One in five developed dementia in the years after a mini-stroke, against 15% of matched peers, in Ontario

A mini-stroke is over within hours and leaves no mark on a brain scan, and the main concern afterward has been that a full stroke may follow. A study of every such case in Ontario over two decades points to a different outcome. Counting from three months after a first mini-stroke, dementia was diagnosed about three times as often as a stroke.

Of 113,081 people followed for an average of seven years, 19.3% were diagnosed with dementia. Among people of the same age, sex and health who had not had a mini-stroke, the figure was about 15%. The study was published on September 23 in Neurology.

It shows that the two go together. It does not show that one leads to the other, and an independent study in the Netherlands, smaller but with more detail on each patient, did not find the same thing.

What a TIA, or mini-stroke, is

The medical name is transient ischemic attack (TIA). A TIA is when blood flow stops to a part of the brain for a short time, causing stroke-like symptoms that go away within a day with no stroke visible on brain scans, in the definition given by the American Academy of Neurology.

Doctors treat it as an emergency because of what can come next. Transient ischemic attack is a strong predictor of subsequent stroke, a statement from the American Heart Association says. The 90-day stroke risk after transient ischemic attack can be as high as 17.8%, with almost half occurring within 2 days of the index event. The index event is the first attack.

Among patients seen quickly at specialist clinics the figures were lower. An international registry of TIA clinics in 21 countries that enrolled 4789 patients with a TIA or minor ischemic stroke followed them for five years. The rate of cardiovascular events including stroke in a selected cohort was 6.4% in the first year and 6.4% in the second through fifth years. The second figure covers the four later years together.

Memory and thinking received less attention. In the Ontario team’s words, the long-term risk of dementia after TIA remains poorly characterized.

How the TIA and dementia study was done

Ontario runs a single public health system, and its records can be linked. The team conducted a population-wide cohort study using linked administrative health data for the province of Ontario, Canada (2002-2022). Adults with a first TIA were identified from emergency department visits or hospital admissions and followed from 90 days after the index event.

That 90-day gap was deliberate. It leaves out dementia picked up during the investigations that follow an attack. Individuals with prior dementia, prior stroke, early death, or long-term care admission were excluded.

Each person was paired with a similar person who had not had a TIA or stroke. Patients with TIA were matched to population controls on age, sex, neighborhood-level deprivation, rurality, and vascular comorbidities. In plain terms, the comparison person was the same age and sex, lived in a similarly poor or well-off, urban or rural area, and had the same recorded conditions of the heart and blood vessels, such as high blood pressure and diabetes. A second comparison paired them with people who had had a first stroke.

Nobody was examined for the study. Dementia was identified using a validated algorithm based on hospital diagnoses, physician claims, and dementia-specific medications. The average age at the time of the TIA was 70, and half the group were women.

What the study found on dementia after TIA

Among the people with a TIA, dementia was diagnosed in 19.3% over a mean follow-up of 7.11 years. “In our study, over an average of seven years, one in five people who had a TIA developed dementia,” said Raed Joundi of McMaster University, the study’s lead author.

Rates of dementia after TIA were consistently higher than in matched controls across all time horizons (overall rate 2.71 vs 2.03 per 100 person-years). A person-year is one person followed for one year. Put another way, among 100 people followed for a year there were about 2.7 new dementia diagnoses after a TIA and about 2.0 without one.

Group	Diagnosed with dementia over about seven years
First TIA	19%
Matched people with no TIA or stroke	15%
Matched people with a first stroke	20%

The American Academy of Neurology, which publishes the journal, set the three figures side by side in its news release: 19% of people with TIA had a new diagnosis of dementia, compared to 15% of those without stroke or TIA and 20% of those with a stroke.

The extra risk was not spread evenly over time. It was highest within the first year, when the rate of dementia was 75% higher than in the matched group, with a plausible range of 66% to 85%. It then narrowed but did not disappear. The rate stayed higher over 20 years of follow-up, at 34% higher across the whole period.

How TIA compared with stroke

A full stroke destroys brain tissue, and its link with dementia is long established. A TIA by definition does not, so the comparison between the two is the most surprising part of the study.

In the first year, people who had a TIA were less likely to be diagnosed with dementia than people who had a stroke, by about a third. In the paper’s words, the risk after TIA was lower early on but converged and became similar after 5 years of follow-up.

The study also compared the two outcomes a TIA patient might face. Dementia incidence substantially exceeded stroke incidence, the paper reports, incidence meaning the number of new cases. The academy’s release puts a number on it: people who had a TIA were three times more likely to develop dementia over the entire follow-up period than to have a stroke.

“After a transient ischemic attack, there is a well-established increased risk of stroke, but less is known about the risk of dementia,” Joundi said. “In addition, dementia occurs far more frequently after a TIA than having a stroke.”

Where this fits: what other TIA studies found

Three independent studies bear on the result, and they do not all agree.

The most detailed comes from Oxford, where researchers tracked every stroke and TIA in a defined population and tested patients’ thinking in person. Dementia was diagnosed on the basis of cognitive testing supplemented by data obtained from hand searches of all hospital and primary care records. During follow-up 432 patients developed dementia. Five years on, 16.2% of those who had a TIA and 33.1% of those who had a stroke had a new diagnosis, a wider gap than in Ontario.

The Oxford team expressed the effect as time. Compared with the general population, the onset of dementia was brought forward by approximately 25 years in those who had severe strokes, 4 years in those who had minor strokes, and about 2 years after a TIA. That is an excess, and a far smaller one than after a serious stroke.

A study from Alabama looked at thinking skills directly. It drew on a national cohort whose members took memory and word tests by telephone twice a year, and so had results from before anything happened. It included 356 individuals with first-time TIA. Unlike scores after a stroke, theirs did not drop at the time of the attack. But the annual decline after the index event was faster than in people who had neither. The authors concluded that despite the quick resolution of stroke symptoms in TIA, there was apparently sufficient impact to be associated with long-term cognitive decline.

The Rotterdam Study points the other way. It has followed residents of one Dutch district for decades with repeated examinations and brain scans. In an analysis published this March, it followed 547 people after a first TIA and matched each to three people of the same age and sex. Compared with reference participants, dementia risk was increased after minor and major stroke, but not after TIA. The rate after a TIA was 0.97 times that of the comparison group, with a plausible range of 0.76 to 1.23.

The Dutch study was 200 times smaller than the Ontario one, so it could have missed a modest excess. It also had something the Ontario records lacked, which was direct knowledge of each person’s brain and thinking before the event.

What the Ontario TIA study leaves unanswered

It cannot show that a TIA leads to dementia. The release is explicit: the study does not prove that a transient ischemic attack causes dementia, it only shows an association.

The likeliest alternative is shared disease. Narrowed and damaged small blood vessels in the brain can produce both a TIA and, over years, dementia. On that reading the TIA is a signal of trouble already under way. The study could not test this. A limitation of the study was that researchers did not have detailed brain imaging to determine if people had hidden disease at the start of or during the study, the release says.

Diagnosis is a second problem. A TIA is hard to confirm, since the symptoms have gone by the time a doctor examines the patient, and some attacks recorded in emergency departments will have been something else. Separately, a person who has had a TIA sees doctors more often, and memory problems may be noticed and recorded sooner than in someone who stays away. That could account for part of the first-year peak.

Dementia was identified from billing codes and prescriptions, not from testing. Mild cases are missed by such methods, and matching on recorded conditions cannot capture education, smoking or how well blood pressure was controlled.

What follows for patients is not settled by this study. Joundi’s view is that care after a TIA has been built around preventing stroke and that dementia belongs on the list. Whether the treatments already given after a TIA, such as blood pressure control, also lower dementia risk was not examined.

Dementia was diagnosed more often after a first TIA than in matched peers across 20 years of Ontario health records, and without brain scans those records cannot separate the attack from the blood vessel disease that may lie behind both.