'Late Window' Paradox: Benefit Increases With Later Thrombectomy?

January 26, 2018

LOS ANGELES — New data from the two trials of patients undergoing late thrombectomy, 6 to 24 hours after stroke onset, suggest the benefits of such treatment can be even larger than seen in trials of similar treatment given earlier.

The two trials of late thrombectomy — on DAWN and DEFUSE 3 — both reported similar findings of unexpected large benefits of the endovascular treatment throughout the whole time window evaluated. And the benefits appeared greater than reported in some of the earlier trials, where endovascular therapy was conducted mainly within 6 hours of stroke onset.

These findings were among the key results of the DAWN and DEFUSE 3 trials presented here at the International Stroke Conference (ISC) 2018.

The authors suggest that this "late window paradox" can be explained by the fact that the trials evaluating late thrombectomy used sophisticated imaging techniques to select appropriate patients with "slow-growing infarcts" because of good collateral circulation, which allowed a large area of salvageable brain tissue to survive; these patients therefore received substantial benefit from removal of the clot even up to 24 hours after stroke onset. In addition, the control group did not benefit from patients having receiving thrombolysis with tissue plasminogen activator (tPA) because they arrived too late for this treatment.

"The 'time is brain' concept requires a 2018 revision that is more generous than the original and provides a reprieve for the fortunate patients who have favorable collaterals and slow infarct growth," DEFUSE 3 lead author, Gregory Albers, MD, Stanford University Medical Center, California, concludes in an "Opinion" piece published online January 24 in Stroke.

Impressive benefits of thrombectomy in patients presenting right out to 24 hours were among some of the newest findings to be reported from the DAWN trial at the meeting here.

The main results of the DAWN trial, reported last year, showed thrombectomy was beneficial in patients presenting between 6 and 24 hours who had "clinical mismatch," defined as a small core stroke volume (unsalvageable brain tissue seen on imaging) but a large clinical deficit (National Institutes of Health Stroke Scale score > 10) signifying an extensive brain area threatened by ischemia yet still potentially salvageable.

The results showed an absolute difference of 2.0 units on the utility-weighted modified Rankin Scale (mRS) in favor of the thrombectomy group.

The latest results from DAWN, an analysis presented as a late breaker, gave more information on patient-level benefits and the effect of time.

"Our results show that half of patients treated with thrombectomy have improved 90-day disability levels, and one third of patients achieve functional independence," said lead investigator, Jeffrey Saver, MD, University of California, Los Angeles.

"The benefit of thrombectomy was also substantial and dramatic in both the 6- to 12-hour and 12- to 24-hour time windows, with about half of patients benefitting in both periods. But earlier patients do better in both the thrombectomy and control groups."

To Medscape Medical News, Dr Saver commented: "We found that in both time windows a similar number of patients — about half — benefitted from thrombectomy. The type of improvement was a little different. In the later window, patients were going from 'really bad' in the control group to 'pretty good' in the thrombectomy group, and in the earlier window, they went from 'somewhat bad' for controls to 'very good' with thrombectomy. But there was still a huge population benefit in both time windows."

He added: "We also showed that selecting patients for thrombectomy who present late using imaging gave similar benefits to treating all patients at 2 hours after stroke onset and much better results than treating all patients at 5 hours after stroke onset.

"This is because when using tissue-based criteria to select patients, every patient who goes on to have the treatment has brain to save, whereas treating all patients at 5 hours from onset will mean that some of the patients will have already finished their infarct — there is no brain left to save."

Dr Saver suggested that some patients presenting even later than 24 hours could benefit from thrombectomy if target mismatch was present, although the likelihood of having salvageable tissue will decline sharply with ongoing time.

"If we continued after 24 hours, I think we would still see benefit — slowly diminishing — but there will be far fewer eligible patients with the right residual tissue the longer the time increases from stroke onset."

In his presentation, Dr Saver reported the DAWN results in terms of benefit per 100 patients treated and the number needed to treat (NNT) to show improvement in one patient, across the whole disability range.

"The number needed to treat is 2," he reported. "So for every 2 patients treated 1 would have a better outcome than if not treated. And for 100 patients treated, 50 would have less disability, including 36 who would be functionally independent."

Table 1. DAWN: Benefit per 100 Patients Treated and Number Needed to Treat

Outcome mRS Score Benefit per 100 Patients NNT
Asymptomatic 0 5 19
Freedom from disability 0 - 1 23 4
Functional independence 0 - 2 36 3
Ambulatory 0 - 3 32 3
Not requiring constant care 0 - 4 11 9
Lower disability over entire mRS range   50 2



"In terms of time of treatment we showed similar benefits in both 6 to 12 hours and 12- to 24-hour time periods, with a number need to treat of 2 in both time windows," he stated.

Table 2. DAWN: Benefits of Procedure by Time Since Stroke Onset

Outcome mRS Score 6 - 12 Hours: Benefit per Hundred 6 - 12 Hours: NNT 12 - 24 Hours: Benefit per 100 Patients 12 - 24 Hours: NNT
Asymptomatic 0 7 14 3 33
Freedom from disability 0 - 1 22 5 22 5
Functional independence 0 - 2 33 3 36 3
Ambulatory 0 - 3 33 3 31 3
Nor requiring constant care 0 - 4 2 44 19 5
Lower disability over entire mRS range   45 2 56 2



"When comparing DAWN to meta-analyses of prior studies with earlier endovascular treatment, patients treated at two hours show similar benefit as in DAWN – with 50 patients improving per 100 treated. But at 5 hours, this drops to 33."

DEFUSE 3 Confirms Time Paradox

Similar results over time were seen in the DEFUSE 3 trial, the second study of late thrombectomy, reported for the first time at the ISC this week.

Presenting these results, Dr Albers reported that the percentage of patents achieving a good outcome (mRS score of 0 - 2) was maintained at 40% to 50% regardless of when they were treated within the 9- to 16-hour window of the study.

"In contrast, control patients have declining rates of good outcome. So the relative benefit of the treatment was actually largest towards the end of our treatment window," he explained.

Table 3. DEFUSE 3: Patients Achieving Good Outcome (mRS Score 0 - 2) by Time of Treatment



Time from Stroke Onset to Randomization (h) Endovascular Therapy (%) Control (%)
<9 40 28
9 - 12 50 17
12 - 16 42 7



Dr Albers made a comparison of the two late thrombectomy trials, DAWN and DEFUSE 3, together with the earlier thrombectomy trials, where treatment was mainly given within 6 hours.

"This shows that the benefits of the endovascular treatment appear similar — close to 50% — whether conducted early or late. But the control group is very different, with much worse results later on. This is probably because in the early trials the vast majority of patients would have received thrombolysis first, whereas the late trials are beyond the thrombolysis time window. In addition, in the late trial patients were carefully selected based on having salvageable tissue, which was not always the case in the early trials," he said.

Table 4. Good Outcome (mRS Score 0 - 2) at 90 Days



Group Early Trials (%) Late Trials: DAWN, DEFUSE (%)
Endovascular 46 47
Control 27 15



Dr Albers elaborates on this "time paradox" in his Stroke "Opinion" paper.

"DAWN randomized patients at a median of 12.5 hours from onset and documented the largest absolute increase in functional independence ever documented in any acute stroke treatment trial, 35.5%," he writes. "DEFUSE 3 randomized patients at a median of 11 hours after onset and documented a 28% increase in functional independence and an additional 20% absolute reduction in death or severe disability, which represents the largest reduction in mortality/severe disability ever achieved.

"By comparison, the pooled analysis of 5 modern early window thrombectomy trials (HERMES) revealed an absolute increase in functional independence of 19.5% and a reduction in mortality/severe disability of 11%. Both early and late window studies included patients of similar ages, baseline National Institutes of Health Stroke Scale scores, and vessel occlusion sites. Patients were treated with the same modern thrombectomy devices and reperfusion was achieved in similar proportions of patients in the endovascular arms of the early and late window studies. So why was the treatment effect larger in the late window trials?" he asks.

Dr Albers explains that a substantial percentage of patients with large-vessel intracranial occlusions have very slow growth of the ischemic core for up to 12 hours or longer, but the favorable collateral circulation that is responsible for keeping the ischemic core size small eventually fails in most patients and infarct volumes ultimately increase.

He notes that the growth of the ischemic area varies substantially between patients. Some patients with very poor collaterals develop very large infarct cores within 2 to 3 hours, whereas other patients have little or no core growth for 12 hours or longer, and it can take 3 days for the maximum core volume to occur.

He points out that the imaging requirements for inclusion into the DAWN and DEFUSE-3 studies meant that these studies selected out patients with slow-growing ischemic cores.

Because these patients had proximal middle cerebral artery or internal carotid artery occlusions, a very large volume of penumbral salvage would be anticipated if substantial or complete reperfusion (as occurs with thrombectomy) is obtained, Dr Albers writes.

However, for the control patients in these late window trials, very poor outcomes would be anticipated for two key reasons: They did not receive tPA because they presented too late for this, and eventually the collateral circulation will fail and the core infarcts expand into the whole region of critical hypoperfusion, he says.

Time Is Still Brain

Expert commentators are, however, concerned that mixed messages may be received about the time issue for thrombectomy on the basis of these results.

"My concern is that the message is going to be confused. Even among providers there may be a feeling that time doesn't really matter anymore — if the patient has the right tissue patterns, then maybe we can all be a bit slower," Bruce Ovbiagele, MD, ISC 2018 chair and , professor and chair of neurology at the Medical University of South Carolina, Charleston, commented to Medscape Medical News.

"That is definitely not the right message to put across," he added. "As we translate these findings into messages for the community, we have to emphasize that time is still brain. Yes, some people have salvageable brain tissue for many hours, but you don't know if you are one of those people or not, and we don't know how long your brain tissue will remain salvageable. The longer you leave it to go the hospital, the less chance you have of having salvageable brain tissue, so it always is the right thing to do to rush."

Ralph Sacco, MD, Jackson Memorial Hospital, Miami, Florida, current president of the American Academy of Neurology and former president of the American Heart Association, agrees.

"The message, when we look at all the studies ever done in stroke, is the quicker we open a vessel the better," Dr Sacco said. "Yes, these new studies are telling us that we can now use mechanical thrombectomy at later hours with an impressive benefit in some patients — and that's great — but it is imperative to understand that these patients have been highly selected."

"For the public the message 'faster is better' is still the most important information," he added. "The good news is that beyond 6 hours with the right imaging we can identify patients who can still benefit for quite some time, but that doesn't mean anyone should wait. We don't have the numbers on how many people do have the right tissue patterns at what time to be eligible for intervention, but it is intuitive that there will be less people eligible the later you go. The message to the public is you still have to get in early to have a better chance of having salvageable brain tissue and be eligible for the procedure."

DEFUSE 3 was funded by the National Institutes of Health. DAWN was funded by Stryker Neurovascular. Dr Albers has an equity interest in iSchemaView and is a consultant for iSchemaView and Medtronic. Dr Saver reports receiving consulting fees and travel expenses for work advising on rigorous trial design and conduct.

International Stroke Conference (ISC) 2018. Presentations LB1, LB3, and LB6. Presented January 24, 2018.

Stroke. Published online January 24, 2018. Abstract

For more Medscape Neurology news, join us on Facebook and Twitter


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.