Pooled Analysis of tPA Stroke Trials Underlines Need for Rapid Treatment

Susan Jeffrey

May 14, 2010

May 14, 2010 — A new pooled analysis of randomized trials of tissue plasminogen activator (tPA) in the treatment of acute ischemic stroke, now expanded to include more recent trials, confirms the benefit of treatment to 4.5 hours but shows for the first time that risk may outweigh benefit after that time.

The new analysis adds data from the recent European Cooperative Acute Stroke Study 3 (ECASS 3) and the Echoplanar Imaging Thrombolysis Evaluation Trial (EPITHET), bringing the total number of pooled studies to 8.

The updated analysis "shows that treatment with thrombolysis until 4.5 [hours] from stroke onset enhances the chance of favorable outcome," the researchers, with lead study author Kennedy R. Lees, FRCP, from the Western Infirmary at the University of Glasgow, Scotland, conclude.

Interestingly, parenchymal hemorrhage rates were independent of time to treatment, but mortality increased with thrombolytic therapy after 4.5 hours, suggesting other mechanisms of mortality may be involved.

"However, across the time window studied, our analysis showed that the greatest benefit comes from earlier treatment, since net benefit is diminishing and is undetectable in our sample beyond 4.5 [hours]," they write.

The analysis is published in the May 15 issue of The Lancet.

Planned Analysis

Since the first trials of tPA in stroke were designed, common data elements have been collected with an eye to pooling the datasets, the study authors note. This pooled analysis includes data from the National Institute of Neurological Disorders and Stroke (NINDS) trial parts 1 and 2, the first 2 ECASS trials, 2 Alteplase Thrombolysis for Acute Noninterventional Therapy in Ischemic Stroke (ATLANTIS) trials, and the more recently reported ECASS 3, as well as EPITHET.

"Initially, we intended not to include [EPITHET], but while preparing to publish our findings we learned that all variables needed for the analysis had been gathered by the EPITHET investigators," the study authors write. EPITHET based inclusion on imaging findings, whereas enrollment in all the other trials was based on clinical diagnosis of a clearly defined time of stroke onset and a computed tomographic scan clear of bleeding.

The 2 additional trials, ECASS 3 with a total of 821 randomized patients and EPITHET with 100 patients, brought the pooled patient total to 3670, with 1850 treated with tPA and 1820 receiving placebo. They used multivariate logistic regression to assess the relationship between the time from stroke onset to the start of treatment and favorable 3-month outcome, defined as a modified Rankin scale score of 0 to 1, mortality, and the occurrence and outcome of clinically relevant parenchymal hemorrhage. Outcomes were examined for patients treated within 360 minutes (6 hours).

They found that the odds of a favorable outcome increased as the time from stroke onset to treatment decreased (P = .0269), with no further benefit seen after 270 minutes (4.5 hours).

Table 1. Adjusted Odds of a Favorable 3-Month Outcome by Time From Stroke Onset to Start of Treatment (OTT)

OTT, min OR (95% CI)
0 – 90 2.55 (1.44 – 4.52)
91 – 180 1.64 (1.12 – 2.40)
181 – 270 1.34 (1.06 – 1.68)
271 – 360 1.22 (0.92 – 1.61)

CI = confidence interval; OR = odds ratio

Mortality risk also increased with time and appeared to outweigh benefit after 4.5 hours.

Table 2. Adjusted Odds of Mortality by Time From Stroke Onset to Start of Treatment (OTT)

OTT, min OR (95% CI)
0 – 90 0.78 (0.41 – 1.48)
91 – 180 1.13 (0.70 – 1.82)
181 – 270 1.22 (0.87 – 1.71)
271 – 360 1.49 (1.00 – 2.21)

CI = confidence interval; OR = odds ratio

Large parenchymal hemorrhage, perhaps the most feared complication of tPA therapy in stroke, occurred in 96 (5.2%) of treated patients and 18 (1.0%) of controls, but with no clear relationship between time of stroke onset and treatment (P = .4140).

"The apparent discrepancy between mortality and parenchymal bleeding deserves comment," the study authors write, because it had been thought previously that the risk of mortality with later treatment lies with increased hemorrhage risk.

....there must be another mechanism underlying the increase in mortality...

The finding might be caused by low power, they speculate, due to a relatively small number of cases. "However, although parenchymal hemorrhage, when it occurs, is associated with a high risk of death, the number of cases is so low that mortality is evidently not a result of such bleeds in most cases: there must be another mechanism underlying the increase in mortality."

Although the effectiveness of thrombolytic therapy is "beyond dispute," they note, treatment still does not help a large proportion of patients. Their analysis shows that about 5 patients need to be treated within 0 to 90 minutes, 9 patients within 91 to 180 minutes, or 15 patients between 181 and 270 minutes after symptom onset for one of them to have a good outcome.

More information is needed to understand better the factors that may prevent treatment from being effective in individual patients.

"The stroke community has many lines of research to improve stroke treatment, such as looking for a possible more effective dose of alteplase, reducing the tendency of thrombolytic drugs to cause intracerebral hemorrhage, testing new thrombolytic agents, exploiting new and faster imaging methods to guide treatment, increasing the extent of lysis with ultrasound, antithrombotic agents, intra-arterial clot manipulation, adding neuroprotective therapies that could be started in the ambulance, and examining cognitive function in clinical trials so that important benefits or risks of treatment are not missed," they conclude.

Renewed Mandate

In a commentary accompanying the publication, Jeffrey L. Saver, MD, from the Geffen School of Medicine at the University of California, Los Angeles, and Steven R. Levine, MD, from Mount Sinai School of Medicine in New York City, write that with this new pooled analysis by Lees et al, "the time profile of benefit and harm for alteplase in broadly selected patients is for the first time well delineated, and the first age of recanalization therapy development in acute stroke therapy draws to a close."

As seen in the previous pooled analysis in 2004, these results showed that benefit decreased substantially over time, they write. "An important new finding is that later onset to start of treatment (more than 4.5 hours from onset) was associated with increased mortality, consonant with findings of a Cochrane study–level meta-analysis of a larger group of 26 trials (7152 patients) that investigated a wider variety of thrombolytics" (Wardlaw et al. Cochrane Database Syst Rev. 2009;4:CD000213).

The mechanism of this increased mortality, though, may be of some debate, they note, because there was no parallel time-related increase in parenchymal hemorrhage.

Still, an "even more important message" from these data is the steep decrease in benefit at the earlier periods, with odds of a favorable outcome appearing to decrease by a factor of 2 for each 90-minute period after symptom onset.

"These findings mandate a renewed commitment by clinicians and policymakers to foster very early intervention," Dr. Saver and Dr. Levine conclude, advocating strategies that include better education of the public of symptom recognition, training prehospital personnel in a policy of "scoop and go," routing ambulances to stroke centers, and decreasing so-called door to needle times within stroke centers.

"In thrombolytic stroke therapy, sooner is better than later, much better," they conclude.

The analysis received no commercial funding. Dr. Lees reports he has received honoraria from Boehringer Ingelheim for his role in conduct of the ECASS trials, as well as from Lundbeck and Thrombogenics. Disclosures for coauthors appear in the paper. Dr. Saver reports he is an employee of the University of California, which holds a patent on retriever devices for stroke; is a scientific consultant regarding trial design and conduct to Concentric Medical, Talecris, and Ev3; has received lecture honoraria from Boehringer Ingelheim; has been a site investigator in multicenter trials sponsored by Vernalis, Paion, Lundbeck, and Neurobiological Technologies, for which the University of California Regents received payments based on the clinical trial contracts for the number of subjects enrolled; is a site investigator in the National Institutes of Health (NIH) IRIS, CLEAR, IMS 3, SAMMPRIS, and VERITAS multicenter clinical trials, for which the University of California Regents receive payments based on the clinical trial contracts for the number of subjects enrolled; is a lead investigator in the NIH MR RESCUE multicentre trial; and is funded by NIH-NINDS Awards. Dr. Levine reports he has been an investigator for an NIH-funded trial of tenecteplase, manufactured by Genentec; has received an honorarium from the National Stroke Association for a Webcast; and is funded by NIH-NINDS Awards. He is the independent medical monitor for NIH clinical trials IMS 3, CLEAR-ER, FAST-MAG, and INSTINCT.

Lancet. 2010;375:1695-1703, 1667-1668.