A new risk prediction tool can accurately identify stroke survivors with the highest risk of developing dementia within a decade after a stroke, according to a large Canadian study to be presented at the American Stroke Association’s International Stroke Conference 2026. The research, which analyzed health records of nearly 50,000 adults hospitalized with stroke, aims to help researchers design better clinical trials and interventions by stratifying patients into different risk categories.
Lead study author Raed A. Joundi, M.D., D.Phil., M.Sc., explained that approximately one in three adults develops dementia after stroke over the long term. The new tool stratifies people into five different levels of dementia risk based on underlying health, stroke characteristics, and risk factors. "The goal is to have a practical, bedside tool that can predict dementia risk after a stroke," Joundi said. "Our tool predicts dementia rates that are very close to the observed rates and may help to enroll high-risk patients who have had transient ischemic attack, ischemic stroke or intracerebral hemorrhage in clinical trials that are focused on reducing the long-term risk of dementia."
Researchers examined data from the Ontario Stroke Registry, including hospital admissions due to stroke between 2002 and 2013. Study participants included 7,554 adults with transient ischemic attack (TIA), 13,833 with ischemic stroke, and 2,340 with intracerebral hemorrhage. All participants were discharged without a dementia diagnosis and were followed for an average of 7.5 years after stroke through March 2024.
The analysis identified key factors associated with increased dementia risk. For people who had a transient ischemic attack, these included older age, needing help with activities of daily living prior to TIA, having diabetes, depression, cognitive symptoms on presentation, and any disability at hospital discharge. For stroke patients, main risk factors were being older, being female, having diabetes, depression, intracerebral hemorrhage (compared to ischemic stroke), cognitive symptoms during hospitalization, or greater disability at discharge.
The risk calculator categorizes individuals into different levels of estimated risk over the next ten years after a stroke. Those in the highest category had a 50% probability of dementia over ten years, compared to just 5% for those in the lowest risk category. The tool was validated using data from the Ontario Stroke Audit, a separate sample of patients from all hospitals in the province.
Joundi emphasized that dementia is more common than recurrent stroke over the long term and requires greater attention. "While our traditional focus has been on preventing another stroke, which is very important, we need to pay more attention to the development of dementia and how to prevent it," he said. "Healthy lifestyle choices and controlling vascular risk factors can lower the risk of dementia, but we need new and effective targeted interventions for dementia prevention."
American Stroke Association volunteer expert Deborah A. Levine, M.D., M.P.H., who was not involved in the study, noted the tool's potential impact. "Dementia after a stroke is very difficult for patients and their loved ones, and there aren’t enough effective treatments to help," Levine said. "This well-done study provides a useful tool that could make research faster, so new treatments can get to stroke survivors sooner."
The study authors clarify that the current focus of the dementia risk prediction tool is to stratify patients for research studies and clinical trials of dementia prevention, rather than for clinical decision-making or treatment. Study limitations include the lack of data about the type of dementia that may develop and the absence of imaging scans that would offer more detailed information about stroke characteristics.
The findings are considered preliminary until published as a full manuscript in a peer-reviewed scientific journal. Abstracts presented at the American Heart Association/American Stroke Association’s scientific meetings are not peer-reviewed. Additional information about the study can be found in the American Stroke Association International Stroke Conference 2026 Online Program Planner.
