Key result
High diastolic blood pressure polygenic risk linked to ~7-fold higher hypertension risk vs lowest decile.
Why the study?
Arterial hypertension is a major contributor to cardiovascular morbidity and mortality, but the performance of polygenic risk scores and their integration with questionnaire-based risk factors had not been evaluated in Russia.
Does integrating polygenic risk scores with questionnaire-based risk factors using machine learning improve the prediction of arterial hypertension in inner Eurasian populations?
Observational (n=175,704)
Does integrating polygenic risk scores with questionnaire-based risk factors using machine learning improve the prediction of arterial hypertension in inner Eurasian populations?
Odds Ratio: 6.71 (95% CI 5.58–8.06)
Integrating polygenic risk scores with conventional risk factors using machine learning significantly improves the prediction of arterial hypertension in diverse inner Eurasian populations.
PRS integration may refine hypertension risk stratification in Eurasian cohorts; leaves open clinical utility pending prospective outcome trials.
Introduction: Arterial hypertension is one of the leading contributors to cardiovascular morbidity and mortality worldwide. This study aimed to evaluate the performance of polygenic risk scores (PRS) for hypertension in Russia and to develop predictive models integrating PRS and questionnaire-based risk factors for disease risk assessment. Methods: We analyzed a cohort of 175,704 individuals from multiethnic inner Eurasian populations. Published PRS for systolic blood pressure, diastolic blood pressure, and pulse pressure were evaluated for association with hypertension across different ancestry groups. Predictive models integrating PRS and questionnaire-derived risk factors were developed using multiple machine learning tools, including neural networks. Results: PRSs for systolic and diastolic blood pressure showed marked differences between the top and bottom deciles of the PRS distribution, with odds ratios of 6.20 (95% CI: 5.22-7.36) and 6.71 (95% CI: 5.58-8.06), respectively. The PRS for pulse pressure was also strongly associated with hypertension, with an odds ratio of 3.71 (95% CI: 3.16-4.35). All evaluated PRSs were consistently associated with hypertension across several ancestry groups represented in Russia and neighboring regions, including East Slavic populations (Russians, Belarusians, and Ukrainians), populations of the Volga-Ural region, such as Tatars, and West Asian-related groups represented by Armenians and Hemshins. A neural network model integrating PRSs with questionnaire-based risk factors achieved a test ROC-AUC of 0.8245 (95% CI: 0.8114-0.8362), demonstrating robust discriminatory performance for arterial hypertension. Conclusion: Our study demonstrates that previously published PRSs for systolic blood pressure, diastolic blood pressure, and pulse pressure retain substantial predictive value across diverse inner Eurasian populations and provide complementary information beyond conventional questionnaire-based risk factors. Among the evaluated scores, the systolic blood pressure PRS showed the most robust and consistent transferability, remaining informative even in several genetically diverse and underpowered cohorts.
No takes yet. Share an insight, caveat, or question.
Tsvetkova et al. (2026) conducted an observational in Arterial hypertension (n=175,704). Polygenic risk scores (PRS) for blood pressure vs. Lowest decile of PRS was evaluated on Arterial hypertension (top vs. lowest decile of PRS_DBP) (OR 6.71, 95% CI 5.58-8.06). Polygenic risk scores for diastolic blood pressure were strongly associated with hypertension (OR 6.71 for top vs. bottom decile), and a neural network integrating PRSs with clinical factors achieved a ROC-AUC of 0.8245.
Synapse has enriched 5 closely related papers on similar clinical questions. Consider them for comparative context: