Objective:We aimed to determine prevalence of pre-stroke frailty in acute stroke and describe validity of a Frailty Index–based assessment.Design:Cross-sectional.Setting:Single UK urban teaching hospital.Subjects:Consecutive acute stroke unit admissions, recruited in four waves (May 2016–August 2018). We performed the assessments within first week and attempted to include all admissions.Main measures:Our primary measure was a Frailty Index, based on cumulative disorders. A proportion of participants were also assessed with the ‘Frail non-disabled’ questionnaire. We evaluated concurrent validity of Frailty Index against variables associated with frailty in non-stroke populations. We described predictive validity of Frailty Index for stroke severity and delirium. We described convergent validity, quantifying agreement between frailty assessments and a measure of pre-stroke disability (modified Rankin Scale) using kappa statistics and correlations.Results:We included 546 patients. A Frailty Index–defined frailty syndrome was observed in 427 of 545 patients (78%), of whom, 151 (28%) had frank frailty and 276 (51%) were pre-frail. Phenotypic frailty was observed in 72 of 258 patients (28%). We demonstrated concurrent validity via significant associations with all variables (all p < 0.01). We demonstrated predictive validity for stroke severity and delirium ( p < 0.01). Agreement between the frailty measures was poor (kappa = –0.06) and convergent validity was moderate (Frail non-disabled ‘Cramer’s V’ = 0.25; modified Rankin Scale ‘Cramer’s V’ = 0.47).Conclusion:Frailty is present in around one in four patients with acute stroke; if pre-frailty is included, then a frailty syndrome is seen in three out of four patients. The Frailty Index is a valid measure of frailty in stroke; however, there is little agreement between this scale and other measurements of frailty.

Background and Purpose— Delirium is a common and serious complication of acute illness. We describe delirium occurrence in an unselected, acute stroke population. Methods— We collected data from consecutive stroke admissions. We performed comprehensive cognitive assessment within the first week including Diagnostic Statistical Manual-5–based delirium diagnosis. We reported proportion with delirium and the clinical and demographic associations with delirium using multiple logistic regression. Results— Of 708 patients, median age of 71 years (interquartile range, 59–80), we recorded delirium in 187 of 708 (26.4%; 95% CI, 23.0–30.0). Across 395 patients with complete risk factor data (105 delirium), factors independently associated with delirium were: age (odds ratio, 1.05; 95% CI, 1.03–1.08), drug/alcohol misuse (odds ratio, 2.64; 95% CI, 1.10–6.26), and stroke severity (odds ratio, 1.22; 95% CI, 1.14–1.31). Conclusions— Delirium is common in acute stroke, affecting 1 in 4. It may be possible to predict those at risk using prestroke and stroke-specific factors. Clinical Trial Registration— URL: researchregistry.com . Protocol: 1147.

Objective: Post-stroke cognitive impairment is common, but mechanisms and risk factors are poorly understood. Frailty may be an important risk factor for cognitive impairment after stroke. We investigated the association between pre-stroke frailty and acute post-stoke cognition. Methods: We studied consecutively admitted acute stroke patients in a single urban teaching hospital during three recruitment waves between May 2016 and December 2017. Cognition was assessed using the Mini-Montreal Cognitive Assessment (min=0; max=12). A Frailty Index was used to generate frailty scores for each patient (min=0; max=100). Clinical and demographic information were collected, including pre-stroke cognition, delirium, and stroke-severity. We conducted univariate and multiple-linear regression analyses with covariates forced in (covariates included were: age, sex, stroke severity, stroke-type, pre-stroke cognitive impairment, delirium, previous stroke/transient ischemic attack) to investigate the association between pre-stroke frailty and post-stroke cognition. Results: Complete data were available for 154 stroke patients. Mean age was 68 years (SD=11; range=32–97); 93 (60%) were male. Median mini-Montreal Cognitive Assessment score was 8 (IQR=4–12). Mean Frailty Index score was 18 (SD=11). Pre-stroke cognitive impairment was apparent in 13/154 (8%) patients. Pre-stroke frailty was significantly associated with lower post-stroke cognition (Standardized-Beta=−0.40; p&lt;0.001) and this association was independent of covariates (Unstandardized-Beta=−0.05; p=0.005). Additional significant variables in the multiple regression model were age (Unstandardized-Beta=−0.05; p=0.002), delirium (Unstandardized-Beta=−2.81; p&lt;0.001), pre-stroke cognitive impairment (Unstandardized-Beta=−2.28; p=0.001), and stroke-severity (Unstandardized-Beta=−0.20; p&lt;0.001). Conclusions: Pre-stroke frailty may be a moderator of post-stroke cognition, independent of other well-established post-stroke cognitive impairment risk factors. (JINS, 2019, 25, 501–506)

Full completion of cognitive screening tests can be problematic in the context of a stroke. Our aim was to examine the completion of various brief cognitive screens and explore reasons for untestability. Data were collected from consecutive stroke admissions (May 2016–August 2018). The cognitive assessment was attempted during the first week of admission. Patients were classified as partially untestable (≥1 test item was incomplete) and fully untestable (where assessment was not attempted, and/or no questions answered). We assessed univariate and multivariate associations of test completion with: age (years), sex, stroke severity (National Institutes of Health Stroke Scale (NIHSS)), stroke classification, pre-morbid disability (modified Rankin Scale (mRS)), previous stroke and previous dementia diagnosis. Of 703 patients admitted (mean age: 69.4), 119 (17%) were classified as fully untestable and 58 (8%) were partially untestable. The 4A-test had 100% completion and the clock-draw task had the lowest completion (533/703, 76%). Independent associations with fully untestable status had a higher NIHSS score (odds ratio (OR): 1.18, 95% CI: 1.11–1.26), higher pre-morbid mRS (OR: 1.28, 95% CI: 1.02–1.60) and pre-stroke dementia (OR: 3.35, 95% CI: 1.53–7.32). Overall, a quarter of patients were classified as untestable on the cognitive assessment, with test incompletion related to stroke and non-stroke factors. Clinicians and researchers would benefit from guidance on how to make the best use of incomplete test data.