The diagnostic accuracy of dobutamine stress echocardiography (DSE) (incremental infused doses of 5, 10, 20 and 30 micrograms/kg/min) was evaluated in 141 patients who underwent coronary arteriography within 2 weeks of DSE. All patients were being evaluated for known or suspected coronary artery disease (CAD). DSE was interpreted blindly as normal or showing evidence of CAD, depending on the presence of resting or inducible wall motion abnormalities. Coronary arteriograms were reviewed in a blinded, quantitative fashion. DSE had a sensitivity of 96% for detecting patients with CAD, and a specificity of 66%. For the 53 patients with normal resting wall motion, sensitivity was 87% and specificity 91%. The protocol was well-tolerated by all patients. In comparison with wall motion analysis, 12-lead electrocardiograms during dobutamine infusion revealed ischemic changes in only 17% of patients with CAD. It is concluded that DSE is a clinically useful and accurate means for detecting CAD, its specificity is hindered in patients with resting wall motion abnormalities, and it can safely be used in patients with known cardiac disease.