Well, good afternoon, everyone. Thank you for making the trek down to the room at the end of the known universe here down in R8 and 9. My name is Mike O'Dell, I'm an emeritus professor at Weill Cornell Medicine in New York City. Welcome to Stroke Journal Club on steroids. I'm actually, will eventually be joined by three really awesome colleagues in the area of clinical and research in stroke rehabilitation. I'll just go ahead and introduce them now, and then we'll just have them come to stage as appropriate. Joel Stein is a professor and chair of the Department of Rehabilitation and Regenerative Medicine at the Columbia Vagelos College of Physicians and Surgeons, as well as the chair of Rehabilitation Medicine at Weill Cornell in New York City. Gerard Francisco is the chair and professor of Physical Medicine and Rehabilitation at the McGovern School of Medicine at UT Houston, and also the chief medical officer of the Rehabilitation Institute of, Rehabilitation, Texas Institute of Rehabilitation and Research in Houston. And Nannika Efechika, thank you very much, is professor of both Rehabilitation Medicine and Neurology at UT Southwestern, and is the director of the Stroke Service there as well. What we're going to be doing today is, and by the way, these are my disclosures. This is a format, Journal Club, that's been used at AAPMNR before. As we all know, this is a commonly used tool and a very fun tool that we've used in our academic practices with residents and fellows and medical students. It's a great mechanism not only for imparting knowledge, but also for imparting critical thinking as well. So in this variation, rather than using residents, we're going to use very seasoned senior clinicians and researchers with the hope of using folks, I don't mean any offense to any residents that are in the audience, having people that actually know what they're talking about in terms of presenting the methodology and the articles. So without further ado, we'll be presenting three articles this morning. Dr. Stein will do the first article, and then followed by Dr. Francisco. Just real briefly, the selection of these articles, I, during the course of the last couple months, looked through the literature in terms of stroke and then various key words. I wanted to choose only articles that were published in 2003, and I actually reviewed several hundred abstracts in coming down with these three articles. Ultimately, quite frankly, it's good to be king. The ultimate choices were really pretty subjective on my part, but I tried to look for articles where there was a physiatrist at least as co-author, in some cases I was successful and others not. I initially looked for authors that could potentially attend and present in person, and I tried to look for a diversity of topics. And I wanted interventional as well as non-interventional articles. So I came up with these three. Our format, real quickly, title, the background, aims and hypotheses, the methodology, the results, discussion per the author, the critique per the presenter. We'll have 12 minutes for the presentation, 5 minutes for questions and answers, and then at the end we'll have a general discussion with all the speakers in terms of general discussion on stroke rehabilitation. So Dr. Stein. Thank you, Mike. So I was asked to present this article. I was not involved in this research, but I did read the article when it first came out. I thought it was really interesting and was pleased to have an opportunity to address it. It is on continuous theta burst stimulation of the contralesional primary motor cortex for promotion of upper limb recovery after stroke, a randomized controlled trial. And this is the reference, which you should have in your slides. And this, I think it's important before we get into the details of the study just to discuss why they did this and what they were trying to accomplish. So this study is based on the theory that overactivity of the homologous motor cortex in the contralesional hemisphere is interfering with recovery through transcolosal inhibition. So we've got a stroke on one side, you've got intact hemisphere on the other side. There are connections between those two sides. And the concern is that there's inhibitory traffic coming from the intact side that is having a negative impact on the recovery process in the damaged hemisphere. The strategy here, and it's one of several studies and projects that have attempted a similar strategy using different techniques, is to inhibit the contralesional motor cortex using repetitive TMS so that you can, by inhibiting the other side, you can reduce this transcolosal input to the damaged motor cortex. For those of you who have been following this literature for a while, the NICHE trial, which was published in Stroke as well, I was a co-investigator in that study. Richard Harvey was the PI on the initial study, and then there was a subsequent follow-up study. This was another attempt to do just this, to use RTMS to inhibit the hemisphere on the other side, and it did not show benefit. So this was another sort of basic same strategy, but without success. This is a diagram that shows you a little bit more what we're talking about here. So normally there is a, let me see if my pointer works here, there are fairly rich connections between the two motor cortices that go through the corpus callosum. They're bidirectional, and they are to some degree inhibitory, which is one of the ways in which we can do unilateral movements, right? You inhibit one side as you're moving the other, because you only want to move one hand at a time, as opposed to a bilateral symmetric movement. If someone has damage to one hemisphere here, shown here schematically, presumably there's less transcolosal traffic going from that area, because the hemisphere is damaged, but the other side ramps up, and so you get this expansion of this inhibitory input into the damaged hemisphere, and the concern is that that may interfere with recovery. So the intervention here then was to take a form of repetitive transcranial magnetic stimulation, which induces a current in the brain. This is a magnetic coil that's held over the surface. It's a non-invasive therapy, in the sense that you don't go through the skin, but you are actively stimulating the brain with this. You hold this coil over the undamaged hemisphere. You can see the stroke is shown schematically here. And then to combine that with exercise therapy for the upper limb. This study was therefore a single center study performed in the Netherlands. They were looking at people with a relatively recent stroke within three weeks. They included both ischemic and hemorrhagic strokes, which is a little unusual. I think oftentimes studies like this will omit hemorrhagic strokes. They enrolled 60 participants, and then they randomized to this other form of stimulation called continuous theta burst stimulation, which is different than stimulation used in the NICHE trial, or sham stimulation in equal numbers. They looked at the ability to extend one or more fingers of the paretic hand and used that to stratify, so that they would have equal numbers of patients that were more or less severely affected in the two groups. They used as their baseline measure something called the motricity index, and had a very broad range of scores that they included, pretty much up to normal, except they excluded people who had a normal pinch grip. So they might have scored normal on the score, but they still had some impairment of pinch. As I mentioned, they started within three weeks of stroke onset, which obviously is a challenge. Bad timing for a cold. The protocol consisted of 10 daily sessions of this stimulation, combined with conventional upper limb exercise therapy. The sham was basically a very low level of stimulation. So they did that at 10% of the resting motor threshold, rather than the more substantial stimulation provided to the treatment group. Then the stimulation is very brief. So the whole thing takes 40 seconds. It's a quick, very fast, multiple stimulation of the affected brain area. In this case, again, the contralesional hemisphere. And then the effects of that are presumed to last as they subsequently do their exercise therapy. So this is not being done during treatment with the exercises, but immediately prior to it. This shows their treatment, patients who were screened, randomized, and treated. They had a little bit of one patient that dropped out before they started, and one that didn't complete the follow-up measure. But in general, their follow-up was pretty good. So this is their primary outcome measure. This is the Action Research Arm Test, or ARAT. They looked at that at three months as their primary outcome. And they also looked at repeated measures using a mixed effect model as a way of assessing change over time, which often gives you more statistical power. And you can see that at three months, they achieved statistical significance, which was a p-value of 0.0244. And that was sustained. Needless to say, it's hard to follow everybody for the entire one year. But to the extent that they did, everybody retained the benefits. This is the upper extremity Fugl-Meyer, and you can see that they showed benefits here as well. Now remember, these are patients early post-stroke, so these are patients who are recovering anyway. You have to be a little careful about not getting too excited about the increases, but certainly the relative differences between the groups are of interest. And you can see here that they did show a significant difference between the two groups on this secondary outcome measure. And a nine-hole PEG test, similar results. And the modified Rankin, which is, of course, a much broader measure of mostly of mobility, it's usually called a disability measure, which is mostly how well can you get around. And they showed a significant shift here as well, that the active treatment group had a less severe disability when measured at three months. The good news is that there were no serious adverse events. There was some increase in headaches in those who were receiving active treatment, which is not entirely shocking. And of course, cranio-magnetic stimulation can stimulate the muscles as well, and so it's not surprising when people may have some headaches, which may be from the actual more superficial effects. And they did see a shorter length of stay. For those of you who are paying attention and are from the U.S., these lengths of stay are quite shocking. Just they're so different than our practice here, it's hard to even talk about it in the same breath. But nonetheless, they saw a shorter length of stay. So there are some limitations, inevitably, in these kinds of research. Despite their efforts to balance the groups, the treatment group was six years younger on average than the control group, and that's, you know, it's hard to get these things perfect. It's not a huge difference, but nonetheless, it's notable. And I did notice that more of the treatment group had their dominant upper limb affected compared to the control group, and that worries me a little bit more. I don't know exactly what the effects are on dominant versus non-dominant arms here, but presumably, people's motivation to recover a function in their dominant arm may be stronger than in their non-dominant arm, because they use it so much, they're doing more exercise with it. So it's always a little bit concerning. The other issue, which is a significant methodologic problem, is that the researchers, and presumably those treating the patients, were not blinded to treatment assignment, and the secondary measures were not blinded, and that's a pretty significant limitation. We just have to acknowledge, you know, especially when we talk about length of stay and things like that, expectations on the part of the caregivers play into this, and if you think somebody's really doing well because they're getting your exciting treatment, then you may, you know, affect their treatment in a variety of ways, including sending them home earlier perhaps, since you'd think they're doing so well. So in conclusions here, I would say that this form of transcranial magnetic stimulation was well-tolerated, which is great news. The benefits that they saw exceeded the minimal clinically important difference for the ARAT, which is encouraging, and I think it's very much worthy of further studies. It is always disappointing, and I think probably a good topic for discussion of why some of these preliminary studies look so promising, and then larger multicenter RCTs fail, and I think that is always a concern here, but I do think a larger multicenter RCT would be appropriate. So I will stop there, and I would welcome questions. I think I'm taking questions now, so please step up to the mic, let me know what questions you may have, if any. Thank you, John, for not leaving me standing here looking so lonely. Go ahead, yeah. The assessors were blinded, but the treating, the research, the clinicians were at least blinded for the outcome measure. So the primary outcome measure, those assessors were blinded. they didn't do those kind of subgroup analyses it was a you know only 60 subjects so they didn't look at that it would be interesting to know if that I think there's a general expectation in these kinds of studies based on prior literature that people who have less severe deficits tend to benefit more and those who are the most severe tend to benefit less which is always frustrating because those are the ones who need it the most but I don't know that they've published that analysis for those who are severely impaired that might be the opposite right I mean there is some thought that that you know just to follow on that that in fact in the most severely impaired patients that some of the restoration of movement may actually come from the uninvolved hemisphere through uncrossed ipsilateral connections and so I think that that's that's a very important you know point it becomes very important to make sure you have a clearly defined population here this was rather broad I have to say in terms of the level of impairment and a little bit concerning that way I would imagine a subsequent multi-center trial would have narrower inclusion criteria in terms of motor deficits certainly I'd like to see that other questions they don't have to be as erudite as John's it was on but now you turned it off are calculated for lots of measures and for most measures in most clinical even meeting the MCID is a real stretch and yet in this trial they achieved almost double the MCID and and that kind of almost sounds too good to be true and so I'm just wondering if you had thoughts on can you explain that away by blinding or by the dominant extremity or or I mean that's really a remarkable achievement yeah I don't want to be too too maybe negative I think that we have to be careful about early stage research like this that are done with small numbers you know there is there are random effects that that come into small samples like this that get washed out in larger samples and so when a trial like this fails everybody sort of shrugs it off doesn't really look maybe doesn't even get published and so there is a publication bias towards more favorable studies and so if you do enough small studies some are going to come out with really nice results just through random you know good luck if you will I'm not saying it doesn't work I don't want to be you know misquoted but I think that we have to be really cautious about overly accepting these results we saw this also in the SSRI studies right the initial studies really looked rather promising and then the subsequent studies were disappointing in that way when they were larger and a lot of those you know big enough study a lot of those random differences between the two populations wash out again this issue of handedness I don't know how that plays in the issue of the patients being six years younger I don't know I mean certainly younger is better in terms of recovery either one you can go wherever you want you can even use the podium please so there are a couple of NIH stroke net studies going on right now that are using TMS so we're trying to get an understanding of the generalizability of this modality when it comes to stroke recovery the problem is the inclusion criteria and exclusion criteria are very strict so you can't have a patient who say is completely plegic and some of the patients the outcome measures are cognitive outcome measures you can imagine a left MCA stroke with cognitive deficits in aphasia couldn't possibly participate in this type of study so that may be some of the reasons why we have these wonderful preliminary studies that show this great efficacy but generalizability is the difficult thing because we simply can't get the sample sizes needed should I sit over there yeah I think that's a very helpful and in a trenchant comment it is it is difficult both to I think conduct these kinds of trials and then you do worry when you succeed if you define things so narrowly how does that apply to your patients who don't necessarily meet criteria you know even with you know now new clinically available treatments like the vagal nerve stimulation we're still trying to figure out well this is how it was studied but what does that mean for my patient that looks like that might be okay but I'm not sure they're not in the in the previous studies criteria thank you so much thank you dr. Stein nice segue what dr. Francisco is making his way up I don't want to get yelled at by a PM&R so welcome to the session please silence your cell phones please fill out your evaluation forms and claim your CMA and visit the pavilion while you're here and visit our folks who are exhibiting weather here remember to wear your physiatry day t-shirt dr. Francisco oh and disclosure I actually am a co-author on the next study that's being presented but honest to God that's not the reason I chose this sure hello everyone so so Mike asked me to present this paper actually this is a post hoc analysis it's a follow-up paper that was published I think published before 2023 but you read it in 2023 yeah yeah okay so this is in order for us to appreciate this particular study we have to go back to the original so this is our paper that was published in the Lancet in 2021 it's the pivotal randomized controlled trial using vagus nerve stimulation and post-stroke upper limb recovery so I will the methods of the paper that we will actually talk about today will be based on on this so shout out to the investigators from the US and the UK some of whom are in this room it's a I'm biased I think but I think it's a well-designed study for a randomized controlled trial it's triple-blinded randomized controlled trial what we did was we recruited the patients who fulfilled the eligibility criteria and consented everyone got the VNS implant so randomization happened after the implantation they were randomized to either one of two groups one to receive active stimulation and the other one got sham stimulation and then after that they all received six weeks of in-clinic therapy under the supervision of a skilled therapist who all were trained on the task-specific training that will be given these subjects also during the time and I will give you more details later it was the therapist who was triggering the vagus nerve stimulator after that six months six weeks we did the first assessment day one after completion of that therapy and that was the main outcome measure following that they underwent the home-based therapy and this time it was the subject who was stimulate who was the triggering the vagus nerve stimulator as he or she performed the activities at home and then they were assessed at 30 and 90 days what this slide doesn't show is that at the completion of the blinded face the group that received the sham stimulation the control group crossed over to the active group so they received the same protocol the same stimulation plus therapy protocol that the original active group did so these are the baseline demographics and characteristics of the population a lot of numbers here but just trust me it was a balanced group in both the VNS and in the control group the age sex race everything else were pretty much comparable in both groups it's a relatively younger population the mean age was in the late 50s early 60s so this is the paired VNS therapy this three two-hour sessions per week for six weeks so it's relatively intense more intense than what we usually give in outpatient therapy and it's also a little bit different because in in the real world outpatient therapy that one hour starts when the therapist says hello to the patient pick them up from the lobby bring them to the therapy gym chit-chat so in the end after that one hour realistically they're getting maybe only about 40 minutes of actual therapy but here the setup is artificial they're in the the clinic and when we say two hours it was fully two hours so it's a pretty intense therapy compared to what we are giving our patients right now I see the therapist here she was holding she's holding that button and that was the trigger so there is a script that I will show you in the next slide when the therapist is supposed to trigger because what what has been found in pre and pre-clinical studies and in the preliminary study the pilot study that we published in 2017 what we found is that it is very important to time lock the stimulation with the task that the patients or the subjects are being taught the key here is repetition repetition repetition the goal is to achieve more than 300 repetitions for a task specific activity and in case you're wondering how many times the stimulation was given again in the control group and in the VNS group there were comparable number I mean stimulation per session we far exceeded the goal of three more than of 300 repetitions so this is an example of a script the therapist can choose the activities on the left the task categories and then there are examples and then you can see in the rightmost column it tells when the therapist is supposed to trigger so that way we can minimize the variability in terms of the temporal relationship between the stimulation and performance of the actual activity so I mentioned earlier after six weeks of in clinic therapy the subjects now performed daily exercises that were prescribed by the via therapists they activated the VNS using a magnet they swiped the generator that's implanted under the chest and we had the ability to track and log the activities within the implanted device so outcome measures the main outcome measures the Fuglemeyer assessment upper extremity component the change from baseline to day one after completing the in-clinic therapy supervised therapy and therapist supervised sessions and then the secondary outcome measures were the responder rate which is defined as at least a six point difference some glad you brought up the clinically meaningful difference because I will bring it up once again in this context also with the Walsh motor function test from baseline to post day 90 and the difference in the Fuglemeyer scores from baseline to 90 days after completion of the in-clinic therapy so let's fast track to the results the red line represents the group that received active VNS blue control and what you see here is the change not the absolute sports but the change in the Fuglemeyer upper extremity score from baseline to day one post completion of the in clinic therapy and you can see the difference between the two so if you look at the blue line only you can say that hey they still improved even though they did not receive the actual VNS stimulation so just shows you the importance of in high intensity task specific exercise but if you compare the difference with the group that received the additional VNS stimulation right before they move the difference in the improvement of the actual of the Fuglemeyer scores were significant both in the main outcome measure post day one and then when it was repeated 90 days later the same observation was found in the secondary outcome measured the wolf motor function test and then the crossover so at the completion of the blinded phase that they were re-baselined and the group the blue line the group the sham group was now crossed over to the active stimulation so they received the same VNS protocol that during this phase and you can see that when we when they were assessed at day one post in clinic therapy and 90 days post the magnitude of improvement or the change in the from baseline to at least the time points were pretty comparable as what was achieved by the group that received the stimulation from the very start of the study same observation with the wolf motor function test now the graph that you have on the right there was a post hoc analysis looking at the clinical clinically meaningful response rate of at least six points on the Fuglemeyer from baseline so six is actually pretty pretty good because most studies would have report only a difference of four or difference of five but I think this is something that's important both for the study and just in general when we perform many of these upper limb upper limb experiments and interpreting the results and as Nneka pointed out just generalizability they may be clinically significant they may be statistically significant but the clinical significance of the clinical impact may not be the case so in this case we're looking at the percent of the responders blue represents the percent of the people who had been at who had at least six six point improvement and the Fuglemeyer scores read for those who received the active VNS but you know then the difference of the rates between the two groups is statistically significant meaning to say that those that receive the active VNS more of them had the response that we were hoping for however pay attention to the y-axis and you will see that only about fifty percent of the people who received the active VNS achieved that at least six point six point improvement on the Fuglemeyer so more on that later now were they really blinded so we surveyed the participants in the study eight percent of the people in the VNS group correctly identified that they were in the VNS group, about 20% in the control group correctly identified that they're in the control group. Conclusion based on this is that it was a pretty good blinding. The majority did not know as to which group they were assigned. Safety and adverse events, I'll just summarize it by saying that there was no new safety signal that was identified in the study, it was pretty much the same as what is expected and what is observed in those who received the vagus nerve stimulator for epilepsy or for severe depression. So back to the study that Mike wanted me to talk about. So this was partly inspired by this. So if we found a statistically significant difference suggesting that the VNS plus therapy is indeed superior to just therapy with no VNS, then how come only about 50% of people benefited? Are there factors that we can pre-identify that can predict who will respond to this paired VNS and therapy protocol? So with that in mind, we looked at some variables of interest, age, we did a cutoff of 62, sex, male or female, country, UK or US, severity and a cutoff of 34 was used for the fugal mire, times in stroke, even though the range is anywhere from nine months to 10 years, we decided in two years so that we will have a more or less equal number of subjects in each group. We compared right versus left and cortical or non-cortical involvement as the etiology of the stroke. Very busy slide, anyway, it's a forest plots, it's relatively easy to interpret the general rule anything to the right of zero, this is the vertical line at zero, if it's to the right, it is in favor of the intervention in that study. And here what we're seeing is that we could not really differentiate the group that received VNS from the group that did not receive VNS, I'm sorry, we could not differentiate who among the participants in each group would benefit from this, meaning to say that we were not able to show or prove that there are factors that can predict who will benefit more than others. So let's talk about some of our observations. Now in general, we have an idea of the post-stroke upper limb recovery outcomes, the publications are usually in general, talking about the natural evolution of stroke and who recovers movement and not, age, sex, lesion location, severity of motor impairment have been used, a nice study by Stenier looked at the motor evoked potential and severity of motor impairment as a predictor of who will benefit and I think they used 11 weeks or that's when they discovered that most people who will recover something in the upper limb would have done so at 11 weeks. The outcomes of paired VNS therapy were also consistent across subgroups, this is what we found in that study, that age, sex, impairment severity, time from stroke, stroke location and the parietic side were not predictive of who will benefit from the paired VNS therapy. This is consistent with the observations in the pilot study, that was a very small study in end of 17, looking at the feasibility and then, excuse me, a paper by Gert Quakl published last, in 2015, also looked at these predictive factors for response to CIMT therapy and again, it's consistent with our observation that it's really hard to pinpoint who will benefit the most. Now, this is curiously also similar to the preclinical evidence in rats, age, cortical or subcortical lesion, ischemic or hemorrhagic type of stroke did not predict who will benefit from this combined therapy. So, limitations of the study, we only studied seven variables. End of 108 seems to be very big but it really, it's relatively small for the purpose of being able to compare different subgroups so everything that we did was binary, male, female, under the age of 62, 62 years and older but what if we had more subjects, we could have had a more complex and more specific categorization. So, for instance, we can compare older males versus older females. So, we could have combined those variables but that was not done in this particular study. Also, we explored only one cutoff point. We did not stratify the severity. We used a midpoint in the Fugl-Meyer but if we had more subjects, then we could have categorized maybe those whose Fugl-Meyer baseline was between 20 and 30 versus those between 30 and 40 and those between 40 and 50. And as with any randomized controlled trial, eligibility criteria is limited. It has to be a clean study. We want to remove as many confounding variables as possible. So, for example, in the inclusion criteria, we only included those with supratentorial ischemic stroke within nine months to 10 years. Does that mean that this is not going to work for someone who's 11 years post-stroke or six months post-stroke? We don't know because that population was not studied. Also, the age, we only studied 22 to 80 and I'm sure you've seen many people who are in their late 80s and early 90s who are still doing quite well after a stroke and perhaps could have benefited from this combined therapy. As I mentioned, it was a narrow Fugl-Meyer range that we used for those who were interested in spasticity. One of the exclusion criteria was not to include those whose ASHRAE score was three or more. But does that mean that they will not benefit from it? Maybe, they could have. We don't even know the effect of Dacus nerve simulation and spasticity. So, it's just quite challenging when you're trying to apply this in the clinical setting. So, the minimum clinically significant difference I think that's one thing that this study has not really answered. But I think it offers an opportunity for us to delve into that into more details. In the study, we use a cutoff point at six. But in some other studies that were not considered statistically significant, the changes that were observed were not different from what we found here. But the fundamental issue is how we define the MCID. When this paper by Steve Page, which is often quoted, was published, it was just based on a study. He surveyed some of the occupational therapists who participated in the Everest study. So, the MCID as we know it, as we're using it right now, is based from the point of view of a group of clinicians. So, we really do not have a universal way or a uniform way to define MCID. Hopefully in the future, MCID will be from the point of view of our patients and the subjects, or perhaps a combination. And that's why we have to be cautious in interpreting those claims that this treatment is fantastic because we reached that MCID that we had hoped for. Always study and always interpret that in the context of the other findings in the study. Still, it's a well-designed study. Some people were questioning whether or not we should consider this to be a really promising therapy for stroke because the findings were not really as robust. But in my opinion, I'm looking at those numbers in the context of having treated many of those patients. And many of my patients who participated in the study, a few of them have reported things that we don't measure because we don't incorporate those in those validated outcome measures that were expected to use in the RCTs. How can I put in the Fugl-Meyer, for instance, that one of our subjects who, I don't think his scores were that great, but functionally he was so much better because now he recovered more movement in the thumb. And that opened a lot of possibilities for him, that small twitch, because it allowed him to navigate the web through his phone even more. So I think we have to revisit how we are defining MCID so that we can better appreciate the significance of the innovative therapies that are introduced. Thank you. Are there any questions? Yes, there's a microphone over there. Thank you. Is it working? Yeah. Why did you exclude patients with a modified Ashworth of three or more? Why did you exclude patients with a modified Ashworth of three or more? Because we wanted the study to be clean. It's going to be hard to interpret if the failure to improve on the Fugl-Meyer was because of weakness or spasticity. So we decided three and four out. Got it. Thank you. Please introduce yourself when you come up to the mic because we want to know who you are. Allison Jones, I'm medical director of inpatient rehab. What is thought to be the mechanism of action of stimulating the nerve, causing improvement of the upper extremity? Great, so the mechanism of action, vagus nerve stimulation has been shown to facilitate the secretion of neuromodulators in particular noradrenaline and acetylcholine, which have also been shown in prior studies to be crucial in motor learning and plasticity. John? Hi, John Che from Case Western Reserve and Metro Health. So I've reviewed this. This is a very nice study, very nicely designed, classic RCT, nice crossover, an important contribution to our literature. Some moderating observations, however. You know, if you look at nearly all motor relearning studies in stroke and you compare the treatment versus control, for all intents and purposes, nearly all these studies will show a Fugl-Meyer difference of three. And this one is no different. And the question is, what does a Fugl-Meyer difference of three mean? Which brings me back to the question of the minimum clinically important difference and how we use that, right? So the way it was used here is from baseline to end of treatment. So those individuals who scored that much, those people got this and you got, so it's basically a responder rate, right? Another way of looking at this is, well, if that's considered clinically important difference for that individual, what do we consider as clinically important difference between the groups? Right? If the clinically important difference for the individual from baseline to end of treatment or follow up is six, I could see in a case series, you can say, oh, this percent of people did this. Another way to use that is, well, what if you compare that to the control group? Is the difference between the treatment group and the control group clinically important? Well, if you use six, then you would say it isn't, right? And so this is just a commentary that this adds to the literature, but it's consistent with everything else we've done in motor relearning. And which begs the question, do we need to move on to a different approach than a classic motor relearning approach? And I say this criticism of my work, because my work also shows no more than three point Fugl-Meier difference. And it's, where are we going in the future when we're getting the same basic result? In this particular case, it's more invasive. Right, thank you, John. Having chills here, I'm presenting an RCT and that John Shea and I did our first RCT together when we were residents. We were co-residents, but he was my research mentor. So I learned a lot from John. If you remember, you asked me to do that Fugl-Meier for those subjects that took a long time. That's my commentary. Well, a few comments on your comments. One, I think Fugl-Meier, when we do the total scoring for the upper limb, it's all tasks that involve proximal and distal. But as you very well know, if all you will recover is a distal task, you might be able to be more functional. You cannot be functional if this is the movement that you're going to recover, if it's more proximal. Whereas if it's distal, even if you're tight here and you cannot move your shoulder, you can bring the object to you and you can manipulate it and therefore you can be more functional. So perhaps that's one opportunity for studies to deconstruct the Fugl-Meier and comparing the proximal and the distal Fugl-Meier scores. So that's one. And number two, I think the MCID really, I think it's a great start. I mean, I think that's Steve Page's greatest contribution that he brought this concept to the forefront. But we have to be more cognizant of what that really means. In this study, for instance, if we pushed the goal of improvement to an MCID of seven rather than six, we may not have shown that significant difference. If it brought it down to five, then it would have been more robust. But what does that mean clinically? And I really think we need to look at, and I think this is the focus of some of our outcomes researcher, perhaps looking at a combination, perhaps a new scale based on currently existing scales that will incorporate objective observations and subjective feedback from the subjects as well. Because what may be statistically significant may not be clinically meaningful. And what may be clinically, you can have a study that will say, no, it doesn't work. But those subjects who participate in the study, some of them said, oh, it worked so well. So I think we have to, we're not at that stage yet, but I think we need to continue having that discussion to come up with a better, more sensible outcome measure that will reflect statistical significance and clinical meaningfulness. Can I ask you to hold your question until the panel discussion at the end? Thank you very much, I appreciate that. Thank you, Mike. Thank you, Gerard. Also a nice setup for the last paper. So I'm gonna bat cleanup today, if I can find. We've had two interventional studies, and so the last paper we're going to cover today is actually a descriptive paper. No? That's not... Hang on a second. I'll be right back. Yep. There we go. So the paper that I chose is Distinguishing Distinct Neural Systems for Proximal and Distal Upper Extremity Motor Control after Acute Stroke. Thank you, Gerard, for setting me up. This was published in Neurology of this year. The lead author was David Lin from the Department of Neurology at Mass General, or as I like to call it, Time to Tweak Twitchel a Tad to Tell a Transformative Tale. So upper extremity motor recovery has just been suggested as absolutely crucial to outcomes of stroke, and that consists of both proximal and distal components. This work has shown that there are two neurosystems serving those proximal and distal muscles. The primary... The proximal muscles are served by primary, pre, and supplemental motor cortices. They're bilateral, and they descend to the medial and ventral columns, as opposed to the distal muscles, which are primarily... The primary motor cortex is much more bilateral, or much more contralateral, and descend to the lateral columns. Mostly those of us of a certain age, mostly the folks that will be setting on the stage in just a few minutes, had beaten to our heads sort of the findings of Twitchel from the 50s, that a stroke is primarily distally weak, and the recovery occurs from proximal to distal. There's been some recent studies that would suggest that this is not always the case, and have countered this, but I think it's wise to point out that both Twitchel and the more recent studies were really actually quite small, and Twitchel's study really looked at only severe strokes. So given the relative absence of the natural history of upper extremity recovery following stroke, the aims of this study were, one, to evaluate proximal and distal deficits in a large representative sample of persons with stroke, and to relate those proximal and distal deficits to functional outcomes. I'm going to touch on some of the neuroimaging data, but time restraints don't allow to go into a whole lot of detail. So these subjects were part of the Stroke Motor Rehabilitation Recovery Study. That was certainly a three-martini lunch that came up with that one. The inclusion criteria were actually quite broad. Within one week of either an ischemic or a hemorrhagic stroke, 18 to 90 years old, able to follow single commands in English, the motor requirements were at least a one on the NIH stroke score, so they had to have weakness of a drift or more, no disorders of consciousness, and admitted to the MGH Stroke Service over a four-year period. The exclusion criteria is that there needed to be no prior developmental, neurologic, or psychiatric disorders, and no visual or auditory problems that would limit testing. Over the four years, there were 3,200 admissions to the service. They found 227 that were eligible, and 141 were consented. This is very important because this study depends very, very heavily on these two assessments. So they define motor strength. They really only looked at strength and movement as the two motor components of motor deficits in this study. They looked at manual muscle testing of the shoulder abductors and finger extensors. This is called the SAFE score. This is very widely used in stroke research. Scored from 0 to 10. And then also sub-scores of the upper extremity, Fugl-Meyer, for 3 to 8, items 3 to 8 for the proximal, and 24 to 30 for the distal. Remember, Fugl-Meyer's score is either 0, you can't do it, 1, you partially do it, or 2, you do it fully, the movement, so the score is 0 to 28. If there didn't happen to be a deficit in those muscle groups, it was not analyzed. The scores were divided by the maximum, so you had a normalized score, and then they looked at, I think rather cleverly, proximal to distal gradients. So it was not you either had it or you didn't have it, it was variations of a gradient. And then they looked at proximal strength and movement versus distal strength and movement, and then they calculated the gradient as essentially the difference over the total score. They also looked at baseline box and blocks as a functional assessment and Barthel index, and then at 90 days, the participants returned with all those things I just mentioned, plus the modified Rankin score. This bled into the pandemic a little bit, and so those folks that weren't able to return, they did modified Rankin and Barthel by phone. Statistics, the main statistics, they used Spearman's to correlate various variables, and used partial correlation. The partial correlation is where you correlate two variables and then hold a third variable as a control. An example of this is looking at a correlation between proximal movement and proximal strength while controlling for the relationship, for example, of proximal movement and distal strength. So results, 141 participants that were assessed, about four days post-stroke, mean age 63 years, NIH stroke's scale of seven, and about half of the sample were female. 62% left hemi, quite very few hemorrhagic strokes in this sample, 96% ischemic. About a quarter of the sample had TPA or thrombectomy, 84% follow-up, which is really quite good, and there was no difference between those who were lost to follow-up and those who were analyzed. Looking at the distribution of severity, it was rather bimodal, about 40% of the sample were considered severe in terms of their motor deficits, and 23% mild. The correlation between the Fugl-Meyer and strength was quite strong, as you might expect. You kind of have to have strength before you're able to move, experiments of 0.9. This is quite important. So when they looked at the correlation between the proximal and distal movement and strength, what they found was a very, very tight relationship between distal movement and distal strength at 0.77. Looking at distal movement and proximal strength, the relationship was very, very much less strong, and looking at the relationship between these two correlations, the partial correlation was quite strong, or was quite significant, meaning that these are very different from one another. They did the same thing looking at proximal movement from distal strength and proximal strength, finding, again, these were not so different from one another. There was not a significant in the correlation, the difference between the correlations. How they interpreted this is that distal movement is a very separate entity from proximal movement and constitutes an independent presentation, or a distinguishable proximal presentation versus a distinguishable distal presentation. This is busy, but we'll get through it. They also, kind of in contraindication to Twitchell, also found that anywhere from 20 to 25% of the sample actually presented with distal deficits actually better than the proximal deficit, the distal motor function better than the proximal motor function. Interesting enough, this had nothing to do with the vascular distribution of the stroke. It didn't matter whether it was a middle or an anterior or a watershed infarct. Looking at the relationship between proximal and distal outcomes, again, this is the gradient. Over here on the left is primarily proximal deficits. On the right-hand side is better distal control. There's a very strong relationship both at baseline and at 90 days. Those folks who presented with better distal function did much better at both baseline and 90 days, both on box and blocks and on the modified Rankin. Again, I just wanted to show you because it's pretty impressive. This is all 141 of the MRIs overlaid with one another. In using various techniques, they were able to overlie just based on whether somebody presented phenotypically with proximal or distal deficits, the blue being the distal and the yellow being the proximal. Just based on those presenting deficits, they were able to overlie the MRIs. Then in technique, they were able to maximize the difference in those voxels. What fell out is exactly what would be predicted. The distal folks had lesions that focused on the primary motor strip and the proximal folks had much more diffuse lesions going through the supplementary motor cortex and the striatum. The conclusion for the authors were that about a quarter of the sample presented with proximal deficits actually greater than the distal deficits and that they were able to make an argument of distinct proximal and distal syndromes after stroke based on their correlation analysis where distal motor deficits seemed to group and correlate much independently from the proximal deficits, that the distal deficits were far more associated with functional outcome at both one week and 90 days, and also, and I think the neuroradiology is really impressive that when you group these folks only by their phenotypic presentation, the combined neuroimaging was exactly as you predicted in terms of where those lesions are. What they conclude is that whereas sort of our teaching by Twitchell that generally you see preservation of the proximal movement following stroke with an evolution of proximal to distal, they're suggesting no, maybe these are syndromes where you start out with a proximal weakness or a distal weakness and those constitute two different syndromes. The authors cite the limitations as you might expect, the simplistic nature of the assessment measures, limitations in the modified rank and the Barthel index that it would have been better to use functional imaging and DTI in terms of imaging and suggested that these results might eventually be able to be used for personalizing neurorehab and there might be, depending on whether somebody presents with proximal or distal deficits, could we develop treatments to target? So from my standpoint, I think there are a couple things that we need to consider. One is it strikes me as a little unusual given how broad the inclusion criteria is that there were 3,200 folks that came through the stroke service and only 7% of them qualified for the study. I would have really liked to have known why the other 100, you know, 33,000 were excluded from the study because it really weren't that exclusive of inclusion and exclusion criteria. There's also something that I think could potentially be a problem. Function is innately, as Gerard just mentioned, function is innately related to what's going on at the hand and it worries me and particularly is very important to the Barthel index and the box and blocks and it worries me that what's going on at the hand could confound the model that they're trying to tease out. Weakness at the hand might reflect what's going on at the outcome measures completely independent than them trying to establish a distal and proximal model. And then the modified rank and score, I mean, gait is what really, what it is dependent on in terms of the score, which has absolutely nothing to do with what they're studying here. And then they really, really depend on strength and movement as the motor constructs that they're looking at to develop these models. And there are other things that can be looked at. What if they use coordination or speed or accuracy as their motor constructs for their model? Would they have come up with different results? Or, I don't know what it would be if they looked at measures that actually measure proximal quantities as their outcome measures, would their model have turned out differently? But despite that, I thought it was very interesting implications of this proximal and distal syndrome and what would be the ramifications on our part? Are there different technologies that we would implement depending on if they presented with one or the other? Would there potentially be medications that we would use differently, different exercise strategies? And particularly for clinical trials, as we try to find the most homogeneous groups as possible, is there a role, if this turns out to actually be true, would we stratify groups by proximal or distal presentation and would that help us come up with more homogeneous groups in clinical trials? But one way or another, I thought it was a fascinating alternative to Twitchell and something for us to consider and at least worth going over today. So guys, if you come on up, we have about 14 minutes and if you've gone up, come on up. We'll start questions and we can broaden the topics if you would like to to stroke rehabilitation recovery clinical trials methodology whatever you see fit so go ahead. Thank you so much and thank you for the panel great presentations I'm Farhan Vahidi chief scientific officer and VP of research at Tier Memorial Hermann MD PhD by training and a stroke outcomes researcher. It was a great present great set of presentations and I would like to maybe seek the opinion of the panel and make a few comments which build upon what has already been discussed. So there was one comment by an esteemed colleague over here is that what does it mean to have a two-point or a three-point or a six-point difference in any scale Fugl-Meyer modified Rankin whatever outcomes of scale it may be. I think the answer to that is it means different to different patients. A one-point improvement in MRS can be huge for a truck driver who is unable to drive and then is able to drive after a stroke but a one-point difference in a bedridden patient may not mean a whole lot of difference and I think that builds into what Gerard was talking about is that we really need to start looking at these outcomes in a more complex patient centric way. So I think that's one area where the field needs to move forward where we start re-evaluating the outcomes that we have. The other sort of aspect of all of this is that we can continue to have these great type 2 efficacy type clinical trials. We will always be challenged by the heterogeneity of the clinical condition. Even if we take ischemic stroke, even if we take supratentorial versus intra-infratentorial, we are not talking about whether it's a large vessel occlusion, whether it's a small vessel disease, whether it is cardioembolic stroke. There is immense amount of phenotypic and genotypic heterogeneity that we would never be able to answer from 100 patients, 200 patients or even 500 patients. So I think what what the expertise in the field really needs to start focusing on is really high-quality longitudinal data, observational longitudinal data. And that builds into your comment, what about those thousands of patients that do not get into clinical trial, right? They screen 3,000 patients, 100 we have some data on, 2,900 patients we don't have any data on, right? So till the time... Hang on just a sec. Nika, just based on that, I'm interested in your take on that because you have great experience in the clinical trial realm. I have to disclose I've been co-authoring papers with this man for about 15 years, so I agree with a lot of what he says. It's exceedingly important to have longitudinal data on all stroke patients. I also think it's prudent to have a conversation about these interventions that we're doing. So for example, for VNS, how many people in this room have had VNS implants into one of their patients yet? Me too. Three people. VNS has been used in epilepsy for decades, it's been used in depression for a very long time, and now we're using an upper extremity function. There are no social determinants of health factored in to the VNS studies for rehabilitation. There's no measures of depression, there's no measures of a promise, which is the toolbox that we use in our patients. So we really have no idea longitudinally about whether these patients were depressed and then their arms got better because their depression improved. We have no idea whether their ability of social roles improved, and therefore that also improved their motor function. So we don't know the chicken or the egg, and I agree with Dr. Vahidi, we need to make sure we have longitudinal data on not just the motor function, but things that affect social determinants of health, quality of life, as well as things that affect their ability to function. So I think we need to make sure that we have longitudinal data on not just the motor function, but things that affect social determinants of health, quality of life, as well as things that affect social determinants of health. VNS and a hemorrhagic stroke patient because I ordered my own CT. I saw hemocederosis and I said, just implant her. That's completely off protocol. I've gone completely off piece, but I can do that because I'm in an academic institution. I know the neurosurgeon. He trusted me. We put it in her and now she's taking care of her baby by herself at home. So I think it's really important to have an idea of the where they stood when it comes to their social terms of health before the intervention, follow them longitudinally, see how things flesh out with things like VNS and have a better idea of what we're really doing outcomes wise for these patients. Thank you. A follow up comment and sort of maybe, nothing there I disagree with. I think that's all right. I do think, however, going back to John's comment earlier, we do have an efficacy problem. And I just think when you think about the upper extremity Fuglmeyer as a 66 point scale, and we take in people who are 20 and we get them three points or six points, they're still a far away from 66. And so I just think as a matter of humility, we need to acknowledge that we have maybe made a few cracks in this nut, but we haven't opened it and we need to keep working on it. Hi, thank you so much for a good presentation. You got to put the mic down. We have people joining virtually. All right. Can you hear me now? Thank you so much for a great presentation. I'm Negar Moradian, MD, MPH by background, currently a postdoc researcher at UTMB working on gauge rehabilitation for post-stroke individuals. So one of the problems that we have by running our RCT there is that we need our PTs to send and refer the patients to go through our protocol. And our protocol is kind of intense. It's like 60 to 80% of maximum heart rate, but we have our own inclusion exclusion criteria. We have great PTs. However, there are many times that they are hesitant to refer the patients for their RCT. So at this point we have patients who are only mild or moderately impaired, but never have had severe patients. Although the protocol is completely safe, the device is completely safe. I would love to learn how you have, if you have any actually any solution to that problem to have more severe patients in your randomized clinical trials and learn more about those people. Great question. Excellent question. I think it all has, I don't know if there's a specific way to solve that. I think if you have a well-designed study, funding is going to be your limiting factor there. And then perhaps you will have to really have a big number of participants. And I would strongly recommend that you look at other factors as well. Neuroimaging, for instance, or neurophysiologic biomarkers, because you might be surprised to find out that even those who are severely impaired may have different neuroimaging or neurophysiologic profiles. So that's one thing. And studies are being done. I know at Johns Hopkins, there is a study recently funded that is looking specifically at severely impaired upper limb people that would typically exclude from our studies. So one of the solutions for that you're not gonna like very much. And so during the course of this year in one of the ethics columns that Johnnie McCurdy edits for the Purple Journal, we kind of address some of the ethical issues around research in neurorehabilitation. And you can make the same argument for folks with cognitive deficits and folks with aphasia as well. How come they're not getting into clinical trials? And one of the solutions is go to where they are, taking the recruitment strategies out into the communities where the potential participants are, and either setting up recruitment strategies out in the community, or for sure, having more persons with disabilities on the research teams and as the PIs of the projects. And if that happens, I guarantee you there's gonna be more strategies and more points of view to get those folks that are underrepresented into the trials. So our study is taking place in patient rehabilitation facility. We already do have those patients. However, so we need PT's approval in order to enroll them in the protocol, gait rehabilitation protocol. So that's the thing that we don't get. We have the funding, we have the device, we have everything, but the referral doesn't really happen. And we only get mild or moderately impaired patients. 50 bucks and happy hour with the PT's. That's true. You know, for a while, we, at least at the NIH, we pursued a single intervention approaches. And when we started to see this glass ceiling, then we went to multi-modal and there was a whole RFP. And so multiple multi-modal approaches have been done with no real difference. I guess this is a question to you as a panel. Is this just a glass ceiling that just needs to be cracked? Or is this a firm, rigid roof that can't be cracked? In which case, we need to start looking at fundamentally different approaches beyond our classic activity-dependent neuroplasticity with some sort of a priming type of approach. So as you look toward the future, your thoughts on this, because I'll tell you at this point, I'm ready to jettison activity-dependent neuroplasticity. I think we probably all have comments on this one. It is sort of the big conundrum in our field. I think, you know, we don't know. We don't know if we really push just harder. If we found the right treatment, we would crack through that glass ceiling, or that's it. That this is the best you can do. You know, we are not fixing the underlying anatomy. And you look at these scans, and you see a lot of damage, some of them, and we're trying to work with what's left. And so maybe there are some biological limits here. There's always been this interest in using stem cells for tissue restoration in some way, but we're still a very long way from that biologically. The notion of figuring out the rest of it in the meantime, though, has a real appeal in terms of better exercise protocols, understanding how to maximize our therapies, because presumably when we get a biological treatment breakthrough, then we'll need to do exercises. Then we'll need to do other interventions to facilitate that. So I think there is a strong argument for working on robotics, trying other stimulation techniques, doing the different things we're doing, in part to get ready for that biological breakthrough. I think the timing of what we're doing is gonna become exceedingly important. The Lynn article goes over specifically MRI sequencing, so DWI, ADC, and looking at the specific tracts that are affected that may portend improved proximal function versus distal function. But this has been going on when it comes to stroke research for about 20 years. Between 1995 and 2015, when the NINDS-TPA study came out, and then the five New England Journal of Medicine studies came out looking at endovascular therapy, stroke has changed even more over the past eight years. So the way that people are having strokes and the evolution of the penumbra and what is surviving and what is not has completely changed from Twitchell. It's completely changed since 2015. And so I think it prudent that we actually start to intervene from a rehab perspective at a different time point. So if we're waiting till we get to inpatient rehab, we might be waiting a little bit too long, but we need to have the ability to understand what's happening in the penumbra, what's happening with neuroinflammation, neuronal damage, to figure out when we should intervene. And I think that will really help a lot of our patients when it comes to determining who's right for early rehabilitation, who's right for certain interventions. But if we don't keep up with the imaging, and in the Linn article, they were very clear. They put up the slide. You have to know where the stroke is. You have to have an idea, if you're gonna have some idea of precision rehabilitation, of what the MRI looks like, what's being affected, what the corticospinal tract looks like. Does it have good integrity? Is it a really large stroke? Do we have new petechial hemorrhage? What's happening after endovascular therapy? And if we don't do that as a field, we won't have the ability to inform our rehabilitation paradigms. The only thing I'll add to that, I think we have to stop just saying that RCTs are the gold standard. I think we have to revisit how we're evaluating different therapies. I think RCTs are very important, but that should not be the only gold standard. Perhaps it will have to be a combination of an RCT and another study validating the findings in a more clinically relevant setting. That's number one. And number two, I just thought about it, but I'll comment on the multimodal approach. I think we have to do it. Function is so complex, we cannot expect a functional improvement by correcting only one impairment. And perhaps that's why some of the treatments have not shown any robust improvement, because we're only looking at one particular aspect. And the other thing that we really haven't looked at is the dose of the therapies or the interventions that we are giving. Many of the upper limb studies now, most of the decision, especially for very new studies, were based on what we learned from the series of the constraint-induced movement therapies. So I think we have to look, we don't have to discover new things. I would prefer that we study what we have now and manipulate how we're doing it. So for instance, TDCS, should we do cathodal stimulation? Should we do anodal stimulation? Should you stimulate the lesion side or the contralateral side? Should you do both sequentially? We don't know those things. I think we have to take a step back and resist being seduced by, yay, new therapies. I would prefer that we study what we have, those that are promising, and see how we can exploit them to be more effective. Okay. Well, first of all, thank you very much for walking 500 miles to come down and join us for the session today. Thank you to the panelists, and have a great afternoon. Thanks. Thank you.

stroke rehabilitation

continuous theta burst stimulation

upper limb recovery

randomized controlled trial

adverse events

vagus nerve stimulation therapy

predictors

patient experiences

clinical trials

motor control

social determinants of health