The association between coronary heart disease (CAD) risk factors with submaximal exercise performance was examined among 12,866 men at high risk in the Multiple Risk Factor Intervention Trial (MRFIT). Men were selected from a risk score based on serum cholesterol level, diastolic blood pressure and number of cigarettes smoked per day. Multivariate analysis using exercise ST depression as the dependent variable showed age, diastolic blood pressure and serum cholesterol level were significant positive predictors of ST depression and cigarettes per day, body mass index and heart rate at rest were significant negative predictors of ST depression. Similarly, multivariate analysis, using exercise duration as the dependent variable, revealed that age, cholesterol level, body mass index and heart rate at rest were significant negative predictors of exercise duration, whereas cigarettes per day and leisure-time physical activity were significant positive predictors. Some of these relationships with exercise performance are consistent with established epidemiologic CAD risk factor associations and others are not. The MRFIT selection process, which resulted in smokers who were significantly younger and who had significantly lower levels of other CAD risk factors than nonsmokers, was partially responsible.