Is pre-existing dementia an independent predictor of outcome after stroke? A propensity score-matched analysis

With an aging population, patients are increasingly likely to present with stroke and pre-existing dementia, which may lead to greater death and disability. The aim of this work was to assess the risk of all-cause mortality and poor functional outcomes after ischemic stroke in patients with and without pre-existing dementia. We conducted a multicenter cohort study of all patients presenting to 12 tertiary care institutions participating in the Registry of the Canadian Stroke Network (RCSN) with a first ischemic stroke between 2003 and 2008. Individuals with pre-existing dementia were matched using propensity-score methods with patients without dementia during their index hospitalization based on the following characteristics: age (within 3 years), sex, stroke severity, stroke subtype (lacunar vs. non-lacunar), level of consciousness, vascular risk factors, dysphagia, glucose and creatinine on admission, Charlson index, residence prior to hospitalization (home vs. other), pre-admission dependency, hospital arrival via ambulance, admission to stroke unit, thrombolysis, and palliative care. A propensity score for all-cause mortality and clinical outcomes was developed. Registry of the Canadian Stroke Network (RCSN) and Registered Persons Database (RPDB). The primary outcome was all-cause mortality at 30 days. Secondary outcomes included mortality at discharge and at 1 year, disability at discharge (modified Rankin scale ≥ 3), medical complications (pneumonia), and discharge disposition. A subgroup analysis assessing the risk of intracerebral hemorrhage among those receiving thrombolysis was also conducted. We matched 877 patients with an acute ischemic stroke and pre-existing dementia to 877 stroke patients without dementia. Patients were well matched. The mean age was 82 years and 58 % were women. Mortality at discharge, 30 days, and 1 year after stroke was similar in patients with and without dementia [for mortality at discharge RR 0.88 [95 % confidence interval (CI) 0.74–1.05]; mortality at 30-days: RR 0.88 (95 % CI 0.75–1.03) and mortality at 1 year: RR 1.01 (95 % CI 0.92–1.11). Patients with pre-existing dementia had similar disability at discharge and home disposition. In the subgroup of patients who received thrombolysis, there were no differences between those with and without dementia in the risk of intracerebral hemorrhage (RR 1.27; 95 % CI 0.69–2.35) and no differences in mortality or disability at discharge. Pre-existing dementia is not independently associated with mortality, disability, or institutionalization after ischemic stroke. Pre-existing dementia may not necessarily preclude access to thrombolytic therapy and specialized stroke care.