SUMMARY We evaluated the ability of ST-segment analysis during submaximal exercise tolerance testing (85% predicted age-adjusted heart rate) to diagnose the presence of significant coronary artery stenosis (2 75% cross sectional area narrowing) in a group of 85 men and 92 women with chest pain syndromes and no previously documented myocardial infarctions. Disease prevalence by selective coronary angiography was 36% for men and 33% for women (NS). Predictive value of a positive exercise test (PV(+ET)) as defined by 1 mm ST-segment depression 0.08 second after the J point was significantly higher for men than for women (77% vs 47%, p < 0.05). Predictive value of a negative test (PV(-ET)) was not significantly different for men (81%) and women (78%). Analysis of the 66 men and 66 women not taking digitalis preparations again showed that PV(+ET) was significantly higher for men than for women (90% vs 45%, p < 0.01).Multivariate analysis showed that patients with angiographically significant coronary disease had significantly lower attained heart rates and shorter exercise duration than those without significant stenosis, independent of ST-segment responses. A discriminant function using ST-segment response, attained heart rate and a sex-dependent ST-segment response factor was developed. Duration of exercise was not an independent predictor by our analysis. This function improved the PV(+ET) and PV(-ET) for the total group and for the women; for men, the PV(-ET) improved, while the PV(+ET) fell slightly. This function has not yet been used prospectively.In patients with chest pain and no previously documented myocardial infarction, men have a significantly higher PV( + ET) than women, which cannot be accounted for simply by a difference in disease prevalence (i.e., Bayes' theorem).EXERCISE TESTING has been a commonly used tool for the evaluation of cardiac status for over 20 years. Despite its extensive use as a screening test for asymptomatic patients, a diagnostic test for patients with symptoms of ischemic heart disease, and as a functional test for patients with known cardiac disease, the usefulness, accuracy, and even basic methods for interpretation of the test remain controversial.Proper use of any test involves definition of the patient populations in which the diagnostic yield will be greatest. The use of exercise testing for diagnosis of coronary atherosclerotic heart disease in women with chest pain has also been controversial. Some authors '-5 Hospital from 1971-1977 were reviewed, and all patients evaluated for chest pain who had both coronary angiography and acceptable exercise tests within 6 months were identified. Patients with valvular heart disease, including mitral valve prolapse, congenital heart disease, idiopathic hypertrophic subaortic stenosis, left bundle branch block on the ECG, a history of chest surgery, or a previously documented myocardial infarction were excluded from the study. The diagnosis of a previous myocardial infarction was based on the finding of at least two of the followin...