This transcript has been edited for clarity.
Thanh Nguyen, MD: Good day, everyone. Thank you very much for joining us today. In this session for Medscape, we're going to be discussing medium and distal vessel occlusions hot off the press after the International Stroke Conference presentations. My name is Thanh Nguyen. I'm a stroke and interventional neurologist, and I'd like to send the microphone over to our colleagues, professor Urs Fischer, who is the principal investigator of the DISTAL trial.
Urs Fischer, MD, PhD: My name is Urs Fischer. I am a clinician scientist, and I am the director and chairman of the Department of Neurology at University Hospital in Bern, Switzerland. I'm president-elect of the European Stroke Organization.
Nguyen: Professor Mayank Goyal, if you could introduce yourself.
Mayank Goyal, MD, PhD: My name is Mayank Goyal. I'm in Calgary. I'm a neuro interventionist, and along with my colleague Michael Hill, we led the ESCAPE-MeVO trial.
The ESCAPE-MeVO Trial
Nguyen: Wonderful. We'll begin with Dr. Goyal about a brief summary of what you learned from the ESCAPE-MeVO trial, and why did you conduct this trial?
Goyal: Starting in 2015, there were all these dramatic changes in stroke care related to large vessel occlusion. And as time went by, it was becoming increasingly difficult to find a subgroup in which there was no benefit of endovascular thrombectomy. In the meantime, what we found was we were getting better at tackling medium vessel occlusions. Technology was getting better, technique was getting better. And starting around 5, 6 years ago, we started working on this idea.
We wrote a couple of editorials. We defined what "MeVO" (medium vessel occlusion) is. We were very fortunate to have a large dataset here in Calgary, and we wrote papers on natural history. We wrote papers on primary vs secondary medium vessel occlusion. And overall, what we realized based on the dataset that was available to us, less than 50% of patients reach modified Rankin scale score of 0 or 1 with so-called standard of care.
So that's when we started exploring the idea that can we improve outcomes further, given that they were not the greatest to begin with, and we started to design what ended up becoming an ESCAPE-MeVO trial. The trial was a standard probe designed, open-label, randomized trial. We conducted across Europe and North America.
A total of 530 patients were randomized 1:1. The treatment arm was endovascular thrombectomy plus standard of care vs standard of care. If the patient was eligible for thrombolysis, they got thrombolysis based on local guidelines. The patient had to be within 12 hours of onset of stroke, have documented medium vessel occlusion, and deficit that was commensurate with the medium vessel occlusion.
We tried to avoid patients with secondary MeVO if possible. For example, if the basal ganglia was down, then the patient was excluded and thought to be endovascularly eligible. We wanted to somewhat standardize the device use. So, the first attempt at endovascular thrombectomy had to be with a Solitaire device.
Subsequent attempts could be whatever device the physician felt comfortable with. There were very few crossovers, very few adverse events, but a summary sentence result was there was no difference in outcomes between the treatment and control arm in the Modified Rankin score or any other outcome that we looked at.
For example, infarct volume on 24-hour follow-up imaging, quality of life; the trial was completely neutral. We were unable to demonstrate any benefit for endovascular thrombectomy. There was a slight increase in adverse events in the treatment arm.
For example, there was increased pneumonia, which we thought was related to a use of general anesthesia. One good thing was there were not too many procedure-related complications. The safety analysis overall wasn't concerning. We didn't have too many intracranial hemorrhages as a consequence of the device or the thrombectomy procedure. We did not have any significant groin complications.
The DISTAL Trial
Nguyen: Outstanding. Many congratulations, Professor Goyal, Professor Hill, Dr Ospel, for this amazing accomplishment. I'd now like to turn the microphone over to Professor Urs Fischer. If you could help us summarize the DISTAL trial.
Fischer: So, the rationale is the same as in the ESCAPE-MeVO trial. We did a survey in Switzerland to see how many people with a medium or distal vessel occlusion have a good functional outcome. And we have the Swiss stroke registry, which is a nationwide registry.
And looking at this registry, we saw that modified Rankin scale of 0 to 2 in people with medium and distal vessel occlusion is roughly 49%. So, this gave us the rationale to do a trial comparing endovascular therapy plus best medical treatment vs best medical treatment alone.
The aim was almost the same as in ESCAPE-MeVO. We started in December 2021, and we finished in July 2024. It is important also to note this: the inclusion/exclusion criteria were slightly different from the ESCAPE-MeVO trial. We were mainly focusing on people with a nondominant or co-dominant M2 segment of the middle cerebral artery (MCA), the M3/4 segment, but also on the A1, A2, or A3 segment of the anterior cerebral artery, or the P1, P2, or P3 of the posterior cerebral artery. We were also excluding people who were bedridden or coming from a nursing facility or people who had an in-hospital stroke. At the end, we were able to randomize 553 people. The good news was that no patient was lost to follow-up.
Median age was 77 years, more than one third of the people were older than 80. If we compare that with the HERMES population, these people were older. The time to randomization was roughly 3.8 hours. The rate of intermediate thrombolysis was around 62%-68% in both treatment arms and, as in the ESCAPE-MeVO, at 90 days.
There was no difference in outcome in the two groups. Overall modified Rankin scale of 0 to 2 was between 55% and 57% in both treatment arms. So good functional outcome, but no difference in both treatment arms. Similar to ESCAPE-MeVO, there were no major device-related complications.
There was a small trend that people with a high NIHSS score at baseline might have a better outcome if they were treated with endovascular therapy. But overall, no clinically significant difference between the two treatment arms.
Why Were DISTAL and ESCAPE-MeVO Neutral?
Nguyen: Many congratulations, Professor Fischer, Professor Psychogios, and the DISTAL team for this amazing accomplishment.
The first question I'd like to open up is, why do you think the results of your trials were neutral? Very astonishingly similar results, similar findings, which you know, really supports the consistency of what we're finding. What do you think it was? Was it selection criteria? Was it the devices, the procedure, the endpoints? Where do you start?
Fischer: First of all, medium and distal vessel occlusion is not the same as large vessel occlusion. Functional outcome in people with a large vessel occlusion who do not receive endovascular therapy is poor.
Modified Rankin scale of 0-2 only occurs in 27% of people with a large vessel occlusion, whereas it's seen in 57% of people with a medium to distal vessel occlusion. To improve the destiny of these people is much more challenging. Second, medium and distal vessel occlusion is not a homogeneous disease. It's quite a heterogeneous disease. We all know that the clinical picture of a patient with an M2 occlusion in the left anterior MCA territory is not the same as one with an M2 in the posterior MCA territory. So, there is a huge heterogeneity, and then there are many other reasons.
The age was significantly higher in people with medium to distal vessel occlusion trials compared to the Hermes collaboration. Treatment was started later in the MeVO trials compared to the Hermes collaboration, IV tPA (tissue plasminogen activator) rates were lower. And last but not least, the inclusion/exclusion criteria were much broader in MDBO trials (mechanical device-based outcome trials) compared to the large vessel occlusion trials. What is still reassuring is we didn’t observe any major harm.
Nguyen: That's great. Over to you, Professor Goyal.
Goyal: I could make it very short and just say ditto, but I'll try to add something because I completely agree with what Dr Fischer said. The other thing that happens in any surgical randomized trial as opposed to a medical randomized trial is the issue of equipoise at the level of the institution, at the level of the collective, vs what you see when you're standing at 10 AM in front of a patient and their family.
The patients most likely to benefit from endovascular thrombectomy are, by definition, the ones that are best for thrombectomy. That's just how it goes. But if you already know who those patients are, then it’s a dilemma for those out in the field, not sitting in an academic enterprise talking about it. And that is a real struggle that happens in all surgical trials, but more so in our field because, based on the last 10 years of cumulative data, those of us down in the trenches have a gestalt about who those patients are [and we’re unlikely to randomize them]. And that may be part of the reason why the average age of patients was older.
That may be part of the reason where average NIHSS in both the trials was 7, 8 as opposed to 10, 12. It's possible that those more severe patients were not enrolled in the trial. We have to recognize that, as we move forward, I'm not sure it'll be easily modifiable. It's possible that as a consequence of these trials, there's an increase in equipoise similar to what we saw in 2013 when the trials IMS-III, etc. finished. We had a very clear idea about where we were going next.
And here it's going to be, in my opinion, multiple small, small things that will need to happen as opposed to a few big things. For example, at that time, we said let's do a CT angiogram. Let's move fast. Let's get the vessel open. That was the big picture summary in 2013; we were so committed to moving fast that the acronym for escape included this idea of a super-fast recanalization, quick workflow. I'm not sure that is that straightforward right now. Opening MeVOs is technically more challenging and more time consuming.
And there's also the part about the technical skill, and then once you go into the M3, is it easy to open or not? And in our trial, the MeVO recanalization rate was 75%, and obviously one would like it to be 95%. But I don't think it's easily achievable to make it go to 95%.
Another thing to keep in mind is that, if you look at the history — even the trials that we ran, say ESCAPE vs ESCAPE-NA-1— there was a substantial jump in 2b/3 recanalization that happened across the world. But at the same time, I don't think this is going be that straightforward.
Fischer: I fully agree that the question of clinical equipoise is most likely one of the major drivers why these trials were neutral. And I would also like to emphasize as a clinician how difficult it is if you do have a patient with NIHSS of 11 with a co-dominant or nondominant M2 occlusion. Should you sacrifice this patient for a trial when you already know that endovascular therapy can open the vessel? And we know that reperfusion is the most relevant predictor of good functional outcome.
Dominant vs Nondominant M2 and Cherry Picking
Nguyen: Thank you. These are great insights. Picking up on the issue of dominant vs nondominant M2. Professor Goyal, this has been a source of a lot of controversy. You know, many trials did not include the dominant M2 and ESCAPE-MeVO did. Could you give us a few insights into what extent were dominant M2 included in ESCAPE-MeVO, and to what extent did that affect the trial results?
Goyal: Generally speaking, when we were doing the site initiation visit, there was no willingness to randomize the so-called M1-looking M2, dominant M2. So, our expectation of having them included in the trial was exceedingly low, even at our own site. And as you know, we wrote an article from the HERMES collaboration where we talked about dominant M2, or the M1-looking M2.
People were also using that article to say that there is lack of equipoise. So essentially, we wanted to keep the inclusion criteria simple: M2, M3, A2, A3, P2, P3. We were expecting that the majority of those patients would be MCA patients as opposed to A2/A3 or P2/P3 patients. And that is how it played out in both of the trials. But we really did not expect much inclusion of M1-looking M2s.
Fischer: Same here. So, we were fully convinced that if we have a patient with an M2 occlusion, which is looking like a distal M1, we could not randomize such a patient. That's why we came up with this quite challenging definition of nondominant or dominant M2. But basically, as Mayank said, the dominant M2 is the one that looks just like an M1, but it’s a little bit more distal. And that was the reason why we did not include those patients.
It is interesting to see that in the subgroup analysis of DISTAL you see a small trend that the higher the NIHSS, it shifts toward benefit of endovascular therapy — highlighting that it is possible that there has been cherry-picking in the trial and people with the severe handicap were not randomized.
Goyal: When we were designing the trial, we looked everywhere in terms of thinking that, if the person has an NIHSS of 5 at presentation, modified Rankin score [was] the best way to assess. We've been working 2, 3 years on this right now, and we created a score called Cosmos: Canadian Outcome Score for Minor Stroke.
It's imminently going to get published. We already have quite a bit of interest in the score and how to evaluate patients presenting with relatively low NIHSS in terms of their outcome measures. The other thing that we focused on was what does the patient find important in terms of limitations of their lifestyle — and make it more patient-centric.
The limitations of the modified Rankin score, especially in this population, is something that we should be thinking about as we go forward. It is unlikely that we will have patients with NIHSS of 12 included in future trials.
Fischer: I fully agree that the modified Rankin scale is a very crude scale. Even for people with large vessel occlusion, I mean, it works. You can show a difference. Nevertheless, with minor and moderate stroke, this is quite challenging. The main problem with all the outcomes scores in stroke is that the brain is affected. So, if we do have people with aphasia, they often cannot report patient-reported outcomes. If the right hemisphere is affected, they don't even realize that they're handicapped; the proxies often see the condition much poorer than the patient themselves. We do not have a global view on the whole mobility/cognition, patient-centered outcomes yet. I’m very keen to look at Cosmos.
Intravenous Thrombolysis Therapy
Nguyen: I'm going pivot gears a little bit and think about how intravenous thrombolysis was highly present in both trials. More than 50% of patients received intravenous thrombolysis. To what extent do you think that affected the neutrality of the trials? And should future trials include patients who receive thrombolytics?
Goyal: What we found somewhat surprising is that, in fact, the early patients showed some degree of benefit of thrombectomy in ESCAPE-MeVO. And, frankly, that came as a little bit of a surprise, because it's the early patients who also got thrombolysis, right?
So, that's one part. The second part was that we had this dataset where we did a 4-hour CT angiogram in these patients, and very often we found — in the first few hours — tPA in those days, now tenecteplase, did not open the vessel.
Our thinking at that time was that we should include these patients. Based on our current datasets, it is not that the benefit was less in patients who had intravenous thrombolysis. In fact, if anything, the benefit was more in the earlier patients who had a higher likelihood of having received thrombolysis.
Fischer: I fully agree. If we look at the thrombolysis rates, it was 65% in DISTAL and I think 57% in ESCAPE-MeVO. But if we go to HERMES, it's 85%. The thrombolysis rates were lower than in the traditional trials, reflecting the fact that we were also including people in a later time window. So eventually, if you want to prove a concept, you should have narrower inclusion/exclusion criteria. DISTAL was designed in a period where we said, well, advanced imaging is no longer necessary, whatever we do with thrombectomy, it has a benefit in the late time window.
And that was the reason these trials had quite liberal inclusion/exclusion criteria compared to the large vessel occlusion (LVO) trials. Eventually, having a trial condition where there are no longer low-hanging fruit and liberal inclusion/exclusion criteria increases the risk of your trial being neutral. If we had focused only on M2 segments of the left MCA territory with an NIHSS of 12 in the first 4.5 hours, [we would have shown a benefit]. However, such a trial is simply not feasible, therefore, you are always struggling between feasibility and proving a clinical concept.
Devices in DISTAL and ESCAPE-MeVO
Nguyen: Certainly, many trade-offs between restriction vs liberalization enrollment criteria. I'm going to change gears and think about the devices. So, DISTAL was pretty liberal that you use whichever device you wanted and in ESCAPE-MeVO the solitaire stent retriever was the first line.
Could either of you comment on the choice of devices and whether that should be changed when you think about the next trial?
Goyal: If you compare the two trials, there's no substantial differences in the complication rate or in the reperfusion rate or recanalization rate. I do think overall, big picture, there's always scope for improvement.
There is this tendency sometimes to talk about the device in isolation without talking about the operator and the technique. There’s a complex interplay between these variables. Not to mention the fact that there's an interplay with the patient and anatomy, tortuosity, and the nature of the clot as well. All of these factors have to be considered together. I don't think that you can say, “Oh, these trials were negative because it used device X, Y, Z, or technique X, Y, Z, and if we jumped over to technique A, B, C, that would be the magic bullet.” I think that is unlikely. At the same time, there is always room for improvement of devices, technique, technology, and training.
Fischer: I couldn't agree more. Yes, devices can be better, they should be better, and they should be more flexible. We should also discuss how far to go with devices. We need other approaches in the very, very distal occlusions, maybe we should exclude them from endovascular therapy. Secondly, opening a large vessel occlusion is not the same as opening a distal vessel occlusion.
Training and experience, in more distal-located vessels, is still lacking, at least in some places. Last but not least: If we do randomized controlled trials, we only are talking about these first few hours and the first few minutes and the devices, but nobody's going to talk about whether the patient was admitted to a stroke unit;
whether the patient had in-hospital rehabilitation; how good was the secondary prevention. You try to minimize differences through randomization, but there's so many other factors which might play a role, and we don't know whether it was equal in both treatment arms. We shouldn't forget that if you have a condition where the overall functional outcome is already quite good and you want to further improve the outcome, it can be challenging — what is happening after the intervention could also play a role.
Observations and Reflections
Nguyen: Having done these trials, what are your take-home messages in terms of what we should do with the next MeVO patient that we face? Are there certain subgroups of patients that we should not touch with endovascular therapy? Are there patients we should say, “Yes, go with endovascular therapy,” like those dominant M2 occlusions that we spoke about.
Goyal: A couple of things. One is, if the patient has significant deficit, and the patient has a more proximal, large-ish occlusion and is early onset, and the imaging looks reasonable, it is tough to say no to thrombectomy. But also, be cognizant of your own skillset. You don't want to produce complications if the patient's deficit is relatively moderate.
Be willing to give up sooner. Don't keep trying and end up on attempt number 7. I'm very cognizant that, as I say this, I'm actually making it tougher for the next trial. That's a completely separate question that we have to tackle.
If I'm on call and I see a 63-year-old with a co-dominant M2 with an NIHSS score of 11 and the imaging looks reasonable, I will probably go ahead and open the vessel. If I see a 79-year-old with an M2 occlusion with an NIHSS of 5, who's 5 hours from onset, I probably won't.
Fischer: Yes, similar here. We need more evidence and, as a clinician scientist, I won't stop here. What we are currently planning is to pool the data. We want to see if we can find any subgroups where there might be a benefit, in order to derive a new hypothesis for future clinical trials. I think that is crucial. We will need more evidence, but it will be challenging, as Dr Goyal has said.
Are people willing to randomize a patient, 62 years old, with an NIHSS of 11, with a core M2? It's unlikely that people are willing to do that. In our clinical practice, there hasn't been a huge change. We continue treating these people given the fact that there are a lot of good reasons for why the trial was neutral, and we couldn't see any signs of harm. I think we wouldn't do it if we would see any harm. Nevertheless, we need better data for future decision-making.
Looking Ahead…but Also Looking Back
Nguyen: Fantastic. So, I'm hearing we need more data. Any insights you can share with us from what you've learned from your respective trials in terms of how you would change your next trial for part 2, presumably presuming that you might plan a part 2 trial?
Goyal: One of the things is that we need to stop doing 500 patient trials. We need to do 2000 patient trials. The low-hanging fruit is gone. I've done analyses in terms of NNT (number needed to treat), what is acceptable, what is MCID (minimal clinically important difference). About 7 years ago. I called 50 people who are well established stroke neurologists and asked for an acceptable MCID. Keeping in mind the technical challenge of thrombectomy, the cost associated with this, the infrastructure associated with it, which also plays a role in the MCID.
It's different if you're talking aspirin vs thrombectomy, and the general figure, which was amazingly consistent across 50 people, was that 5% figure, which means all of us collectively believe that, if we are able to demonstrate an absolute benefit of 5%, that's good enough from the perspective of implementation of that procedure. Which means we have to design a trial that is capable of detecting the 5% difference based on the results of DISTAL and ESCAPE-MeVO. And we haven't talked that much about the third trial that was presented that actually showed a little bit of harm, the DISCOUNT trial. I do think we cannot be over-optimistic for the next trial in terms of how it's all going play out.
If you aim for detection of a 5% effect size, that translates to a 2000 patient trial, which means it has to be a multi-country, multi-center effort across probably Western Europe, North America, Australia. There is this idea that is sometimes promoted that, “Oh yeah, don't worry about it. We'll do five small trials and merge it all together at the end.” My personal view is that is not the appropriate thing to do. We have to set up the trial design so that we are able to demonstrate a 5% benefit.
People are putting a lot of hope on this Bayesian analysis, but the reality is the Bayesian analysis, for the most part, is dependent on your posterior probability. And here, if you use DISTAL and ESCAPE-MeVO, we are back to nil.
There is this concept of adaptive trial design, which probably makes sense. But the adaptive trial design may go head-to-head with equipoise.
Fischer: So before running into the next trial, we have to clearly look at the data, try to formulate a hypothesis. What might be potential cases that might benefit? And also, what are the cases who clearly don't benefit from endovascular therapy?
If you want to prove a concept, advanced imaging will be key. But we need big data. We need the big trials, and this will need to be a joint effort for many interventionalists.
Goyal: What Urs mentioned in passing is we have a clear plan to merge the three trials. We are in an advanced stage of working out the details. We should be able to put the data of the three trials together, or for that matter, any other trial that is imminently coming to a conclusion. I know, Dr Nguyen, you're involved in Oriental MeVO; we have been in conversations with them as well. The DISCOUNT trial is not, so to speak, completed as of right now. So we are waiting for that, and that will give us 1300-1800 patients in total.
Which, as was mentioned, will not only try to answer the question of which subgroup shows benefit but also to answer the question of which subgroup clearly shows no benefit, so that we can exclude it the next time around. And I think that second part of that Urs mentioned is actually critical.
Final Thoughts
Nguyen: Wonderful. Any closing remarks?
Fischer: As a clinician scientist after the trial is before the trial. But we have to really be cautious [with] how we design these types of trials, because trials are extremely expensive. They need a lot of effort. You have to plan these things wisely. And I'm looking forward to the next endeavor.
Goyal: I like to talk about this term, which you'll soon see in print as well. And we just had an article on this accepted in Stroke, which I called W-I-I-F-U, What's In It For Us? And we have to think about this from the perspective of younger people who are our junior colleagues, who are the boots on the ground doing the hard work. Are we able to collectively create enough of a WIIFU for them so that they feel motivated?
Nguyen: Thank you. Thank you both very much indeed. That was an outstanding discussion.
COMMENTARY
Thrombectomy in Distal Stroke Overreach or Fine Tuning Needed?
DISCLOSURES
| April 10, 2025This transcript has been edited for clarity.
Thanh Nguyen, MD: Good day, everyone. Thank you very much for joining us today. In this session for Medscape, we're going to be discussing medium and distal vessel occlusions hot off the press after the International Stroke Conference presentations. My name is Thanh Nguyen. I'm a stroke and interventional neurologist, and I'd like to send the microphone over to our colleagues, professor Urs Fischer, who is the principal investigator of the DISTAL trial.
Urs Fischer, MD, PhD: My name is Urs Fischer. I am a clinician scientist, and I am the director and chairman of the Department of Neurology at University Hospital in Bern, Switzerland. I'm president-elect of the European Stroke Organization.
Nguyen: Professor Mayank Goyal, if you could introduce yourself.
Mayank Goyal, MD, PhD: My name is Mayank Goyal. I'm in Calgary. I'm a neuro interventionist, and along with my colleague Michael Hill, we led the ESCAPE-MeVO trial.
The ESCAPE-MeVO Trial
Nguyen: Wonderful. We'll begin with Dr. Goyal about a brief summary of what you learned from the ESCAPE-MeVO trial, and why did you conduct this trial?
Goyal: Starting in 2015, there were all these dramatic changes in stroke care related to large vessel occlusion. And as time went by, it was becoming increasingly difficult to find a subgroup in which there was no benefit of endovascular thrombectomy. In the meantime, what we found was we were getting better at tackling medium vessel occlusions. Technology was getting better, technique was getting better. And starting around 5, 6 years ago, we started working on this idea.
We wrote a couple of editorials. We defined what "MeVO" (medium vessel occlusion) is. We were very fortunate to have a large dataset here in Calgary, and we wrote papers on natural history. We wrote papers on primary vs secondary medium vessel occlusion. And overall, what we realized based on the dataset that was available to us, less than 50% of patients reach modified Rankin scale score of 0 or 1 with so-called standard of care.
So that's when we started exploring the idea that can we improve outcomes further, given that they were not the greatest to begin with, and we started to design what ended up becoming an ESCAPE-MeVO trial. The trial was a standard probe designed, open-label, randomized trial. We conducted across Europe and North America.
A total of 530 patients were randomized 1:1. The treatment arm was endovascular thrombectomy plus standard of care vs standard of care. If the patient was eligible for thrombolysis, they got thrombolysis based on local guidelines. The patient had to be within 12 hours of onset of stroke, have documented medium vessel occlusion, and deficit that was commensurate with the medium vessel occlusion.
We tried to avoid patients with secondary MeVO if possible. For example, if the basal ganglia was down, then the patient was excluded and thought to be endovascularly eligible. We wanted to somewhat standardize the device use. So, the first attempt at endovascular thrombectomy had to be with a Solitaire device.
Subsequent attempts could be whatever device the physician felt comfortable with. There were very few crossovers, very few adverse events, but a summary sentence result was there was no difference in outcomes between the treatment and control arm in the Modified Rankin score or any other outcome that we looked at.
For example, infarct volume on 24-hour follow-up imaging, quality of life; the trial was completely neutral. We were unable to demonstrate any benefit for endovascular thrombectomy. There was a slight increase in adverse events in the treatment arm.
For example, there was increased pneumonia, which we thought was related to a use of general anesthesia. One good thing was there were not too many procedure-related complications. The safety analysis overall wasn't concerning. We didn't have too many intracranial hemorrhages as a consequence of the device or the thrombectomy procedure. We did not have any significant groin complications.
The DISTAL Trial
Nguyen: Outstanding. Many congratulations, Professor Goyal, Professor Hill, Dr Ospel, for this amazing accomplishment. I'd now like to turn the microphone over to Professor Urs Fischer. If you could help us summarize the DISTAL trial.
Fischer: So, the rationale is the same as in the ESCAPE-MeVO trial. We did a survey in Switzerland to see how many people with a medium or distal vessel occlusion have a good functional outcome. And we have the Swiss stroke registry, which is a nationwide registry.
And looking at this registry, we saw that modified Rankin scale of 0 to 2 in people with medium and distal vessel occlusion is roughly 49%. So, this gave us the rationale to do a trial comparing endovascular therapy plus best medical treatment vs best medical treatment alone.
The aim was almost the same as in ESCAPE-MeVO. We started in December 2021, and we finished in July 2024. It is important also to note this: the inclusion/exclusion criteria were slightly different from the ESCAPE-MeVO trial. We were mainly focusing on people with a nondominant or co-dominant M2 segment of the middle cerebral artery (MCA), the M3/4 segment, but also on the A1, A2, or A3 segment of the anterior cerebral artery, or the P1, P2, or P3 of the posterior cerebral artery. We were also excluding people who were bedridden or coming from a nursing facility or people who had an in-hospital stroke. At the end, we were able to randomize 553 people. The good news was that no patient was lost to follow-up.
Median age was 77 years, more than one third of the people were older than 80. If we compare that with the HERMES population, these people were older. The time to randomization was roughly 3.8 hours. The rate of intermediate thrombolysis was around 62%-68% in both treatment arms and, as in the ESCAPE-MeVO, at 90 days.
There was no difference in outcome in the two groups. Overall modified Rankin scale of 0 to 2 was between 55% and 57% in both treatment arms. So good functional outcome, but no difference in both treatment arms. Similar to ESCAPE-MeVO, there were no major device-related complications.
There was a small trend that people with a high NIHSS score at baseline might have a better outcome if they were treated with endovascular therapy. But overall, no clinically significant difference between the two treatment arms.
Why Were DISTAL and ESCAPE-MeVO Neutral?
Nguyen: Many congratulations, Professor Fischer, Professor Psychogios, and the DISTAL team for this amazing accomplishment.
The first question I'd like to open up is, why do you think the results of your trials were neutral? Very astonishingly similar results, similar findings, which you know, really supports the consistency of what we're finding. What do you think it was? Was it selection criteria? Was it the devices, the procedure, the endpoints? Where do you start?
Fischer: First of all, medium and distal vessel occlusion is not the same as large vessel occlusion. Functional outcome in people with a large vessel occlusion who do not receive endovascular therapy is poor.
Modified Rankin scale of 0-2 only occurs in 27% of people with a large vessel occlusion, whereas it's seen in 57% of people with a medium to distal vessel occlusion. To improve the destiny of these people is much more challenging. Second, medium and distal vessel occlusion is not a homogeneous disease. It's quite a heterogeneous disease. We all know that the clinical picture of a patient with an M2 occlusion in the left anterior MCA territory is not the same as one with an M2 in the posterior MCA territory. So, there is a huge heterogeneity, and then there are many other reasons.
The age was significantly higher in people with medium to distal vessel occlusion trials compared to the Hermes collaboration. Treatment was started later in the MeVO trials compared to the Hermes collaboration, IV tPA (tissue plasminogen activator) rates were lower. And last but not least, the inclusion/exclusion criteria were much broader in MDBO trials (mechanical device-based outcome trials) compared to the large vessel occlusion trials. What is still reassuring is we didn’t observe any major harm.
Nguyen: That's great. Over to you, Professor Goyal.
Goyal: I could make it very short and just say ditto, but I'll try to add something because I completely agree with what Dr Fischer said. The other thing that happens in any surgical randomized trial as opposed to a medical randomized trial is the issue of equipoise at the level of the institution, at the level of the collective, vs what you see when you're standing at 10 AM in front of a patient and their family.
The patients most likely to benefit from endovascular thrombectomy are, by definition, the ones that are best for thrombectomy. That's just how it goes. But if you already know who those patients are, then it’s a dilemma for those out in the field, not sitting in an academic enterprise talking about it. And that is a real struggle that happens in all surgical trials, but more so in our field because, based on the last 10 years of cumulative data, those of us down in the trenches have a gestalt about who those patients are [and we’re unlikely to randomize them]. And that may be part of the reason why the average age of patients was older.
That may be part of the reason where average NIHSS in both the trials was 7, 8 as opposed to 10, 12. It's possible that those more severe patients were not enrolled in the trial. We have to recognize that, as we move forward, I'm not sure it'll be easily modifiable. It's possible that as a consequence of these trials, there's an increase in equipoise similar to what we saw in 2013 when the trials IMS-III, etc. finished. We had a very clear idea about where we were going next.
And here it's going to be, in my opinion, multiple small, small things that will need to happen as opposed to a few big things. For example, at that time, we said let's do a CT angiogram. Let's move fast. Let's get the vessel open. That was the big picture summary in 2013; we were so committed to moving fast that the acronym for escape included this idea of a super-fast recanalization, quick workflow. I'm not sure that is that straightforward right now. Opening MeVOs is technically more challenging and more time consuming.
And there's also the part about the technical skill, and then once you go into the M3, is it easy to open or not? And in our trial, the MeVO recanalization rate was 75%, and obviously one would like it to be 95%. But I don't think it's easily achievable to make it go to 95%.
Another thing to keep in mind is that, if you look at the history — even the trials that we ran, say ESCAPE vs ESCAPE-NA-1— there was a substantial jump in 2b/3 recanalization that happened across the world. But at the same time, I don't think this is going be that straightforward.
Fischer: I fully agree that the question of clinical equipoise is most likely one of the major drivers why these trials were neutral. And I would also like to emphasize as a clinician how difficult it is if you do have a patient with NIHSS of 11 with a co-dominant or nondominant M2 occlusion. Should you sacrifice this patient for a trial when you already know that endovascular therapy can open the vessel? And we know that reperfusion is the most relevant predictor of good functional outcome.
Dominant vs Nondominant M2 and Cherry Picking
Nguyen: Thank you. These are great insights. Picking up on the issue of dominant vs nondominant M2. Professor Goyal, this has been a source of a lot of controversy. You know, many trials did not include the dominant M2 and ESCAPE-MeVO did. Could you give us a few insights into what extent were dominant M2 included in ESCAPE-MeVO, and to what extent did that affect the trial results?
Goyal: Generally speaking, when we were doing the site initiation visit, there was no willingness to randomize the so-called M1-looking M2, dominant M2. So, our expectation of having them included in the trial was exceedingly low, even at our own site. And as you know, we wrote an article from the HERMES collaboration where we talked about dominant M2, or the M1-looking M2.
People were also using that article to say that there is lack of equipoise. So essentially, we wanted to keep the inclusion criteria simple: M2, M3, A2, A3, P2, P3. We were expecting that the majority of those patients would be MCA patients as opposed to A2/A3 or P2/P3 patients. And that is how it played out in both of the trials. But we really did not expect much inclusion of M1-looking M2s.
Fischer: Same here. So, we were fully convinced that if we have a patient with an M2 occlusion, which is looking like a distal M1, we could not randomize such a patient. That's why we came up with this quite challenging definition of nondominant or dominant M2. But basically, as Mayank said, the dominant M2 is the one that looks just like an M1, but it’s a little bit more distal. And that was the reason why we did not include those patients.
It is interesting to see that in the subgroup analysis of DISTAL you see a small trend that the higher the NIHSS, it shifts toward benefit of endovascular therapy — highlighting that it is possible that there has been cherry-picking in the trial and people with the severe handicap were not randomized.
Goyal: When we were designing the trial, we looked everywhere in terms of thinking that, if the person has an NIHSS of 5 at presentation, modified Rankin score [was] the best way to assess. We've been working 2, 3 years on this right now, and we created a score called Cosmos: Canadian Outcome Score for Minor Stroke.
It's imminently going to get published. We already have quite a bit of interest in the score and how to evaluate patients presenting with relatively low NIHSS in terms of their outcome measures. The other thing that we focused on was what does the patient find important in terms of limitations of their lifestyle — and make it more patient-centric.
The limitations of the modified Rankin score, especially in this population, is something that we should be thinking about as we go forward. It is unlikely that we will have patients with NIHSS of 12 included in future trials.
Fischer: I fully agree that the modified Rankin scale is a very crude scale. Even for people with large vessel occlusion, I mean, it works. You can show a difference. Nevertheless, with minor and moderate stroke, this is quite challenging. The main problem with all the outcomes scores in stroke is that the brain is affected. So, if we do have people with aphasia, they often cannot report patient-reported outcomes. If the right hemisphere is affected, they don't even realize that they're handicapped; the proxies often see the condition much poorer than the patient themselves. We do not have a global view on the whole mobility/cognition, patient-centered outcomes yet. I’m very keen to look at Cosmos.
Intravenous Thrombolysis Therapy
Nguyen: I'm going pivot gears a little bit and think about how intravenous thrombolysis was highly present in both trials. More than 50% of patients received intravenous thrombolysis. To what extent do you think that affected the neutrality of the trials? And should future trials include patients who receive thrombolytics?
Goyal: What we found somewhat surprising is that, in fact, the early patients showed some degree of benefit of thrombectomy in ESCAPE-MeVO. And, frankly, that came as a little bit of a surprise, because it's the early patients who also got thrombolysis, right?
So, that's one part. The second part was that we had this dataset where we did a 4-hour CT angiogram in these patients, and very often we found — in the first few hours — tPA in those days, now tenecteplase, did not open the vessel.
Our thinking at that time was that we should include these patients. Based on our current datasets, it is not that the benefit was less in patients who had intravenous thrombolysis. In fact, if anything, the benefit was more in the earlier patients who had a higher likelihood of having received thrombolysis.
Fischer: I fully agree. If we look at the thrombolysis rates, it was 65% in DISTAL and I think 57% in ESCAPE-MeVO. But if we go to HERMES, it's 85%. The thrombolysis rates were lower than in the traditional trials, reflecting the fact that we were also including people in a later time window. So eventually, if you want to prove a concept, you should have narrower inclusion/exclusion criteria. DISTAL was designed in a period where we said, well, advanced imaging is no longer necessary, whatever we do with thrombectomy, it has a benefit in the late time window.
And that was the reason these trials had quite liberal inclusion/exclusion criteria compared to the large vessel occlusion (LVO) trials. Eventually, having a trial condition where there are no longer low-hanging fruit and liberal inclusion/exclusion criteria increases the risk of your trial being neutral. If we had focused only on M2 segments of the left MCA territory with an NIHSS of 12 in the first 4.5 hours, [we would have shown a benefit]. However, such a trial is simply not feasible, therefore, you are always struggling between feasibility and proving a clinical concept.
Devices in DISTAL and ESCAPE-MeVO
Nguyen: Certainly, many trade-offs between restriction vs liberalization enrollment criteria. I'm going to change gears and think about the devices. So, DISTAL was pretty liberal that you use whichever device you wanted and in ESCAPE-MeVO the solitaire stent retriever was the first line.
Could either of you comment on the choice of devices and whether that should be changed when you think about the next trial?
Goyal: If you compare the two trials, there's no substantial differences in the complication rate or in the reperfusion rate or recanalization rate. I do think overall, big picture, there's always scope for improvement.
There is this tendency sometimes to talk about the device in isolation without talking about the operator and the technique. There’s a complex interplay between these variables. Not to mention the fact that there's an interplay with the patient and anatomy, tortuosity, and the nature of the clot as well. All of these factors have to be considered together. I don't think that you can say, “Oh, these trials were negative because it used device X, Y, Z, or technique X, Y, Z, and if we jumped over to technique A, B, C, that would be the magic bullet.” I think that is unlikely. At the same time, there is always room for improvement of devices, technique, technology, and training.
Fischer: I couldn't agree more. Yes, devices can be better, they should be better, and they should be more flexible. We should also discuss how far to go with devices. We need other approaches in the very, very distal occlusions, maybe we should exclude them from endovascular therapy. Secondly, opening a large vessel occlusion is not the same as opening a distal vessel occlusion.
Training and experience, in more distal-located vessels, is still lacking, at least in some places. Last but not least: If we do randomized controlled trials, we only are talking about these first few hours and the first few minutes and the devices, but nobody's going to talk about whether the patient was admitted to a stroke unit;
whether the patient had in-hospital rehabilitation; how good was the secondary prevention. You try to minimize differences through randomization, but there's so many other factors which might play a role, and we don't know whether it was equal in both treatment arms. We shouldn't forget that if you have a condition where the overall functional outcome is already quite good and you want to further improve the outcome, it can be challenging — what is happening after the intervention could also play a role.
Observations and Reflections
Nguyen: Having done these trials, what are your take-home messages in terms of what we should do with the next MeVO patient that we face? Are there certain subgroups of patients that we should not touch with endovascular therapy? Are there patients we should say, “Yes, go with endovascular therapy,” like those dominant M2 occlusions that we spoke about.
Goyal: A couple of things. One is, if the patient has significant deficit, and the patient has a more proximal, large-ish occlusion and is early onset, and the imaging looks reasonable, it is tough to say no to thrombectomy. But also, be cognizant of your own skillset. You don't want to produce complications if the patient's deficit is relatively moderate.
Be willing to give up sooner. Don't keep trying and end up on attempt number 7. I'm very cognizant that, as I say this, I'm actually making it tougher for the next trial. That's a completely separate question that we have to tackle.
If I'm on call and I see a 63-year-old with a co-dominant M2 with an NIHSS score of 11 and the imaging looks reasonable, I will probably go ahead and open the vessel. If I see a 79-year-old with an M2 occlusion with an NIHSS of 5, who's 5 hours from onset, I probably won't.
Fischer: Yes, similar here. We need more evidence and, as a clinician scientist, I won't stop here. What we are currently planning is to pool the data. We want to see if we can find any subgroups where there might be a benefit, in order to derive a new hypothesis for future clinical trials. I think that is crucial. We will need more evidence, but it will be challenging, as Dr Goyal has said.
Are people willing to randomize a patient, 62 years old, with an NIHSS of 11, with a core M2? It's unlikely that people are willing to do that. In our clinical practice, there hasn't been a huge change. We continue treating these people given the fact that there are a lot of good reasons for why the trial was neutral, and we couldn't see any signs of harm. I think we wouldn't do it if we would see any harm. Nevertheless, we need better data for future decision-making.
Looking Ahead…but Also Looking Back
Nguyen: Fantastic. So, I'm hearing we need more data. Any insights you can share with us from what you've learned from your respective trials in terms of how you would change your next trial for part 2, presumably presuming that you might plan a part 2 trial?
Goyal: One of the things is that we need to stop doing 500 patient trials. We need to do 2000 patient trials. The low-hanging fruit is gone. I've done analyses in terms of NNT (number needed to treat), what is acceptable, what is MCID (minimal clinically important difference). About 7 years ago. I called 50 people who are well established stroke neurologists and asked for an acceptable MCID. Keeping in mind the technical challenge of thrombectomy, the cost associated with this, the infrastructure associated with it, which also plays a role in the MCID.
It's different if you're talking aspirin vs thrombectomy, and the general figure, which was amazingly consistent across 50 people, was that 5% figure, which means all of us collectively believe that, if we are able to demonstrate an absolute benefit of 5%, that's good enough from the perspective of implementation of that procedure. Which means we have to design a trial that is capable of detecting the 5% difference based on the results of DISTAL and ESCAPE-MeVO. And we haven't talked that much about the third trial that was presented that actually showed a little bit of harm, the DISCOUNT trial. I do think we cannot be over-optimistic for the next trial in terms of how it's all going play out.
If you aim for detection of a 5% effect size, that translates to a 2000 patient trial, which means it has to be a multi-country, multi-center effort across probably Western Europe, North America, Australia. There is this idea that is sometimes promoted that, “Oh yeah, don't worry about it. We'll do five small trials and merge it all together at the end.” My personal view is that is not the appropriate thing to do. We have to set up the trial design so that we are able to demonstrate a 5% benefit.
People are putting a lot of hope on this Bayesian analysis, but the reality is the Bayesian analysis, for the most part, is dependent on your posterior probability. And here, if you use DISTAL and ESCAPE-MeVO, we are back to nil.
There is this concept of adaptive trial design, which probably makes sense. But the adaptive trial design may go head-to-head with equipoise.
Fischer: So before running into the next trial, we have to clearly look at the data, try to formulate a hypothesis. What might be potential cases that might benefit? And also, what are the cases who clearly don't benefit from endovascular therapy?
If you want to prove a concept, advanced imaging will be key. But we need big data. We need the big trials, and this will need to be a joint effort for many interventionalists.
Goyal: What Urs mentioned in passing is we have a clear plan to merge the three trials. We are in an advanced stage of working out the details. We should be able to put the data of the three trials together, or for that matter, any other trial that is imminently coming to a conclusion. I know, Dr Nguyen, you're involved in Oriental MeVO; we have been in conversations with them as well. The DISCOUNT trial is not, so to speak, completed as of right now. So we are waiting for that, and that will give us 1300-1800 patients in total.
Which, as was mentioned, will not only try to answer the question of which subgroup shows benefit but also to answer the question of which subgroup clearly shows no benefit, so that we can exclude it the next time around. And I think that second part of that Urs mentioned is actually critical.
Final Thoughts
Nguyen: Wonderful. Any closing remarks?
Fischer: As a clinician scientist after the trial is before the trial. But we have to really be cautious [with] how we design these types of trials, because trials are extremely expensive. They need a lot of effort. You have to plan these things wisely. And I'm looking forward to the next endeavor.
Goyal: I like to talk about this term, which you'll soon see in print as well. And we just had an article on this accepted in Stroke, which I called W-I-I-F-U, What's In It For Us? And we have to think about this from the perspective of younger people who are our junior colleagues, who are the boots on the ground doing the hard work. Are we able to collectively create enough of a WIIFU for them so that they feel motivated?
Nguyen: Thank you. Thank you both very much indeed. That was an outstanding discussion.
Any views expressed above are the author's own and do not necessarily reflect the views of WebMD or Medscape.
TOP PICKS FOR YOU