Before 1981, the neurologic morbidity and mortality associated with carotid endarterectomy (CEA) in the Los Angeles County/USC Medical Center public teaching institution was 20 per cent, similar to results from other hospitals. In 1981, a standardized protocol was adopted in an attempt to improve surgical outcome following CEA. Between 1981 through 1990, 89 patients with a mean age of 60.9 years (range 38 to 80 yrs) had 100 consecutive CEAs. Atherosclerotic risk factors included hypertension in 57 patients (61.8%), tobacco use in 57 (64.0%), and diabetes mellitus in 28 (31.5%). Forty-nine patients had a history of ischemic heart disease. Indications for CEA were stroke in 40 cases, transient ischemic attack in 33 cases, and asymptomatic, high-grade stenosis of the internal carotid artery (greater than 85 per cent) in 19. Perioperative and surgical management are detailed in the authors' protocol. Intraluminal shunts were routinely used (99 cases) and 24 arteriotomies were patched. Completion arteriograms were performed in 99 cases, four of which were revised because of arteriographic abnormalities. Three patients sustained postoperative ipsilateral neurologic events. Ten patients had cranial nerve palsies, six of which were transient. Two patients had nonfatal postoperative myocardial infarctions. There were no deaths. The combined stroke and transient ischemic attack (TIA) mortality rate was 3 per cent. In conclusion, the audit and quality assurance process identified unacceptable results following CEA. A protocol was developed that addressed pre, intra, and postoperative details of patient selection, operative technique, and postoperative care. By adhering to the protocol, the major neurologic morbidity and mortality rate has been reduced to 3 per cent.