The purpose of this study was to review experience with carotid artery surgery based on findings obtained solely from duplex scanning with special regard to unexpected findings during surgery and the early outcome. From January 1993 through December 1999, 271 consecutive patients underwent 287 carotid endarterectomies (CEAs), 229 (80%) of which were performed solely based on duplex scan findings. During the study period 5,932 carotid artery duplex scans were performed in 4,466 patients. Of 589 patients with internal carotid artery (ICA) stenosis 70%, 246 underwent CEA compared to 25 of 156 with 50-69% ICA stenosis. The indications for CEA were transient ischemic attack (TIA) in 88 (30.7%), amaurosis fugax in 60 (20.9%), previous stroke in 91 (31.7%) and asymptomatic disease in 48 (16.7%) cases. There were no statistically significant differences between the groups operated on with and without preoperative angiography with respect to the indications for surgery, associated risk factors, or the degree of stenosis on the contralateral side. In patients undergoing surgery without angiography, there were no unexpected findings that influenced the performance of surgery, in all except 1. There were no significant differences in perioperative morbidity and mortality in patients undergoing surgery with and without conventional angiography. The combined mortality and major stroke rates were 3.4% and 2.2%, respectively. It is concluded that CEA can safely be performed without preoperative angiography in cases with conclusive duplex scan findings.