Objective: The durability of carotid endarterectomy (CEA) may be affected by carotid restenosis. The data from randomized trials show that the highest incidence of restenosis after CEA occurs from 12 to 18 months after surgery. The optimal CEA technique to reduce perioperative complications and restenosis rates is still undefined. This study examines the long-term clinical outcome and incidence of recurrent stenosis in patients who undergo eversion CEA. Previously published perioperative results of this study did not show statistically significant differences in study endpoints between the eversion and standard techniques. Methods: From October 1994 to March 1997, 1353 patients with surgical indications for carotid stenosis were randomly assigned to undergo eversion (n = 678) or standard CEA (n = 675; primary closure, 419; patch, 256). Withdrawal from the assigned treatment occurred in 1.6% of the patients (in 13 assigned to eversion CEA, and in nine assigned to standard CEA). The clinical and duplex scan follow-up examination was 99% complete, and the mean follow-up interval was 33 months (range, 12 to 55 months). The primary outcomes were perioperative and late major stroke and death, carotid restenosis (stenosis ≥ 50% of the lumen diameter detected at duplex scanning), and carotid occlusion. The primary evaluation of study outcomes was conducted on the basis of an intention-to-treat analysis. Results: Restenosis was found at duplex scanning in 56 patients (19 in the eversion group, and 37 in the standard group). Within the standard group, the restenosis rates were 7.9% in the primary closure population and 1.5% in the patched population. Of the patients with restenosis, 36% underwent cerebral angiography that confirmed restenosis in all cases. The cumulative restenosis risk at 4 years was significantly lower in the group that underwent treatment with eversion CEA as compared with the standard group (3.6% vs 9.2%; P = .01), with an absolute risk reduction of 5.6% and a relative risk reduction of 62%. Eighteen patients would have had to undergo treatment with eversion CEA to prevent one restenosis during the 4-year period. The incidence rate of ipsilateral stroke was 3.3% in the eversion population and 2.2% in the standard group. There were no significant differences in the cumulative risks of ipsilateral stroke (3.9% for eversion, and 2.2% for standard; P = .2) and death (13.1% for eversion, and 12.7% for standard; P = .7)) in the two groups. Of the 18 variables that were examined for their influence on restenosis, eversion CEA (hazard ratio, 0.3; 95% confidence interval, 0.2 to 0.6; P = .0004) and patch CEA (hazard ratio, 0.2; 95% confidence interval, 0.07 to 0.6; P = .002) were negative independent predictors of restenosis with multivariate Cox proportional hazards regression analysis. Conclusion: The EVEREST (EVERsion carotid Endarterectomy versus Standard Trial) showed that eversion CEA is safe, effective, and durable. No statistically significant differences were found in late outcome between the eversion and standard techniques at the available follow-up examination. (J Vasc Surg 2000;31:19-30.)