD4 counts, and hepatitis C virus. We compared the status quo to a scenario where, from 2025, the ART start rate is 1.5x higher, the rate of stopping ART is halved, and HIV infections decrease by 1,000 each year from 2025 until reaching zero. Results: By 2050, 1.1 million PWH are predicted to be living in Western Europe, up from 950,000 in 2025. Assuming current intervention coverage, 88% of PWH would be on ART in 2050. The percentage of PWH aged ≥76 years is predicted to rise from 10% in 2025 to 29% in 2050. The all-cause mortality rate is estimated to increase from 1.09 (95% Credibility Interval: 1.04-1.15) in 2025 to 1.53 (1.51 1.57) per 100 person-years in 2050. The estimated decline in the AIDS mortality rate between 2010 and 2025 is 80%. From 2025-2050, projected AIDS mortality rates remain at 0.27 per 100 person-years (Figure). Meanwhile, cardiovascular related and non-AIDS defining cancer mortality rates would increase from 0.17 (0.16-0.18) to 0.33 (0.31-0.35), and 0.23 (0.22-0.24) to 0.37 (0.36-0.39) per 100 person-years, respectively, to be the two leading causes of death among PWH by 2050. In 2050, 42% of deaths among PWH would occur among PWH aged ≥76 years. Over the next 25 years, if there is an increase in ART coverage (reaching 91% in 2030) and a steady decrease in HIV incidence, there would be 32,387 (29,757-35,569) fewer deaths. Conclusions: We estimate that by 2025 the UNAIDS target of reducing AIDS deaths by 75% from 2010 levels will be met in Western Europe. However, subsequently, non-AIDS-related deaths would increase over time due to the ageing population of PWH offsetting the AIDS mortality reduction.

1165 Causes of Mortality Among Persons Living With HIV in Thailand, 2007-2022 Wiphawee Kiatchanon 1 , Niramon Punsuwan 2 , Sanny Northbrook 3 , Suchunya Aungkulanon 4 , Apiratee Kanphukiew 1 , Jennifer Favaloro 5 , Poonchana Wareechai 6 , Cheewanan Lertpiriyasuwat 2 1 Thailand Ministry of Public Health–US CDC Collaboration, Nonthaburi, Thailand, 2 Ministry of Public Health, Nonthaburi, Thailand, 3 US Centers for Disease Control and Prevention Nonthaburi, Nonthaburi, Thailand, 4 National Health Security Office, Chiang Mai, Thailand, 5 Centers for Disease Control and Prevention, Atlanta, GA, USA, 6 National Health Security Office, Bangkok, Thailand Background: Over 1,200 government and private healthcare facilities provide free HIV services under Universal Health Coverage (UHC) for >90% of persons living with HIV (PLHIV) in Thailand. The National AIDS Program (NAP) extracted death information from the National Death Registration System for all PLHIV since 2005. We analyzed mortality trends and causes among reported PLHIV from 2007 to 2022. Methods: We conducted a retrospective analysis of PLHIV ≥15 years registered in NAP as of January 31, 2023. Causes of death were categorized using an algorithm designed by clinicians, following Thai HIV case surveillance definitions and WHO HIV clinical staging guidelines. A text clustering model was used to automatically classify causes of death from 2007 to 2022. We applied a Cox proportional hazards regression model to calculated cause-specific mortality rate ratios (MRR) across two-year intervals adjusting for the last reported CD4 category, age group at death, and time-period. Results: Among 673,635 registered PLHIV ≥15 years, 162,051 (24%) died from various causes. The median age at HIV diagnosis was 35 years (IQR: 28-42 years), while it was slightly higher at 38 years (IQR: 31-46 years) for those who died. Mortality rates varied, from 9.5% (15,390) in 2007-2008 to 14% (22,736) in 2021-2022. Half of the deceased had a baseline CD4 count < 100 cells/µL and 30% (49,157) died before initiating treatment. The most common causes of death were AIDS (59,573 deaths; 36%), unspecified AIDS-related causes (44,369 deaths; 27%), and non-AIDS infections (11,918 deaths; 7%). The proportion of AIDS-related deaths increased slightly from 32% (4,870) in 2007-2008 to 35% (7,962) in 2021-2022. Using a Cox model, we found that AIDS-related mortality (adjusted MRR [aMRR]: 1.03, 95% CI: 1.02-1.04), unspecified AIDS-related mortality (aMRR: 1.05, 95% CI: 1.04-1.07) and deaths from suicide or accidents (aMRR: 1.07, 95% CI: 1.04-1.10) were associated with higher mortality risk compared to death from non-AIDS infections, after adjusting for factors such as CD4 count and age. Conversely, deaths from non-AIDS non-hepatitis malignancies (aMRR: 0.97, 95% CI: 0.95 0.99) and CNS-related causes (aMRR: 0.90, 95% CI: 0.86-0.94) were associated with a lower mortality risk. Conclusions: AIDS-related causes accounted for the majority of deaths among PLHIV >15 years. Targeted interventions to strengthen early case detection and improve treatment initiation could help reduce AIDS-related deaths. 1166 Text Clustering Model for Defining Cause of Death Among PLHIV in Thailand (2020-2022) Apiratee Kanphukiew 1 , Wiphawee Kiatchanon 1 , Niramon Punsuwan 2 , Khunkansasi Pimpakhan 2 , Suchunya Aungkulanon 3 , Cheewanan Lertpiriyasuwat 2 , Poonchana Wareechai 4 , Natdanai Khosanam 2 , Anunthawip Chaimao 2 , Sanny Northbrook 5 1 Thailand Ministry of Public Health–US CDC Collaboration, Nonthaburi, Thailand, 2 Ministry of Public Health, Nonthaburi, Thailand, 3 National Health Security Office, Chiang Mai, Thailand, 4 National Health Security Office, Bangkok, Thailand, 5 US Centers for Disease Control and Prevention Nonthaburi, Nonthaburi, Thailand Background: In Thailand, cause of death is stored in free text format on death certificates and classified manually based on clinical expertise, Thai HIV surveillance definitions, and WHO clinical staging guidelines. Limitations, particularly in extracting and categorizing causes of death from text-based records in the Ministry of Interior's death registry, are significant obstacles in monitoring HIV-related mortality. We developed a text clustering model to automatically classify cause of death and differentiate between AIDS-related and other causes of death among HIV-positive individuals who died between 2002 and 2022. Methods: We developed and applied a text clustering model to automatically categorize cause of death into 3 main groups: AIDS-related causes (e.g., opportunistic infections), non-AIDS related causes (including non communicable and non-natural causes), and ill-defined causes. Clustering was based on similarity ratios, keywords, and >85% fuzzy string-matching score in Python. Additionally, we created a confusion matrix to measure the text

Poster Abstracts

CROI 2025 381