Reliable data on the prevalence and predictors of post-stroke dementia are needed to inform patients and carers, plan services and clinical trials, ascertain the overall burden of stroke, and understand its causes. However, published data on the prevalence and risk factors for pre-stroke and post-stroke dementia are conflicting. We undertook this systematic review to assess the heterogeneity in the reported rates and to identify risk factors for pre-stroke and post-stroke dementia.

Studies published between 1950 and May 1, 2009, were identified from bibliographic databases, reference lists, and journal contents pages. Studies were included if they were on patients with symptomatic stroke, were published in English, reported on a series of consecutive eligible patients or volunteers in prospective cohort studies, included all stroke or all ischaemic stroke, measured dementia by standard criteria, and followed up patients for at least 3 months after stroke. Pooled rates of dementia were stratified by study setting, inclusion or exclusion of pre-stroke dementia, and by first, any, or recurrent stroke. Pooled odds ratios were calculated for factors associated with pre-stroke and post-stroke dementia.

We identified 22 hospital-based and eight population-based eligible cohorts (7511 patients) described in 73 papers. The pooled prevalence of pre-stroke dementia was higher (14·4%, 95% CI 12·0–16·8) in hospital-based studies than in population-based studies (9·1%, 6·9–11·3). Although post-stroke (≤1 year) dementia rates were heterogeneous overall, 93% of the variance was explained by study methods and case mix; the rates ranged from 7·4% (4·8–10·0) in population-based studies of first-ever stroke in which pre-stroke dementia was excluded to 41·3% (29·6–53·1) in hospital-based studies of recurrent stroke in which pre-stroke dementia was included. The cumulative incidence of dementia after the first year was little greater (3·0%, 1·3–4·7) per year in hospital-based studies than expected on the basis of recurrent stroke alone. Medial temporal lobe atrophy, female sex, and a family history of dementia were strongly associated with pre-stroke dementia, whereas the characteristics and complications of the stroke and the presence of multiple lesions in time and place were more strongly associated with post-stroke dementia.

After study methods and case mix are taken into account, reported estimates of the prevalence of dementia are consistent: 10% of patients had dementia before first stroke, 10% developed new dementia soon after first stroke, and more than a third had dementia after recurrent stroke. The strong association of post-stroke dementia with multiple strokes and the prognostic value of other stroke characteristics highlight the central causal role of stroke itself as opposed to the underlying vascular risk factors and, thus, the likely effect of optimum acute stroke care and secondary prevention in reducing the burden of dementia.

None.