iScore Predicts tPA Benefit and Risk After Ischemic Stroke

Daniel M. Keller, PhD

October 31, 2012

BRASILIA, Brazil — Researchers have developed a simple, validated tool, the iScore, that uses clinical parameters and comorbid conditions that are easily assessed at the bedside after an ischemic stroke to allow clinicians to estimate near- and long-term response and the risk for intracerebral hemorrhage (ICH) from using intravenous tissue plasminogen activator (tPA).

Gustavo Saposnik, MD, associate professor of medicine and director of the Stroke Research Unit at St. Michael's Hospital of the University of Toronto, Canada, told Medscape Medical News that the iScore can predict clinical outcomes, and specifically 30-day mortality, 1-year mortality, death or disability, or death or institutionalization at discharge.

"But moreover, the iScore is probably one of the single scores that has shown that there is an interaction with thrombolytic therapy...which means the clinical response to tPA may predict the clinical outcomes and the final infarct volume," he said.

He presented their findings here at the 8th World Stroke Congress (WSC).

iScore Calculator

The iScore, essentially an ischemic stroke risk score, is calculated using 8 variables — age, sex, stroke severity using the Canadian Neurological Scale, stroke subtype, cardiovascular risk factors, comorbid conditions, preadmission disability, and blood glucose on admission — with each variable accounting for a portion of the total score. The tool has been validated using internal and external cohorts that included more than 16,000 patients with acute ischemic stroke.

The Stroke Outcomes Research Canada (SORCan) working group, of which Dr. Saposnik is a member, has developed a Web site ( for calculating the iScore by entering values for the variables. It shows the risk for 30-day mortality, 30-day death or disability, 30-day death or institutionalization, and 1-year mortality.

With the possibility of residual confounding in observational studies, Dr. Saposnik and colleagues retrospectively applied the tool to data from patients (n = 624) in the National Institute of Neurological Disorders and Stroke randomized, controlled trial on tPA, stratifying the patients by iScore, to determine its ability to predict clinical response to tPA and the risk for all or symptomatic ICH.

A favorable composite outcome was defined as a modified Rankin Scale (mRS) score of 0 or 1, a National Institutes of Health Stroke Scale score of 1 or less, a Barthel Index of 95 or greater, or a Glasgow outcome scale score of less than 1 at 3 months.

An mRS of 0 to 2 at 12 months was another favorable outcome. Patients' data were stratified by iScore of less than 200 (n = 507) vs 200 or greater (n = 117).

At the lower iScore, tPA was beneficial at 3 months. Of the patients receiving tPA, 58.7% (145/247) had a favorable outcome at 3 months vs 41.9% (109/260) for patients getting placebo (P < .001). A shift analysis at 3 months showed that tPA was associated with a shift toward a lower mRS (P = .002) for the patients with iScores less than 200.

There did not appear to be any benefit of tPA at 3 months for patients with the higher scores — 15.3% (10/65) had a good outcome with tPA vs 13.5% (7/52) with placebo (P = .77). Similarly, a shift analysis showed that no benefit accrued to patients with iScores greater than 200 (P = .80).

ICH of any type occurred significantly more often with tPA regardless of the iScore but was more prevalent with the higher scores. At iScores of less than 200, ICH occurred in 11.3% with tPA vs 5.4% with placebo (P = .011).

At iScores of greater than 200, ICH affected 30.8% of patients receiving thromobolysis vs 11.5% of patients getting placebo (P = .014). The number needed to harm using tPA in the lower iScore group was 17 but was only 5 if tPA was used with higher iScores.

For the risk of ICH, "it's not the 6% we quote [for tPA], and I think that that's important," Dr. Saposnik said. "So families may want to make a decision. I'm not sure if those patients may be more amenable for the less or more aggressive interventions, such as endovascular [interventions]. I'm not sure, and this is something that needs to be explored further." He noted that he is now trying to obtain data from endovascular therapy trials to apply an analysis to them based on iScores.

Final infarct volumes by computed tomography at 3 months increased with rising quartiles of iScores (P < .001) but did not differ according to the use of tPA or not.

The online iScore calculator provides suggestions for the use of tPA on the basis of an adjusted relative risk of using tPA vs not using it. The adjusted relative risk has a value of 1 at an iScore of about 140, meaning that for scores below 140, the relative benefit of tPA therapy outweighs the risks. At scores greater than 140, risks outweigh benefits, as determined on the basis of data from the large cohorts from which the iScore was derived.

In use online, the calculator shows that iScores of 130 or less predict good outcomes with the use of tPA and suggest to proceed with it; scores between 130 and 200 suggest "proceed with caution," and scores above 200 advise "caution." A good clinical outcome is considered an mRS of 0 to 2.

Dr. Saposnik said the tool is not meant to replace good clinical judgment, and a disclaimer on the Web site makes this point clear. The iScore calculator is available as a free application (app) at the iPhone App Store.

Not for Patient Selection

Session moderator Joseph Broderick, MD, professor and chair of the Department of Neurology at the University of Cincinnati Neuroscience Institute in Ohio, commented to Medscape Medical News that in his opinion, the iScore should not be used for selecting patients for thrombolytic therapy, and Dr. Saposnik agreed.

"You certainly could use it as a way of maybe talking with the families" about the predicted futility of using tPA if their family member had a stroke, Dr. Broderick said. "So you can use it as a reality check for families, but at the same time, you don't want to take away people's hope, because there's always [a patient] who defies what you said."

Even without the iScore, clinicians are quite good at predicting poor outcomes after stroke, based on several well-recognized variables, he said. "The question is, does tPA actually improve your chances of having a good outcome compared to placebo, even if both groups are going to do relatively poorly?"

So far, the iScore has been validated using retrospective data. "What you would really need to do is do a trial where you said, ''We'll have an iScore of 200, and we'll randomize people to tPA or not and see if there's a difference. That's how you'd answer that question" of whether the iScore predicts benefit from tPA, Dr. Broderick advised.

In this vein, Dr. Saposnik suggested that "some clinical trials may benefit from using the iScore to identify and to target specific patients — those who may have a prespecified or poor predicted outcome before [they] entered into a trial. Some investigator may decide to specifically target those patients. Some other groups — it depends on the intervention — may target the lower-risk patients."

The study received no commercial funding. Dr. Saposnik has disclosed no relevant financial relationships. Dr. Broderick was not involved in the trial and has disclosed no relevant financial relationships.

8th World Stroke Congress (WSC). Abstract 697. Presented October 11, 2012.