Background and Objective: This article aims to address two issues: first, whether risk scores derived from Caucasian cohorts can accurately predict coronary heart disease (CHD) in a Chinese population; second, derivation and validation of a new risk prediction score for cardiovascular disease (CVD) from a Chinese cohort. Methods: A cohort of 4,400 male steelworkers (18-74 years old) recruited between 1974-1980 was followed up for an average 13.5 years. A Cox model was used to predict 10 year risk of CHD, ischemic, and hemorrhagic stroke based on blood pressure, age, total cholesterol, body mass index, and smoking status. Results: The prediction rule from the European Recommendations greatly overestimates the risk of CHD in Orientals. The new models had good predictive ability, with the area under receiver operating characteristic curves (AUCs area) in the derivation set being 0.76, 0.72, and 0.82 for CHD, ischemic, and hemorrhagic stroke separately; AUCs in the validation set were equally good, at 0.76, 0.78, and 0.82, respectively. Conclusion: The risk stratification rules derived from Caucasian cohorts overestimates the CHD risk in Orientals, and we derive a specific risk stratification rule for CHD, ischemic and hemorrhagic stroke for Chinese men.
Journal of Clinical Epidemiology Vol. 58, no. 9, p. 951-958