Polygenic risk scores improve stroke risk stratification in Chinese adults: Validation from the Chinese multi-provincial cohort study

Objective: To validate whether incorporating existing polygenic risk scores (PRSs) derived from East Asian or trans-ancestry populations into clinical risk equations improves stroke risk stratification in Chinese adults. Methods: Participants from the Chinese Multi-provincial Cohort study with genotyped data (n = 2931) were included. Four well-established PRSs (i.e., PRS-GBMI, PRS-GIGA, PRS-ChinaPAR, and PRS-MEGA) from either the predominantly Chinese or trans-ancestry populations were constructed and evaluated by assessing their associations with stroke and its subtypes. We tested the incremental predictive capability of the four PRSs for the 10- and 20-year risk of stroke and its subtypes after adding PRSs to recalibrated China-PAR stroke risk equations, based on discrimination, calibration, and reclassification. Results: Over a median follow-up period of 28.2 years, 340 stroke events were recorded. Higher PRSs were generally associated with a higher stroke risk, though only the highest quantile group of PRS-GIGA showed statistical significance (hazard ratio (HR): 1.79, 95% confidence interval (CI): 1.05–3.07). Adding PRS-GIGA to the recalibrated China-PAR stroke risk equations (i.e., the base model) yielded a moderate improvement in 20-year stroke risk, with 17.2% (95% CI: 3.8%–30.6%) more participants correctly categorized into their corresponding risk groups. However, for ischemic stroke, adding PRS-GIGA, PRS-ChinaPAR, and PRS-MEGA to the base model could correctly categorize 18.7%–23.8% more participants into their corresponding 10-year risk groups and 27.8%–32.5% more participants into their corresponding 20-year risk groups. Adding PRSs did not improve prediction for hemorrhagic stroke. Conclusion: Adding existing PRSs, particularly PRS-GIGA, to clinical risk equations can improve all stroke and ischemic stroke risk stratification in Chinese adults.