OBJECTIVE: To determine whether carotid endarterectomy is superior to best medical therapy in preserving cognition, and whether low Mini-Mental State Examination (MMSE) scores predict TIA, stroke, myocardial infarction, or death. METHODS: Subjects participating in the Asymptomatic Carotid Atherosclerosis Study were administered the MMSE at periodic intervals. Group means were calculated at randomization, 1 and 3 months later, and every 6 months thereafter. The group means were compared by treatment and over time. A proportional hazard regression model incorporating postrandomization MMSE score as a predictor variable was used to estimate risk of death, stroke, or other outcome events. RESULTS: There was no intergroup difference in mean MMSE score during 5 years of observation. For individual patients, the relationship between a low postrandomization score on the MMSE and increased risk of death was significant (p = 0.0001). Patients who experienced stroke after randomization also had a significant and persistent reduction in MMSE score (p = 0.0001). CONCLUSIONS: Carotid endarterectomy had no impact on MMSE score in this study. Patients with low postrandomization MMSE scores had a greater likelihood of death. Stroke reduced MMSE scores and may portend cognitive impairment. The authors recommend the routine inclusion of cognitive testing in future clinical trials designed to evaluate prophylaxis or acute therapy of stroke.