We studied the value of dipyridamole-echocardiography test in comparison with clinical, resting electrocardiogram and echocardiogram variables in predicting cardiac events occurring in 539 consecutive patients referred for dipyridamole-echocardiography test from 1984 to 1987. There were 118 cardiac events: 11 cardiac deaths, 12 nonfatal myocardial infarctions, and 95 coronary revascularization (bypass or angioplasty) procedures. A Cox survival analysis identified echocardiographic positivity after dipyridamole administration as the best predictor of cardiac events (relative risk ratio, 2.7). The next most powerful predictor was angina after dipyridamole administration (relative risk ratio, 1.9). Cardiac events occurred in 14 (6%) of 253 patients with normal high-dose dipyridamole echocardiographic test results, in 21 (26%) of 82 patients with high-dose dipyridamole echocardiographic positivity (0.84 mg/kg during 10 minutes), and in 83 (41%) of 204 patients with low-dose dipyridamole echocardiographic positivity (0.56 mg/kg during 4 minutes) (p less than 0.0001). In a subset of 341 patients, exercise electrocardiography stress test and coronary angiography were also available. A Cox survival analysis again identified echocardiographic positivity after dipyridamole as the best predictor of cardiac events (relative risk ratio, 1.9) followed by a pathologic coronary arteriography (relative risk ratio, 1.2). We conclude that the presence and timing of a transient dyssynergy during dipyridamole stress are useful predictors of subsequent cardiac events.