Share this article on:

Neuroimaging Found to Improve Prognosis on TIA and Stroke

Rukovets, Olga

doi: 10.1097/
Back to Top | Article Outline


INVESTIGATORS wanted to know if the presence or absence of brain infarction on CT or MRI would provide a reliable prognosis for stroke after transient ischemic attack

INVESTIGATORS wanted to know if the presence or absence of brain infarction on CT or MRI would provide a reliable prognosis for stroke after transient ischemic attack

Investigators report findings that support a tissue-based definition of TIA for prognosis.

For nearly a decade, neurologists have been debating how best to assess transient ischemic attacks (TIA) and determine the likelihood of a stroke. Is the resolution of symptoms within 24 hours — time-based criteria — sufficient? Will the presence or absence of brain infarction on CT or MRI — a tissue-based definition — provide a reliable prognosis? And how predictive are ABCD2 scores, a seven-point scale that assesses five clinical measures — including Age (60 years or older), Blood pressure, Clinical features of TIA, Duration, and Diabetes — in forecasting stroke risk after TIA? [See the “ABCD2 Scoring System.”]

In the Aug. 24 online edition of Neurology, investigators tried to answer these questions by analyzing data from 12 centers that provided ABCD2 scores, MRI-diffusion-weighted imaging (DWI) or CT brain imaging, and follow-up stroke rates from 4,574 patients with TIA diagnosed by time-based criteria — at seven and 90 days.



“The studies [so far] have been reasonably small, so the intention was to do a big, multicenter study and to try and draw together clinical features and investigation results in terms of infarction on brain imaging,” the lead author Matthew F. Giles, DPhil, of the Stroke Prevention Research Unit at the National Institute of Health Research Biomedical Research Centre of Oxford University, told Neurology Today. “We found the ABCD2 score to be predictive in TIA both with and without evidence of infarction on brain imaging.” [For more specific data, see “By the Numbers: What Predicts Stroke After TIA.”]

But the study also showed that identifying a small area of brain injury on the MRI was much more powerful than only using the ABCD2 score in predicting early stroke risk, said co-author Gregory Albers, MD, director of the Stanford Stroke Center and Coyote Foundation Professor of Neurology and Neurological Sciences. “The one week stroke risk is at least ten times higher in those who have a small area of tissue injury on the MRI versus those who have a normal MRI.”

This allows you to identify which patients should be evaluated in the hospital — or very close by — so tissue plasminogen activator (tPA) can be administered rapidly if they do have the stroke, Dr. Albers said. “An important point to emphasize is that tPA is not contraindicated in these patients, despite the small DWI lesion,” he noted.

For the patients who showed no infarction on MRI, the 7 day stroke risk, even when the ABCD2 score was on the higher side (a score of 4-7 indicates intermediate to high risk for stroke) was less than 1.0 percent. Dr. Albers said it is possible that many of those patients didn't really have a TIA, but had experienced TIA-like symptoms from non-ischemic causes. The MRI provides a step towards being able to confirm whether or not the transient symptoms were truly caused by an ischemic event, he said.

Dr. Albers said part of the problem in diagnosing a TIA is the lack of “objective gold standard evidence because by the time the patient gets to you, the symptoms are gone, the exam is normal, and it can be very difficult even for a stroke specialist to say whether this was or wasn't a cerebral ischemic event.”

When the MR scan is positive, he said, “we can say this is a very high-risk patient, who is 10-15 times more likely to go on to have a subsequent stroke than if the MRI was negative. This is important because the highest risk time is the first 24 to 48 hours. So at Stanford, we admit the DWI-positive patients into the hospital for about 48 hours — that way, if they have recurrent symptoms, we can evaluate them immediately and, if appropriate, start tPA treatment.”

Back to Top | Article Outline


Kerstin Bettermann, MD, PhD, associate professor of neurology at the Penn State College of Medicine Milton S. Hershey Medical Center, was one of several stroke experts who praised the paper. On the plus side, she said, “it's very representative of different populations, looks at clinical settings taking into account the different practice styles — community-based, EDs, specialty centers…and it's a pretty compelling result.”

But she was less certain about how it would change clinical management of TIA. The findings from the study could be helpful for future research — for identifying patients at risk for recurring events and enrolling them into studies, she said. But, she added, there will still be some clinical ambiguity.

“People who present with TIAs or stroke-like symptoms require emergent work-up,” Dr. Bettermann said. “Even if the patient falls within a relatively small risk group by tissue-based definition and ABCD2 score, it's individually impossible to say who is at a high risk and who isn't. Even if an individual patient is in a group that only has 0.4 [percent] chance of a recurring event based on the new criteria, you don't want to miss that one patient who is going to develop a stroke later,” she explained. “This means we still need to individualize risk better, and be more aggressive with risk prevention in general.”





But, said another commentator, the study underscores the utility of using DWI in assessing patients with TIA. “I think that the combination of both the imaging information as well as the ABCD2 score is important in assessing long-term prognosis and the risk of stroke so that that can help us to triage the higher-risk patients, determine what their diagnostic evaluation needs to be, and guide secondary prevention,” said Chelsea Kidwell, MD, medical director of the Georgetown Stroke Center in Washington, DC.

The biggest challenge, she said, is getting the larger medical community to understand the importance of early imaging and prognosis because the time-based definition is so ingrained in the medical field. And then there is the question about access to timely imaging. Dr. Bettermann said she practices at a tertiary medical care center where advanced imaging is readily available, but she noted that family practices or those in rural areas might not have that same level of access.

“Certainly in the UK, the new definition has not been widely taken up,” Dr. Giles acknowledged. “In a lot of patients, MR imaging is not tolerated or is contra-indicated if they have pacemakers, so that limits how the definition can be applied. Having said that, CT scanning, although it's not recommended, does give some useful information in terms of early prognosis, and that's a result which had not really been shown before in large trials.”

Back to Top | Article Outline


In an accompanying editorial in Neurology, Kevin M. Barrett, MD, assistant professor of neurology at the Mayo Clinic in Jacksonville, FL, wrote: “More studies of TIA prognosis using a tissue-based definition will be necessary to better understand the clinical importance of differentiating between TIA and minor stroke with rapid and complete recovery.”

There is a definite need for more research, particularly on the use of perfusion CT imaging, Dr. Bettermann agreed. “We also need to study strokes that are clinically present but are not image-able,” she said, adding that this is where biomarkers may soon play a significant role.

Back to Top | Article Outline


The ABCD2 score, which calculates stroke risk after two days, involves assessment of five independent factors: age 60 or older (1 point); blood pressure: systolic BP greater or equal to 140 mmHg OR diastolic BP greater or equal to 90 Hg (1 point); clinical features of TIA (choose one) — unilateral weakness with or without speech (2 points) or speech impairment without unilateral weakness (1 point); TIA duration greater than 60 minutes (2 points) or TIA duration between 10 and 59 minutes (1 point); history of diabetes (1 point). Scores 0-3 are considered low risk; 4-5, intermediate risk, and 6-7, high-risk for subsequhhent stroke 2, 7, 30, and 90 days after TIA.

—Source: Johnston SC, Rothwell PM, Sidney S, et al. Validation and refinement of scores to predict very early stroke risk after transient ischemic attack. Lancet 2007; 369:283-292.

Back to Top | Article Outline


In 2002, Gregory Albers, MD, and other cerebrovascular experts called for a revision in the definition of TIA — from time-based (the resolution of symptoms in 24 hours) — to the presence or absence of brain infarction — a tissue-based definition — on neuroimaging.

Dr. Albers, director of the Stanford Stroke Center and Coyote Foundation Professor of Neurology and Neurological Sciences said that in 2002, “some people would wait around hoping the symptoms would disappear because there was this arbitrary 24 hour limit — for example, a patient might present to an ER with a neurologic deficit that looked like a stroke, but instead of treating them immediately with tPA, the reaction was often, ‘well, it might be a TIA, why don't we wait?’”

Once the MRIs were more routinely obtained for these patients, he said, clinicians realized that in about a third of the cases, there is a small region of acute brain injury, even with the events as short as 30 minutes or an hour. In fact, he added, for transient episodes lasting beyond six hours, the chances of a positive MRI are about 50 percent. Basically, “the definition that we grew up with was made outdated by technology,” Dr Albers explained. In 2009, the American Stroke Association also released a guideline endorsing this change in the definition of TIA.

Back to Top | Article Outline


  • The study included a total of 4,574 patients with a TIA based on time-based criteria.
  • Among the 3,206 patients who had an MRI, 884 had a DWI infarct and 2,322 did not have an infarct. Recurrent stroke rates at seven days were 7.1 percent (95% confidence interval 5.5–9.1) in those patients with an infarct and 0.4% (0.2–0.7) in those without, an 18-fold difference (p diff<0.0001).
  • In DWI tissue-positive patients, the seven-day stroke rate was 1.8 percent in those with ABCD2 scores of 0-3 (low risk); 7.5 percent with ABCD2 scores of 4-5 (intermediate risk), and 12.5 percent for scores of 6-7 (high-risk). The ABCD2 score was predictive of stroke risk at seven days in those with and without infarction.
  • The recurrent stroke rates in 1,368 CT-imaged patients were 12.8 percent (9.3–17.4) in tissue-positive patients and 3 percent (2.0–4.2) in tissue negative patients (p diff<0.0001).
  • Tissue-positive events with low ABCD2 scores and tissue-negative events with high ABCD2 scores had similar stroke risks, especially after a 90-day follow-up.
Back to Top | Article Outline




Have a comment on this Neurology® article you would like to share? Now it is easier and more convenient. has launched WriteClick and the Editors encourage remarks and debate among users. WriteClick is restricted to comments about studies published in Neurology within the last eight weeks.

Scan this code with your smartphone or go to to join the debate now!

Back to Top | Article Outline


Giles MF, Albers GW, Amarenco P, et al. Early stroke risk and ABCD2 score performance in tissue- vs. time-defined TIA. Neurology 2011; E-pub 2011 Aug. 24
    Barrett KM. Early stroke risk: Tissue is the issue. Neurology 2011; E-pub 2011 Aug. 24.
      Albers GW, Caplan LR, Easton JD, et al. Transient ischemic attack—proposal for a new definition. N Engl J Med. Nov 21 2002;347(21):1713-6.
        ©2011 American Academy of Neurology