Abstract
OBJECTIVES: To determine whether the association between impaired cognition and greater risk of incident stroke is also observed when cognitive impairment is defined using different criteria for mild cognitive impairment (MCI). DESIGN: Prospective cohort study with 10 years of follow-up. SETTING: Large multicentre study in the United Kingdom. PARTICIPANTS: Individuals (aged 64-105) from the Medical Research Council Cognitive Function and Ageing Study (N = 13,004). From this, a subsample of 2,640 individuals was selected based on age, center, and cognitive ability to undergo a detailed cognitive assessment. MEASUREMENTS: Information on sociodemographic characteristics, health, cognition, and functional ability was collected in an interview. The Geriatric Mental State Automated Geriatric Examination for Computer Assisted Taxonomy and the Cambridge Cognitive Examination were used to determine cognitive status. Stroke incidence was derived from self-report, informant report, and death certificates. Participants were divided into no, mild, moderate, and severe cognitive impairment according to their baseline Mini-Mental State Examination (MMSE) score. MCI criteria were used to classify persons into four groups: no cognitive impairment, MCI, severe impairment (i.e. other cognitive impairment no dementia: OCIND) and dementia. RESULTS: Over 10 years, 703 incident strokes occurred. Lower MMSE score at baseline was associated with greater risk of incident stroke. When cognitive status was determined according to MCI criteria, those with severe impairment (odds ratio (OR) = 1.5, 95% confidence interval (CI) = 1.0-2.2) and dementia (OR = 2.6, 95% CI = 1.6-3.4) had a significantly greater risk of stroke than those with no cognitive impairment. CONCLUSION: Criteria for MCI, defined using MMSE scores or clinical criteria, can capture individuals at greater stroke risk. The results highlight the need to focus on stroke risk in individuals even with MCI.