Cognitive impairment following stroke was found to improve most within the first 6 months but continued to evolve during long-term follow-up, with recovery trajectories differing substantially across language, memory, attention, executive function, and number processing domains, according to a longitudinal cohort study.
In the study, researchers recruited patients from the John Radcliffe Hospital stroke unit in England between 2012 and 2019. Patients underwent cognitive screening acutely within 2 weeks of stroke, at 6 months, and 2 years poststroke through the OX-CHRONIC long-term follow-up study.
Among 866 patients assessed acutely, 430 completed 6-month follow-up, 105 enrolled in long-term follow-up, and 98 completed Oxford Cognitive Screen (OCS) testing at the final timepoint. The mean age at stroke occurrence was 69 years, 41% of the patients were female, and about 84% of them had an ischemic stroke. The median acute National Institutes of Health Stroke Scale (NIHSS) score was 7.
In mixed-effects models, the researchers assessed global cognitive impairment by calculating the proportion of impaired OCS subtasks. Across the cohort, patients had impairment on an average of 23% of tasks acutely, 12% at 6 months, and 8% at chronic follow-up, corresponding to approximately three impaired tasks acutely, one to two impaired tasks at 6 months, and one impaired task at long-term follow-up.
Acute cognitive impairment severity was the strongest predictor of long-term cognitive outcomes. Adding acute impairment severity explained more than half of the variance in outcomes, while age, sex, education, lesion hemisphere, NIHSS score, atrial fibrillation, hypertension, diabetes, smoking status, stroke recurrence, and days poststroke added limited explanatory value.
Domain-specific models showed reductions in impairment over time across language, memory, attention, executive function, and number processing. These models were adjusted for time and baseline impairment status. Recovery was greatest for memory and language, while attention and executive function showed more limited improvement.
Latent class growth analysis identified four global cognitive trajectories. The researchers detailed that 48% of patients had stable mild or no impairment, 32% had moderate acute impairment with improvement, 15% had severe acute impairment with large improvement, and 5% had moderate acute impairment followed by decline.
Domain-specific trajectories showed additional heterogeneity. Language impairment was mild or absent over time in 82% of the patients, while 11% had early full recovery and 7% had later improvement. Memory impairment was mild or absent in 78% of the patients, with 15% showing ongoing recovery and 7% showing worsening. Attention trajectories included mild or no impairment in 61% of patients, early recovery in 19%, delayed recovery in 10%, and progressive decline in 11%. Executive function showed stable performance in 66% of the patients, fluctuating performance in 13%, and late recovery in 21%.
The researchers noted that acute language and memory impairments were more common with left hemisphere lesions, while attention impairments were more common with right hemisphere lesions during the acute phase. These differences were not observed at long-term follow-up.
The study had several limitations, including substantial attrition over time, lack of prestroke cognitive data, and possible underrepresentation of patients with more severe stroke because the participants needed to concentrate for 20 minutes. Long-term testing was performed remotely because of disruptions caused by COVID-19, and the OCS praxis domain could not be assessed beyond 6 months.
“Distinct trajectories across all domains, ranging from early and delayed recovery to persistent deficits and late decline, underscore the need to move beyond static models of cognitive outcome,” wrote lead study author Elise Milosevich, DPhil, of the Investigative Medicine Division in the Radcliffe Department of Medicine at the University of Oxford, and colleagues.
The study received support from the Stroke Association, Alzheimer’s Research UK fellowship, National Institute for Health and Care Research (NIHR) Development and Skills Enhancement Award, NIHR Program Grant, NIHR Oxford Biomedical Research Centre, NIHR Advanced Fellowship.
Dr. Milosevich reported grants from Alzheimer’s Research UK. Co–study author Andrea Kusec, PhD, reported grants from the NIHR. Senior study author Nele Demeyere, PhD, reported being a developer of the OCS and disclosed NIHR grant support, employment with Elsevier Publishing and the University of Edinburgh, and consultancy payments from Brain Stimulation. Co–study author Sarah T. Pendlebury, DPhil, reported honoraria from Trondheim, Sydney, and LaTrobe Universities and royalties from Oxford University Press and Cambridge University Press.
Source: Stroke
