Dementia risk after TIA stays raised for years and nears the post-stroke risk

Design

Population-wide matched cohort study using linked administrative data (Ontario, 2002 to 2022)

Population

113,081 adults with first TIA matched to controls and to patients with stroke

Primary outcome

Incident dementia at 1, 5 and 10 years and overall, with competing risk of death

Effect

Overall HR 1.34 (95% CI 1.31 to 1.36); first-year HR 1.75 (1.66 to 1.85); 2.71 vs 2.03 per 100 person-years

This Ontario population cohort used linked administrative data from 2002 to 2022. It followed 113,081 adults with a first TIA, mean age 70, from 90 days after the event, and matched each to population controls on age, sex, deprivation, rurality and vascular comorbidities. A separate match compared TIA with first stroke. People with prior dementia, prior stroke, early death or long-term care were excluded.

Dementia was diagnosed in 19.3% over a mean follow-up of 7.1 years. The rate after TIA was 2.71 per 100 person-years vs 2.03 in controls. The hazard ratio was highest in the first year (1.75, 95% CI 1.66 to 1.85) and remained raised across 20 years (overall 1.34, 1.31 to 1.36). Compared with stroke, the dementia hazard after TIA was lower in year one (HR 0.67, 0.64 to 0.70) but became similar after about 5 years. Adjusting for later stroke changed results little, and dementia was far more frequent than recurrent stroke.

This is an association from administrative data: dementia was ascertained from diagnoses, claims and prescriptions, and cognitive testing was not done. It cannot show that preventing stroke alone prevents dementia, which the authors say needs strategies beyond stroke prevention.

• After a TIA, ask about memory and function at follow-up visits, not only about recurrent events.
• The risk is highest in the first year (HR 1.75) and persists for decades.
• Document baseline cognition after TIA so later change can be judged.
• Address vascular risk factors that overlap with dementia risk, such as blood pressure, lipids and glucose.
• Counsel patients that the long-term cognitive risk is an association from population data.

Why it matters

It challenges the idea that the main long-term risk after TIA is another stroke.

Don't overread it

Dementia was identified from coded health data, so mild or undiagnosed cases are missed and the findings show association rather than cause.

The statistics, in plain English

A hazard ratio of 1.34 means a third higher rate of dementia than matched controls at any given time, not that one in three patients develops it. The absolute rate difference was 0.68 per 100 person-years. Administrative coding can miss mild dementia and over-capture treated cases.

Read the rest in the app

You have read your two free briefings this month. The app carries all 27 specialties, every morning, free — and this finding is waiting in it.

Get Daily Doctor — free

Scan to keep reading on your phone. No account needed to start.