Between 1 January 1991 and 31 December 1994, 215 carotid endarterectomies were performed at the authors' institution, which utilized a clinical pathway. Prior to May 1992, arteriography was performed routinely. A near perfect correlation was found between the arteriograms and duplex scans when they were compared as part of the Intersocietal Commission for the Accreditation of Vascular Laboratories (ICAVL) vascular laboratory accreditation process. A policy of selective arteriography was instituted in May 1992. Only 11 arteriograms were performed on the next 148 patients (7%) who underwent carotid endarterectomies. Arteriography was performed on two patients with extremely high bifurcations, and five patients when an exact degree of stenosis could not be determined. Two patients with simultaneous mid common carotid and bifurcation stenoses had arteriography to confirm the duplex findings. Arteriography confirmed a long, high-grade internal carotid artery stenosis, which was felt to be operable by duplex, and a simultaneous bifurcation and suspected left common carotid orificial stenosis in one patient each. Arteriograms were performed on three non-operated patients felt to have occluded internal carotid artery on duplex scanning. Two had string-like internal carotid arteries that extended intracranially from the bifurcation, and one patient had an internal carotid artery dissection. Duplex results were grossly confirmed at operation and pathologically. There were no neurological complications in those patients undergoing carotid endarterectomies based on the results of duplex scanning alone. Carotid endarterectomies can be safely performed based on the results of a duplex scan performed in an accredited vascular laboratory. This approach eliminates the risk and cost of arteriography. Approximately 10% of patients will require arteriography if the duplex scan is equivocal or shows disease at other areas than the carotid bifurcation.