Expert Conversations: Neuroprotection for Acute Ischemic Stroke

In this podcast, Neurology Learning Network’s Stroke & Vascular Section Editor Amrou Sarraj, MD, interviews Michael Hill, MD, FAHA, FCAHS, FESO, FRCPC, University of Calgary, on the past, present, and future of neuroprotection treatments for acute ischemic stroke.

Dr. Amrou Sarraj:  Greetings. This is Dr. Amrou Sarraj. I'm an associate professor of neurology at UT McGovern. In this podcast, we will focus on neuroprotection as an adjunctive treatment for acute ischemic strokes. No one better than an expert in the subject to be our guest for this podcast. We welcome Dr. Michael Hill, professor of neurology at University of Calgary, and the senior medical director of stroke at Alberta Health Service.

Dr. Hill is an expert in the subject as I mentioned, and had been involved in several clinical trials for more than 10 years now with ALIAS I and II. Then most recently, ESCAPE‑NA‑1, and then ESCAPE‑NEXT. Thank you for joining us Dr. Hill.

Dr. Michael Hill:  Thanks, Amrou. It's a pleasure to be here with you.

Dr. Sarraj:  Wonderful. We'll jump right into it and as our listeners know, the mainstay for acute ischemic stroke treatment is reperfusion with thrombolysis or with thrombectomy. The adjunctive treatment is which we all hope for to improve the outcomes in neuroprotection. It has been tried in the past. Let's start with some of the mechanisms and hypotheses that we hope this works through apart from reperfusion and protecting tissue from dying through reperfusion.

Dr. Hill:  I think, Amrou, that the start of this was the evolution of the excitotoxicity hypothesis. That's 30 years ago that this mechanism of cell death was uncovered and it related to ischemia producing excessive release of glutamate in the extracellular space. Glutamate, of course, is a neurotransmitter. It binds to transmitter receptors or the NMDA receptors and also the AMPA receptors. NMDA is N‑methyl‑D‑aspartate. Glutamate was the prototypic neurotransmitter. In the setting of ischemia, what happened is all these receptors got flooded, their voltage‑gated calcium channels. Calcium comes into the cell as a massive influx and that resulted in a number of cell‑signaling pathways which ultimately produced cell death. There are two different pathways that were elucidated. One was the direct cell death, a necrotic, rapid cell death, probably associated with things like free‑radical production inside the cell and this kind of thing. Also an apoptotic cell death where sub‑lethal damage was done and then the cell wasn't able to recover and then underwent a program cell death. This hypothesis then stimulated a whole bunch of efforts to try to block the MDA receptors. We had lots of different compounds that the prototypic one was called MK‑801. It came from Merck. It was clearly shown in rodent models to reduce infarct volume substantially in rodent models of stroke, turned out to be pretty toxic in humans.

It never really got to substantial human trials, but other compounds did and unfortunately...I think Jeff Donnan's group...What's her name? Victoria O'Connor did a nice meta‑analysis of all these compounds that have been checked and it's published in...I think it's in "Annals of Neurology." The title is "1,026 Molecules Have all Essentially Failed." It's not really a 1,026 because only about a 100 got to her or 10 percent even got to human trials. Of the ones that got to human trials, only a proportion of those got to phase II or phase III level trials so it's not...It's a little bit of an exaggeration but... The bottom line was that all of these compounds never got to the clinic and some of them were toxic, some of them were tragic. Sulfatel which was a tragic story because there was a massive...The trial was stopped early by the DSMC because there was quite a large affect size of increasing mortality in the treatment group, and so that's an example.

It ultimately ended up, I think the last one that was the final nail in the coffin was the SAINT II trial. Looked at this compound called NXY‑059. It was done globally, massive trial, like 3,000 patients...Absolutely plum neutral and then that was it. Everybody just faded away from neuroprotection.

Dr. Sarraj:  Actually, this is a sad fact but very good leeway to our next question which is a lot of people are interested in what you alluded to several attempts were done to test neuroprotection in the setting of acute ischemic stroke, whether good promising results in the brief clinical stroke animal models. However, using the transnational research, the results did not translate positive outcomes when testing in clinical trials. What are some of these? You name, maybe some of these agents were not the right ones.

I'm not sure. Science is always great, and I am not trying to criticize others, but was their lack of rigor and testing them in the pre‑clinical setting? We mentioned that the mainstay of this drug treatment is reperfusion, and you have to deliver the agent to the tissue in order for it to work. We will talk more about ESCAPE‑NA 1 and the promising results that you all had. The lack of re‑canalization, you're not delivering the agent, so the agent will not work. Is it the fact that it's not plausible? There is no clear mechanism? Some of the people comment on hypothermia for that. What are your thoughts on why we never got hold?

Dr. Hill:  Like many things, I suspect multi‑factorial reasons, and specific to individual agents. Amrou, I think there's been a whole literature on thinking of reasons why these things didn't work. One was, for sure, toxicity. In the NMB receptors and the cell signaling that goes with it and action potentials that are developed when you have a synaptic transmission using glutamate are important for human function. When you blocked that mechanism in certain ways, it was pretty negative for people. For example, they got hypothermic, they became psychotic, became comatose. Toxicity was one. That's a drug thing. Drug might work, but if it's so toxic, then you can't. It doesn't work. That's one.

Another one, and you alluded to it in your discussion, is that the question of whether we got the right models. Interestingly, the models that have looked at nerinetide or NA‑1 are similar to when used for other previous models, looking at these molecules, like selfotel or MK-801.

The question is, what's different? One of the key differences that in the intervening time, we now have much more reliable true ischemia reperfusion models. The vast majority of studies done in animal setting are done with ischemia reperfusion, ischemia for three hours, four hours, two hours, whatever, and then blood flow is returned. These compounds seem to work on top of that. If you have permanent ischemia, there is a much smaller affect size of these compounds in rodents and rabbits and other animal models. The fact that, for example, IV thrombolysis will open an M‑1 artery at four hours after delivery, only a third of the time, tells you that we didn't have a human model, which was mimicking what was shown in the rat, and mouse and rabbit and other models. That's important.

There was a big, strong argument made in the past about the fact that we should use primates. We should stop with the rabbits and rats. We should use older rats and all this kind of stuff. Certainly, for nerinetide, one of the key in NA‑1 that is right, the key things that have been done, it's really impressive is that the primates have been used. Was it in retrospect, absolutely necessary? I don't know. Time will tell. The fundamental difference between what happened in the early 2000s, and what's happening now is that we have a true ischemia reperfusion situation in the humans and we just didn't have that before. If we're getting TICI 2B-3 rates of 90 percent now, 85 to 90 percent, compared to say, 30 percent with IV thrombolysis. We can see that there's just a huge difference. That to me is the dominant difference now.

Dr. Sarraj:  Absolutely. I agree, just thinking of the prior even thrombectomy attempts and the trials that did not show positive results at 20 and 25 percent reperfusion rates and now jumping to 80, 90 percent. It's fast reperfusion.

Before, the cascade that you mentioned earlier happened with neuroinflammation and toxicity occur, we can deliver the agent. Wonderful.

Dr. Hill:  Can I jump on that? I wanted to say that's a really, really nice point. The speed is important, and I think...We don't know this for sure yet...But I suspect that if we can show that there's an adjunctive agent, say it's nerinetide in this case, onset or door to nerinetide time will be as important as door to thrombolysis, door to groin puncture, door to reperfusion. All of those matrice are going to be super important. We didn't do that. Interestingly, if you look at the first three endovascular trials, synthesis, MR RESCUE, and IMSIII, they both all‑important trials, but in fact, in those trials, we did not stress the workflow, the interval times, as we do now.

Dr. Sarraj:  It took forever to reperfuse if reperfusion happened. Hoping for a future success, but it's always good to go back to the past and look at what might have happened. You were involved in ALIAS 1 and 2 in testing albumin. I would say that I was a resident at the time...

Dr. Hill:  [laughs]

Dr. Sarraj:  ...and enrolled some of the patients here. I was very optimistic and then was stopped for adverse event, cardiac adverse event. In brief some thoughts on the trials.

Dr. Hill:  First of all, there was a large body of preclinical evidence. Myron Ginsberg from Miami was the lead scientist and then the overall PI for the ALIAS trials. Myron's work, affirmed by others, showed that giving albumin was useful in the rodent models of stroke, different types, and different mechanisms, and different durations. It looked pretty robust. I have to say the evidence was very strong. We certainly got into trouble in humans. One suspects that it was at least a significant portion of it was the toxicity problems. That by giving the dose of albumin that we gave at the speed that we gave it, people got into trouble. They got into trouble.

Dr. Sarraj:  Just heart failure?

Dr. Hill:  Yeah. It would have been a heart failure, or we caused a bit more stress on the left ventricle, and they got ischemia or whatever. I have to tell you that looking at all the data afterward, it was very difficult to definitively show that from adverse events. It's in part because older folks with a bit of heart failure, they got into different problems with stroke. It wasn't the adverse events were all heart failure. It was not clear. Nevertheless, toxicity was one side. We changed the trial. We, as you know, and continued on in ALIAS part 2. We tightened up on the restrictions. We were able to avoid some of the safety problems, and then never were again able to show an effect size.

I have to wonder now if part of the problem is what we've seen in stroke over the years, which is that there's going to be specific therapy for specific stroke types. When you got an LVO, the current solution is give thrombolysis and go get the clot with EVT and open it up. We know what that is. When you've got a lacune, the solution is not endovascular therapy. It's an obvious example. In ALIAS, of course, we had all comers being enrolled in the trial. I have to wonder whether perhaps albumin is effective but only maybe in large vessel occlusion. I don't know. I suspect we're never going to know.

Well, maybe another trial will be possible, but increasingly in stroke, you see that the specificity of what you're doing is important. You can apply some therapy to all stroke.

Dr. Sarraj:  Definitely. We always learn from those things. As you mentioned, stroke patients are heterogenous group patients and moving from that, they had heterogenous group patients to that homogenous group of patients with large vessel occlusions and moving from ALIAS to what I believe was a successful trial in escaping NA‑1 through 30 promising results where the lead author on the escaping NA‑1 that tested the potential adjunctive benefit of nerinetide in patients with large vascular occlusions receiving thrombectomy.

If we can discuss agent mechanism and the design and the results of the trials. We're going to dig deeper into some of the specifics.

Dr. Hill:  The story behind the story so to speak. There's a couple of interesting things. I'll speak a little bit about the pharmacology if that's all right. Then...

Dr. Sarraj:  Sure.

Dr. Hill:  ...I'd like to make a few comments about how the trial was designed and what the history was...

Dr. Sarraj:  Sure.

Dr. Hill:  ...because it's helpful in terms of thinking about interpretation. First of all, NA‑1 now has a name. It's called nerinetide. That's going to be its formal generic name. It's a peptide. It was designed in 1999 in the lab to inhibit a mechanism of cell death and understand a mechanism of cell death. The fact that it's made it all the way to human testing from that perspective is quite remarkable. It could've been highly toxic, and lots of things could have gone wrong. In that way, it's highly serendipitous. The molecule is an example of rational drug design. It's a peptide which inhibits the binding of a cytosolic protein called PSD‑95 to the cytosolic surface of NMDA receptor, one of the NR2B subunits. What it does is that it inhibits a cell signaling pathway that occurs when the glutamate binds and calcium comes in through the channel. In this case, what's you're doing is inhibiting cell signaling that results in overproduction of nitric oxide. You inhibit free radical production inside the cell. One important conceptual difference is that this molecule's mechanism of action is cytosolic intracellular.

Whereas previous so‑called cytoprotective or neuroprotective agents, their mechanism of action was extracellular at the outside surface of the cell at the channel binding sites. Lots of great data to show that this works in animals, including a terrific primate model which mimics what happens in human strokes. We did this great data that this thing works. When we were designing the trial, the way this had all come together, is that we'd been involved with the group developing this. The compound was developed in an academic lab in Toronto at the Toronto Western Hospital. Mike Tymianski and his colleague Mike Salter were the supervising props. Michelle Arts was the PhD student who first developed and reported the compound. It's published in science in '99. They showed this mechanism of cell death. Then, they incorporated a company to try to take the intellectual property forward.

The interesting thing was that this compound also is an analgesic compound. It was initially thought that it would be ideal for pain and pain control. One of the things we've seen in the adverse events profile in the stroke trial is that it results in less pain for the people treated. It's interesting that there are other things going on with the compound. Nevertheless, the stroke side was one bent avenue of this. As we were in Calgary, I have interacted with the group in Toronto for at least 15 years. We've been helping them on the clinical side. I'm not a lab scientist by any means, but on the clinical side, we have been helping.

We did first the EMAX trial which was looking at this compound in coiling of aneurysms. We used MRI to show that you had reduced MRI‑defined strokes when in the NA‑1 treated group. In Calgary, we were part of multiple groups around the world looking at endovascular therapy. We did the ESCAPE trial, and then that was obvious that the natural marriage would be to look at nerinetide in this. That's how it began to develop. As we go through this, we were designing ESCAPE‑NA‑1, thinking about the paradigm of how we done it in ESCAPE, and in particular, with alteplase. I remember we had a great interaction with, in this case, the US FDA. We were interacting with Health Canada, the European Medicines Agency and the Korean FDA.

With the US FDA, one of the reviewers of the application said, "Listen, what about a drug‑drug interaction between alteplase and nerinetide." We said, "Yeah, good point." We've done some experiments on this and we've shown that if you mix alteplase and nerinetide together in a test tube, nothing happens. It's inert. Nothing happens to alteplase, nothing happens to nerinetide. We know that if you give nerinetide before alteplase in the Rat model, you still preserve the neuroprotective effect. We think we're going to be OK, but what we're going to do is stratify the randomization on alteplase or not, because that effectively a priori allows you to have a randomized comparison in the alteplase group and a randomized comparison in the no alteplase group, and then we're going to run the trial. We knew a priori that there might be drug‑drug interactions but we didn't...As you know we've shown a dramatic drug‑drug interaction. We didn't know it was going to be this bad or this significant. We did design the trial to look at the two groups separately. When we ran the trial, it was fun. We got 48 centers around the world, mostly in North America but also Europe, and South Korea and Australia contributed.

I think a couple of things that were good about it were we had things that we stressed where we wanted high reperfusion rates, same paradigm because we want ischemia‑reperfusion then we add the drug on top. We had 87 percent TICI 2b-3, which for 48 sites around, we thought that was good. I'd love to see it better for the next trial. The other thing we wanted was, we wanted to get the drug in, on average half an hour before reperfusion because that's what had been shown in the primate model that if we got the drug in early it had a chance to get in and get into the brain. Then you reperfuse, that seemed to be the best model. If people recall, one of the things we stressed was you had to have pretty good collaterals. Now some new interesting data from Mirko Pham, in Germany, and he's looked at doing intra‑arterial sampling of alteplase and measuring alteplase levels, you give alteplase intravenously at the clot and then through the clot.

You can identify that...If you have good collaterals, tPA gets to the back end of the clot quite well, and the concentrations are therapeutic. We can infer the same thing that if you have good collaterals drug gets to where it needs to get to. If you have no collaterals all bets are off, right on this kind of thing. I think that the fortuitous, and that's a fortuitous thing because that paper has only been published this year. The concept that it's the collaterals that bring compounds to the ischemic tissue is the pial collaterals was a critical one. The main result of the study is that we found no overall effect in the study, or it was a nominal 2 percent benefit. There's a qualitative interaction that if you get alteplase there's no benefit at all, and maybe even negative. But if you don't get alteplase, there's a 9.5 percent absolute effect size difference, which's an NNT of 11, essentially to get one more patient with a rank and 02. Quite large, and it was bolstered by a mortality reduction by an infarct volume reduction on day and then the pharmacokinetics, we showed this terrific pharmacology in a subset, mind you. Only a small group of patients that the ones who got alteplase had no detectable or substantially reduced detectable nerinetide in their blood. Whereas those who didn't have the expected rise of nerinetide in their blood. We think that one of the things that clearly happened here was that if you got alteplase as just by routine care, then it chewed up the nerinetide. Recently, got the pharmacology papers published. It's in "Science Translational Medicine," just this month.

What's been shown interesting as alteplase and nerinetide are inert if you mix them together, but as soon as you add plasma. Then nerinetide gets chewed up rapidly in 10 minutes there's no detectable nerinetide left. The cleavage sites appear to be all at the carrier molecule end. The active moiety is preserved, but the carrier molecule is critical for getting it out of the blood into the brain. If you cleave the carrier molecule, that's useless essentially, and this is true for both alteplase and tenecteplase.

Then finally, the final bit of interesting pharmacology, and this is what is coming out, just to give you a tidbit, is that if you change the conformational structure of the amino acids in nerinetide, you can produce a version which is thrombolysis resistant. NoNO Inc., Mike Tymianski's group, now has in development, and it's going into the first trials now, a purely thrombolysis‑resistance version of nerinetide. If we can prove that this is useful, if nerinetide works in the current trial called ESCAPE‑NEXT, the trial after that will be to look at this again in the thrombolysis population but with this new compound. The future looks really exciting.

Dr. Sarraj:  That is really exciting. We spoke about several attempts with unsuccessful results in toxicity, and now we're talking about almost 10 percent improvement in the outcome, an adjunctive benefit with nerinetide. Actually, that was going to be my follow‑up question about. If ESCAPE‑NEXT show benefit in patients who are not eligible for IV tPA, then we will have to deal with that. The issue is, the patient, first of all, should not be eligible for IV tPA, and then what do you do next with the potential that IV tPA will go to 24 hours or tenecteplase or a thrombolytic agent? It's great news that you shared that there will be a version of the agent that would be resistant to thrombolysis.

Dr. Hill:  I hope so. I think it is potentially exciting. Honestly, it's still possible that what we saw in ESCAPE‑NA‑1 is the effect size is larger/smaller than what we've seen, but interestingly, if we get more reperfusion and we get the drug in faster, in ESCAPE‑NA‑1, it looks like the effect size is even larger in that as you cut the population. We'll see. We'll see how it goes, but if we can show it and then you get a compound which could be given in any situation, even better, and we're also experimenting. There's another study ongoing in Vancouver, in Canada in which we're giving nerinetide in the ambulance.

Conceptually, if the problem is we've got to freeze the stroke or protect the brain until such time as there's an interventional team can get the artery open, then we're hopeful that we may be able to improve outcomes by giving it even earlier.

Dr. Sarraj:  Right. That model would be good especially with recent mobile stroke units and all of that.

Dr. Hill:  That's right. Imagine having it stocked on a mobile stroke unit where you can then treat, a dual treat, and then come to cath lab. Absolutely.

Dr. Sarraj:  Just need multiple IVs, but...

Dr. Hill:  [laughs] That's right.

Dr. Sarraj:  [Laughs] Less of a problem. Then we have to talk here about the, especially with ESCAPE‑NA‑1 and ESCAPE‑NEXT. Thrombectomy is a very powerful treatment with a huge treatment effect in LDL, and the group of patients is mostly with good imaging profiles in ESCAPE‑NA‑1, ESCAPE‑NEXT. Do you see that it's possible that if we want to have adjunctive benefit, we'll have better chances or even more other sub‑populations, say those with, it's just an interest of mine, large core strokes? Those who are fast progression that you can slow down, robust the collaterals, have less probably benefit of thrombectomy or any treatment, and you bring a neuroprotection agent there. What are your thoughts?

Dr. Hill:  You addressed it by bringing out the idea of the mobile stroke unit. There are a couple of avenues to think about. One is that Mark Fisher in Massachusetts did a neat study with nerinetide where they looked at, it was a rodent model, and what they showed is at the point they gave NA‑1, they did serial MR imaging and they had a control versus NA‑1, it seemed like the NA‑1 would freeze the penumbra based on MR volumes. The volume of fusion image would stop growing. This is this idea that the faster you can get the treatment in and you halt the stroke, then you could proceed with reperfusion at some later time point as you're getting people in, whether it's with a thrombolytic, or with endovascular thrombectomy or both. That's going to be important.

One issue is the time addressing that you're treating early. A second issue is to think about, as you say, that the so‑called large core and that kind of stuff, those patients. We're pretty good at estimating it, but we're not good at figuring out around the edges, what's eloquent tissue and what's not. That's probably where treating large cores are useful. The depth of ischemia in a given stroke is going to vary across a whole volume of tissue. In some circumstances, you're going to be able to save the tissue that's still potentially viable. Obviously, tissue that's hyperdense on CT, or has very low ADC on MR, it's done and you can't bring it back to life. You can't reanimate that tissue. You're going to tie, it's not going to reanimate that tissue any more than reperfusion is, but it's around the edges where you can potentially help. Clinically, I suspect the issue is that it'll be this little judgment.

Eric Sauvageau always tells me this from Florida, he's a surgeon, he says, "Look, if I get a whole bunch of people with subdural hemorrhages, I just got to take out the bleeding because I don't know who's going to get better, and I'm going to relieve the pressure, and some of them are just not going to do well, and I understand that, but it's better for me to just treat every subdural."

We're going to have to get there with ischemic stroke, we're going to have to be more aggressive, with the patients who already have a bit of damage. See whether we gradually, by saving some of the tissue around the edges, on average, we get better outcomes amongst our patients, get more patients home or prevent more disability. I suspect that'll happen.

Interestingly, if you look at ESCAPE‑NA‑1, the population of patients with superb imaging, ASPECTS of 10, a blocked artery, excellent robust collaterals. They would have no CBV drop. If you looked on RAPID, they'd be all green for go. Those patients don't have a big effect size if they reperfuse. In other words, as you point out, all you need to do is get the blood flow back to those patients. The challenge, of course, in terms of thinking about how the flow of treatments might go is that at the time you want to give the person an adjunctive agent like nerinetide or, I'm sure we'll have other ones potentially, or other compounds will be approved. You don't know how it's going to play out, so you're going to end up treating everybody.

Dr. Sarraj:  Which might be the ultimate answer, certifying the benefit and likelihood of benefit. It's a good problem to have.

Dr. Hill:  It's a good problem to have.

Dr. Sarraj:  When we were talking a few years ago about no benefit of thrombectomy and reperfusion a large vessel. It's amazing that we're talking about that now. The science is wonderful and interesting, getting clinical trials and the role of evolution till we find the agent or the subgroup. How do the generalizability and the clinical applicability of those findings to the population? Let's say moving forward, we hopefully, find that red tide is effective, and there is an agent that works with TPA. What is the model? Did you see to work in the clinical trial because you want it 30 minutes before then should we give it in the remote hospital and transfer the patients for thrombectomy?

Especially, now we have trials that testing, should we give TPA to start with prior to thrombectomy, prevention, and the other ones?

Dr. Hill:  Great questions. It's speculative. It's fair to say that, for ESCAPE‑NA‑1 trial, the target audience is researchers. We're still at that research stage. We don't have an approved drug, and I don't want to count the chickens before they hatch. I don't know how the next trial is going to go. You're always surprised in clinical research. In all research, you discover new things. It never follows an entirely predictable path. I'm sure that one certainty is once we finished the current trial ESCAPE next in a year and a half, or so, or a couple of years, depending on how it goes, we'll have more questions. I hope that we'll have enough evidence to say, "Yes, we can use this compound in this situation." We're going to have a whole bunch more questions and try to figure out how to use it. It would be great...Not only is this compound under development, there are others being looked at. There's a variation of the activated protein C, protein 3K, 3C, ABC being developed by a company called ZZ Pharma.

Pat Lyden, from Los Angeles, is the key person driving that development from the medical side. There are others, hyperthermia needs to be looked at. I guess that my hope more than anything, is that if we can show that there's a specific effect here, then all of a sudden, we open up the possibility of adjuvant therapy. How it will play out in the long run, rather than the door being slam shut as it is, now. We open it up a bit, it's a jar, and now people can try to push through and develop new things and we can make things better. The other thing that will drive this is ease and applicability and safety of use. A safe drug, this one appears to be pretty safe, but a safe drug can be given in the periphery, can be given in a primary hospital, and be given in an ambulance. That's possible. Difficult things, like hypothermia, you can't do that in an ambulance. You need an ICU setting. All of these things are going to come into play as we look at these compounds.

Dr. Sarraj:  Hopefully, we'll learn more from the trial. When you're doing research, you don't just assess the scientific question, but you also assess what works in terms of delivering the drug and all of that, and which hospitals can do this model and cannot. Hopefully, we'll learn all of that from the trial.

We talked a lot about your experience and this, and you have seen it through different stages and learned a lot. As you mentioned there are people who are doing this research so we can't leave this without asking you, what do you see for future development? Another more important question for young researchers and clinical trialists, who are trying to delve into neuroprotection, what would be your advice and suggestion from this for a long time?

Dr. Hill:  Probably the most important thing I would say that's been relevant in clinical research is quality, paying a lot of attention to the quality of the people you're working with, the sights that they have high‑quality care. There's nothing like a whole bunch of adverse events occurring in a trial, which then obliterates any possibility of seeing an effect of what you're doing, or what you're trying to study. The quality of data matters a lot. There's often I find that people seem to be immune to the idea that there's misclassification bias and lots of data. Just because the data is in a spreadsheet doesn't mean that it's good data. You got to dig in and understand the source of your data, the quality of it, and make sure that you're happy that it's consistent, reproducible, all that kind of stuff.

That attention to quality, all‑around quality of your trial process, your data, but also the quality of care. Clinical research is all about building on high‑quality care systems. If you don't have high‑quality care systems, you're not going to be able to demonstrate an additional effect size over and above, because you lose all the effect and the error that occurs when care quality is poor. That would be my biggest lesson. How do you assure that? Well, fundamentally, medicine at all of its levels is a human‑human interaction process. You've got to get to know the people you're working with. You got to go see them and talk to them. See how their systems work, establish quality feedback loops, and try to make sure things are working well in your trial. Otherwise, you won't get the result you want. People won't be engaged and won't make it. That's the biggest lesson.

As we go forward, there will be challenges of this. As you know, we all are getting all of our medical systems into the cloud, right? It's all good. It's great data or not necessarily great data. There's volumes of data. It's not always good data, and it's not always...You have to figure out where you want to tackle. That's probably the best part of it. The number one thing I've always been worried about is quality of data and quality of care, and I continue to worry about it. I continue to worry that I'm going to get burned in the next trial because we haven't paid attention enough to quality. Well, we'll have to see it.

Dr. Sarraj:  Hopefully, you won't.

Dr. Hill:  I know. We'll work hard not to.

Dr. Sarraj:  The infrastructure is great. Back to neuroprotection agents. What are some of the ongoing trials, NIH? There is the NIH SPAN project. Looking into the future of neuroprotection, what could happen?

Dr. Hill:  There are people working on this on different things. I don't know all of the member and I don't claim to. The SPAN project that NIH is run from Pat Lyden's group, and they're trying to develop...It's a really interesting model. They're trying to apply the lessons from multi‑center clinical trials down to the bench level, and looking at foundational science, applying rigor that's learned from multi‑center clinical trials to that level, and that's cool. It may well be important in trying to show that there's consistency of effect across multiple labs, and then hopefully, take those compounds or those ideas forward to the next level.

The one I mentioned before, which is a modified variant of activated protein C, which is quite a large molecule, but also interestingly a biological. In the same way that nerinetide is a peptide, a biological, although parenthetically classified as a drug because it's a small peptide. Activated protein C is a much larger molecule and it's also a biological. It follows along the...all the stuff we've seen with cancer immunotherapies and all the stuff for Crohn's disease and rheumatoid arthritis and psoriasis. All of these biological agents have come through from the advances in microbiology and they've had dramatic effects on treatment. I bet you that is a key step for us to take in stroke. There's a lot of immunology that goes on in inflammation, and that won't apply just to ischemic stroke, that'll apply to hemorrhage as well.

You can foresee the things that happen after the ischemic insult or the hemorrhagic insult can be potentially addressed by targeted mechanisms. I think we'll start to see more of that science being translated in the vascular medicine. That'll be exciting. That'd be really exciting for us.

Dr. Sarraj:  Wonderful. Definitely, we moved from chances of no benefit really to powerful reperfusion benefit that we're trying to expand the subgroups and trying to start even before with neuroprotection. Hopefully, this translates to positive results then all the way afterwards for rehabilitation and all of that.

That was wonderful. Thank you for sharing the great experience, the journey of long time doing successful research. Hopefully this comes to a great success with ESCAPE‑NEXT for you and your colleagues.

For the field, I'm sure that the vascular neurologist, the neurologist will look forward to these results and will enjoy this conversation we had with you.

Dr. Hill:  Thanks, Amrou. Nice to talk to you guys.

Dr. Sarraj:  Same here. Hopefully next time we talk is about ESCAPE‑NEXT's result. Thank you so much. Take care. Bye.