Objective: The net benefit for patients undergoing carotid endarterectomy is critically dependent on the risk of perioperative stroke and death. Information about risk factors can aid appropriate selection of patients and inform efforts to reduce complication rates. This study identifies the clinical, radiographic, surgical, and anesthesia variables that are independent predictors of deaths and stroke following carotid endarterectomy. Methods: A retrospective cohort study of patients undergoing carotid endarterectomy in 1997 and 1998 by 64 surgeons in 6 hospitals was performed (N = 1972). Detailed information on clinical, radiographic, surgical, anesthesia, and medical management variables and deaths or strokes within 30 days of surgery were abstracted from inpatient and outpatient records. Multivariate logistic regression models identified independent clinical characteristics and operative techniques associated with risk-adjusted rates of combined death and nonfatal stroke as well as all strokes. Results: Death or stroke occurred in 2.28% of patients without carotid symptoms, 2.93% of those with carotid transient ischemic attacks, and 7.11% of those with strokes (P < .0001). Three clinical factors increased the risk-adjusted odds of complications: stroke as the indication for surgery (odds ratio [OR], 2.84; 95% confidence interval [CI] = 1.55-5.20), presence of active coronary artery disease (OR, 3.58; 95% CI = 1.53-8.36), and contralateral carotid stenosis <50% (OR, 2.32; 95% CI = 1.33-4.02). Two surgical techniques reduced the risk-adjusted odds of death or stroke: use of local anesthesia (OR, 0.30; 95% CI = 0.16-0.58) and patch closure (OR, 0.43; 95% CI = 0.24-0.76). Conclusions: Information about these risk factors may help physicians weigh the risks and benefits of carotid endarterectomy in individual patients. Two operative techniques (use of local anesthesia and patch closure) may lower the risk of death or stroke.