Background and Purpose-We sought to investigate the association of cigarette smoking with high-grade carotid artery stenosis in Hispanic, black, and white patients with cerebral ischemia in two independent samples. Methods-Prospectively collected data from the Northern Manhattan Stroke Study (NOMASS) (nϭ431) and the Berlin Cerebral Ischemia Databank (BCID) (nϭ483) were used separately for a cross-sectional study estimating the association between cigarette smoking and high-grade carotid stenosis (defined as a luminal narrowing of Ն60%, diagnosed by duplex and/or Doppler ultrasound). In both studies, cerebral ischemia patients with normal sonographic findings or nonstenosing plaques of their carotid arteries served as a comparison group. Multivariate logistic regression models were used for statistical tests to determine the association between smoking and the dependent variable for high-grade carotid stenosis. Age, sex, hypertension, diabetes, hypercholesterolemia, and race/ethnicity were considered potential confounders. Further analyses of the NOMASS data estimated the effect of the amount of cigarette use and the impact of race/ethnicity. Results-High-grade carotid stenoses were found in 14% of the NOMASS and in 21% of the Berlin patients. In Berlin the entire sample was white, whereas in New York only 19% of the cohort were white. In both samples, smoking was independently associated with severe carotid stenosis (NOMASS: odds ratio [OR], 1.5; 95% confidence interval [CI], 1.1 to 2.0; BCID: OR, 3.9; 95% CI, 2.4 to 6.4). Patients smoking 20 pack-years or more showed a significant association (OR, 2.0; 95% CI, 1.1 to 3.9), whereas no significant effect was found for lower amounts of cigarette use. In NOMASS, white smokers displayed a significant (OR, 3.2; 95% CI, 1.1 to 8.9) association with high-grade carotid stenosis, the association for black smokers was less strong, and no association was found among Hispanics. Conclusions-Smoking is an independent determinant of severe carotid artery stenosis in patients with focal cerebral ischemia. The association differs by race/ethnicity, with the greatest effect observed among whites. (Stroke. 1998;29:908-912.)