Prediction of coronary heart disease (CHD) is based on multivariable risk equations developed from population-based observational studies in which people without clinical CHD at the initiation of study were examined and followed until their first CHD events. The risk equations from the Framingham Heart Study have been widely used in our clinical practice1-3 and research.4,5 The recent report of the third National Cholesterol Education Program-Adult Treatment Panel (NCEP-ATP) incorporated the Framingham risk equations to predict ten-year absolute CHD risk and to identify certain patients who are at high risk and more likely to benefit from primary prevention with aggressive lipid-lowering treatment.1 In addition to CHD prediction, population-based observational studies also provide the clue to understand how much of CHD can be prevented by modifying major cardiovascular risk factors such as serum cholesterol level, blood pressure level, and current smoking.4,6,7

In this narrative review, we described how CHD prediction works and how it can be improved by including nontraditional cardiovascular risk factors. We also discussed about how likely it is to prevent the majority of CHD.