An innovative model for predicting coronary heart disease using TyG-index: A machine learning-based cohort study

Abstract

Background Various coronary heart disease (CHD) predictive models have been developed for predicting CHD incidence, but none of them has optimal predictive value. Although these models consider diabetes as an important CHD risk factor, they did not consider insulin resistance or Triglyceride. Methods Two-thousand participants of a community-based Iranian population, aged 20–74 years, were investigated with a mean follow-up of 9.9 years (range: 7.6 to 12.2). The association between TyG-index (a logarithmised combination of fasting blood glucose and triglyceride that demonstrates insulin resistance) and CHD was investigated using multivariate Cox proportional hazard models. Diabetes was substituted for TyG-index in CHD prediction models developed using machine learning. CHD-predicting TyG-index cut-off points were calculated. Results The incidence of CHD was 14.5%.As compared to the lowest quartile of TyG-index, the fourth quartile had a fully adjusted hazard ratio of 2.32 (CI: 1.16–4.68, p-trend 0.04). In order to predict coronary heart disease, TyG-index > 8.42 had the highest negative predictive value. Machine learning models that predicted CHD based on TyG-index performed significantly better than those based on diabetes. TyG-index was not only more important than diabetes in prediction of CHD; it was the most important factor in machine learning models. Conclusion TyG-index can be used in clinical practice and predictive models as a highly valuable index for predicting and preventing CHD.