Abstract TU184: Polygenic Risk Scores That Uncouple Obesity From Cardiometabolic Risk Protect Against Longitudinal CHD and T2D Risk

Abstract

Body mass index (BMI) defined obesity is an established risk factor for cardiometabolic diseases, yet its cardiometabolic consequence is highly heterogeneous. We recently conducted genome-wide association analyses (GWAS) of 24 metabolically healthy obesity (MHO) traits in the UK Biobank, combining three adiposity (body fat percentage [BFP], BMI, and waist–hip ratio) and eight cardiometabolic (lipids, glycemic, and blood pressure) traits. We hypothesized that polygenic risk scores (PRSs) derived from these MHO traits capture genetic predisposition to higher adiposity with protective cardiometabolic profiles, predicting MHO status and lower coronary heart disease (CHD) and type 2 diabetes (T2D) risk. MHO status was defined as (abdominal) obesity without elevated BP, glucose, or lipids. Individual MHO trait PRSs were constructed using SBayesRC, integrating summary statistics with ancestry-specific LD and functional annotations ( Fig. 1 ). Associations between 24 MHO PRSs and MHO status were evaluated in the REasons for Geographic and Racial Differences in Stroke (REGARDS) study (12,084 European ancestry [EUR]; 8,672 African Americans [AAs]). To optimize prediction, we trained a composite PRS (PRS MHO ) using an elastic net model with 100x 10-fold cross-validations to combine individual MHO PRSs in REGARDS. The resulting composite PRS MHO and individual MHO PRS were tested for predicting MHO status, incident CHD and T2D in the Atherosclerosis Risk in Communities (ARIC) study participants (9,241 EURs; 2,822 AAs). Descriptive characteristics are presented in Table 1. Most MHO trait PRSs were positively associated with MHO in both ancestries, but the strongest associations differed: ‘BMI–lipid’ PRSs in EURs and ‘BFP–SBP’ PRS in AAs. The odds ratio (OR) per SD increase in PRS BFP-SBP was over twofold greater in ARIC AA ([95% CI]: 4.01 [2.41–6.68]) vs. EUR (1.63 [1.51–1.76]). Individual MHO PRS explained up to 5.8% of MHO variance, while the composite PRS MHO explained 9.6% in EUR and 5.5% in AA ( Fig. 2 ). PRS MHO was associated with lower 20-year risk of CHD and T2D, particularly in individuals with obesity. The protective association with CHD remained after adjusting for baseline MHO (HR [95% CI]: 0.92 [0.86–0.99] in EUR; 0.89 [0.76–1.06] in AA). This study leverages genetics that uncouple obesity from its cardiometabolic comorbidities to predict MHO and its protection against CHD and T2D, providing potential targets for precision prevention.