Stroke Alert May 2022

Stroke Alert

On Episode 16 of the Stroke Alert Podcast, Dr. Negar Asdaghi highlights two articles from the May issue of Stroke: “Number of Affected Relatives, Age, Smoking, and Hypertension Prediction Score for Intracranial Aneurysms in Persons With a Family History for Subarachnoid Hemorrhage” and “Endovascular Treatment for Acute Ischemic Stroke With or Without General Anesthesia.” She also interviews Dr. Patrick Lyden on “The Stroke Preclinical Assessment Network: Rationale, Design, Feasibility, and Stage 1 Results.”

Dr. Negar Asdaghi: Let's start with some questions.

1) How is it that stroke can be cured in rodents but not in humans?

2) Are we wasting time or gaining time with general anesthesia before endovascular thrombectomy?

3) My father had an aneurysmal subarachnoid hemorrhage, Doctor. What is my risk of having an aneurysm, and how often should we check for one?

We're back here with the Stroke Alert Podcast to tackle the toughest questions in the field because this is the best in Stroke. Stay with us.

Dr. Negar Asdaghi: Welcome back to the May 2022 issue of the Stroke Alert Podcast. My name is Negar Asdaghi. I'm an Associate Professor of Neurology at the University of Miami Miller School of Medicine and your host for the monthly Stroke Alert Podcast. For the May 2022 issue of Stroke, we have a number of papers that I'd like to highlight. We have seven articles as part of our Focused Update on the topic of neuroimmunology and stroke, organized by our own Stroke editors, Drs. Johannes Boltze and Miguel Perez-Pinzon. We also have an interesting study by Dr. David Saadoun and colleagues from Sorbonne University in Paris, where we learn that in patients with Takayasu disease, how the delay in diagnosis, as defined by the time from symptom onset to the diagnosis being over one year, was significantly associated with development of ischemic cerebrovascular events. In the Comments and Opinions section, we have an interesting study by Dr. Goldenberg and colleagues from University of Toronto on the benefits of GLP-1 receptor agonists for stroke reduction in type 2 diabetes and why should stroke neurologists be familiar with this new class of diabetic medication.

Dr. Negar Asdaghi: Later, in the interview section of the podcast, I have the great honor of interviewing Dr. Patrick Lyden, one of the founding fathers of thrombolytic therapy in stroke, as he walks us through the Stroke Preclinical Assessment Network and what his hopes are for the future of stroke therapy. I also ask him for some advice, and he did tell us about the view from the top, as he truly stands on the shoulder of giants. But first with these two articles.

Dr. Negar Asdaghi: In a landmark population-based study out of Sweden that was published in Brain in 2008, we learned that the odds of development of aneurysmal subarachnoid hemorrhage for individuals with one first-degree relative with a prior history of aneurysmal subarachnoid hemorrhage was 2.15. For individuals with two affected first-degree relatives, the odds ratio was 51. So, it's not surprising that a great deal of anxiety is caused within a family when a relative has an aneurysmal subarachnoid hemorrhage, especially if that family member was young or another member of the family had the same condition before. This scenario is commonly followed by a number of inevitable questions: Should all family members of the affected individual be screened for presence of an intracranial aneurysm? If yes, how often should vascular imaging be performed, and should other aneurysmal risk factors, such as age, sex, smoking, and hypertension, be also considered in the screening decision-making? In this issue of the journal, as part of a derivation-validation study, a group of investigators, led by Dr. Charlotte Zuurbier from University Medical Center at Utrecht Brain Center in the Netherlands, studied the ability of a simple scoring system that was developed in their derivation cohort to predict the presence of an intracranial aneurysm on vascular imaging.

Dr. Negar Asdaghi: They then tested the scoring model in their validation cohort. So, for their development cohort, they used data on 660 persons who were screened at the University Medical Center for presence of an intracranial aneurysm because they had two or more affected first-degree relatives with a prior history of aneurysmal subarachnoid hemorrhage. The median age of participants at the time of first screening was 40, and 59% were female.

Dr. Negar Asdaghi: So, in this cohort, the investigators simply looked at factors that were independently associated with finding an aneurysm on vascular screening by their multivariate analysis. And they identified the following factors; the first factor was the number of affected relatives. Now, a reminder that all of these people in the cohort had at least two first-degree relatives with an aneurysmal subarachnoid hemorrhage. And they found that amongst these people, those that had three or more family members with aneurysmal subarachnoid hemorrhage were significantly more likely to have a positive screening test for intracranial aneurysm. The next factor was older age — the older that relative, the more likely their screening imaging was positive for an aneurysm — and the other independent factors were smoking and hypertension. So they created the NASH acronym; N for number of relatives, A for age, S for smoking, and H for hypertension. When assigning points for each of these factors, the NASH scoring system had a C statistics of 0.68 in predicting whether or not someone would have a positive test, which is an intracranial aneurysm.

Dr. Negar Asdaghi: And now a reminder for our listeners that C statistics gives us the probability that a person with a certain condition, in this case, a certain NASH score, will have the outcome of interest, in this case, an aneurysm found by vascular imaging. In general, for C statistics, the closer we get to 1, the more robust is our predictive model. Values over 0.7 indicate that we have a good model, but values over 0.8 indicate a very strong model. So the NASH score, at 0.68, has a reasonably good capability in predicting who will or will not have an intracranial aneurysm if we complete the vascular imaging. But it's not a very strong model, and this should be kept in mind. Let's look at some of their numbers. In their development cohort, the probability of finding an intracranial aneurysm for a person who scored low on NASH, that is a young person who never smoked and is not hypertensive, was only 5%, whereas the probability of finding an intracranial aneurysm in a person who scored high on NASH, that is an older person in their 60s or 70s, with three or more affected relatives, who is hypertensive and a smoker, was 36%.

Dr. Negar Asdaghi: So, then they tested this NASH score in their external validation cohort and found that the likelihood of identifying an aneurysm increased as expected along the range of predicted probabilities of NASH. That is, the higher the score, the more likely to find an aneurysm on screening with vascular imaging. And the C statistics in the validation cohort was slightly lower than the C statistics in the derivation cohort. So, the important lesson we learned from this study is that the risk of having an intracranial aneurysm in a person who has a first-degree family member with a prior history of aneurysmal subarachnoid hemorrhage is substantially different depending on their NASH score, and this should be taken into consideration when deciding on screening and counseling various family members of the affected patient or prioritizing who should be screened first in routine practice.

Dr. Negar Asdaghi: The ideal anesthetic management during endovascular therapy is still unknown. A number of studies have compared the different anesthetic options available during thrombectomy, which include general anesthesia, or GA, conscious sedation, use of local anesthesia, and no sedation at all. The main argument for doing endovascular therapy under general anesthesia is that although this procedure will take some precious pre-thrombectomy time, it does result in strict immobility. And that is really ideal in the sense that it improves catheter navigation and interpretation of angiography, in addition to obviously providing a secure airway and, of course, avoiding the need to have to do an emergency intubation in case of procedural complications. The argument against general anesthesia is not only the issue of time but also the risk of hypotension and hemodynamic compromise, especially during induction, and the loss of very valuable neurological examination in a completely sedated patient during the procedure.

Dr. Negar Asdaghi: The question is, does general anesthesia improve or worsen neurological and functional outcomes post-thrombectomy? Several smaller randomized trials have looked at this very question, mainly comparing GA to all other forms of sedation during thrombectomy, but they have yielded inconsistent findings regarding the three-month functional outcome.

Dr. Negar Asdaghi: Some of them showed that patients under GA ended up doing better. Some showed no difference in the overall outcomes. But overall, their pooled analysis suggested that GA might be superior to the competing counterpart, which is the conscious sedation, and associated with better functional outcome. But these centers had highly specialized anesthesia teams, and it's possible that their findings may not be generalizable to routine practice. So, in this issue of the journal, using the Swiss Stroke Registry, Dr. Benjamin Wagner from the Department of Neurology at the University Hospital in Basel and colleagues report on the outcomes of endovascularly treated patients in the Swiss Stroke Registry receiving thrombectomy for an anterior circulation stroke with or without general anesthesia. The primary outcome was disability on the modified Rankin Scale after three months. For this study, they excluded one out of the nine centers in the registry that had lots of missing data on their three-month follow-up.

Dr. Negar Asdaghi: And so, from 2014 to 2017, 1,284 patients across eight stroke centers in the registry were included in this study. Sixty-six percent received thrombectomy under general anesthesia. On baseline comparison, the patients in the GA group were older, had a higher NIH Stroke Scale on admission, had worse preclinical functional status, and more likely to have presented with multi-territorial ischemic stroke. So, many reasons as to why people who underwent general anesthesia would have a worse clinical outcome in this study. So, now let's look at their primary outcome. In the unadjusted model, the three-month modified Rankin Scale was significantly worse in the GA group as compared to the non-GA group, which is obviously expected given the differences in their baseline characteristics.

Dr. Negar Asdaghi: But what was surprising was that the odds of having a higher mRS score was significantly greater still in the adjusted models. They also did propensity score matching analysis, and they found that the NIH Stroke Scale after 24 hours, and the odds of dependency and death and mortality were all higher in the adjusted model in the GA group. They also looked at a number of secondary outcomes and found that door-to-puncture time was longer in the GA group.

Dr. Negar Asdaghi: And also these patients were more likely to be transferred to ICU after treatment as compared to the non-GA treated counterparts. The authors point out that these real-world data are in keeping with the findings from the HERMES meta-analysis, which included over 1,700 endovascularly treated patients, and two previously published large registry data, one from Italy, which included over 4,000 endovascularly treated patients, and one from Germany, including 5,808 patients, all of them showing a worse functional outcome in endovascular therapy if the treatment was performed under general anesthesia, as compared to all other forms of sedation or no sedation at all. Again, these findings are in contrast with the reassuring results of the randomized trials on this topic, specifically in contrast to the AnStroke, SIESTA, and GOLIATH randomized trials, which compare GA to conscious sedation, showing either neutral or positive results in favor of general anesthesia pre-thrombectomy.

Dr. Negar Asdaghi: So, in summary, what we learned from this real-world, observational study is that general anesthesia was associated with worse functional outcome post-endovascular thrombectomy, independent of other confounders, which means that the jury is still out on the ideal form of anesthesia for an individual patient prior to endovascular therapy, and we definitely need larger, multicenter studies on this topic.

Dr. Negar Asdaghi: There are over a thousand experimental treatments that have shown benefit in prevention of neurological disability in animal models of ischemic stroke but have failed to show the same efficacy in human randomized trials. In fact, to date, reperfusion therapies, either in the form of intravenous lytic therapies or endovascular treatments, are the only successful treatments available to improve clinical outcomes in patients who suffer from ischemic stroke, and stroke remains a leading cause of death and disability worldwide. How come stroke can be cured in rodents but not in humans? Are neuroprotective therapies, or as more correctly referred to, the cerebroprotective therapies, the epitome of bench-to-bedside translational research failure? And if this is true, what are the key contributors to the scientific conundrum, and how can this be averted in the future? This is the question that a remarkable group of neuroscientists, led by Dr. Patrick Lyden from University of Southern California, are hoping to answer.

Dr. Negar Asdaghi: In this issue of the journal, these investigators describe the rationale, design, feasibility, and stage 1 results of their multicenter SPAN collaboration, which stands for the Stroke Preclinical Assessment Network. I'm joined today by Professor Lyden himself to discuss this collaboration. Now, Professor Lyden absolutely needs no introduction to our stroke community, but as always, introductions are nice. So, here we go. Dr. Lyden is a Professor of Physiology, Neuroscience, and Neurology at Zilkha Neurogenetic Institute, Keck School of Medicine, at USC. He has truly been a leader in the field of preclinical and clinical vascular research with over 30 years of experience in conducting studies and randomized trials, including conducting the pivotal NINDS clinical trial that led to the approval of the first treatment for acute ischemic stroke in 1996. Throughout his exemplary career, he has accumulated many accolades and is the recipient of multiple awards and honors, including the prestigious 2019 American Stroke Association William Feinberg Award for Excellence in Clinical Stroke. Good morning, Pat, it's truly an honor to welcome you to our podcast today.

Dr. Patrick Lyden: Thanks, I'm glad to be here.

Dr. Negar Asdaghi: Well, in the era of successful reperfusion therapies, it seems that the new generation of stroke neurologists and interventionalists have their eyes, so to speak, on the clock and are interested in opening the blood vessels and opening them fast. In the age of reperfusion treatments, why do we still need to talk about the role of cerebroprotective treatments?

Dr. Patrick Lyden: Well, not to sound too glib about it, but not everybody gets better after a thrombectomy. So, thrombectomy is good, it's more effective than anything else that we've tried before, but there are a remaining number of patients with a residual disability. Not only that, and from a more scientific standpoint, thrombectomy offers us the opportunity now to combine cerebroprotective therapy with known reperfusion. Remember, before, we didn't know when the artery had opened, but now we do an embolectomy, we know there's reperfusion. It gives us the opportunity to know that we're combining our treatment with reperfusion.

Dr. Negar Asdaghi: So, in the paper, you discussed how hundreds of treatments have been studied and shown efficacy in reducing neurological disability in animal models of stroke, and yet failed in human studies. In your opinion, what were the top two most disappointing studies in terms of clinical failure despite pre-clinical encouraging data?

Dr. Patrick Lyden: Well, the first one I mentioned was personal because it was the first one that I led, and it was a molecule called clomethiazole that I had helped establish the rationale for in my very first grant. So, it was the first trial I led, it was multinational, and, of course, I firmly believed we were going to hit a home run, and we failed. But to the field, the real watershed moment in neuroprotective therapy was the so-called SAINT II Trial. SAINT II was a study of a drug called NXY-059, and it was the first drug that purportedly had satisfied all of the so-called STAIR criteria. The STAIR criteria came out of a roundtable between academics and industry on how to best qualify drugs preclinically before going to human trials. And the idea was, if you were a 10 out of 10 on the STAIR criteria, then you should win when you come to human clinical trials. And the SAINT II Trial, which I was a co-leader, a co-investigator, on, also failed.

Dr. Patrick Lyden: And so many, many, many drugs had failed by that point. Tens of millions, if not a hundred million dollars, had been spent by industry, and SAINT II really caused the field to stop. Industry stopped investing in stroke; academic investment in stroke dried up. NIH funding became more difficult to get after SAINT II, and that really was sort of the really historical low moment in the development of treatment for stroke.

Dr. Negar Asdaghi: I was a resident when SAINT II came out, and I remember that somber feeling.

Dr. Patrick Lyden: It was a sad day.

Dr. Negar Asdaghi: Yeah. So, in the paper, you outline a number of potential causes as to why this translational failure may have occurred. But you highlighted the absence of preclinical scientific rigor as the most responsible source. And you already alluded to this a little bit. Can you please tell us a bit more?

Dr. Patrick Lyden: Absolutely. And first, of course, we have to say that the ideal clinical trial design is not available. We really don't know the absolute best way to test the drugs in human clinical trials. But leave that for another day.

Dr. Patrick Lyden: On the preclinical side, what can we say we're doing wrong? We're not sure, but one thing that has been highlighted over and over is that we don't approach preclinical characterization with as much rigor as we should. What do I mean by that? Animal models recapitulate for us some of the biology of a stroke, but not all. For example, many, many times we test a drug in a young model, an animal that's quite young, corresponding to a late teenager in human terms. Well, that's ridiculous. Stroke occurs in elderly people, and so on. So, the NIH called in a landmark conference for additional rigor, enhanced rigor. And I should mention the STAIR criteria were a first attempt at this. STAIR put out guidelines that said animals should be elderly, the animals should be randomized, et cetera, et cetera. And so that didn't happen. Although the STAIR criteria were out there, very few laboratories really committed to full rigor. And so the NIH funded the Stroke Preclinical Assessment Network, SPAN, to implement every aspect that we could think of that would add the best possible scientific design, the utmost rigor. So, we implemented true blinded assessment, true randomization, complete case ascertainment where we follow every single subject in the study and account for dropouts and subjects that don't complete the treatment, and, most importantly, a proper statistical design with adequate power and very large numbers. And the hypothesis that we're testing is that additional rigor in SPAN will lead to a better positive predictive value when we think about drugs that should go forward for testing in human stroke trials.

Dr. Negar Asdaghi: So, I think you already answered my next question, which was basically, why do you think SPAN is going to achieve what all others have failed to achieve? But I wanted to simplify and repeat what you mentioned. So, in simple terms, what SPAN is trying to do is to bring all preclinical research to a level of scientific rigor that was not necessarily present and make it a multicenter effort. And can you a little bit tell us about the different stages, again, of SPAN?

Dr. Patrick Lyden: Well, I'm not arguing that all preclinical research needs to be done following a SPAN type of model. Where SPAN fits in is at the end of a development project. So, if you want to characterize the cellular and molecular mechanisms, you don't need to do all of this rigor that we're doing. Just study the drug in the lab and do the mechanistic studies that need to be done. If you want to do dose finding, it doesn't need to be done this way. But at the end of that, OK, first we establish the mechanism, that's the first stage. Then we establish the toxicity. Then we establish target engagement. At the end, we are looking for some evidence that the drug will have a beneficial effect on outcomes. And in previous animal models, the only outcome, generally, the most common outcome that was studied, was size of the stroke. But in humans, the FDA does not recognize stroke size as a valid outcome.

Dr. Patrick Lyden: We look at function, most often measured with the Rankin score and the NIH Stroke Scale. So, we had to create a functional outcome, and then we had to study it at multiple laboratories to make sure we could replicate the effect across multiple sites. And we chose what's called a multi-arm, multi-stage (MAMS) statistical design. All the drugs start out in the experiment at the end of the first interim analysis, which is 25% of the sample size. We cull any compounds or treatments that appear futile are removed. Any that appear effective move on. At the end of the second stage, there's more culling. There's a total of four stages, and we're about to enter stage four, by the way. That's starting next week. And in stage four, there will be, at most, two, maybe only one treatment that has appeared non-futile and possibly effective for final characterization.

Dr. Negar Asdaghi: So, really interesting. I just want to highlight two important comments that you mentioned for our listeners again. So this is multi-layer, as you mentioned, multi-arm, multi-stages. It's sort of filter by filter, just ensuring that what we're seeing, the efficacy we're seeing in preclinical studies, will potentially be replicated in clinical studies. And what you mentioned that's very important is outcomes that classically is measured in animal models are infarct volume that are obviously very important but not necessarily may translate to exactly what we look at in clinical studies, which is functional outcomes, modified Rankin score and NIH Stroke Scale. So, with that, I want to then come back to the treatments that are actually being studied as part of SPAN. You have six very different agents as part of SPAN, from tocilizumab to uric acid. Why do you think these therapies will work?

Dr. Patrick Lyden: Well, my job as the PI of the coordinating center is to remain completely agnostic to the treatments. So, everybody's equal, and they all come in on an equal playing field. We actually have a mechanical treatment called remote ischemic conditioning, as well, and then five drugs. And these were selected through a peer review process at NIH. And then we were informed at the coordinating center what drugs we would be studying. Five drugs and one treatment. And then, of course, the challenge to us was to somehow create a blinded, randomized situation. Now, this turned out to be a fascinating, it's more mechanical, but how do you blind when some of the drugs are given orally, some are given intraperitoneally, some are given intravenously, some are given once, some are given multiple times? So, we had to work with the manufacturers and inventors of these drugs and figure out a way to package them, and in the paper, actually, there's a photograph in the appendix that shows we had to find these bottles that were amber-colored and how to load them and lyophilize the drug.

Dr. Patrick Lyden: And it's actually pretty fascinating how we were able to get all of these different, wildly different therapies, as you say, into a paradigm where they could be tested one against another in a truly blinded, truly randomized way.

Dr. Negar Asdaghi: Do you think you can go on record and say which one is your favorite?

Dr. Patrick Lyden: My favorite drug's not even in SPAN. I am truly agnostic because where my heart is, is with a drug that I've been studying in my laboratory completely separately and not part of SPAN.

Dr. Negar Asdaghi: All right, so we don't have a favorite. So, in a recent review article in Stroke, you commented on treatments used by ancient Persians, Greeks, and Romans to remedy the brain affected by stroke and how the future generation of physicians will look back at our current practices of stroke with the same, how you said, awe and bemusement we hold for Galen, Aristotle, and Avicenna. How do you think stroke will be treated in the year 2222?

Dr. Patrick Lyden: Well, first of all, and to be serious for just one moment, 200 years from now, I worry more about the climate than about medicine. And I really believe our biggest efforts need to be spent on saving the planet. But assuming we make it that long, obviously diagnostic methods will be completely different. Using ionizing radiation to scan the body will be laughed at by physicians in the future. There'll be detection technologies that aren't even on our radar yet today. And then treatments will be cellular focused and regionally focused. We give a drug through a vein and it circulates throughout the entire body, and I'm sure physicians in the future will find a way to somehow get treatment into the part of the body that's injured, not the whole body. And then, who knows? All we can say is they will laugh at us in the same way that we laugh at Theodoric the Barber of York.

Dr. Negar Asdaghi: Let's move on from the future to the past. You're arguably one of the founding fathers of reperfusion therapies. You were instrumental in getting intravenous lytic therapy approved in 1996. It literally took the field 20 years for the next treatment to be approved, that's endovascular treatment. If you could go back in time and give your young self an advice on the subject of research, of course, design and execution, what advice would you give yourself?

Dr. Patrick Lyden: Don't listen to old guys. We got a lot of advice from gray-bearded folks back when we were putting together the tPA trial, and fortunately we ignored some very bad advice and did what we imagined was the right thing to do as young, headstrong up-and-comers do. The other thing is, we really believed that by publishing our science very objectively, without editorial comment, we would be listened to. And that was dead wrong. So, the data was printed in the New England Journal in a very neutral tone, and we felt people would read that data and they would start using tPA the day after the publication. And, as you say, it took 20 years for tPA to really gain widespread acceptance, thrombolytic therapy. Today, people view it as standard, but it wasn't that way at the beginning. And I would say to myself and my colleagues at that time, "Don't be afraid to promote a positive result." Yes, it has to be done with the utmost rigor, but once you have a positive result, there will be plenty of people around pretending they know more than you and telling the world why you are wrong. And it's very important to stand up for your science and stand up for your results and say, no, no, no, no, that interpretation is wrong. The data says what we said it says, and this is an effective treatment and should be used, as an example.

Dr. Negar Asdaghi: What a great advice. Just be bold and say it loud and stand up for your science. Pat, it's been a pleasure interviewing you and having you on the podcast. We really look forward to watching your research. Bring, let me say it again, 2222 closer to now.

Dr. Patrick Lyden: Thank you. Glad to be here.

Dr. Negar Asdaghi: Thank you.

Dr. Negar Asdaghi: And this concludes our podcast for the May 2022 issue of Stroke. Please be sure to check out this month's table of contents for the full list of publications, including two articles on quality improvement in stroke and neurohospitalist—inpatient teleneurology, which comes as part of our Advances in Stroke series prepared by our section editors. And as we close our podcast today, let's take a moment and ask ourselves the same question that I asked Dr. Lyden earlier. What is the next frontier in stroke treatment? Past reperfusion therapies, we have to find ways to preserve the neurons and not just the neurons, all components of the brain. So, is the future of stroke therapy cerebroprotection? Ever since the dawn of history, humanity has lived alongside of death with the conscious apprehension that as we age, we lose the very gift of life. But unlike our ancestors, the search for immortality isn't the quest to find a fountain of youth anymore. We learned that death is inevitable, but with medicine, we can reduce illness and suffering to prolong a life worth living, one with a healthy brain. And today we're closer than ever to this modern immortality with cerebroprotection in stroke, as we stay alert with Stroke Alert.

Dr. Negar Asdaghi: This program is copyright of the American Heart Association, 2022. The opinions expressed by speakers in this podcast are their own and not necessarily those of the editors or of the American Heart Association. For more, visit AHAjournals.org.