Exercise treadmill testing is frequently performed to screen for coronary artery disease (CAD) in asymptomatic individuals; however, its clinical value is unclear. We examined a consecutive cohort of asymptomatic adults undergoing exercise treadmill testing at the Cleveland Clinic Foundation between September 1990 and December 1993. End points included (1) identification of subjects with severe CAD and (2) performance of any second diagnostic study within 90 days of the index exercise treadmill test. Screening exercise treadmill testing was performed in 4,334 adults (median age 51, 89% men); only 34% had > or = 1 cardiac risk factor and 15% exhibited an abnormal response to exercise. A second test after treadmill testing was performed in 215 patients (in 110, coronary angiography; in 105, stress thallium scintigraphy, followed by coronary angiography in 16). The strongest predictor of referral for a second test was an ischemic ST-segment response (adjusted odds ratio [OR] 34, 95% confidence intervals [Cl] 24 to 47, p < 0.0001). The only clinical variable independently associated with referral for a second test was female gender (adjusted OR 0.35, 95% CI 0.21 to 0.60, p <0.0001). Of the 126 patients who underwent coronary angiography, severe CAD was identified in only 19 individuals (10.44% of the original cohort, 95% CI 0.26% to 0.62%); coronary artery bypass grafting was performed in 14 of these patients. The estimated cost of exercise treadmill testing to identify 1 case of severe CAD for which surgical revascularization may provide a survival benefit was $39,623. The estimated cost per year of life saved was at least $55,274. Thus, as used in actual practice in 1 center, screening exercise treadmill testing has a low yield and is costly. This is perhaps in part because of the low-risk population that was selected and the failure to incorporate pretest variables, increasing probability of disease into post-test clinical decision making.