Evidence supporting the use of thrombolysis for selected stroke patients with unknown time of onset has been strengthened with data from a new individual patient meta-analysis.
The meta-analysis combined data from four clinical trials that included a total of 483 patients with stroke of unknown time of onset. MRI or CT-perfusion imaging was used to identify those patients with salvageable brain tissue. They had been randomly assigned to receive either thrombolysis with intravenous tissue plasminogen activator (tPA, alteplase) or placebo/standard of care.
Results showed that use of intravenous alteplase resulted in better functional outcome, as measured on Modified Rankin Scale (mRS), at 90 days. Although there was an increase in risk for symptomatic intracranial hemorrhage and more deaths occurred among patients treated with alteplase than placebo, there were fewer cases of severe disability or death.
The new data were presented by Götz Thomalla, MD, University Medical Center Hamburg, Germany, at the European Stroke Organisation–World Stroke Organization (ESO-WSO) Conference 2020 on November 8. They were also simultaneously published online in The Lancet.
“These results provide level 1a evidence for the use of MRI or CT-perfusion imaging to guide treatment with intravenous alteplase in unknown onset stroke,” Thomalla concluded.
He explained to Medscape Medical News that alteplase is not approved for treating stroke of unknown time of symptom onset, but American Heart Association/American Stroke Association guidelines and updated ESO guidelines recommend its use on the basis of results of the WAKE-UP trial.
“This meta-analysis refutes formal doubts resulting from the fact that previously, just a single trial (WAKE-UP) demonstrated the benefit of intravenous alteplase in unknown onset stroke,” Thomalla said. “The meta-analysis, based on four trials, now provides level 1a evidence, which is a clear basis for guideline recommendations.”
Moreover, the large number of patients enabled subgroup analyses, which did not identify any treatment heterogeneity for relevant subgroups. “To summarize: intravenous alteplase guided by MRI or perfusion CT is effective in unknown onset stroke across all subgroups, including [those based on] age, severity, and vessel occlusion,” he added.
In an ESO-WSO press conference at which the meta-analysis was discussed, Patrik Michel, MD, Lausanne University Hospital, Lausanne, Switzerland, who was not involved in the study, said: “This is very important data, given that 20% to 25% of ischemic stroke patients have an unknown time of symptom onset. It shows us who will benefit from thrombolysis and when, based on advanced imaging.”
Asked how many of the patients with unknown time of stroke onset may be eligible for thrombolysis, Thomalla estimated that about 1 in 3 of such patients could now be considered. “In the WAKE-UP trial, we screened all wake-up patients with MRI and were able to randomize one third of them. But the meta-analysis included trials with two different technologies ― MRI and CT perfusion ― and they are not completely overlapping, so using either of these criteria could extend the number of patents eligible,” he noted.
For the meta-analysis, the researchers combined individual patient data from four randomized trials (WAKE-UP, EXTEND, THAWS, and ECASS-4) of intravenous alteplase vs standard of care or placebo for adults with stroke with unknown time of onset who had been selected on the basis of advanced brain imaging.
Either penumbral imaging (perfusion-diffusion MRI or perfusion CT) or MRI-based tissue clocking were used. MRI-based tissue clocking utilizes a mismatch on MRI between a visible lesion on diffusion-weighted imaging (DWI) and an absence of marked parenchymal hyperintensity on fluid-attenuated inversion recovery (FLAIR). DWI shows high signal intensity early after stroke, whereas FLAIR signal changes are delayed.
The four trials provided individual patient data for 843 patients, of whom 429 (51%) were assigned to received alteplase, and 414 (49%) were assigned to receive placebo or standard care.
The primary outcome was favorable functional outcome (mRS score of 0–1, indicating no disability) at 90 days. This was achieved by 47% of patients who received alteplase, vs 39% of those in the control groups, giving an adjusted odds ratio of 1.49 (P = .011).
Alteplase was also associated with positive results on the two secondary outcomes: mRS shift toward a better functional outcome, and independent outcome (mRS score of 0–2) at 90 days.
Alteplase was associated with a significant shift toward better functional outcome, with an adjusted common odds ratio of 1.38 (P = .019). Among the patients who received thrombolysis, there was a greater likelihood of an independent outcome, with an adjusted odds ratio of 1.50 (P = .022).
In the alteplase group, 21% of patients experienced severe disability or died (mRS score, 4–6), compared with 25% of patients in the control group (adjusted odds ratio, 0.76; P = .15).
Death occurred in 6% of alteplase patients, vs 3% of control patients (adjusted odds ratio, 2.06; P = .040). The prevalence of symptomatic intracranial hemorrhage was also higher in the alteplase group than among control patients (11 patients vs two patients), giving an adjusted odds ratio of 5.58 (P = .024).
“The observed slight increase in mortality is consistent with findings from previous studies and likely relates at least in part to the increased risk of symptomatic intracranial hemorrhage, which is known for all thrombolysis trials,” Thomalla said of these outcomes. “There is, however, a net benefit of intravenous alteplase on outcome even if the higher rates of death are included…and the higher mortality is offset by a lower rate of very poor outcomes, such as nursing home outcomes (mRS, 5).”
Thomalla believes these new data should lead to consistent recommendations in all national and international guidelines and should lead to changes in clinical practice.
“In expert stroke centers, patients are usually already treated this way. However, in many smaller stroke centers, advanced imaging may not yet be available 24/7, or there may have been doubts in using it to treat patients with unknown time of symptom onset,” he said. “The results of our meta-analysis will hopefully help to make this treatment available more widely and also in smaller stroke centers.”
Asked how the inclusion criteria for late thrombolysis differ from those for late thrombectomy, Thomalla noted that the DWI-FLAIR mismatch concept used in these studies of thrombolysis is based on the use of MRI for determining stroke lesion age, and this approach has not been used for thrombectomy, whereas the use of penumbral imaging based on CT perfusion imaging is very much the same as in the thrombectomy trials in the late time window.
He pointed out that patients in the unknown time of onset thrombolytic trials and in the thrombectomy trials differ with regard to the severity of symptoms. Thrombectomy is appropriate only for large-vessel occlusion, whereas thrombolysis is also effective for patients with minor and moderate strokes.
“Intravenous alteplase was also effective in patients with large-vessel occlusion on baseline imaging in trials mainly conducted before thrombectomy was standard of care. Thus, I would conclude that in unknown onset stroke with large-vessel occlusion, IV tPA should be started immediately, and patients then should be transferred to thrombectomy,” he added.
During the discussion that followed the presentation of these new results at the ESO-WSO 2020 meeting, many questions were raised as to which of the imaging modalities was preferable for identifying patients suitable for receiving thrombolysis.
“The main point is that most centers should have some form of advanced imaging available ― MRI or CT perfusion ― and my recommendation would be to use the one you are familiar with,” Thomalla responded. “MRI has some superiority in detecting lacunar strokes, but both imaging modalities work well for guiding treatment decisions on the most severely affected patients. Plain CT alone is not enough for this situation.”
Asked whether there was a quantitative method of assessment for MRI mismatch, Thomalla replied, “I don’t think we need this. The FLAIR images don’t have a real quantitative signal. It is a very simple idea ― you get this visually very easily. If there is a clear marker of FLAIR hyperintensity, then the patient is not a candidate for thrombolysis; if there is no hyperintensity or if you are in doubt, then you should go for thrombolysis.”
There was no funding for the meta-analysis. Thomalla has received grants and personal fees from Bayer, personal fees from Acandis, Boehringer Ingelheim, Bristol-Myers Squibb, Pfizer, Daiichi Sankyo, Portola, and Stryker, outside of the submitted work.
European Stroke Organisation-World Stroke Organization (ESO-WSO) Conference 2020: Presented November 8, 2020.
Lancet. Published online November 8, 2020. Abstract