Polygenic Scores Research Articles

Abstract 4139199: Quantifying the Burden of High Polygenic Risk for Cardiometabolic Disease in a US-based Healthcare System Biobank

Background: Healthcare systems are increasingly collecting genotype data on patients. Polygenic risk scores (PRS), which quantify inherited risk of a given disease, have the potential to leverage this data to identify high risk individuals. While PRS for single traits have often been evaluated, the same data can be used to calculate PRSs across traits and evaluate the overall burden of high cardiometabolic polygenic risk using a single test. Methods: We examined the association of cardiometabolic PRS from the Polygenic Score Catalog among 236,393 All of Us (AoU) participants for 6 target traits of interest, using all available PRS for coronary artery disease (CAD; 57 PRS), atrial fibrillation (AF; 30 PRS), heart failure (HF; 9 PRS), type 2 diabetes mellitus (T2DM; 133 PRS), venous thromboembolism (VTE; 9 PRS), and thoracic aortic aneurysm (TAA; 2 PRS). PRSMix is a tool that incorporates information across several PRS for a given trait to improve prediction accuracy for a target population. We generated a PRSMix for each trait and externally benchmarked its performance against individual PRS in the Mass General Brigham Biobank (MGBB). Phenotypes were ascertained using hospital diagnosis and procedural codes. Logistic regression was performed, adjusting for age, sex, and the first 10 principal components of inferred genetic similarity. Results: Of the 53,306 genotyped MGBB participants, 55.6% are female, mean enrollment age is 53 ± 17 years. Each trait-specific PRSMix demonstrated generally stronger associations with the target trait compared to any individual contributing score. This included trait-specific PRSMix associations (OR/SD, [95%CI]) of 2.07 [1.99-2.15] for CAD, 1.88 [1.82-1.95] for AF, 1.47 [1.43-1.51] for HF, 2.07 [2.00-2.14] for T2DM, 1.43 [1.39-1.47] for VTE, and 1.40 [1.34-1.46] for TAA. Findings were consistent across inferred genetically similar ancestry groups, with stronger PRSMix associations for CAD detected in the EUR group (2.14 [2.05-2.22]) compared to the AFR group (1.43 [1.21-1.70]). Approximately 37% of people in MGBB have 3-fold greater risk of at least one cardiometabolic disease, with 22% at risk for CAD, 14% for T2DM, 13% for Afib, 3% for VTE, 2% for HF and <1% for TAA. Conclusion: We developed PRSMix scores for cardiometabolic traits in AoU and validated them in MGBB. We identified 37% of biobank participants with at least 3-fold greater disease risk associated with high polygenic scores for at least one cardiometabolic outcome.

Abstract 4137717: Atrial Fibrillation in Valvular Heart Disease: Assessing the Impact of Genetics in the UK Biobank

Background: Atrial fibrillation (AF) in valvular heart disease (VHD) is typically linked to pressure or volume overload resulting in atrial remodeling. However, the independent contribution of genetics to AF in VHD beyond abnormal hemodynamics and other risk factors remains unclear. Objective: To study the role of a previously developed Polygenic Risk Score (PRS) for AF in predicting AF across different VHDs in the UK Biobank cohort. Methods: We analyzed UK Biobank data between 2000 and 2022. ICD codes were used to define incident AF and VHD status, including mitral valve (MV) disease [prolapse (MVP), regurgitation (MR) and stenosis (MS)] and aortic (AV) disease [stenosis (AS) and regurgitation (AR)]. Prior AF was excluded. Participants were divided into quintiles based on the PRS: the top quintile was defined as high genetic AF risk, second through fourth as intermediate genetic risk, and bottom quintile as low genetic risk. Results: Among 452841 participants with available PRS, 2238 had at least one VHD (385 MVP, 1058 MR, 81 MS, 421 AR, and 509 AS). AF incidence was 9.9%, 13.6%, and 19% in the low, intermediate and high genetic risk groups respectively, compared to 1.4%, 2.8%, and 5.7% in the general population (all p<0.05) - Figure 1A . Cox regression models adjusted for cardiovascular risk factors showed different genetic effects on AF across VHD types. Censoring patients at death, including valve intervention as a time-dependent variable, and excluding AF within 30 days post-intervention, PRS independently predicted AF in MV disease (hazard ratio [HR] 1.7, p=0.03 for the intermediate risk/mid PRS; HR 2.3, p=0.002 for the high risk PRS), particularly in those with MVP (HR 3.1, p=0.03 for the high PRS). However, in those with AV disease, PRS was not significantly associated with AF (p=0.93 and p=0.07) - Figure 1B. Conclusion: In patients with VHD, AF PRS is significantly associated with an increased risk of developing AF. However, the genetic impact varies across different VHD. In AV disease, there is a minimal genetic contribution, and the elevated AF risk may be predominantly due to pressure or volume overload. In contrast, in MV disease, particularly in MVP, there is a substantial genetic influence on AF risk.

Abstract 4136873: The Role of Socioeconomic Status in the Association Between Cardiovascular Polygenic Risk Score and Cardiovascular Events

Introduction: Polygenic risk scores (PRS) provide a summation of small-effect genetic variants that predict cardiovascular (CV) disease risk. Socioeconomic status (SES) can also impact CV risk, but it is unknown whether SES modifies the effects of PRS on CV events. Hypothesis: The association between CV PRS and CV events is modified by SES. Methods: Participants without CV disease at baseline from the UK Biobank prospective cohort study were followed from recruitment (2006-2010) through 2023. We included participants with genetic data (for CV PRS) and CV risk factors assessed at baseline (hemoglobin A1c, total cholesterol, blood pressure, body mass index, physical activity, tobacco use, and sleep). The CV PRS was centered at zero and standardized to one unit variance within each ancestry group. SES was defined by tertiles of index of multiple deprivation (from housing, crime, income, education, employment, health deprivation and disability, and neighborhood environment). The primary outcome was CV events (myocardial infarction, heart failure hospitalization, stroke, or death). Using Cox proportional hazard models, we estimated the total effect of PRS on CV events, tested for interaction of SES in the association between PRS and CV events, and estimated the direct effect of PRS on CV events by adjusting for age, sex, CV risk factors, and SES. Results: There were 201,374 participants (mean age 55.9 (± 8.1), 53.4% female). CV events occurred in 21,916 (10.9%) participants: 5,754 (2.9 %) myocardial infarctions, 4,061 (2.0%) strokes, 4,312 (2.1%) heart failure hospitalizations, and 12,782 (6.4%) deaths. Each unit of PRS was associated with 1.17 times the risk of CV events (HR 1.17, 95%CI 1.15 to 1.18, p-value < .001, adjusted for age and sex). People with the lowest SES had 1.22 times the risk of CV events (HR 1.22; 95%CI 1.18 to 1.26; p-value <.001) and the middle SES had 1.07 times the risk of CV events (HR 1.07; 95%CI 1.03 to 1.10; p-value <.001) compared to the highest SES, adjusted for age, sex, and CV risk factors. PRS remained associated with CV events (HR 1.15, 95%CI 1.13 to 1.16, p-value<.001) after adjusting for age, sex, CV risk factors, and SES. There was no significant multiplicative interaction between SES and PRS for CV events (interaction p-value = 0.45). Conclusions: SES is associated with CV events but does not modify the effect of PRS on CV events. Genetic risk remains independently important in CV event risk beyond traditional CV risk factors and SES.

Abstract 4137518: The implications of polygenic versus phenotypic characterization of obesity for cardiovascular-kidney-metabolic health

Introduction: The overall amount and the distribution of body fat, proxied by body-mass index (BMI) and waist-to-hip ratio (WHR), are two important measures of obesity and have significant polygenic heritability. We aimed to investigate whether polygenic scores (PGS) for these traits have additional predictive utility, beyond phenotypic characterization of obesity, for cardiovascular-kidney-metabolic (CKM) health. Methods: We derived PGS for BMI (PGS BMI ) and WHR adjusted for BMI (PGS WHRadjBMI ) using PRS-CS-auto on summary statistics from ~700k individuals of European ancestry (EA) in the GIANT consortium and selecting ~1M common variants for each score. We conducted association analyses in the Atherosclerosis Risk in Communities study, a prospective cohort including 8,631 EA individuals with ascertainment of incident diabetes, hypertension, chronic kidney disease, atrial fibrillation, coronary heart disease, ischemic stroke, and heart failure. We used nested Cox models to measure the added predictive utility of PGS to BMI and WHR for incident CKM phenotypes. Results: The study cohort (age 54.4±5.7 years, 52.4% women, follow-up 24.5±7.6 years) had a baseline BMI of 26.4 kg/m 2 (median [IQR]; 21.6-29.6) and WHR of 0.94 (0.88-0.98). PGS BMI and PGS WHRadjBMI were associated with their phenotypes (partial R 2 ; BMI 0.11 and WHR 0.04, p <0.001), although with a large phenotypic variability across the distribution of PGS (IQR for 1 st and 10 th deciles of PGS; BMI 19.6-28.3 and 22.6-35.9, WHR 0.79-1.03 and 0.87-1.05). Anthropometric measures of obesity were associated with a wider gradient of risk compared to PGS for obesity–most notable for diabetes and cardiovascular phenotypes (Table 1a). Addition of PGS BMI and PGS WHRadjBMI to BMI and WHR resulted in limited improvement in risk discrimination across incident CKM phenotypes (Table 1b). Conclusions: Observed anthropometric measures of obesity are sufficient, and for certain conditions far superior, to capture the risk of incident CKM phenotypes. These findings suggest PGS for adiposity traits likely have limited clinical utility to refine risk stratification for CKM syndrome among middle aged adults.

Abstract 4136501: Machine Learning Uncovers the Contribution of Rare Genetic Variants and Enhances Risk Prediction for Coronary Artery Disease in the Japanese Population

Background: Genome-wide association studies (GWASs) have enhanced our understanding of the genetic basis of coronary artery disease (CAD), and polygenic risk scores (PRSs) have facilitated the assessment of genetic risk. However, these methods predominantly focus on common variants due to statistical power, potentially leaving rare variants insufficiently analyzed and thus limiting the predictive performance of PRS. Methods: We conducted whole genome sequencing (WGS) of 1,752 Japanese early-onset myocardial infarction (MI) patients and 3,019 controls from Biobank Japan (BBJ). We performed case-control association studies including GWAS and gene-based tests, as well as a novel machine learning-based framework. In this framework, we developed a penalized regression model to predict the CAD status from genome-wide rare nonsynonymous variants. The model identified the minimal set of most distinguishing features (genes) and generated a rare variant-based risk score (RVS). The RVS was evaluated on an independent validation WGS cohort of 200 cases and 824 controls. We also derived a PRS based on CAD-GWAS (25,668 CAD cases vs 141,667 controls from BBJ) to compare the properties and performance between the RVS and PRS. Results: In the case-control studies, only two common variants in chromosome 12 were identified in GWAS, with no genes in gene-based analysis (SKAT-O), suggesting the challenges in rare variant analysis. On the other hand, our machine-learning framework identified 59 CAD-related genes, including LDLR , a causal gene of familial hypercholesteremia. Functional analyses revealed that various biological pathways, including lipid metabolism, immune system, and vessel development, are involved in CAD. For the genetic risk prediction, RVS significantly predicted CAD (area under the curve [AUC], 0.58; p=0.001, pseudo-R 2 , 0.051; p=7.92*10 -9 ). RVS was significantly associated with LDL cholesterol levels and coagulation function (Pearson’s r, 0.21; p=4.5*10 -5 and 0.10; p=0.03, respectively) and MI patients with high RVS (top 5%) showed higher cardiovascular mortality rate (p=0.03, log-rank test), highlighting the clinical importance of RVS. Finally, the combined risk score (CRS) of RVS and PRS significantly improved CAD prediction compared to PRS (AUC, 0.66 (CRS) vs 0.61 (PRS); p=0.007, pseudo-R 2 , 0.093 (CRS) vs 0.040 (PRS); p=0.0018, Figure ). Conclusions: Our machine learning framework successfully characterized rare variants and enhanced genetic risk prediction in CAD.

Abstract 4118183: Anxiety disorder, depression and coronary artery disease: associations and modification by genetic susceptibility

Background: Associations of anxiety disorder and depression with coronary artery disease (CAD) are heterogeneous between populations. Research questions: How does genetic susceptibility to CAD alter the associations between anxiety and depression and incident CAD? Aims: To investigate how genetic susceptibility to CAD alters the associations of anxiety and depression with incident CAD, comparing and combining anxiety and depression. Methods: This is a prospective cohort study using UK Biobank. Diagnoses of anxiety disorder and depression before the baseline assessment were ascertained through linked hospital admission data. Incident CAD was ascertained through hospital admission and death certificate data after baseline. CAD polygenic risk score (PRS) was obtained from CARDIoGRAMplus4 and categorised into low, intermediate, and high. Cox proportional hazard models were used to examine associations between anxiety and depression and CAD. The product term of PRS and these mental health conditions, and the relative excess risk due to interaction (RERI), were used to investigate the presence of multiplicative and additive interactions, respectively. Results: Our study included 288,152 participants of whom 54.4% were female, with a mean age of 56.4 years. Both anxiety disorder (HR: 2.24, 95% CI: 1.86–2.70) and depression (HR: 2.04, 95% CI: 1.80–2.32) were associated with CAD after adjusting for sociodemographic confounders. There was an addictive interaction between depression and PRS (RERI: 0.86; 95% CI: 0.06–1.65) such that the association between depression and CAD was strongest among those with a high PRS whilst there was no such evidence for anxiety disorder. Anxiety only (HR 1.69, 95% 1.17–2.45), depression only (HR 2.01, 95% CI 1.62–2.49) and concomitant anxiety and depression (HR 3.51, 95% CI 2.26–5.45) were associated with CAD even among people with a low PRS. Lifestyle mediators attenuated all these associations across PRS categories. Conclusions: CAD genetic susceptibility might partly contribute to the clustering of depression and CAD but does not provide a full explanation, nor does it explain the association between anxiety disorder and CAD. Therefore, other mechanisms should be explored.

Abstract 4121555: Integrated Polygenic Score Stratifies Risk of Peripheral Artery Disease and Major Adverse Limb Events

Background: Peripheral artery disease (PAD) is a heritable atherosclerotic condition. With growing knowledge of the genetic basis for PAD and related risk factors, there is opportunity to identify high-risk individuals for prevention and potentially intervention based on polygenic background. Aims: To develop and validate an integrated genome-wide polygenic score for PAD (GPS PAD ), evaluate association of polygenic risk with major adverse limb events (MALE), and compare predictive performance in diverse ancestral populations to previously published polygenic scores. Methods: We developed GPS PAD by integrating genetic association summary statistics for PAD and related traits stratified by ancestry. GPS PAD was trained in a sample of 96,239 European individuals from the UK Biobank, and validated in multi-ancestry cohorts, including a holdout validation UK Biobank dataset (N=304,294), and external All of Us (AoU; N=237,173) and Mass General Brigham Biobank datasets (N=37,017). GPS PAD performance was benchmarked against previously published polygenic scores and clinical risk factors. Results: GPS PAD was comprised of 603,595 variants with weights informed by association with PAD and ten PAD risk factors across five ancestry groups. GPS PAD had an OR-per SD of 1.63 in the holdout UK Biobank dataset (95% CI 1.60-1.68). GPS PAD outperformed previously published scores in non-European subgroups in the UK Biobank and achieved superior cross-population portability in external validation cohorts. GPS PAD was associated with incident PAD in the UK Biobank (HR 1.66; 95% CI 1.61-1.71) and achieved improvements in discrimination (ΔC-statistic 0.030) that were nearly equivalent to the additive performances of diabetes (ΔC-statistic 0.029) and smoking (ΔC-statistic 0.034) to the baseline model including age, sex, and 10 principal components of ancestry. Among individuals with prevalent PAD, GPS PAD was associated with incident MALE in the UK Biobank (HR 1.48; 95% CI 1.24-1.77). GPS PAD consistently identified individuals with PAD at high MALE-risk in the Mass General Brigham Biobank (HR 1.34; 95% CI 1.12-1.60), and AoU (HR 1.33; 95% CI 1.12-1.58). Conclusions: These findings contribute a polygenic score for PAD that predicts disease and MALE with cross-cohort performance and transferability to diverse ancestries. Integrated polygenic scores may be used to stratify PAD risk and target more aggressive management, lifestyle modification, and surveillance efforts.

Abstract 4127792: Age-stratified Monogenic and Polygenic Contributions for Atrial Fibrillation in the All of Us Research Program

Background: Rare and common genetic variation both contribute to atrial fibrillation (AF), but the effects may not be uniform across ages. The All of Us Research Program contains whole-genome sequencing of 100,574 adult participants with electronic health records. Hypothesis: Rare genetic variants are associated with early-onset AF; common genetic variants are associated with late-onset AF. Aims: To quantify the diagnostic yield provided by rare and common genetic variation in AF across age. Methods: The clinical, monogenic, and polygenic contributions to AF were assessed in a cross-sectional group of participants in the NIH-funded All of Us Research Program, stratified by three age groups: <45 years (n=22,290), 45-60 years (n= 26,805), >60 years (n=51,659). Variant discovery was performed for pathogenic or likely pathogenic (P/LP) variants among 145 candidate cardiac genes with dominant inheritance for cardiomyopathy and arrhythmia. A previously established polygenic risk score (PRS) was calculated. After adjusting for known clinical factors, multivariable analysis quantified the associations between monogenic and polygenic factors versus AF status in each age group. Results: There were 100,574 participants: 7,811 with AF and 92,763 without AF. The presence of ≥1 P/LP variant(s) were associated with AF across all age groups, with stronger associations in the age <45 group (OR 2.3 [95% CI 1.4-4.1]) compared older groups (OR 1.5 [95% CI 1.1-2.1] in age 45-60, and OR 1.4 [95% CI 1.2-1.7] in age >60). In contrast, PRS was not associated with AF in participants <45 years (OR 1.02 [95% CI 0.88-1.17]) but was associated with AF in the older groups: OR 1.54 [95% CI 1.31-1.80] in age 45-60 and OR 1.49 [95% CI 1.40-1.59] for age >60. Overall, clinical risk factors are highly associated with AF (AUC 0.84 [0.83-0.85]); adding polygenic and monogenic factors contribute only marginal improvement (AUC 0.86 [0.85-0.87]). Conclusion: In this cross-sectional dataset, monogenic and polygenic factors have opposite associations with AF. Monogenic variants are associated with early-onset AF, whereas PRS has no association with AF at younger ages. Clinical risk factors continue to explain the majority of the variation in AF status.

Abstract 4124655: Multi-trait analysis identifies 307 genomic loci for thoracic aortic diameter and disease

Introduction: Thoracic aortic aneurysm and dissection (TAAD) is a morbid cardiovascular disease with 21 known common-variant genomic risk loci. Aneurysm represents an extreme of diameter, so we sought to jointly analyze TAAD and aortic diameter to enhance discovery power. Methods: Genome-wide association studies (GWAS) were performed for UK Biobank MRI-derived aortic diameter measured at the aortic root (N=62,919), ascending aorta (N=61,156), and descending aorta (N=62,412). Separately, TAAD GWAS summary statistics from UKB (1,076 cases, 416,263 controls), FinnGen (3,880 cases, 381,977 controls), and the Million Veteran Program (8,626 cases, 453,043 controls) were meta-analyzed. Genetic correlation was assessed with ldsc . Multi-trait analysis was then performed for the aortic measurements and TAAD with MTAG . A 1.1-million-variant polygenic risk score (PRS) was produced and tested in the All of Us biobank. Results: The genetic correlation between ascending aortic diameter and TAAD was 0.83 (P=6.6E-129). The TAAD meta-analysis had an effective sample size of 53,516, augmented to 147,377 after multi-trait analysis. We identified 59 risk loci for TAAD (189 in the multi-trait analysis). These loci included one on the X-chromosome near MIR222HG , previously reported to be regulated by angiopoietin-2; others were near TGF-β-superfamily members ( FBN1 , FBN2 , GDF6 , GDF7 , LTBP4 , TGFB2 , SMAD3 ); genes involved in vasoconstriction ( ADRA1D , ADRB1 , ANGPT1 , EDN1 , EDN2 , EDNRA , PDE3A ); and cell cycle regulators ( CCND2 , CCNE1 , CDC27 , CDK6 , CDKN1A , CDKN1B , CENPW , DSCC1 , MAD2L1 , TP53 ). 307 distinct loci were significantly associated with at least one of the four traits; all but 7 TAAD loci were genome-wide significant for one of the three aortic measurements, and all loci had at least sub-significant signal in those traits. Gene-set analyses highlighted contributions from fibroblasts, pericytes, and vascular smooth muscle cells. In All of Us , a one standard deviation (SD) greater PRS was associated with a hazard ratio (HR) of 1.88 for TAAD (P=1.2E-78) and 1.62 for thoracic aortic dissection (P=5.7E-05). Individuals in the top 5% for the PRS had HR 3.77 for TAAD (P=2.7E-48) and HR 2.87 for thoracic aortic dissection (P=3.3E-03). Discussion: The multi-trait analysis yielded an eight-fold expansion of loci for TAAD risk. The findings underscore that sporadic TAAD is a complex, common disease—intricately linked with the processes associated with normal variation in aortic diameter.

Abstract 4114417: Polygenic Risk in Heart Failure – Evaluating the Ability of a Polygenic Risk Score to Identify Risk of Future Heart Failure Events in 60,000 Patients from 6 TIMI Randomized Trials

Background: The ability of polygenic risk scores (PRS) to predict heart failure (HF) beyond classic risk factors is unclear. Aims: To evaluate the risk of a HF PRS with incident hospitalization for HF (HHF) in patients with established cardiovascular disease. Methods: We analyzed a 59 SNP HF PRS (HERMES collaboration, Henry et al) in genotyped patients from six multinational TIMI randomized controlled trials (DECLARE, ENGAGE AF, FOURIER, PEGASUS, SAVOR, SOLID). Patients were stratified to low (quintile, Q1), intermediate (Q2-Q4), or high (Q5) genetic risk. We investigated the association of the HF PRS with incident HHF (median of 2.5 years) i) stratified by prior HF, and ii) across NTproBNP concentrations at baseline, using Cox-proportional hazard models. Results are reported as HR [95% CI], per 1-SD or with Q1 as a reference, and adjusted for ancestry (principal components 1-5), age, sex, trial, BMI, smoking, systolic BP, prior CAD, DM, AF, and eGFR. Results: In 59,906 pts (median age 66 years; 71% male; 26% prior HF), the HF PRS was associated with incident HHF (HR adj per 1-SD 1.08 [1.03, 1.13], p<0.001). Compared to low genetic risk pts, the HR adj for HHF was 1.12 [1.00, 1.26] in intermediate and 1.17 [1.02, 1.35] in high-risk pts. There were significant interactions between the HF PRS and prior HF (P int 0.001), and NTproBNP (P int 0.038), respectively. In pts without prior HF (Fig A), compared to low-risk pts, the risk was increased in high- (HR adj 1.56 [1.25, 1.94], p<0.001) and intermediate-risk pts (HR adj 1.27 [1.05, 1.54], p=0.013). In contrast, there was no significant association in pts with known HF at baseline ( Fig B ). Similarly, the HF PRS was significantly associated with HHF in pts with low but not elevated NTproBNP at baseline ( Fig C ). Conclusion: Our study confirms the role of polygenic risk for HF and demonstrates that a 59 SNP PRS can identify pts at increased risk of future HF events beyond traditional clinical factors, specifically in patients without established HF or elevated NTproBNP. These findings suggest a potential application of a HF PRS for screening and early identification of pts at increased risk for HF.

Abstract 4113263: Serum Metabolomics Adds Predictive Value to Clinical and Polygenic Risk Scores for Atrial Fibrillation: A UK Biobank study of >240,000 Patients

Background: Atrial fibrillation (AF) poses substantial morbidity and mortality burdens globally. It is critical to identify individuals at high-risk for AF to implement both lifestyle modifications and close monitoring for initiation of oral anticoagulants. Currently, risk prediction in AF relies on clinical scores such as CHARGE-AF. Small molecule metabolites are thought to play a role in AF pathogenesis, but the additional predictive value of metabolomics over clinical and genetic risk factors is unknown. Using the UK Biobank (UKBB), we sought to perform this evaluation. Methods: We included all UKBB participants with complete 1 H-NMR metabolite measurements. We excluded participants with prior history of AF or with incomplete AF Polygenic Risk Score (AF-PRS) or CHARGE-AF components. The final cohort included 240,628 patients. In-patient records, primary care notes, and death registries were queried to identify patients who had developed incident AF within 5 years of enrollment. The cohort was divided into a 80/20:Train/Test split. We compared the performance of an Elastic-Net regularized Cox Proportional Hazards (EN-CPH) model trained on CHARGE-AF, AF-PRS, and the full 170 metabolite panel to a CPH model trained only on CHARGE-AF and AF-PRS. Results: Within 5 years, 4,174 (1.7%) patients developed AF. Compared to the control population, the incident AF cases were more likely to be older (62 vs. 56, p<0.001), male (65% vs 45%, p<0.001), and have higher CHARGE-AF and AF-PRS scores (p<0.001). After training the EN-CPH model on CHARGE-AF, AF-PRS, and 170 metabolites, the final model included 8 metabolites on top of CHARGE-AF (HR: 1.28) and AF-PRS (HR: 1.11). After adjusting for CHARGE-AF and AF-PRS, creatinine level was associated with increased risk of AF (HR: 1.01) while linoleic acid level (HR: 0.985) was associated with decreased risk. Furthermore, total cholesterol, esterified cholesterol, free cholesterol, cholesteryl esters in IDL, omega-6 fatty acids, and cholesterol in large LDL were associated with HR 0.993-0.999. The EN-CPH model out-performed the CPH model without metabolomics on the test set (AUC 0.786 vs 0.753, p < 0.001). Conclusions: The addition of metabolomics to clinical and genomic risk scores improves prediction of 5-year incident AF within a general population. This study highlights the significance of the metabolome as an independent prognosticator of AF risk. Further study of the mechanisms by which selected metabolites alter AF risk is needed.

Type 2 diabetes pathway-specific polygenic risk scores elucidate heterogeneity in clinical presentation, disease progression and diabetic complications in 18,217 Chinese individuals with type 2 diabetes.

Type 2 diabetes is a complex and heterogeneous disease and the aetiological components underlying the heterogeneity remain unclear in the Chinese and East Asian population. Therefore, we aimed to investigate whether specific pathophysiological pathways drive the clinical heterogeneity in type 2 diabetes. We employed newly developed type 2 diabetes hard-clustering and soft-clustering pathway-specific polygenic risk scores (psPRSs) to characterise individual genetic susceptibility to pathophysiological pathways implicated in type 2 diabetes in 18,217 Chinese patients from Hong Kong. The 'total' type 2 diabetes polygenic risk score (PRS) was summed by genome-wide significant type 2 diabetes signals (n=1289). We examined the associations between psPRSs and cardiometabolic profile, age of onset, two glycaemic deterioration outcomes (clinical requirement of insulin treatment, defined by two consecutive HbA1c values ≥69 mmol/mol [8.5%] more than 3 months apart during treatment with two or more oral glucose-lowering drugs, and insulin initiation), three renal (albuminuria, end-stage renal disease and chronic kidney disease) outcomes and five cardiovascular outcomes. Although most psPRSs and total type 2 diabetes PRS were associated with an earlier and younger onset of type 2 diabetes, the psPRSs showed distinct associations with clinical outcomes. In particular, individuals with normal weight showed higher psPRSs for beta cell dysfunction and lipodystrophy than those who were overweight. The psPRSs for obesity were associated with faster progression to clinical requirement of insulin treatment (adjusted HR [95% CI] 1.09 [1.05, 1.13], p<0.0001), end-stage renal disease (1.10 [1.04, 1.16], p=0.0007) and CVD (1.10 [1.05, 1.16], p<0.0001) while the psPRSs for beta cell dysfunction were associated with reduced incident end-stage renal disease (0.90 [0.85, 0.95], p=0.0001) and heart failure (0.83 [0.73, 0.93], p=0.0011). Major findings remained significant after adjusting for a set of clinical variables. Beta cell dysfunction and lipodystrophy could be the driving pathological pathways in type 2 diabetes in individuals with normal weight. Genetic risks of beta cell dysfunction and obesity represent two major genetic drivers of type 2 diabetes heterogeneity in disease progression and diabetic complications, which are shared across ancestry groups. Type 2 diabetes psPRSs may help inform patient stratification according to aetiology and guide precision diabetes care.

Towards Risk Stratification in Clinical Care for IgA Nephropathy: Genetic Risk Scores

Towards Risk Stratification in Clinical Care for IgA Nephropathy: Genetic Risk Scores

Abstract 4143289: Contemporary and Genomic Cardiovascular Risk Factors Have Explanatory Power Similar to Traditional Risk Factors for Incident Myocardial Infarction

Background: Modifiable cardiovascular risk factors (CRF) have been proposed to be responsible for 80-90% of the risk for incident coronary artery disease (CAD). However, these studies were conducted prior to the era of preventive medications, novel biomarkers, and genetic risk scores. The relative contributions of traditional and contemporary CRF in light of secular trends in worsening cardiometabolic health globally have not been assessed. Hypothesis: Including genetic risk scores and contemporary biomarkers will enhance discrimination and explainability of myocardial infarction (MI) incidence prediction. Methods: The UKBiobank was used to identify traditional CRF (hypertension, diabetes, dyslipidemia, smoking, waist-to-hip ratio (WHR), diet, exercise, alcohol intake and socioeconomic deprivation), and contemporary/genetic CRF (lipoprotein(a), high-sensitivity C-reactive protein [hsCRP], familial hypercholesterolemia [FH] variants, and polygenic risk score for CAD [PRS CAD ]). Incident MI was defined as first-time MI diagnosis or coronary revascularization. Base model discrimination was assessed using C-statistics from Cox proportional hazards models. Percent contribution of each risk factor was calculated by explanatory power lost via Nagelkerke R 2 after removal of the CRF from the full model. Population attributable risks (PAR) were additionally assessed for each model and CRF individually. Results: Over a median [IQR] follow-up of 11.0 [9.6, 12.5] years, 17409/299707 (5.8%) of participants developed incident CAD. C-statistics sequentially increased from base model to traditional CRF to contemporary/genetic CRF model with PAR of 84.3% (95% CI 82.4%-86.5%) ( Table 1 ). Among CRFs, hypertension (C 0.74, R 2 loss 15.2%, PAR 32.5%) and PRS CAD (C 0.72, R 2 loss 12.4%, PAR 38.4%) most strongly explained MI incidence by all 3 indices. Based on discriminability, ApoB:ApoA1 ratio (C 0.71, R 2 loss 3.4%), presence of diabetes (C 0.71, R 2 loss 2.2%), and log(hsCRP) (C 0.71, R 2 loss 1.82%) were subsequently prioritized. PAR analyses included prevalence in prioritization where WHR, presence of diabetes, and log(lipoprotein(a)) levels rose higher. Conclusions: The addition of genetic risk factors and contemporary biomarkers to explanatory models for CAD shows previously underappreciated importance of contemporary CRFs such as PRS, hsCRP, and lipoprotein(a) alongside traditional CRFs such as hypertension, dyslipidemia and presence of diabetes.

Abstract 4138564: Fitbit-Measured Physical Activity Is Inversely Associated With Incident Atrial Fibrillation Among All Of Us Participants

Background: Studies show that moderate to vigorous physical activity (MVPA) reduces atrial fibrillation (AF) risk. However, these studies used subjective questionnaires or short-term (&lt;1 week) accelerometry assessments. Objective: This study investigated the link between MVPA levels and AF incidence using long-term (1-year) accelerometry data from the NIH All of Us Fitbit database. Methods: The study included All of Us participants who agreed to share electronic health record (EHR) and Fitbit data between January 1, 2016 to July 1, 2022. Among participants without an existing diagnosis of AF, the average weekly minutes of MVPA over the first year of Fitbit use was computed. Incident AF events were identified using validated ICD-10 codes in a five-year follow-up period. The risk of incident AF was evaluated using Cox proportional hazards regression models, for both MVPA as a continuous and a categorical variable (&lt;30, 30-150,151-300, and &gt;300 minutes MVPA per week; aligned with WHO guidelines). Covariates included age, sex, race/ethnicity, BMI, smoking status, and diagnoses of hypertension and diabetes. In a subset of participants with whole genome data, we evaluated if the association between MVPA and AF incidence varied by a participant’s AF genetic risk score (GRS) Results: 6086 participants (mean age 50 years, 71% female, 83% White) had EHR and at least 1 year of Fitbit data. Over the 1-year MVPA assessment period, participants had 41±12 weeks of complete data and a median 202 [241] minutes of MVPA per week. 55 participants (0.9%) experienced incident AF in the five-year follow-up period. In adjusted analysis, every additional hour of MVPA was associated with an 11% reduced risk of AF (HR = 0.89 [0.81,0.98], p = 0.02). There was a step-wise reduction in risk of AF by MVPA category with &lt;30 min/week as the reference: 30-150 (HR = 0.62 [0.24,1.6], p = 0.31), 151-300 (HR = 0.40 [0.15,1.07], p = 0.07), and &gt;300 (HR = 0.35 [0.13,0.95], p = 0.04) ( Figure 1 ). These associations also held after adjusting for GRS. Conclusion: Higher amounts of objectively measured MVPA were inversely associated with risk of incident AF, independent of clinical and genetic risk factors.

Abstract 4135624: A Hypertrophic Cardiomyopathy Polygenic Score Modifies Penetrance of Pathogenic Hypertrophic and Dilated Cardiomyopathy Variants in Opposite Directions

Background: Hypertrophic (HCM) and dilated (DCM) cardiomyopathy have been postulated to represent morphologically opposing cardiac disease traits, with genetic pathways implicated in HCM demonstrating inverse relationships with DCM. The degree to which polygenic risk for HCM protects from DCM in individuals at genetic risk of DCM remains unexplored. Research Question: Is polygenic risk for HCM protective against developing DCM in carriers of titin truncating variants? Methods: Associations between polygenic risk for HCM, monogenic variants, and cardiomyopathy outcomes were modeled among Penn Medicine BioBank (PMBB) participants genetically similar to European reference populations using logistic regression. Polygenic risk for HCM was determined using a previously published HCM polygenic score (PGS; PGS Catalog PGS000739; Harper et al. 2021); scores were PCA adjusted, and the mean and variance were normalized using pgsc_calc. Pathogenic variants in definitive HCM genes were identified using ACMG criteria. Titin truncating variants in high percentage spliced in exons (TTNtv) were considered pathogenic. HCM and DCM outcomes were assigned based on diagnosis codes +/- clinical notes in the electronic health record. Results: Of 26,728 PMBB participants, 205 (0.77%) had HCM and 589 (2.2%) had DCM. 289 had a TTNtv, and 144 had a pathogenic HCM variant. Pathogenic HCM variant status (n=144) conferred a 76-fold increase in HCM risk (p=1.8 e-87). A one standard deviation change in HCM PGS had a significant positive association with HCM (OR 1.6; p=1.5e-9) and negative association with DCM (OR 0.72; p=1.e-12). TTNtv status (n=289) was associated with an increased risk of DCM (OR 5.0; p=1.4e-12). An interaction term between TTNtv status and HCM PGS had a negative association with DCM (OR 0.61; p=4.5e-2). Among the top HCM PGS quintile, there was no statistically significant difference in DCM risk between those with and without TTNtv (Figure 1). Conclusions: These results provide additional evidence that HCM and DCM are influenced by polygenic background. We observe that polygenic risk of HCM additively increases risk of HCM among pathogenic HCM variant carriers and may synergistically protect from DCM among TTNtv carriers.

The NHGRI-EBI GWAS Catalog: standards for reusability, sustainability and diversity.

The NHGRI-EBI GWAS Catalog serves as a vital resource for the genetic research community, providing access to the most comprehensive database of human GWAS results. Currently, it contains close to 7000 publications for >15000 traits, from which more than 625000 lead associations have been curated. Additionally, 85 000 full genome-wide summary statistics datasets-containing association data for all variants in the analysis-are available for downstream analyses such as meta-analysis, fine-mapping, Mendelian randomisation or development of polygenic risk scores. As a centralised repository for GWAS results, the GWAS Catalog sets and implements standards for data submission and harmonisation, and encourages the use of consistent descriptors for traits, samples and methodologies. We share processes and vocabulary with the PGS Catalog, improving interoperability for a growing user group. Here, we describe the latest changes in data content, improvements in our user interface, and the implementation of the GWAS-SSF standard format for summary statistics. We address the challenges of handling the rapid increase in large-scale molecular quantitative trait GWAS and the need for sensitivity in the use of population and cohort descriptors while maintaining data interoperability and reusability.

Variants in LPA are associated with Familial Hypercholesterolaemia: whole genome sequencing analysis in the 100,000 Genomes Project.

Familial Hypercholesterolaemia (FH) is an inherited disease of high LDL-cholesterol (LDL-C) caused by defects in LDLR, APOB, APOE and PCSK9 genes. A pathogenic variant cannot be found in ∼60% of clinical FH patients. Using whole genome sequencing (WGS) we examined genetic determinants of FH. WGS data generated by the 100,000 Genomes Project (100KGP) included 536 FH patients diagnosed using the FH Simon Broome criteria. Rare variants in known FH genes were analysed. Genome-wide association study (GWAS) between 443 FH variant-negative unrelated FH cases and 77,275 control participants of the 100KGP was run using high coverage WGS data. Polygenic risk scores for LDL-C (LDL PRS) and lipoprotein(a) (Lp(a) PRS) were computed. An FH-causing variant was found in 17.4% of FH cases. GWAS identified the LPA gene locus being significantly associated (p<1x10-8). FH variant-negative participants had higher LDL and Lp(a) PRSs in comparison to the controls (p<1.0×10-16 and p<4.09×10-6, respectively). Similar associations were found in the monogenic FH with both LDL and Lp(a) PRSs being higher than in controls (p<4.03×10-4 and p<3.01x10-3, respectively). High LDL PRS was observed in 36.4% of FH variant-negative cases, whereas high Lp(a) PRS in 18.5%, with 7.0% having both high LDL and Lp(a) PRSs. This genome-wide analysis of monogenic and polygenic FH causes confirms a complex and heterogenous architecture of hypercholesterolaemia, with the LPA gene playing a significant role. Both Lp(a) and LDL-C should be measured for precision FH diagnosis. Specific therapies to lower Lp(a) should be targeted to those who will benefit most.

Type 1 diabetes genetic risk contributes to phenotypic presentation in monogenic autoimmune diabetes.

Disease-causing variants in key immune homeostasis genes can lead to monogenic autoimmune diabetes. Some individuals carrying disease-causing variants do not develop autoimmune diabetes, even though they develop other autoimmune disease. We aimed to determine whether type 1 diabetes polygenic risk contributes to phenotypic presentation in monogenic autoimmune diabetes. We used a 67 SNP type 1 diabetes genetic risk score (T1D-GRS) to determine polygenic risk in 62 individuals with monogenic autoimmune diabetes and 180 non-autoimmune neonatal diabetes (NDM) controls. We used population-based controls (n=10,405) and individuals with type 1 diabetes (n=285) as a comparator. Individuals with monogenic autoimmune diabetes had higher T1D-GRSs compared to non-autoimmune NDM (mean 11.3 vs. 9.8; P=1.7×10-5) and controls (mean 10.3; P=7.5×10-6) but were markedly lower than type 1 diabetes (14.9; P= 3.3 × 10-21). These differences were explained by monogenic autoimmune diabetes cases having higher Class II HLA genetic risk, specifically from the DRB1*03:01-DQA1*05:01-DQB1*02:01 haplotype (DR3-DQ2) (P<0.01). In the presence of monogenic autoimmunity, the polygenic class II HLA susceptibility contributes to development of autoimmune diabetes. This suggests a role of class II HLA in targeting the dysregulated immune response towards the beta-cell.

Functional and Practical Insights Into the Genetic Basis of Takayasu Arteritis.

Takayasu arteritis (TAK) is a rare vasculitis characterized by inflammation of large arteries. Although the exact etiology of TAK remains unclear, a genetic predisposition to the disease has been established. Large-scale genetic studies have significantly contributed to the identification of genetic variation associated with immune-mediated diseases. To date, five genome-wide association studies (GWAS) have been performed in TAK, identifying multiple genetic susceptibility loci across the genome. Here, we summarize the major findings from GWAS in TAK and provide an in silico functional evaluation of the associated loci (P < 5 × 10-8) and variants in high linkage disequilibrium with them (r2 > 0.8). By exploring gene expression and chromatin interaction data, we identified candidate causal genes in TAK, some of them with well-known functional implications. The analysis of transcription factor motifs within TAK-associated loci revealed enrichment of the STAT and RUNX families, both characterized by their role in immune functions and inflammatory responses. The enrichment in biological processes in susceptibility loci confirmed the involvement of specific immune-related pathways in TAK. Further, we devised and calculated a cumulative genetic risk score for TAK and confirmed differences in genetic risk for the disease among ancestries. Finally, we provide a practical guide to communicate genetic information for TAK to patients and families.