EPITHET: Strictly Negative Using tPA Beyond 3 Hours in Stroke, but Reason to Hope?

Susan Jeffrey

February 22, 2008

February 22, 2008 (New Orleans, Louisiana) — Results of a phase 2 trial using intravenous tissue plasminogen activator (tPA) beyond 3 hours in patients with mismatch on diffusion- and perfusion-weighted imaging are negative for the primary outcome of lower infarct growth measured as a ratio of geometric means but nevertheless show some positive trends in secondary end points.

Other infarct growth measures, including relative growth and the proportion of patients with no infarct growth, showed positive trends, and reperfusion was significantly increased in patients with mismatch on imaging who received treatment.

"Because reperfusion was associated with improved clinical outcomes, phase 3 trials beyond 3 hours after treatment are warranted," the researchers, with first author Stephen M. Davis, MD, from the Royal Melbourne Hospital, University of Melbourne, in Victoria, Australia, conclude.

The results, from the Echoplanar Imaging Thrombolytic Evaluation Trial (EPITHET), are published online February 22 in Lancet Neurology to coincide with their presentation here at the American Stroke Association International Stroke Conference 2008.

Opening the Window for Treatment

Currently, tPA is used in patients only up to 3 hours after symptom onset. However, there has been some thought that by imaging the ischemic penumbra — that is, the region of critically hypoperfused but potentially viable tissue around an irreversibly damaged infarct core — it might be possible to select patients who could still benefit from reperfusion therapy beyond the 3-hour time window. The penumbra can be imaged as a mismatch between perfusion-weighted magnetic resonance imaging (MRI) (PWI) and diffusion-weighted MRI (DWI) and is present in up to 80% of patients within 3 hours of symptom onset, although it diminishes rapidly with time.

The primary hypothesis of EPITHET was that tPA treatment would attenuate infarct growth in patients with a mismatch between PWI and DWI lesions. "However," the authors write, "we did not plan to use mismatch in the selection of patients, because rapid online detection of mismatch was not feasible, and we were keen to include a proportion of patients without mismatch for an exploratory analysis."

In this phase 2 trial, Dr. Davis and the EPITHET investigators randomly assigned 101 patients with acute ischemic stroke who were imaged with serial echoplanar MRI to receive tPA or placebo between 3 and 6 hours after symptom onset; 52 were randomized to tPA and 49 to placebo. The mean age of patients was 71.6 years, and the median National Institute of Health Stroke Scale (NIHSS) score was 13. A baseline screening noncontrast computed tomography (CT) was done to exclude hemorrhage.

The primary outcome measure was infarct growth attenuation in mismatch patients with tPA vs placebo between baseline and day 90 using a ratio of geometric means as the primary analysis. Secondary analyses of infarct growth attenuation included relative growth, absolute growth, and the difference in cube-root lesion volumes. Other secondary outcomes of interest included differences among mismatch patients in reperfusion, good neurological outcomes, and good functional outcomes between treatment groups.

PWI and DWI were done before and then 3 to 5 days after therapy, with T2-weighted MRI at around day 90; 85 of 99 evaluable patients (86%) had mismatch of DWI and PWI.

Results of the primary analysis, the geometric mean infarct growth, were not statistically significantly different between the groups, although there was a nonsignificant reduction of about 30% with tPA. The median relative infarct growth, a secondary analysis, approached significance, with a 61% reduction seen with tPA.

EPITHET: Geometric Mean Infarct Growth and Median Relative Infarct Growth With tPA Treatment vs Placebo

End Point tPA Placebo Ratio (95% CI) P
Geometric mean infarct growth (primary) 1.24 1.78 0.69 0.38 – 1.28 .239a
Median relative infarct growth 1.18 1.79 0.66 0.36 – 0.92 .054b
a. t
b. Wilcoxon test

Reperfusion occurred more often with tPA than placebo and was associated with less infarct growth (P = .001), better neurologic outcome (P < .001), and better functional outcome (P = .010), they write.

They speculate that their study was underpowered to test the primary end point of growth attenuation, despite sample size calculations based on pilot data. "On the basis of results presented here, 224 patients would be needed in each arm, at 80% power and P < .05, for differences in the primary outcome of geometric mean growth."

The researchers point out that although the aim of thrombolytic therapy is to reduce infarct growth, "growth measurements are complex, particularly in patients with small lesions. Patients with very large baseline lesions have a poor outcome irrespective of therapy."

More evidence is needed about thrombolytic therapy in nonmismatch patients, of whom only 11 in this trial could be evaluated for the primary outcome, they add. (Growth, major perfusion, and clinical outcomes did not differ between those with and without mismatch, they note, and in the nonmismatch patients, infarct growth did not differ between those receiving and not receiving tPA.)

"EPITHET provides further evidence that the time window for thrombolysis treatment might be extended beyond 3 hours in some patients," they conclude. "These results emphasize the need for phase 3 trials with primary clinical end points in this time window, such as the Third International Stroke Trial (IST-3) and the third European Cooperative Acute Stroke Study (ECASS-3), before any change in clinical practice is recommended."

"Failed Chance or New Hope?"

In a Reflection and Reaction commentary accompanying the paper in Lancet Neurology, Peter D. Schellinger, MD, from the University of Erlangen, in Germany, discusses whether EPITHET is a "failed chance or new hope."

For example, the investigators in this trial carried out MRI but did not use the result to select patients, instead using CT criteria. In addition, reperfusion was assessed at days 3 to 5. Many patients already have spontaneous recanalization by that time, so the effect of the drug was not really assessed, Dr. Schellinger writes.

"A more rigorous randomized controlled trial of intravenous thrombolytic therapy on the basis of [PWI]-DWI mismatch would have used MRI to select patients and randomize them between alteplase and placebo," he writes.

The value of PWI-DWI mismatch has been criticized by some researchers who have called for more standardized definitions of mismatch and perfusion, as well as more data on the role of mismatch. "Two phase 2 trials of another thrombolytic drug, desmoteplase, were positive for surrogate and clinical end points, but the paramount study, the second Desmoteplase in Acute Ischemic Stroke Trial (DIAS-2), was negative, supposedly owing partly to the use of selection methods other than MRI," Dr. Schellinger notes.

After the setback of DIAS-2, the planned combined analysis of EPITHET and a similar but nonrandomized trial, the Diffusion and Perfusion Imaging Evaluation For Understanding Stroke Evolution(DEFUSE) trial, "provides new hope that the effects of alteplase beyond 3 hours after stroke will be better understood," although great care must be taken in pooling end points that are not consistent in these trials, he notes. "Finally, I hope that after the promising hypotheses tested in the secondary end points of EPITHET, and after the results of [the The European Cooperative Acute Stroke Study] ECASS-3 become available in autumn 2008, a randomized controlled trial based exclusively on MRI mismatch will be designed and initiated, whether it looks at administration of alteplase beyond 3 hours or 4.5 hours," he concludes. "The time is right for a joint international effort to plan and undertake such a trial."

American Stroke Association program committee chair Philip B. Gorelick, MD, from the University of Illinois College of Medicine, in Chicago, moderated the session where this study was presented. Commenting on the study for Medscape Neurology & Neurosurgery, he said that even though the primary outcome of this trial was strictly negative, he would nevertheless interpret the findings in a positive light.

"With the 30% attenuation of mean growth, a 61% reduction in median relative infarct growth that approached significance, I think we have to view this in a more positive way, in that in a larger study, a phase 3 trial, we might be able to show that this truly going to be positive," he said. "And that would be a great thing, if we could extend the window of tPA from 3 to 6 hours."

The study was funded by the National Health and Medical Research Council, Australia; National Stroke Foundation, Australia; and the Heart Foundation of Australia. Dr. Davis is a member of the PAION, Servier, and Novo Nordisk advisory boards and has received honoraria for lectures from Novo Nordisk, Sanofi-Aventis, Pfizer, and Boehringer Ingelheim. Disclosures for other coauthors appear in the paper.

Lancet Neurol. Published online February 22, 2008.

American Stroke Association International Stroke Conference: Late-Breaking Clinical Trials Session. Presented February 22, 2008.


Comments on Medscape are moderated and should be professional in tone and on topic. You must declare any conflicts of interest related to your comments and responses. Please see our Commenting Guide for further information. We reserve the right to remove posts at our sole discretion.