Development And Evaluation Of A Polygenic Risk Score For Stroke In Large Chinese Prospective Cohorts

Background and objectiveStroke is the leading cause of death and disease burden in China and worldwide.Early identification of high-risk individuals and advocation of healthy lifestyles or drug interventions for major stroke risk factors(hypertension,diabetes,dyslipidemia,etc.)are essential for primary prevention of stroke.Stroke is a complex disorder caused by both genetic and environmental factors.Genome-wide association studies(GWAS)have identified 42 genetic loci associated with stroke,and hundreds of genetic loci associated with a range of stroke-related traits(blood pressure,diabetes,obesity,etc.).The integration of these genetic variants into polygenic risk score(PRS)would facilitate early risk prediction and precise prevention for stroke.However,the prediction accuracy of PRS for stroke needs to be improved.In addition,almost all existing PRS for stroke were developed in European-descent populations.Their applicability to Chinese population is questionable due to marked differences in environmental risk factors and genetic background across populations of different ancestries.We aimed to generate a PRS for stroke in Chinese population and evaluate its potential in clinical utility for risk prediction and primary prevention of stroke in large prospective population-based cohorts.Subjects and methodsWe used a training set of 5366 individuals(2872 stroke cases and 2494 controls)to develop the PRS and externally validated its association with incident stroke in large prospective cohorts comprising 41,006 individuals.We selected 588 single-nucleotide polymorphisms(SNPs)that were of genome-wide significance in association with stroke and a range of stroke-related traits.Participants from both the training and validation set were genotyped using multiplex polymerase chain reaction(PCR)targeted amplicon sequencing technology.In the training set,we generated 16 candidate PRSs for stroke and stroke-related traits,using different linkage disequilibrium(LD)metric r2 and significance thresholds based on trait-specific summary statistics from East Asian ancestry.The score with the largest odds ratio(OR)(per standard deviation increase of PRS)for stroke estimated from a logistic regression model was selected as the optimal trait-specific PRS.Finally,the meta-polygenic risk score(metaPRS)of stroke was generated by integrating these trait-specific PRSs.Validation population were from three cohorts of the China-PAR(Prediction for Atherosclerotic cardiovascular disease Risk in China)project.At the baseline survey,participants were interviewed and underwent physical examinations and laboratory tests.A series of demographic characteristics,lifestyle factors and cardiometabolic conditions were collected by well-trained investigators according to standard protocols.We defined 5 major stroke clinical risk factors at baseline,including hypertension,diabetes,family history of stroke,obesity,and dyslipidemia.Seven cardiovascular health(CVH)metrics at baseline were also defined based on previous studies comprising blood pressure,total cholesterol,fasting blood glucose,body mass index(BMI),smoking,physical activity,and diet.We collected information on incident stroke cases by interviewing participants or their proxies during the follow-up period,and further obtained medical records and death certificates for verification.We defined incident stroke as a confirmed diagnosis of first-ever fatal or nonfatal stroke event during follow-up.Subtypes of stroke were classified as ischemic stroke and hemorrhagic stroke.The polygenic risk of cohort participants was categorized into low(lowest quintile of metaPRS),intermediate(2-4 quintiles of metaPRS),and high(highest quintile of metaPRS).CVH profiles were classified as favorable(6 to 7 ideal CVH metrics),intermediate(4 to 5 ideal CVH metrics),and unfavorable(0 to 3 ideal CVH metrics).Age-as-time-scale and stratified Cox proportional hazards regression models with strata defined as cohorts were used to estimate the hazard ratios(HRs)and lifetime risk(cumulative incidence of stroke by age of 80 years)for incident stroke associated with polygenic risk,major clinical risk factors and CVH categories.ResultsTrait-specific PRSs were derived and combined to construct the metaPRS in the training set.The mean(SD)age of stroke cases was 66.6(9.8)years,which was similar to that of controls of 66.1(10.3)years.We generated the stroke PRS using summary statistics derived from different ancestries and further compared their association magnitude with stroke in the training set.We observed that the ORs decayed markedly when using those from European population compared with those from East Asian population.Also using effect sizes from East Asian population,we generated other trait-specific PRSs.Finally,12 trait-specific PRSs were included in the calculation of metaPRS after adjusting for correlations by elastic-net logistic regression model.These procedures resulted in 534 SNPs included in the metaPRS finally.We then evaluated the performance of the metaPRS alone or in combination with major clinical risk factors in stroke prediction and risk stratification in prospective cohorts.The mean(SD)age of participants in the validation cohorts at baseline was 51.9(10.6)years old,and 43.1%were males.During a total of 367,750 person-years of follow-up(mean follow-up 9.0 years),1227 participants developed stroke before age of 80 years,including 769 ischemic stroke and 355 hemorrhagic stroke.The HR for incident stroke was 1.28(95%CI:1.21-1.36)per standard deviation increase in metaPRS,which showed similar predictive effects for both ischemic stroke(HR:1.29,95%CI:1.20-1.39)and hemorrhagic stroke(HR:1.30,95%CI:1.17-1.45).Individuals with high polygenic risk had an about 2-fold higher risk of incident stroke compared with those with low polygenic risk(HR:1.99,95%CI:1.66-2.38),with the lifetime risk of stroke being 25.2%(95%CI:22.5%-27.7%)and 13.6%(95%CI:11.6%-15.5%),respectively.Polygenic risk and major clinical risk factors including family history were independent predictors of incident stroke.Individuals with both high polygenic risk and family history displayed relatively high lifetime risk of stroke(41.1%,95%CI:31.4%-49.5%).and those with both high polygenic risk and≥3 clinical risk factors showed lifetime risk of 45.5%(95%CI:36.0%-53.6%).We finally investigated the benefits of adherence to ideal CVH individually or jointly across different polygenic risk categories.Among the seven CVH metrics,adherence to ideal blood pressure displayed the strongest protective effects with the HRs(95%CIs)of 0.39(0.33-0.46),which remained essentially unchanged across the three polygenic risk categories.There were significant interactions between polygenic risk and ideal fasting blood glucose or total cholesterol(P for interaction=0.003 and 0.002.respectively).Individuals with high polygenic risk(HR:0.62,95%CI:0.50-0.77)derived greater relative risk reductions from ideal fasting blood glucose than those with low polygenic risk(HR:0.89,95%CI:0.64-1.24).We also observed similar results for ideal total cholesterol(HR:0.62,95%CI:0.50-0.78 vs.HR:0.93,95%CI:0.67-1.29).Moreover,we noted significant increasing trends in absolute risk reductions(ARRs)along with the increase of polygenic risk categories for the above two ideal cardiometabolic measurements(both P for trend<0.007).The benefits in terms of ARRs from ideal fasting blood glucose and total cholesterol across the low,intermediate,and high polygenic risk categories were 1.6%,6.5%,11.6%and 0.9%,2.1%,11.2%,respectively.When assessing the benefits of ideal CVH jointly,participants with favorable CVH(6-7 ideal metrics)in comparison with unfavorable CVH profiles showed 64%,59%,and 66%lower risk of stroke in low,intermediate,and high polygenic risk categories,respectively.Individuals with high polygenic risk(from 34.6%to 13.2%,ARR=21.4%)received a 1.81-fold greater absolute benefit from favorable CVH than those with low polygenic risk(from 19.4%to 7.6%,ARR=11.8%),substantially mitigating the increased risk in the high polygenic risk category to the level of the low polygenic risk category.ConclusionsIn summary,we developed a metaPRS for stroke with good performance in stroke prediction and risk stratification in Chinese population.Risk prediction by metaPRS for stroke was independent of major clinical risk factors including family history.The metaPRS alone or in combination with major clinical risk factors could effectively identify individuals at high risk for stroke.Individuals with high polygenic risk derived greater benefits from adherence to ideal fasting blood glucose and total cholesterol.Maintaining favorable cardiovascular health(CVH)profile could substantially mitigate the increased risk conferred by high polygenic risk to the level of the low polygenic risk.The practical applications of polygenic risk information for guiding primary prevention in the clinical setting remain to be defined in further studies.