Background About 10% of patients develop new dementia after first stroke and about 30% after recurrent stroke, and milder cognitive impairment is probably even more prevalent. There is some evidence that optimal acute stroke care may reduce these risks. We aimed to determine whether measures of dependency and disability commonly used in assessing stroke outcome adequately reflect cognitive burden.

Methods In a population-based study of consecutive strokes, functional outcome was assessed with the Rankin and Barthel scores and cognitive outcome with the Mini-Mental State Examination (MMSE) and Montreal Cognitive Assessment (MoCA) at least 6 months after index event. Standard cut-offs of MMSE≥27 (normal cognitive function), and MoCA≥26 (normal cognitive function) were used.

Results Among 808 patients (mean age 72.3, 50% male), 591 were nondependent according to the Rankin score (0–2), of whom 181 (31%) had MMSE<27 and 58 (10%) had MMSE<24. Of 512 patients with normal Barthel (20), 144 (28%) had MMSE<27 and 45 (9%) had MMSE<24. In 349 patients who received the MoCA, 256 were nondependent on Rankin score, of whom 174 (68%) had MoCA<26, and 234 had a normal Barthel score, of whom 151 (64%) had MoCA<26.

Conclusion Up to 60% of stroke patients defined as nondependent on standard functional outcome scales have some degree of cognitive impairment indicating that such functional outcome scales alone may be inadequate to assess the full effects of interventions in acute stroke.