The safety of any diagnostic test is a major issue in deciding its practicability and cost-effectiveness. The aim of this study was to evaluate the safety of various stress echocardiographic modalities in the "real world." From February 1998 to January 2004, a simple written questionnaire was distributed to echocardiography laboratories across the world known to perform stress echocardiography. The following categories of major complications (known to occur during stress testing) were indicated: sustained ventricular tachycardia (> 30 beats/min), ventricular fibrillation, myocardial infarction, third-degree atrioventricular block, severe hypotension requiring therapy, cardiac asystole, cardiac rupture, stroke, and death. Three hundred centers were polled, from which 71 co-investigators responded and reported on 85,997 patient examinations. Exercise was used in 26,295 cases, dobutamine in 35,103, and dipyridamole in 24,599 cases. Life-threatening events occurred in 86 cases: during exercise in 4 patients (event rate 1 in 6,574), during dobutamine infusion (small dose for viability and/or large dose for ischemia) in 63 patients (event rate 1 in 557), and during dipyridamole stress testing in 19 patients (event rate 1 in 1,294). Of the 86 patients with complications, 5 died during dobutamine stress testing (ventricular fibrillation, n = 2; cardiac rupture, n = 3) and 1 after dipyridamole testing (cardiogenic shock). In conclusion, stress echocardiography is a safe method in the real world, but serious complications may occur. Exercise seems safer than pharmacologic stress and dipyridamole safer than dobutamine, possibly because of preselection criteria.