Sarah Tamsin Pendlebury

BACKGROUND: Cognitive impairment is common after stroke and linked to poor outcomes, yet long-term recovery or decline, particularly across specific cognitive domains, remains unclear. Most studies use brief global screeners with short follow-up, limiting insight into recovery patterns. This study aimed to characterize domain-specific cognitive trajectories over ≥2 years poststroke and identify predictors of persistent impairment. METHODS: Participants were recruited at a regional acute stroke unit (John Radcliffe Hospital, Oxford, United Kingdom; 2012–2019) and assessed acutely, at 6 months, and ≥2 years poststroke. The Oxford Cognitive Screen was administered at all timepoints. Global impairment severity was quantified by the proportion of Oxford Cognitive Screen subtasks impaired. Logistic mixed-effects models examined longitudinal change and predictors of domain-specific impairments (language, memory, attention, executive function, and number processing). Latent class growth analysis identified distinct cognitive trajectories. Models were adjusted for acute cognitive impairment severity and time. RESULTS: Of 866 patients assessed acutely, 105 were followed up at ≥2 years (98 with complete Oxford Cognitive Screen data; median, 4.1 [interquartile range, 3.3] years; mean age, 69 years, 41% female). Cognitive impairment severity improved substantially by 6 months (β=−0.11; P P P <0.001), while demographic and vascular factors explained minimal variance. Latent class growth analysis identified 4 overall trajectories: no or mild acute impairment with stability (47.6%), moderate-improving (32.3%), large improvement (15.2%), and decline (4.8%). Domain-specific improvements were greatest in memory (odds ratio, 16.40 [95% CI, 5.52–48.7]) and language (odds ratio, 8.17 [95% CI, 3.17–21.1]), more limited in attention (odds ratio, 5.41 [95% CI, 2.52–11.6]) and executive function (odds ratio, 4.14 [95% CI, 1.96–8.75]). Domain models revealed additional classes of persistent or delayed recovery, particularly in executive function and attention. CONCLUSIONS: Cognitive recovery is most pronounced within 6 months and continues across domains though executive dysfunction often persists. Acute impairment severity best predicted long-term outcomes, while vascular and demographic factors were less informative. Distinct trajectory classes highlight the need for individualized, long-term cognitive monitoring to guide rehabilitation and prognostication. These findings underscore the importance of long-term cognitive follow-up in stroke care and provide empirical benchmarks for recovery across domains.

Introduction: Out-of-hours (OOH) services provide urgent primary care outside normal General Practice (GP) hours, serving patients who cannot wait for routine service. Delirium is commonly associated with acute illness, causes distress and leads to poor outcomes. However, little is known about delirium presentations in OOH services. We aimed to investigate using records from an OOH service in South-West England. Methods: The database contained 33 345 contacts with patients ≥65 years attending the OOH service. We screened consultations during April and July 2019, and January 2020 using an automated search then clinical review by two independent GPs. We validated our search strategy by reviewing a random sample of 100 ‘search-negative’ consultations initially and assessed inter-rater reliability. Patient characteristics were compared using Chi-squared tests. Results: Of 4288 consultations with patients ≥65 years in the study periods, 402 (9.4%) involved possible or probable delirium. A further 74 (1.7%) had end-of-life delirium and were excluded from further analysis. Patients with delirium were older (mean age 84.4 years vs. 80.1 years), and more often had dementia (46.6% vs. 10.4%, P < .001). 67.9% of delirious patients required home visits, compared to 22.2% without delirium (P < .001). Patients with delirium were admitted to hospital twice as often as those without (20.6% vs. 8.5%, P < .001). Conclusions: Delirium is a common OOH presentation, representing ~10% of consultations with patients ≥65 years. These patients often have cognitive impairment, require home visits and are more likely to be admitted to hospital. These findings are important for planning urgent care services tailored to the needs of older people.

BACKGROUND: Depression is common after stroke and is associated with increased mortality. However, there are few data on the prevalence, predictors, or prognosis of depression after transient ischemic attack (TIA). Although poststroke depression is often assumed to be due to the brain lesion or disability, patients with TIA may also be susceptible due to shared risk factors, lifestyle/medication changes, or worries about stroke. METHODS: The prevalence of depression 1 and 12 months after TIA was assessed in a population-based cohort (OXVASC [Oxford Vascular Study]) from 2014 to 2020. Depression was related to risk factors (including infarction on brain imaging) in a multivariable logistic regression model and outcome (disability, quality of life, institutionalization, recurrent events, and mortality) in a Cox proportional hazard regression model during 5-year follow-up after adjustment for covariates that were found to be significant ( P P =0.022). Depression was independently associated with younger age (adjusted odds ratio/decade, 0.74 [95% CI, 0.60–0.90]), low baseline mood (4.06 [95% CI, 2.31–7.15]), past depression (1.81 [95% CI, 1.09–3.03]), multimorbidity (1.19 [95% CI, 1.02–1.39]), living alone (1.94 [95% CI, 1.14–3.32]), disability (3.53 [95% CI, 1.89–6.59]), and deprivation (1.28 [95% CI, 1.03–1.59]). After adjustment for confounders, depression did not predict risk of recurrent vascular events (adjusted hazard ratio [aHR], 1.42 [95% CI, 0.76–2.64]; P =0.27) but did predict increased 5-year all-cause mortality (aHR, 2.27 [95% CI, 1.21–4.27]; P =0.011), independently of acute lesions on brain imaging (aHR, 2.18 [95% CI, 1.09–4.34]; P =0.027), and particularly when persistent (present at 1- and 12-month follow-up; aHR, 4.58 [95% CI, 2.07–10.13]; P P P =0.003) within 5 years and with reduced quality of life (coefficient, −0.245; P <0.001). CONCLUSIONS: Depression is common early after TIA, independent of acute ischemia on brain imaging, and persistent depression is also strongly associated with adverse outcomes.

Background: Cognitive impairment is common after stroke and linked to poor outcomes, yet long-term recovery or decline, particularly across specific cognitive domains, remains unclear. Most studies use brief global screeners with short follow-up, limiting insight into recovery patterns. This study aimed to characterize domain-specific cognitive trajectories over ≥2 years post-stroke and identify predictors of persistent impairment. Methods: Participants were assessed acutely, at 6 months, and ≥2 years post-stroke (Median = 4.1 years [IQR = 3.3]). The Oxford Cognitive Screen (OCS) was administered at all timepoints. Global impairment severity was quantified by the proportion of OCS subtasks impaired. Logistic mixed-effects models examined longitudinal change and predictors of domain-specific impairments (language, memory, attention, executive function, number processing). Latent class growth analysis (LCGA) identified distinct cognitive trajectories. Results: Of 105 patients assessed, 98 had complete OCS data. Cognitive impairment severity improved substantially by 6 months (β = −0.11) and further long term (>2 years; β = −0.15). Acute impairment severity strongly predicted long-term outcomes (β = 0.50; p < 0.001), while demographic and vascular factors explained minimal variance. LCGA identified four overall trajectories: no or mild acute impairment with stability (47.6%), moderate-improving (32.3%), large improvement (15.2%), and decline (4.8%). Domain-specific improvements were greatest in memory (OR = 16.40) and language (OR = 8.17), more limited in attention (OR = 5.41) and executive function (OR = 4.14). Domain models revealed additional classes of persistent or delayed recovery, particularly in executive function and attention. Conclusions: Cognitive recovery is most pronounced within 6 months and continues across domains, though executive dysfunction often persists. Acute impairment severity best predicted long-term outcomes, while vascular and demographic factors were less informative. Distinct trajectory classes highlight the need for individualized, long-term cognitive monitoring to guide rehabilitation and prognostication. These findings underscore the importance of long-term cognitive follow-up in stroke care and provide empirical benchmarks for recovery across domains.

Quantification of brain atrophy currently requires visual rating scales which are time consuming and automated brain image analysis is warranted. We validated our automated deep learning (DL) tool measuring the Global Cerebral Atrophy (GCA) score against trained human raters, and associations with age and cognitive impairment, in representative older (>65 years) patients. CT-brain scans were obtained from patients in acute medicine (ORCHARD-EPR), acute stroke (OCS studies) and a legacy sample. Scans were divided in a 60/20/20 ratio for training, optimisation and testing. CT-images were assessed by two trained raters (rater-1=864 scans, rater-2=20 scans). Agreement between DL tool-predicted GCA scores (range 0-39) and the visual ratings was evaluated using mean absolute error (MAE) and Cohen's weighted kappa. Among 864 scans (ORCHARD-EPR=578, OCS=200, legacy scans=86), MAE between the DL tool and rater-1 GCA scores was 3.2 overall, 3.1 for ORCHARD-EPR, 3.3 for OCS and 2.6 for the legacy scans and half had DL-predicted GCA error between -2 and 2. Inter-rater agreement was Kappa=0.45 between the DL-tool and rater-1, and 0.41 between the tool and rater- 2 whereas it was lower at 0.28 for rater-1 and rater-2. There was no difference in GCA scores from the DL-tool and the two raters (one-way ANOVA, p=0.35) or in mean GCA scores between the DL-tool and rater-1 (paired t-test, t=-0.43, p=0.66), the tool and rater-2 (t=1.35, p=0.18) or between rater-1 and rater-2 (t=0.99, p=0.32). DL-tool GCA scores correlated with age and cognitive scores (both p<0.001). Our DL CT-brain analysis tool measured GCA score accurately and without user input in real-world scans acquired from older patients. Our tool will enable extraction of standardised quantitative measures of atrophy at scale for use in health data research and will act as proof-of-concept towards a point-of-care clinically approved tool.

Background: Delirium is underdiagnosed in older inpatients. We determined clinically ascertained delirium prevalence, characteristics and outcomes after implementation of mandatory cognitive/delirium screening via the electronic patient record (EPR) compared to previously acquired prospective cohorts. Methods: Acute general medicine patients aged ≥70 years in two separate cohorts were included (i) the Oxford Cognitive Comorbidity, Frailty and Ageing Research Database (ORCHARD-EPR, 2017–2019) and (ii) prospective cohorts (2010–2018). The ORCHARD-EPR cognitive screen included the Confusion Assessment Method, and 10-point Abbreviated Mental Test (AMT) with delirium diagnosis = certain/uncertain/no. In the prospective cohorts, delirium was diagnosed by Diagnostic and Statistical Manual of Mental Disorders, 4th Edition criteria. Odds ratios for in-hospital mortality were adjusted for age, sex, comorbidity and illness severity. Results: Among 18 614 ORCHARD-EPR patients (mean/SD age = 82.9/7.4 years, dementia = 20%), certain delirium (delirium = yes or ICD-10 delirium code) was present in 3077/18 614 (17%,95%CI 16%–17%) of whom 1199/3077 (39%) had dementia, and uncertain delirium (delirium = uncertain and no ICD-10 code) in 2007 (11%,10%–11%) of whom 773/2007 (39%) had dementia. In the prospective cohorts (n = 731, mean/SD age = 82.7/7.1 years, dementia = 21%), 277 (38%, 34%–42%, both P < 0.001) had delirium of whom 97 (35%) had dementia. Frailty, AMT < 8, infection and hyponatraemia (Standardised Mean Difference-SMD all >0.1) but not comorbid dementia (SMD = 0.009) were more common in certain vs uncertain delirium. Excess mortality was similar in certain and uncertain delirium and the prospective cohorts. Discussion: Delirium diagnostic uncertainty was frequent particularly in the absence of key delirium risk factors and delirium prevalence, even combining both certain and uncertain cases, was lower than in the prospective cohorts. Nevertheless, co-occurrence with dementia and outcomes were similar indicating validity of screening and research potential of ORCHARD-EPR.

This cohort study assesses the frequency of screen completion and the prevalence and outcomes of delirium diagnosis for older patients with unplanned admission to acute general (internal) medicine.