In a recent study published in The Lancet Healthy Longevity, researchers performed a meta-analysis to analyze aspects that increase post-stroke cognitive impairment (PSCI) and dementia (PSD) risks.

The increasing variety of stroke survivors worldwide has highlighted the long-term consequences of stroke, particularly cognitive impairments and dementia. This burden on patients, caregivers, and healthcare systems is important.

Understanding the aspects predisposing individuals to stroke-related complications is crucial for patient counseling and prevention trials. Nevertheless, systematic reviews and meta-analyses are scarce, and there may be an absence of pooled estimates for PSCI independent of well-recognized risk aspects equivalent to stroke severity and patient age.

Concerning the study

In the current meta-analysis, researchers explored PSCI and PSD risk aspects aside from stroke severity and age.

The team searched the Cochrane and MEDLINE databases for articles on risk aspects for post-stroke cognitive deficits published in English through September 15, 2023. A researcher screened the information, and one other researcher resolved discrepancies.

Studies included observational studies, retrospective and prospective cohort-type research, post-hoc evaluation from randomized clinical trials, and case-control-type studies of acute stroke patients (hemorrhagic, ischemic, and transient ischemic attack [TIA]), exploring aspects contributing to PSD or PSCI risks after 3 months follow-up.

Included records reported summary estimates for binary outcomes (PSD or PSCI) based on predetermined diagnostic criteria, or thresholds in neuropsychological tests, or each; include ≥30 adult patients; and assess risk aspects inside the first three months after stroke. The researchers also searched the reference lists of the included articles to discover additional records.

They used the Newcastle-Ottawa scale (NOS) to evaluate the standard of the included studies. They determined the pooled values for relative risks (RRs) by random-effects-type meta-analytical research, including meta-regressions, sensitivity, and subgroup analyses.

Exclusion criteria included animal studies, cross-sectional studies, randomized clinical trials, stroke patients with specific brain regions, subjective studies, cohorts with pre-stroke dementia, cognitive impairment or diseases which will alter cognitive function, and studies without stroke severity and age adjustments. 

The team also excluded studies specializing in cognitive impairment, genetic predispositions, stroke-free controls, continuous cognitive outcomes, trajectories of cognitive performance, domain-specific performance or impairment, and meta-analysis studies without stroke severity and age adjustments.

Results

Out of 13,127 initially identified records, 162 were eligible for the systematic review and 113 (89 studies, including 160,783 stroke patients) for the meta-analysis. The median NOS rating for the included studies was 5. 

Most studies were hospital-based (73 studies, 29,341 patients), while fewer were population- or registry-based. PSCI effect estimates were correlated significantly with those for PSD, with a beta regression coefficient of 0.7 indicating larger effect sizes for PSD than PSCI.

Cognitive impairment at baseline contributed essentially the most to PSCI (relative risk [RR], 2.0) and PSD risks (RR, 3.1). The researchers identified diabetes (RR, 1.3), atrial fibrillation history or presence (RR, 1.3), moderate to severe hyperintensities in white matter (RR, 1.5), and white matter hyperintensity severity (RR, 1.3) as modifiable-type PSCI risk aspects, regardless of stroke severity and age.  

The researchers identified diabetes (RR, 1.4), moderate to severe white matter hyperintensities (RR, 1.6), and white matter hyperintensity severity (RR, 1.6, 1) as modifiable risk aspects for PSD. Other risk aspects were a lower level of education, prior stroke history, brain atrophy, stroke within the left hemisphere of the brain, the presence of a minimum of three lacunes, and low functional status at baseline. The researchers found considerable heterogeneity and reporting bias within the included studies.

Within the subgroup evaluation, studies on diabetes and atrial fibrillation showed larger effect sizes for post-dementia stress disorder (PDS) when published before 2009 in a hospital-based setting using a neuropsychological test battery relatively than cognitive screening tools.

Meta-regressions showed that a later recruitment date attenuated the association of the National Institute of Health Stroke Scale/Rating (NIHSS), educational attainment, and white matter hyperintensity severity with PSD, and a later publication date attenuated the association of NIHSS, educational attainment, and atrial fibrillation with PSD.

Conclusions

Overall, the study findings showed that cognitive impairments within the acute period following stroke are the strongest predictor of post-stroke stroke complications (PSCI) and stroke-related stroke complications (PSD). Diabetes was identified because the essential cardiovascular risk factor for cognitive impairments and dementia after stroke, while the role of atrial fibrillation in PSD development is less clear.

Other aspects for post-stroke cognitive decline include lower education, prior stroke history, cerebral small vessel disease (cSVD)-associated neuroimaging markers, left hemisphere stroke, urinary incontinence, medial temporal lobe atrophy (MTA), and lower functional status and cognitive performance. Screening for cognitive dysfunction through the acute period after stroke may aid in identifying individuals at increased risks of long-term post-stroke cognitive impairments and dementia.