We evaluated whether a preoperative clinical algorithm allows an adequate stratification in cardiac risk and the predictive value of dipyridamole thallium-201 scintigraphy and rest echocardiography for postoperative adverse cardiac outcomes. Three hundred twenty patients undergoing 338 vascular surgery procedures were prospectively stratified into low, intermediate, and high risk. The low- and intermediate-risk patients underwent surgery without further diagnostic evaluation. In 7 high-risk patients the vascular procedure was canceled (1 died of myocardial infarction at 6-month follow-up), 9 underwent presurgical myocardial revascularization (1 died of myocardial infarction), and 49 underwent vascular surgery with perioperative intensive care treatment. Hospital mortality was 3.8%. Cardiac mortality and morbidity were 1.5% and 10.4%, respectively. We observed a significant difference in "hard" (death, myocardial infarction, pulmonary edema, major arrhythmias) and "soft" (myocardial ischemia, minor arrhythmias) events between groups, p <0.001. Previous pulmonary edema was a predictive variable of cardiac outcomes (multiple logistic regression analysis). Ninety-nine of 220 intermediate-risk patients randomly underwent dipyridamole thallium-201 scintigraphy: 37 had redistribution, 10 persistent, and 52 no defects; 7 of 13 soft and hard cardiac events occurred in patients without redistribution defects. Sensitivity, specificity, and positive and negative predictive values of redistribution defects for postoperative adverse outcomes were 38%, 63%, 14%, 87%, respectively. This algorithm may provide a safe and cost-effective approach (average cost saving per patient $1,500) to cardiac risk stratification. These results suggest that routine use of dipyridamole thallium-201 scintigraphy for screening of intermediate-risk patients may not be warranted.