External Validation of Two 10-Year Stroke Risk Prediction Models Using Korean Genome and Epidemiology Study Data.

Area of Science:

• Epidemiology
• Cardiovascular Research
• Biostatistics

Background:

• Stroke remains a leading cause of disability and mortality worldwide.
• Accurate 10-year stroke risk prediction is crucial for preventative strategies.
• Existing prediction models require external validation in diverse populations.

Purpose of the Study:

• To externally validate two 10-year stroke risk prediction models (Model 1: Korean, Model 2: Chinese).
• To assess the performance of these models using community-based cohort data.
• To determine the effectiveness of simplified models with fewer risk factors.

Main Methods:

• Utilized community-based cohort data for 8432 participants (Model 1) and 8915 participants (Model 2).
• Calculated 10-year stroke risk using established model equations.
• Assessed model performance via the area under the receiver operating characteristic curve (AUC).

Main Results:

• Age, blood pressure, and diabetes mellitus were common significant stroke risk factors.
• Model 1 AUCs: 0.72 (men), 0.68 (women); Model 2 AUCs: 0.70 (men), 0.66 (women).
• Simplified Model 2 showed reduced performance (0.68 men, 0.63 women); Model 1 performance was largely unchanged.

Conclusions:

• Both validated stroke risk prediction models demonstrated insufficient performance in this cohort.
• Simplified models did not consistently maintain predictive accuracy.
• Future stroke prediction models should be developed incorporating significant risk factors specific to the target population.