Background and ObjectivesReports assessing the association of stroke risk factors with incident stroke have generally assumed a uniform magnitude of associations across the age spectrum, an assumption we assess in this report.MethodsParticipants enrolled 2003–2007 in the REasons for Geographic And Racial Differences in Stroke (REGARDS) cohort study who were stroke free at baseline were followed for incident stroke. Associations of traditional stroke risk factors with incident stroke were assessed using (1) proportional hazards analysis based on the baseline age of the participant and (2) Poisson regression analysis assessing associations based on the changing age of the participant during their follow-up (age at exposure). In each analysis, age strata were selected to have a similar number of strokes in each stratum, specifically 45–64, 65–73, and 74+ years for the proportional hazards analysis and 45–69, 70–79, and 80+ years for Poisson regression.ResultsA total of 1,405 ischemic stroke events occurred among 28,235 participants over a median follow-up of 11.3 years, with a total of 276,074 person-years exposure. For both analytic approaches, the magnitude of the association with stroke was significantly less at older ages for diabetes (hazard or relative risk decreasing from ≈2.0 in younger strata to ≈1.3 in older strata), heart disease (from ≈2.0 to ≈1.3), and hypertension defined at a threshold of 140/90 mm Hg (from ≈1.80 to ≈1.50); however, there was no age-related difference in the magnitude of the association for smoking, atrial fibrillation, or left ventricular hypertrophy.DiscussionHypertension and diabetes are 2 of the more important risk factors for stroke; however, their association with stroke risk appears substantially less at older ages. That the magnitude of the association for smoking, atrial fibrillation, and left ventricular hypertrophy does not decrease with age suggests their relative importance in determining stroke risk likely increases with age.