CROI 2025 Abstract eBook

Abstract eBook

Poster Abstracts

836

Agrarian Diet Intervention Improves Metabolic Health in Men With HIV Who Have Sex With Men John O'Connor, Jennifer Fouquier, Charles Neff, Amy Noe, Catherine Lozupone University of Colorado Anschutz Medical Campus, Aurora, CO, USA Background: People living with HIV (PLWH) suffer from increased incidence of metabolic disease and chronic systemic inflammation, even with effective Antiretroviral therapy (ART). Prevotella -rich, Bacteroides -poor (PRBP) microbiomes are commonly found in men who have sex with men (MSM), which make up a large portion of PLWH in the United States. A high-fiber agrarian diet (AD) has been found to be metabolically beneficial in PRBP microbiomes and is also seen in agrarian populations. Furthermore, a high-fat western diet (WD) can exacerbate disease in HIV animal models. There is thus potential for high fiber/low-diets diets to have health benefits in HIV positive MSM (HIV(+)MSM). Methods: We conducted a 4-week diet intervention in HIV(+)MSM and HIV(-) control groups, with subjects randomized to an AD or WD arm. Blood samples were collected for measurement of plasma inflammatory markers, as well as low-density lipoprotein cholesterol (LDL-C), high-density lipoprotein cholesterol (HDL-C), and triglycerides. Fecal and colonic biopsy samples were collected for 16S rRNA sequencing and cytometry by time-of-flight mass spectrometry (CyTOF). Additional host and microbial RNA sequencing was performed on biopsy samples. Results: Individuals randomized to the AD had reductions in levels of both HDL-C and LDL-C, with the sustained reductions in LDL-C being strongest in HIV(+)MSM. The AD was also associated with reduced PD1+ and CD40+ T cells and reduced monocytes in the blood, independent of HIV or MSM status. While the fecal microbiome did not significantly change on the AD, a baseline microbiome with higher Prevotella and lower Bacteroides abundance (lower Weighted Unifrac Axis 2) was observed in MSM groups ( Figure 1A and B ). Prevotella richness and Bacteroides poorness predicted greater decreases in LDL-C with the AD ( Figure 1B ). Analysis of host transcriptional changes in biopsy samples of individuals highly compliant to the AD-intervention showed an increase in immune activation and changes in small RNA signaling molecules with the AD. Conclusions: Our findings suggest that matching PRBP microbiomes to an AD has the power to reduce dyslipidemia and immune activation and exhaustion in PLWH, providing clinical utility in HIV-associated metabolic disease. The figure, table, or graphic for this abstract has been removed. Decreased Anti-Inflammatory Tryptophan Metabolites Correlate With Higher CVD Risk in PWH Manuela Ceccarelli 1 , Alejandro Abner Garcia Leon 2 , Eoin R. Feeney 2 , Mary Horgan 2 , Aoife Cotter 2 , Eoghan de Barra 3 , Corinna Sadlier 4 , Padraig McGettrick 2 , Giuseppe Nunnari 5 , Jane A. O'Halloran 2 , Mallon W. G. Paddy 2 , for the All-Ireland Infectious Diseases Cohort Study 1 Università Kore di Enna, Enna, Italy, 2 University College Dublin, Dublin, Ireland, 3 Royal College of Surgeons in Ireland, Dublin, Ireland, 4 Cork University Hospital, Cork, Ireland, 5 University of Catania, Catania, Italy Background: Distinct inflammatory profiles have been associated with increased prevalence of cardiovascular disease (CVD) in people with HIV (PWH). Alterations in tryptophan metabolism through the kynurenine pathway (KP) have also been observed in PWH and are associated with CVD. We aimed to explore relationships between KP metabolites and inflammatory profiles of PWH. Methods: In the All-Ireland Infectious Diseases (AIID) cohort, we previously identified 3 distinct inflammatory clusters using principal components analysis and hierarchical clustering on 38 plasma biomarkers. Cluster 3 (C3), denoted by higher markers of T-cell inflammation and gut endothelial dysfunction, was associated with 7-fold increased prevalence of CVD. Within this cohort, we matched (1:1:1) participants in C3 to Cluster 1 (C1) (non inflamed) and Cluster 2 (C2) (systemically inflamed) by sex-at-birth, age and ethnicity using propensity score with the nearest neighbor method. We measured the relative concentrations of ten KP metabolites (AU/mL) in plasma by UHPLC-MS/MS. We used Kruskall-Wallis to identify differences between groups followed by Dunn's multiple comparison test (with correction) to determine specific pairwise differences between clusters. Data are median (IQR) unless specified. Results: We included 87 PWH, 29 from each cluster. Overall, 34.5% were female, 52.3% Caucasian, age 45 (39.5-49) years, time since HIV diagnosis

the predictive potential for CVD of the studied parameters. Enrichment and Pathway analysis was performed by MetaboAnalyst 5.0 software. Results: Lysophosphatidylcholines (LPC-16:0e, -16:1p, and -O18:1p) were less abundant in PWH before CVD, while 29 lipids, mainly triacylglycerols, were more abundant. Pyruvic acid and d-fructose levels were significantly increased in CVD group. Pyruvic metabolism was the pathway with major impact. We observed increased levels of regulatory and senescent T cells, higher percentages of more differentiated monocyte subsets and lower percentages of myeloid and plasmacytoid DCs in PWH before suffering CVD, comparing with controls. An integrative predictive model combining LPC-O 18:1p, d-fructose, myeloid DCs, and non-classical monocytes achieved high predictive accuracy (AUC 0.831, PPV 0.762, NPV 0.824). The introduction of this combination increased the predictive accuracy of FRS (FRS, AUC 0.742, PPV 1.000, NPV 0.630; FRS + LPC-O 18:1p + d-fructose + myeloid DCs + non-classical monocytes, AUC 0.975, PPV 0.949, NPV 0.919). Conclusions: Significant alterations in lipidomic, metabolomic and immune profiles precede CVD in PWH. Moreover, the integration of these parameters increase the predictive accuracy of a common risk score for CVD in PWH. These findings underscore the potential for early identification of CVD risk in PWH using a multi-biomarker approach. Patterns of Social and Biological Risk Factors in People With HIV and Risk of Cardiovascular Disease Cynthia N. Ramirez 1 , Stacey Alexeeff 1 , Andrea Altschuler 1 , Nicola Bulled 2 , Michael Horberg 3 , Alexandra Lea 1 , Tory Levine 1 , Elise Riley 4 , Haihong Hu 3 , Alan Go 1 , Derek Satre 4 , Michael Silverberg 5 1 Kaiser Permanente Division of Research, Oakland, CA, USA, 2 University of Connecticut, Storrs, CT, USA, 3 Kaiser Permanente Mid-Atlantic States, Rockville, MD, USA, 4 University of California San Francisco, San Francisco, CA, USA, 5 Kaiser Permanente Northern California, Oakland, CA, USA Background: Despite extensive research, it is unclear why the high cardiovascular disease (CVD) burden in people with HIV has endured for decades. One potential explanation is that elevated CVD risk is the result of the combined effects of social and biological factors in people with HIV which has received limited attention. Methods: We identified all adult (≥18 years) people with HIV with at least 1-year of continuous membership enrolled in Kaiser Permanente (KP) Northern California and KP Mid-Atlantic between 2000 – 2022. The study sample excluded people with HIV with prior history of a Major Adverse Cardiovascular Events (MACE). We used latent class analysis (LCA) to identify classes of people with HIV with similar patterns of healthcare access, social vulnerability (census based CDC social vulnerability index themes), mental health (MH) (depression or anxiety) and substance use (SU) (smoking status and SU disorders), and CVD risk factors (see Figure for individual risk factors). Bayesian information criterion was used to determine model complexity and fit. We then used Cox proportional hazard models to estimate the hazard ratio (HR) of incident MACE by LCA class, adjusting for age, race, sex, start of follow up year, years of prior enrollment, and KP region. Results: The study sample included 25,510 people with HIV separated into 6 classes (Figure). Probabilities for individual risk factors from LCA were used to qualitatively name the 6 classes (see labels in Figure). In adjusted models, Class 5 (high CVD risk) [HR: 1.90; 95% CI 1.57-2.30] and Class 6 (high CVD and social vulnerability risk) [2.62; 1.94-3.52] had a higher risk of incident MACE compared with Class 1 (low overall risk). Class 2 (mixed social vulnerability risk) [1.11; 0.95-1.31], Class 3 (high social vulnerability risk) [1.16; 0.99-1.36], and Class 4 (high MH/SU risk factors) [1.22; 0.99-1.51] also had a higher, but not statistically significant, risk of incident MACE compared with the low overall risk class. Conclusions: We identified 2 classes with increased risk of MACE, both of which included people with HIV with a higher prevalence of CVD risk factors. What distinguished these classes was that one had a high prevalence of social vulnerability risk factors while the other did not, suggesting distinct targeted efforts may be needed to close the CVD gap among people with HIV.

835

Poster Abstracts

837

The figure, table, or graphic for this abstract has been removed.

CROI 2025 253