Background: Little is known about patterns in the use of carotid revascularization since a 2004 Medicare national coverage decision supporting carotid artery stenting. We examined geographic variation in and predictors of carotid endarterectomy and carotid stenting. Methods: Analysis of claims from the Centers for Medicare& Medicaid Services from January 1, 2003, through December 31, 2006. Patients were 65 years or older and had undergone carotid endarterectomy or carotid stenting. The main outcome measures were annual age-adjusted rates of carotid endarterectomy and carotid stenting, factors associated with the use of carotid revascularization, and mortality rate at 30 days and 1 year. Results: The rate of endarterectomy decreased from 3.2 per 1000 person-years in 2003 to 2.6 per 1000 person-years in 2006. After adjustment for demographic and clinical characteristics, there was significant geographic variation in the odds of carotid revascularization, with the East North Central region having the greatest odds of endarterectomy (odds ratio, 1.60; 95% confidence interval, 1.55-1.65) and stenting (1.61; 1.46-1.78) compared with New England. Prior endarterectomy (odds ratio, 3.06; 95% confidence interval, 2.65-3.53) and coronary artery disease (2.12; 2.03-2.21) were strong predictors of carotid stenting. In 2005, mortality was 1.2% at 30 days and 6.8% at 1 year for endarterectomy and 2.3% at 30 days and 10.3% at 1 year for stenting. Conclusions: Significant geographic variation exists for carotid endarterectomy and carotid stenting. Prior end-arterectomy and coronary disease were associated with greater odds of carotid stenting.