Mitral valve prolapse is sometimes associated with chest pain, but this symptom may also be caused by coexisting coronary disease. The accuracy of exercise echocardiography in diagnosing coronary disease in these patients and the most cost-efficient diagnostic approach are unclear. We studied 96 patients (aged 59 ± 12 years; 70 men) with mitral valve prolapse who underwent exercise electrocardiography, exercise echocardiography, and coronary angiography. The accuracies of seven diagnostic strategies based on the current and expected use of exercise electrocardiography and exercise echocardiography in patients with mitral valve prolapse were examined, with the costs calculated based on Medicare reimbursement. Thirteen (13.5%) patients had significant coronary artery disease. The sensitivity and specificity of exercise electrocardiography in the 71 patients with interpretable electrocardiograms were 50% and 72%, respectively. For the 73 patients who achieved >85% of maximal heart rate, 52 had interpretable exercise electrocardiograms (sensitivity 50%; specificity 69%). Exercise echocardiography had a sensitivity of 69% and a specificity of 98% in the 96 patients and a sensitivity of 82% and a specificity of 96% in patients who achieved >85% of maximal heart rate. Approaches adopting Bayes' theorem and restricting further investigation to patients with at least intermediate pretest probability of coronary disease were the least costly. When combined with exercise echocardiography as the initial test, clinical stratification was associated with a false-negative rate of 21%. The utility of exercise electrocardiography is limited by the high prevalence of resting electrocardiographic abnormalities and suboptimal sensitivity and specificity. The best balance of cost and diagnostic accuracy is to perform exercise echocardiography in patients with at least intermediate probability of coronary artery disease. (Am Heart J 1997;134:1052-7.)