Abstract 106: Low Cognitive Scores in Domains of Learning, Memory or Executive Function Predict Incident Stroke: The REGARDS Study
Objective: Low scores on tests of learning, memory, and executive function have been associated with subclinical cerebrovascular pathology in the absence of stroke. In a population-based, stroke-free cohort, we hypothesized that lower performance in these cognitive domains would predict incident stroke.
Methods: From 2003-2007, REGARDS enrolled black and white adults aged ≥45 throughout the U.S. Participants are followed by phone twice a year for stroke events and biennially for longitudinal cognitive assessments. Word list learning and recall tasks are used to measure learning and memory, and semantic and letter fluency tasks to measure executive function. Included participants were stroke-free at enrollment and remained so prior to their cognitive assessments. We assessed incident stroke risk through September 1, 2012 using standard survival analyses adjusted for age, education, race, and geographic region.
Results: Sample sizes up to 20,422 varied by cognitive task, and follow-up time ranged up to 7 years. Adjusted hazard ratios (95% CI) for stroke associated with the lowest versus highest quintiles of pre-stroke cognitive scores were 1.90 (1.24, 2.93) for learning, 2.23 (1.45, 3.42) for memory, and 1.69 (1.10, 2.61) and 1.98 (1.12, 3.52) for executive function as assessed by semantic and letter fluency, respectively (Figure).
Conclusions: In a stroke-free, population-based cohort, lower baseline function in each cognitive domain was associated with increased risk of incident stroke. Low scores may signal the progression and severity of stroke risk factors.
Author Disclosures: V.G. Wadley: None. E.L. Thacker: None. A.J. Letter: None. F.W. Unverzagt: None. C.S. Moy: None. D.G. Clark: None. L.A. McClure: None. V.J. Howard: None. M. Cushman: None. D.A. Levine: None. B.J. Kelley: None. D. Kleindorfer: None. B. Kissela: None. G. Howard: None.
- © 2014 by American Heart Association, Inc.