|| Checking for direct PDF access through Ovid
Introduction: Guidelines recommend screening all stroke admissions for cognitive impairment. Delirium may be an important contributor to cognitive problems, but available data are limited by potential selection biases and imperfect delirium assessment. We aimed to describe robust estimates of delirium occurrence (incidence and prevalence) and risk factors in an acute stroke unit (ASU).Methods: We collected data from sequential admissions to our University Hospital ASU over a 20 week period (Feb-July 2016). The only exclusion was where the clinical team felt that any form of assessment was inappropriate. We aimed to perform cognitive assessments at 48 hours post stroke. Cognitive assessments were based on mini Montreal Cognitive Assessment (m-MoCA) and GP-Cog informant interview to assess for pre-stroke dementia. We assessed for delirium based on DSM-V criteria. We described univariable and multivariable associations with delirium occurrence.Results: Across the study period, 184 strokes were admitted; median age 71 years (IQR:61-79). Some form of cognitive assessment was available for 178 (97%). In total. 152 had evidence of cognitive impairment on m-MOCA, 55 (31%) met criteria for delirium and 25 (15%) had pre-stroke dementia. Significant univariable associations with delirium were:age and pre-existing cognitive decline. On logistic regression:age, history of drug or alcohol misuse and pre-stroke mRS were all independently associated with delirium.Discussion: In an unselected ASU population a substantial proportion will have cognitive impairment based on screening. Delirium is common and contributes to this cognitive problems seen in acute stroke. Independent associations with delirium occurrence were identified but were non-modifiable. Our results have implications for policy and practice. As post-stroke cognitive impairment is multifactorial, screening delirium and pre-existing cognitive issues should form part of the initial assessment.