To assess the value of dobutamine stress echocardiography (DSE) to detect coronary artery disease (CAD) and to compare the diagnostic accuracy between DSE and treadmill exercise test (TXT), 104 patients (mean age 58 +/- 12 years) presenting for coronary angiography were prospectively studied. TXT was performed according to the Bruce protocol. Dobutamine (5-40 micrograms/kg/min) was infused in 3-min stages. Digital echocardiograms were recorded on-line at baseline, during low- and peak-dose dobutamine infusion, and at recovery. An echocardiogram positive for CAD was defined as the one showing a new wall motion abnormality (WMA) induced by dobutamine. There were no major complications during the study. Significant CAD (> or = 50% diameter stenosis) was present in 17 of 30 patients who had normal echocardiograms at baseline. The sensitivity for detecting CAD was 76% by TXT and 94% by DSE, and the specificity was only 38% by TXT and 92% by DSE, respectively. Seventy-four patients had localized rest WMAs. Twenty-four had no significant CAD or lesions only confined to regions with abnormal rest wall motion, and 50 had disease remote from these regions. The sensitivity for detection of remote disease was 60% by TXT and 76% by DSE, and the specificity was 75% by TXT and 96% by DSE, respectively. In conclusion, DSE is a safe and accurate diagnostic tool for identifying CAD and for predicting the extent of disease in those who have localized rest WMAs.