sumption, smoking, exercise, hypertension, diabetes, history of high cholesterol, history of angina, parental history of premature myocardial infarction, and randomized treatment assignment did not attenuate this estimate (relative risk, 1.24 [95% confidence interval, 0.80-1.92]). It is not entirely surprising that the effect estimate is somewhat smaller in the elderly because individuals with high BMIs have greater risks of cardiovascular events at a younger age and thus may die earlier. 2 Those who survive may be less susceptible to those events. In addition, the elderly have higher baseline risks of stroke and relative risks of even modest to moderate magnitude can translate into substantial differences in absolute risk.The combined evidence from our study and that of Milionis and colleagues suggests that excess weight may be an independent risk factor for ischemic stroke regardless of geographic location and lifestyle factors. Since the diet of participants in the 2 studies are different, the results also suggest that it does not matter how the excess weight was gained with respect to the association between BMI and ischemic stroke.