EXCR initiation after ischaemic stroke or TIA was associated with a lower hazard of incident dementia, and this association remained consistent across the 3-year, 30-day landmark, 1-year lagged and all-stroke-severity sensitivity analyses.
Key Findings
Results
Initiating exercise-based cardiac rehabilitation (EXCR) within 6 months after ischaemic stroke or TIA was associated with a lower 5-year hazard of incident dementia.
At 5 years, the hazard ratio for incident dementia was 0.70 (95% CI, 0.59-0.82).
After propensity score matching, 4138 patients remained in each cohort.
Matching balanced demographics, comorbidities, medication use and available stroke-severity measures.
Data came from the TriNetX global federated network; the study was observational.
Results
The dementia association was consistent across multiple sensitivity analyses.
Sensitivity analyses included a 3-year follow-up, a 30-day landmark, a 1-year lag, and inclusion of all stroke severities.
The 30-day landmark and 1-year lag analyses address possible reverse causation or immortal time bias.
Patients with NIHSS ≥16 (identified by ICD-10-CM codes) were excluded from the primary analysis but included in the all-severity analysis.
Results
EXCR was associated with substantially lower 5-year all-cause mortality.
Hazard ratio 0.50 (95% CI, 0.45-0.55) at 5 years.
Compared with matched patients without documented post-index EXCR.
Results
EXCR was associated with lower 5-year risk of recurrent stroke.
Hazard ratio 0.75 (95% CI, 0.61-0.91) at 5 years.
Results
EXCR was associated with a higher hazard of rehospitalisation at 5 years.
Hazard ratio 1.21 (95% CI, 1.06-1.38).
Results
No significant association was observed between EXCR and acute myocardial infarction.
The abstract reports no significant association at 5 years; no effect estimate is given in the abstract.
Results
Before matching, the EXCR and non-EXCR cohorts differed markedly in size and baseline characteristics.
Unmatched cohorts: 4226 patients initiating EXCR (mean age 64.7 years; 60.8% male) vs 1,889,796 without EXCR (mean age 66.3 years; 48.2% male).
Exposure was EXCR initiation within 6 months of the index event; comparators had no documented post-index EXCR.
Discussion
The authors conclude that benefits of post-stroke EXCR may extend beyond cardiovascular and functional outcomes to cognition.
The suggestion rests on the observed association with lower dementia risk.
Proposed mechanisms include vascular and inflammatory pathways relevant to cognitive decline.
What This Means
This study used a large international database of health records to look at whether people who started an exercise-based cardiac rehabilitation programme after a stroke or mini-stroke (TIA) were less likely to develop dementia. Researchers compared about 4,100 people who started rehabilitation within 6 months with an equal number of similar people who did not, matching them on age, sex, other health conditions, medications and stroke severity measures.
Over 5 years, those who took part in rehabilitation had about a 30% lower rate of new dementia diagnoses. They also had lower rates of death from any cause and of repeat stroke. Heart attack rates did not differ significantly. However, they were more likely to be hospitalised again, which may reflect closer contact with healthcare. The dementia finding held up in several extra checks designed to reduce bias, such as accounting for early events and including people with more severe strokes.
Because this is an observational study, it cannot prove that rehabilitation itself prevents dementia. People who join rehabilitation may be healthier or more motivated in ways that matching cannot fully capture. Still, this research suggests that exercise-based rehabilitation after stroke might have benefits for brain health in addition to heart health, which could be tested in clinical trials.
Serio L, Askarinejad A, Wang C, Ahmadzadeh K, Gu T, Proietti R, et al.. (2026). Exercise-based cardiac rehabilitation after stroke and the risk of incident dementia: an observational study from a global federated network.. European stroke journal. https://doi.org/10.1093/esj/aakag127