Drawing test 'predicts stroke death risk'

Thursday May 10 2012

“A simple drawing test may help predict the risk of older men dying after a first stroke,” says BBC News. The test asks participants to draw lines between a series of ascending numbers in as short a time as possible. The aim of the test is to indicate how well their minds are working.

In a new study published this week researchers looked at whether performance in the test could predict the risk of dying after suffering a stroke. In the study, the test, known as the Trail Making Test, was given to 919 older men at the start of the research. The participants were then followed using medical records for the next 14 years. In total, 155 participants had a stroke, of whom 84 died. When researchers examined the risk of dying in relation to men's cognitive test scores they found that doing poorly on the test was associated with an increased risk of dying following a stroke. The researchers say that the Trail Making Test offers an easy-to-use option for predicting death after a stroke.

Overall, this small study suggests that the simple test may offer an additional tool for identifying individuals at high risk of death from stroke. Given that the test does not require specialised equipment or extensive training it may, in theory, be helpful when used alongside other diagnostic techniques such as brain scans. However, the mechanism that explains the predictive power of this test is still uncertain, and the idea would benefit from testing in a more diverse group.

Where did the story come from?

The study was carried out by researchers from Uppsala University in Sweden and was funded by Uppsala University and the Swedish Stroke Association (STROKE-Riksforbundet). The study was published in the peer-reviewed American Journal of Cardiology.

The media reported the story appropriately, with the BBC pointing out that the study was relatively small and that the underlying causes of poor performance on the test are not known.

What kind of research was this?

This was a prospective cohort study in which a group of men were given a cognition test. The results were then analysed to assess how they related to the participants’ risk of dying from a stroke in the years that followed.

The researchers initially recruited 919 white men who had never had a stroke and asked them to complete the Trail Making Test (TMT), a simple cognitive test that involves drawing lines between numbers and letters in ascending order as quickly as possible. Participants performed two slightly different versions, A and B (TMT-A and TMT-B). TMT-A simply involves joining up ascending numbers scattered randomly across a page, while TMT-B adds letters to the task, and involves alternating between letters and numbers in ascending order, again, as quickly as possible. Requiring a long time to complete the tests is considered to reflect impairment in movements associated with mental activity.

The participants were then followed over time to see how their performance in the TMT tests related to their risk of dying of a stroke.

A prospective cohort study is necessary to determine the predictive or prognostic ability of a test. During research of this type, researchers can require participants to complete the test while healthy, and then follow them up to assess how their health changes. In this study that means the researchers were able to assess how well the TMT-A and TMT-B test predicted the participants’ risk of future stroke.

What did the research involve?

The research included 919 men between the ages of 69 and 75. At the beginning of the study information was also collected on medical history, alcohol habits, demographic factors and physical health status. Participants who had had a previous stroke were excluded from entering the study. The participants then completed the TMT-A and TMT-B, and their times were recorded.

The researchers followed the men for up to 13.6 years (median follow-up 11.2 years) and, using hospital discharge records and cause of death registries, recorded:

  • how many participants had a stroke during follow-up
  • how many of those who had a stroke died within two and a half years of it occurring

The researchers then compared the risk of dying among stroke patients according to TMT-A and TMT-B test performances at the start of the study. To do this, they split the cohort into three groups (or tertiles), with tertile 1 comprised of the men with the best (fastest) scores on the TMT-A and TMT-B tests, tertile 2 comprised of the men with intermediate scores and tertile 3 comprised of the men who performed the worst (slowest) on the tests.

During this analysis they controlled for multiple variables that had the potential to distort or influence the relationship between test performance and death risk, including age, education, social group and health status.

What were the basic results?

In all, 155 (16.9%) of the participants suffered a stroke or ‘mini-stroke’ during the follow-up period. A mini-stroke, also known as a transient ischaemic attack or TIA, occurs when the blood supply to parts of the brain is momentarily restricted. The event generally lasts a few minutes and causes similar symptoms to a stroke. Although a TIA can cause lasting effects, most symptoms generally resolve within a day or so. Having a TIA can be a warning sign that a person is at risk of having a stroke in the future.

On average (median) participants were followed for 2.5 years after their first ever stroke or TIA, and during this time 84 of the 155 men who suffered a stroke or TIA died (equating to 54% of patients who had a stroke dying). Twenty-two of the deaths occurred within the first month after the stroke or TIA.

The researchers found that diabetes (Hazard Ratio [HR] 1.67, 95% CI 1.04 to 2.69) and treatment for high blood pressure (HR 1.56, 95% CI 1.02 to 2.40) at the start of the study (the baseline) were significantly related to the risk of death after first ever stroke or TIA. No other variables at baseline were significantly associated with risk of death following stroke.

The researchers first assessed the relationship between performance on the TMT-A test (involving joining numbers only) and fatal stroke. They found that, overall, for each standard deviation increase in test time (about 20 seconds), the risk of dying after a first ever stroke or TIA increased by 88% (HR 1.88, 95% CI 1.31 to 2.71).

When comparing mortality between test time groups, the researchers found that the men who performed worst on the test were nearly three times more likely to have died following a stroke than those who performed the best (HR 2.90, 95% CI 1.24 to 6.77). There was no significant increase in mortality between men in the middle group and the best performers.

The researchers then examined the association between performance in the TMT-B test (involving both letters and numbers) and mortality after a stroke. They found that a standard deviation increase in test time (about 45 seconds) was associated with a significantly increased risk of mortality after a stroke (HR 2.01, 95% CI 1.28 to 3.15).

Compared with the fastest test performers, the slowest group were more than three times more likely to have died after a stroke (HR 3.53, 95% CI 1.21 to 10.34). Once again, there was no significant difference in mortality between the middle group and the fastest group.

How did the researchers interpret the results?

The researchers conclude that levels of cognitive functioning before a stroke, assessed using a simple test, predicted survival following a stroke in a sample of elderly men.


This study suggests that the results of a relatively simple test given at age 70 could predict the likelihood of dying after a stroke. This particular study did not, however, assess whether or not the TMT-A or TMT-B could itself predict the likelihood of having a stroke, as some internet coverage has suggested.

At present there are many risk factors that are used to identify individuals at an increased risk of having a stroke, including age, family history, ethnicity and medical history, as well as lifestyle factors such as smoking, excessive alcohol consumption and diet. This research suggests that the Trail Making Test may be useful for predicting outcomes after a stroke, although this particular paper did not provide data on the test’s ability to predict who will have a stroke in the first place. That said, the researchers highlight that previous research has shown that the TMT-B is also useful for predicting stroke in elderly men.

The study did have several strengths that allow us to be fairly confident in its results. First, the researchers were able to follow up on all of the study participants, which limits the likelihood of people dropping out of the study, biasing the results. Second, both the TMT test and the outcome of interest (death following a first stroke) were measured in a consistent manner across all the participants, and important potential confounders were accounted for in the data analysis.

The study sample was, however, not completely representative of the people that are likely to take such a test. While the age range of the sample is likely to be similar to that of patients who would be given this test, all of the study participants were white men. It is unclear whether the same or similar results would be seen in women or other ethnicities.

Also, as a test of observation and manual dexterity the TMT test may not be suitable for people with certain conditions such as sight problems or joint problems, which may hinder performance in the test. These are generally more prevalent among older people, who also have a greater risk of stroke, further complicating the issue.

The mechanism underlying this association is unclear, the researchers point out that having dementia before a stroke is known to be a predictor for stroke severity and death from stroke. So it is possible that this cognitive test is identifying early cases of sub-clinical dementia. This theory will, however, need further research.

Analysis by Bazian
Edited by NHS Choices