The relative prognostic value of widely accessible resting two-dimensional echocardiographic ventricular function data has not been compared with recognized clinical and scintigraphic risk markers in patients who are unable to exercise before major nonvascular surgery. To this end, 53 consecutive patients aged 67 +/- 13 years undergoing preoperative evaluation (intraabdominal, 23%; orthopedic, 30%; thoracic, 9%; other, 38%) for known or suspected coronary artery disease were followed up to evaluate the prognostic value of these studies for the perioperative cardiac events (cardiac death [n = 4], myocardial infarction [n = 2], unstable angina [n = 3], and pulmonary edema [n = 8]) that occurred in 13 of the 53 patients (25%). Dipyridamole thallium-201 myocardial redistribution defects occurred in 15 (28%) patients. Resting echocardiographic left ventricular dysfunction was present in 21 (40%) patients. Multivariate analysis of clinical, echocardiographic, and scintigraphic risk predictors revealed that cardiac events were not predicted by clinical variables, including Goldman class or score. Cardiac events were independently predicted only by the presence of significant left ventricular dysfunction on resting two-dimensional echocardiography (p < 0.042) and dipyridamole thallium-201 defect redistribution (p < 0.026). A dipyridamole-induced reversible thallium-201 perfusion defect was predictive of subsequent cardiac death or myocardial infarction (p < 0.02), whereas left ventricular dysfunction on resting echocardiography was predictive of perioperative pulmonary edema (p < 0.023). We conclude that stress thallium-201 perfusion imaging and resting two-dimensional echocardiography provide independent prognostic information in patients undergoing major nonvascular surgery who are at significant risk for ischemic cardiac events and who are unable to perform standard exercise stress tests.(ABSTRACT TRUNCATED AT 250 WORDS)