Trajectories of cognitive change following stroke: stepwise decline towards dementia in the elderly

Stroke events increase the risk of developing dementia, 10% for a first-ever stroke and 30% for recurrent strokes. However, the effects of stroke on global cognition, leading up to dementia, remain poorly understood. We investigated: (i) post-stroke trajectories of cognitive change, (ii) trajectories of cognitive decline in those who develop dementia over periods of follow-up length and (iii) risk factors precipitating the onset of dementia. Prospective cohort of hospital-based stroke survivors in North-East England was followed for up to 12 years. In this study, we included 355 stroke survivors of ≥75 years of age, not demented 3 months post-stroke, who had had annual assessments during follow-up. Global cognition was measured annually and characterized using standardized tests: Cambridge Cognition Examination-Revised and Mini-Mental State Examination. Demographic data and risk factors were recorded at baseline. Mixed-effects models were used to study trajectories in global cognition, and logistic models to test associations between the onset of dementia and key risk factors, adjusted for age and sex. Of the 355 participants, 91 (25.6%) developed dementia during follow-up. The dementia group had a sharper decline in Cambridge Cognition Examination-Revised (coeff. = -1.91, 95% confidence interval = -2.23 to -1.59, P < 0.01) and Mini-Mental State Examination (coeff. = -0.46, 95% confidence interval = -0.58 to -0.34, P < 0.01) scores during follow-up. Stroke survivors who developed dementia within 3 years after stroke showed a steep decline in global cognition. However, a period of cognitive stability after stroke lasting 3 years was identified for individuals diagnosed with dementia in 4-6 years (coeff. = 0.28, 95% confidence interval = -3.28 to 3.8, P = 0.88) of 4 years when diagnosed at 7-9 years (coeff. = -3.00, 95% confidence interval = -6.45 to 0.45, P = 0.09); and of 6 years when diagnosed at 10-12 years (coeff. = -6.50, 95% confidence interval = -13.27 to 0.27, P = 0.06). These groups then showed a steep decline in Cambridge Cognition Examination-Revised in the 3 years prior to diagnosis of dementia. Risk factors for dementia within 3 years include recurrent stroke (odds ratio = 3.99, 95% confidence interval = 1.30-12.25, P = 0.016) and previous disabling stroke, total number of risk factors for dementia (odds ratio = 2.02, 95% confidence interval = 1.26-3.25, P = 0.004) and a Cambridge Cognition Examination-Revised score below 80 at baseline (odds ratio = 3.50, 95% confidence interval = 1.29-9.49, P = 0.014). Our unique longitudinal study showed cognitive decline following stroke occurs in two stages, a period of cognitive stability followed by rapid decline before a diagnosis of dementia. This pattern suggests stroke may predispose survivors for dementia by diminishing cognitive reserve but with a smaller impact on cognitive function, where cognitive decline may be precipitated by subsequent events, e.g. another cerebrovascular event. This supports the assertion that the development of vascular dementia can be stepwise even when patients have small stroke lesions.