Dobutamine stress echocardiography has increasingly been used to assess patients for coronary artery disease. Despite the popularity of this test, the optimal dose of dobutamine has not been established. The objective of this study was to assess the accuracy of dobutamine stress echocardiography at various infusion doses and its utility as a predictor of perioperative risk in patients undergoing a noncardiac surgical procedure. One hundred thirteen consecutive patients underwent dobutamine stress echocardiography, subsequent cardiac catheterization and/or a noncardiac surgical procedure. Three patient groups were analyzed on the basis of peak dobutamine infusion rates (17 +/- 4, 29 +/- 2, and 40 +/- 0 micrograms/kg/min, respectively). The three groups were comparable with regard to age, sex, ejection fraction, and severity of coronary artery disease. In group I, the sensitivity and specificity of dobutamine stress echocardiography were 74% and 33%, respectively, with a positive predictive value of 78%. In group II, the sensitivity and specificity improved to 84% and 78%, with a positive predictive value of 89%. In group III, the sensitivity and specificity were 86% and 80%, respectively, with a positive predictive value of 86%. In the noncardiac surgical group there was only one nonfatal cardiac complication among the 50 patients with a dobutamine echocardiogram, which was negative for evidence of inducible ischemia. In conclusion, this study demonstrates that dobutamine stress echocardiography should use an infusion rate of > or = 30 micrograms/kg/min to optimize diagnostic accuracy relative to angiographic coronary artery disease. A test that shows no evidence of new, inducible ischemia predicts a low risk of perioperative cardiac events in patients undergoing noncardiac surgery, even at an infusion rate as low as 20 microns/kg/min.