Comparison of ASCVD Risk Prediction Models in STEMI: Insights From a South Asian Cohort

Area of Science:

• Cardiology
• Public Health
• Epidemiology

Background:

• Cardiovascular (CV) risk prediction models are crucial for primary prevention but often lack generalizability to South Asian populations due to derivation in Western cohorts.
• Existing models may not accurately reflect the CV risk profile of diverse ethnic groups, potentially leading to suboptimal preventive strategies.

Purpose of the Study:

• To compare the risk classification, agreement, and discriminative performance of major atherosclerotic cardiovascular disease (ASCVD) risk prediction models in a South Asian population presenting with first acute myocardial infarction (AMI).

Main Methods:

• Retrospective analysis of 4,975 patients (aged 40-79) with first AMI, extracting pre-event data.
• Calculated 10-year CV risk using Framingham Risk Score (FRS), ASCVD 2013, WHO CVD charts, JBS-3, and PREVENT equations.
• Assessed risk categories (low, intermediate, high) and statistical agreement using correlation, intraclass correlation, and Cohen's kappa.

Main Results:

• Models showed varied high-risk classifications, with PREVENT (19.8%) and FRS (20.2%) classifying the highest proportions.
• PREVENT exhibited the widest risk distribution (0.2%-91%), indicating diverse risk levels within the cohort.
• Poor agreement was observed between PREVENT and other models (e.g., Cohen's kappa with ASCVD = 0.228), despite moderate correlations.

Conclusions:

• Major CV risk algorithms demonstrate significant differences in classifying Indian AMI patients, leading to potential misclassification.
• A substantial majority (80%) of patients were not identified as high risk prior to their event, emphasizing limitations of current models.
• There is a critical need for the development and validation of South Asia-specific risk scores and alternative preventive strategies for cardiovascular disease.