BACKGROUND AND PURPOSE Recurrent stenosis after carotid endarterectomy (CEA) has been reported to vary between a few percent and 30%. Regression of recurrent stenosis has been reported sporadically in the literature, but studies analyzing the factors affecting regression are lacking. This study analyzed factors affecting the regression of postoperative stenosis from a prospective randomized trial of CEA comparing primary closure (PC) versus patching. PATIENT POPULATION AND METHODS Three hundred ninety-nine CEAs were randomized into three groups: 135 PCs, 135 polytetrafluoroethylene patch closures (PTFE), and 130 vein patch closures (VPC). Postoperative duplex ultrasounds were done at 1, 6, and 12 months, and then yearly. The subgroup of these CEAs that exhibited postoperative stenosis was followed for possible regression of the stenosis. Analyses of various risk factors were examined for possible association with regression of recurrent stenosis. Mean follow-up was 46 months. RESULTS Of 105 postoperative stenoses, regression was noted in 6/64 (9%) in PC, 6/13 (46%) in PTFE, and 10/28 (36%) in VPC. Overall, 22 recurrent stenoses regressed; 19 regressed to normal and 3 regressed from 50% to 80% stenosis to 20% to <50% stenosis. The mean time to regression was 383 days. Regression was more common in patching than PC. Both VPC and PTFE had significantly more regression than PC. When stenoses of 50% to 80% were analyzed, patching had more regression than PC. None of the recurrent stenoses > or = 80% regressed. There was no association between regression and other factors, including gender, hypertension, diabetes mellitus, coronary artery disease, smoking, internal carotid artery diameter, hyperlipidemia, hypercholesterolemia, or aspirin intake. CONCLUSIONS Regression of recurrent stenosis was associated more strongly with patching than with PC. There was no association between regression and other factors.