Ryosuke FUJII

BACKGROUND: Although utility of composite trait-specific polygenic risk score (multi-trait PRS) has been examined among European ancestries, few studies investigated among East Asians and incorporated modifiable risk factors. We examined the associations of multi-trait PRS for cardiometabolic factors with cardiovascular disease mortality by integrating nongenetic determinants. METHODS: A total of 14 086 Japanese participants (mean age, 55±9; 55.8% women) of the J-MICC (Japan Multi-Institutional Collaborative Cohort) study were analyzed in this study. We calculated 6 PRSs for cardiometabolic traits (systolic blood pressure, body mass index, triglycerides, low-density lipoprotein cholesterol, estimated glomerular filtration rate, and hemoglobin A1c). Based on these PRSs, we developed multi-trait PRS and considered as a primary exposure. Three nongenetic factors (smoking, alcohol drinking, and educational attainment) from the self-reported questionnaire were also examined. RESULTS: During a median 12.1-year follow-up period, a total of 472 all-cause and 79 cardiovascular disease mortality cases were documented. Compared with 0% to 90% of multi-trait PRSs, an adjusted hazard ratio (HR) among the top 10% of multi-trait PRSs was 1.32 (95% CI, 1.00-1.73) for all-cause death and 2.63 (95% CI, 1.48-4.67) for cardiovascular disease death. Incorporation of educational attainment with multi-trait PRSs showed null associations in those who went beyond high school (HR, 2.07 [95% CI, 0.44-9.66]) even in the top 10% of multi-trait PRS. CONCLUSIONS: Our analysis combining both genetic and nongenetic determinants highlighted that lifestyle factors and educational attainment can slightly reduce an individual's composite genetic risk for cardiovascular disease death.

Although previous polygenic risk score (PRS) studies for cardiovascular disease (CVD) focused on incidence, few studies addressed CVD mortality and quantified risks by environmental exposures in different genetic liability groups. This prospective study aimed to examine the associations of blood pressure PRS with all-cause and CVD mortality and to quantify the attributable risk by modifiable lifestyles across different PRS strata. 9,296 participants in the Japan Multi-Institutional Collaborative Cohort Study without hypertension at baseline were analyzed in this analysis. PRS for systolic blood pressure and diastolic blood pressure (PRSSBP and PRSDBP) were developed using publicly available Biobank Japan GWAS summary statistics. CVD-related mortality was defined by the International Classification of Diseases 10th version (I00-I99). Cox-proportional hazard model was used to examine associations of PRSs and lifestyle variables (smoking, drinking, and dietary sodium intake) with mortality. During a median 12.6-year follow-up period, we observed 273 all-cause and 41 CVD mortality cases. Compared to the middle PRS group (20-80th percentile), adjusted hazard ratios for CVD mortality at the top PRS group ( > 90th percentile) were 3.67 for PRSSBP and 2.92 for PRSDBP. Attributable risks of CVD mortality by modifiable lifestyles were higher in the high PRS group ( > 80th percentile) compared with the low PRS group (0-80th percentile). In summary, blood pressure PRS is associated with CVD mortality in the general Japanese population. Our study implies that integrating PRS with lifestyle could contribute to identify target populations for lifestyle intervention even though improvement of discriminatory ability by PRS alone is limited.

BACKGROUND: Although many polygenic risk scores (PRS) for cardiovascular traits have been developed in European populations, it is an urgent task to construct a PRS and to evaluate its ability in non-European populations. We developed a genome-wide PRS for blood pressure in a Japanese population and examined the associations between this PRS and hypertension prevalence. METHODS: We performed a cross-sectional study in 11 252 Japanese individuals who participated in the J-MICC (Japan Multi-Institutional Collaborative Cohort) study. Using publicly available GWAS summary statistics from Biobank Japan, we developed the PRS in the target data (n=7876). With >30 000 single nucleotide polymorphisms, we evaluated PRS performance in the test data (n=3376). Hypertension was defined as systolic blood pressure of 130 mm Hg or more, or diastolic blood pressure of 85 mm Hg or more, or taking an antihypertensive drug. RESULTS: Compared with the middle PRS quintile, the prevalence of hypertension at the top PRS quintile was higher independently from traditional risk factors (odds ratio, 1.73 [95% CI, 1.32-2.27]). The difference of mean systolic blood pressure and diastolic blood pressure between the middle and the top PRS quintile was 4.55 (95% CI, 2.26-6.85) and 2.32 (95% CI, 0.86-3.78) mm Hg, respectively. Subgroups reflecting combinations of Japanese PRS and modifiable lifestyles and factors (smoking, alcohol intake, sedentary time, and obesity) were associated with the prevalence of hypertension. A European-derived PRS was not associated with hypertension in our participants. CONCLUSIONS: A PRS for blood pressure was significantly associated with hypertension and BP traits in a general Japanese population. Our findings also highlighted the importance of a combination of PRS and risk factors for identifying high-risk subgroups.