This study prospectively evaluated 75 consecutive patients (mean age 69 +/- 9 years) undergoing major vascular surgery to test the hypothesis that dobutamine stress echocardiography can be used to predict perioperative cardiac events. A positive test was defined as a new or worsening wall motion abnormality in at least 2 of 18 wall segments. Up to 40 micrograms/kg/min of dobutamine was administered. All readings were done by physicians unaware of the patients' symptoms and electrocardiographic response. In addition, physicians caring for the patients were unaware of the test result. End points of the study were unstable angina with documented electrocardiographic changes, nonfatal myocardial infarction or cardiac death. The perioperative ischemic event rate was 7% (5 of 75 patients). Three patients developed unstable angina and 2 sustained nonfatal myocardial infarctions. All of these patients had positive results on dobutamine stress echocardiography (sensitivity 100%). However, 22 patients who also had positive results on dobutamine stress echocardiography did not have perioperative events (specificity 69%). The corresponding positive predictive value was 19%. None of the 48 patients who had negative results on dobutamine stress echocardiography had events (negative predictive value 100%). In conclusion, dobutamine stress echocardiography can be used to predict perioperative events with great sensitivity, but its positive predictive value in this patient population in low, likely due to the low incidence of perioperative events in patients with known coronary artery disease and the imperfect specificity of dobutamine stress echocardiography in identifying significant coronary stenosis. Dobutamine stress echocardiography is most useful in this setting when negative, because it predicts safety from complications with confidence.