What is the clinical evidence from this study?

Study design: Cross-Sectional. Population: Coronary heart disease (n=16997). Intervention: Support Vector Machine (SVM) model vs. Other machine learning models (DT, AdaBoost, XGBoost, NN, MLP, NB, GLM). Primary outcome: Coronary heart disease prediction accuracy (AUC 0.909, 95% CI 0.828-0.841).

What does this research mean for the field?

A Support Vector Machine model using demographic and lifestyle data predicts coronary heart disease with 83.4% accuracy, while Mendelian randomization confirms causal links for body mass index, total cholesterol, insomnia, diastolic blood pressure, and smoking. Novelty: ClaimNovelty.CONFIRMATORY. Consensus alignment: ConsensusAlignment.NEUTRAL.

October 26, 2025International Journal of Cardiology Cardiovascular Risk and Prevention1 citationsOpen Access

Machine learning and Mendelian randomization identify key lifestyle factors in coronary heart disease: A NHANES-Based study

YCYang CuiChina Agricultural University YZYonghong ZhangSouth China Normal University LZLang ZengTongji Hospital

Key Result

A Support Vector Machine model using lifestyle data predicted coronary heart disease with 83.4% accuracy (AUC 0.909), and Mendelian randomization confirmed causal links for five lifestyle factors.

Study Design

Type

Cross-Sectional (n=16,997)

Structured PICO

Does the integration of machine learning and Mendelian randomization accurately predict coronary heart disease and identify causal lifestyle factors in the NHANES population?

Population

29,400 individuals from the National Health and Nutrition Examination Survey (NHANES) database (2013–2018)

Intervention

Machine learning models (including Support Vector Machine) and Mendelian randomization

Comparator

Comparison among 8 machine learning models (SVM, NN, NB, XGBoost, MLP, GLM, AdaBoost, DT)

Outcome

Predictive performance for coronary heart disease (accuracy, AUC) and causal association of lifestyle factors

An integrated approach using an SVM machine learning model and Mendelian randomization accurately predicts coronary heart disease risk and confirms causal links for BMI, cholesterol intake, sleep duration, diastolic blood pressure, and smoking.

Main Result

Effect estimate: AUC 0.909 (95% CI 0.828-0.841)

Limitations

Diagnosis of CHD was based on self-reported questionnaire data, which may introduce information or recall bias.
Missing metrics for certain participants required statistical imputation, leaving potential residual bias.
Results are based on the US NHANES database and GWAS gene repository, which may not be directly generalizable to other global populations.

Abstract

This study aims to bridge the gap between predictive modeling and causal inference by utilizing lifestyle data from the National Health and Nutrition Examination Survey (NHANES) database to compare the predictive performance of multiple machine learning models for coronary heart disease (CHD). By incorporating Mendelian randomization, the study seeks to validate and identify the lifestyle variables with both predictive power and causal impact on CHD. We extracted variables related to demographic characteristics and lifestyle from the NHANES database (2013–2018; n= 29,400). Recursive feature elimination (RFE) was employed to rank variable importance and determine the optimal feature subset. Subsequently, eight machine learning models-including Support Vector Machine (SVM), Neural Network (NN), Naive Bayes (NB), Extreme Gradient Boosting (XGBoost), Multilayer Perceptron (MLP), Generalized Linear Model (GLM), Adaptive Boosting (AdaBoost), and Decision Tree (DT)-were developed for CHD prediction. Model performance was evaluated using metrics such as accuracy, precision, sensitivity, specificity, recall, F1-score, and the Receiver Operating Characteristic (ROC) curve, with variable contributions visualized using Shapley Additive Explanations (SHAP). Additionally, Mendelian randomization (MR) was applied to distinguish associative from causal relationships, validating top predictors via Genome-Wide Association Study (GWAS)-derived genetic instruments. RFE identified age, sex, fasting blood glucose, body mass index (BMI), total cholesterol (TC) intake, sleep duration, diastolic blood pressure, and smoking as the most significant predictors of CHD. Among the models, SVM outperformed DT, AdaBoost, XGBoost, NN, MLP, NB, and GLM. The SVM model achieved the highest performance with an accuracy of 83.4% and an AUC value of 0.909, demonstrating clinically actionable predictive power. MR confirmed causal associations for five variables: BMI (OR: 1.01, P 0.05). The SVM machine learning model, based on NHANES data, enables faster and more efficient prediction of CHD. The study identified age, sex, BMI, TC intake, sleep duration, diastolic blood pressure, and smoking as the lifestyle variables with the greatest impact on CHD. This dual approach advances precision prevention by combining predictive accuracy with genetic evidence. • This study integrates machine learning and Mendelian randomization to improve coronary heart disease (CHD) risk prediction and causal inference using NHANES lifestyle data. • Recursive feature elimination identified key predictors of CHD, including age, sex, BMI, total cholesterol intake, sleep duration, diastolic blood pressure, and smoking. • Among eight evaluated machine learning models, Support Vector Machine (SVM) demonstrated the highest predictive accuracy (83.4%) and AUC (0.909). • Mendelian randomization confirmed causal effects for five lifestyle factors, strengthening the evidence beyond association for BMI, cholesterol intake, sleep duration, diastolic blood pressure, and smoking. • This integrated approach enhances precision prevention of CHD by combining robust predictive modeling with genetic validation of causal risk factors.

问 AI

Bookmark

View Full Paper