Praxis Precision Medicines, Inc. (PRAX) Earnings Call Transcript & Summary

All right. I think we can get started now. Good afternoon, everyone, and thanks for joining us for our virtual fireside chat with Marcio De'Souza, CEO of Praxis Precision Medicine. Before we begin, I need to read our safe harbor statement. This call has been arranged by Truist Securities Research for use by institutional investors as defined under FINRA rules. If you're not an institutional investor, please disconnect at this time, and please see our website for our equity research library and the required disclosures. If you have any questions? Can you hear me, Marcio?

I can hear you.

Also, if you have any questions, you'd like me to ask management, just e-mail me at [email protected], and I'd be more than happy to ask that for you.

And I hope you can hear me now, Joon.

I can hear you. Let me just ask -- because one of my associates said that they can't hear me. Okay. I think they can hear me now.

Awesome.

All right. So -- all right. Good. Thank you. Thank you, everyone, for telling me that you can hear me. We had some audio. So let me just quickly read the disclaimer. This call has been arranged by Truist Securities Research for institutional investors. And if you're not -- as defined on the FINRA rules, if you're not an institutional investor, please disconnect at this time and see our website for the required disclosures at the Truist Research Library. And if you have any questions, e-mail me at [email protected], and I'll be more than happy to ask them for you. So Marcio, thanks for making the time. You had a busy day yesterday and today.

Just a bit.

Yes. Just a little bit. And you top line data from the open-label trial of RADIANT evaluating vormatrigine for focal epilepsy. I want to dig into the patient disposition, the efficacy that you toplined and the tolerability and background therapy and upcoming data at IEC and POWER1, 2 and 3. But I'd like to start off with maybe a 5-minute recap of what you shared yesterday.

Yes, absolutely. I appreciate being here today after like a great data that we presented on RADIANT yesterday. Maybe we start talking a little bit about the market itself, right, because maybe that's where this always started and then how RADIANT and POWER1, 2 and 3 would play a role here. We know -- I think we all know that how large the epilepsy market is like about 3 million, 3.5 million depending on how we look into the number in the U.S., of course, proportionately more outside of the U.S. Many drugs have been attempted, many drugs are approved to treat epilepsy patients. But unfortunately, I would say on one side, that is really not a solution right now, not a drug or a set of drugs that address the majority of these patients or even the minority of these patients. So on that side, fairly unfortunate, there's a lot of opportunities there. On the other side, fairly fortunate, right? Because what we know is, you can come in on to this market and as we're looking for vormatrigine, make a real impact with a drug and it's by no stretch of imagination, the only drug or the only set of drugs like you had 3 other epilepsy drugs for focal epilepsy. It would be interesting to develop all of them in a sense because the markets really can accommodate a number of drugs. And I think that is a very interesting motivation because what I see a lot going on in conversations that we have way less with the neurologists treating these patients, but a lot more with people like us and conversations like this. Is this impression that there's a race for the last few patients, right? And while maybe one day, that's going to be the case, and it's going to be like one way to cure epilepsy. I think right now, we're really trying to get better and better as these drugs and treating more and more of those patients. So in that regard, incredibly happy on making a step forward. I think the second incredibly positive thing about yesterday is it's a little bit of a water like divider for us, right, moments or shared moments where we were very bullish about everything that happened before yesterday, before vormatrigine in terms of the PPR results of the preclinical data or the Phase I studies. But we always said, and until yesterday, I would still have said that, we need that data in focal patients to really be able to, like say, this is a drug that is highly effective in focal onset seizures. And I think we are now at the moment that we can turn that page towards really saying that quite confidently. Now from this point on, then it's how to bring this drug to patients in the broader markets. Looking to the profile of this drug is one that really sets itself to keep moving lower and lower on the adoption, right? I think everyone starts. One could argue rightly so. I'm not sure if I agree with that on these hyper-refractory patients. They are important, but they are a fraction of the market. They are a small fraction of the market. They're incredibly valuable from a revenue perspective, but there's still a small part of the market. But then moving lower and lower would be ideal. If you go from an external reference and then I'm going to bring an internal reference by looking to the human epilepsy project, one that focus on focal onset seizures. That was the 2, the focal and generalized. There are 2 major conclusions there that I think are super interesting. They have many, many conclusions, but I'm going to focus on 2 of them. One is, there is no doubt that modulating sudden channels is the most effective way of treating focal onset seizures, that's from that set of publication for that large study that was done and is still ongoing. And the second is the cycle of ASMs is not the best thing for these patients. When you put those things together, it's quite nice to have vormatrigine to be able to add to that, to replace many of those drugs right now and to move towards really eventually as our vision, and we're going to talk about that on POWER3 to really be the drug of choice at the end. Do we need to get there to be a multibillion-dollar drug? No, as we don't need to show many things that people oftentimes ask us to. But that is the vision, and that's the path towards that position. And then the second piece, and I'm going to get out of the soapbox here is the fact that we completely replicated that with our own data, right? We run a fairly large claims analysis about 0.5 million or so patients in the U.S. and it maps almost one to one in terms of like the use of drugs, what they stay the longest are on sodium channel blockers, what we see patients like cycling through like about 60% or so of the patients at a second drug and then a third drug and a fourth drug, and we see them cycling a lot. So it creates a nice market dynamic from a revenue perspective that you can always add another drug doesn't really matter in a sense, but a huge opportunity for a drug like vormatrigine. So we're going to talk about the results, but I wanted to set up a little bit to the market opportunities that is in front of us.

Thank you, Marcio for that. And we will use some slides that you sharedyesterday to go through some of the details of the results. But to your point, the centrality of sodium channel modulation as the cornerstone of epilepsy management was -- I mean, 81% of the patients had background therapy sodium channel blocker and actually managed to show pretty remarkable seizure reduction even on top of that. I don't know how your vormatrigine, which modulates sodium channel could even work on top of an existing sodium channel blocker because the receptor must be already occupied. So that's for another conversation, right, because -- but that was important. And I mean -- so I think -- I don't want to like run this too much, but are you going to need POWER3 to register? Or is that something that you might after submission with...

Yes. No, not at all. But I do want to take a little bit of a segue from what you just said, right? The priority, if you think about like "rationally," we're going to pretend to rational animals as humans, like adding same mechanism, right? People talk about orthogonal versus non-orthogonal mechanisms and stuff like that, it wouldn't make sense, right, that you actually have so much more efficacy out of that. But take a step back and ask the questions like, if this thing, I'm going to call vormatrigine a thing for a second, that we are putting on top of other things that are similar are not really dissimilar. It's not really different. Why would you show? Let's take, from a more basic level. So we see that as yet another proof point on how different this drug is. And therefore, its ability to continue to grow the market and to replace certain drugs and to add to other drugs because at the end of the day, it's not going to be a magic or silver bullets. This thing there is space for a lot of people and that there's a space for vormatrrigine for certain.

Got it. All right. That's -- I'm going to share the screen so we can actually discuss with some slides. This is a slide from yesterday's data presentation that describes the patient disposition in the RADIANT. 99 patients were screened, of which 61 were dosed to date, and you presented data from 37 patients yesterday. Now how many -- of the 61 patients that were dosed and remaining 24, how many of those are still focal epilepsy and how many are generalized epilepsy?

Yes. So -- and I'll be the first one to say we probably could have being a little bit more clear here. So here is the chance. So the slide is entirely about focal onset seizures, focal epilepsy, right? So the way I'm going to walk from left to right because I think it's going to be a little bit simpler. When we set up to do this, we're like -- we're going to give about 35 or so patients, right, 50x 0.7. And we saw accounting, right, 1, 2, 3, 37. That's how that 37 was defined, was the 37 completers at the time that we did the data cutoff on the 25th of July. So that is the first 37. Now, between the time that we actually said to the sites, we're pretty close like we're getting there to the 35. And when we actually stop completely recruiting for focal and I'm going to go back to generalize, there are significantly more patients on that process. Now the process takes anywhere between 28 days and 56 days to screen these patients, so you see that you can just not sometimes stop a running train. And that's on the periods that were these other patients that have been started dose. So it's anywhere between a little bit over a week to like about 4, 5 weeks that these patients would be dosed. So it was a mix of exposures, but pretty significant exposures for the other patients here on the difference between 61 and 37. That's why we included that on the safety because we thought it was very important. Of those at that point in time, 99 had completed the screening. And what do we expect is like within those and a few more that are not included here for generalized, we would have like a general update at AES later in the year, and that is the box on the left most side of this slide. where we expect with the focal like 61, 62, 63, whatever the final number is there. And the difference between that and 75, that's where would be the overall. And the difference is the generalized patients, right, the primary generalized patients. So I think the bottom line here is we're not only delivering what we said we would for focal, but really show the incredible interest for this drug and for this study. And just the ability to recruit, I think that is one of the many urban legends out there is just how hard those patients are and how difficult it is to recruit patients. And of course, it's not easy and requires a lot of attention, requires attention to quality. But it is completely doable like as we are showing here.

No. I mean so just to nail down on the details here. By -- in fourth quarter, you'll have data maybe at AES on about 75 patients that will include the 37 patients top line yesterday with longer follow-up beyond 8 weeks. And there may be more focal epilepsy patients included in the AES presentation maybe. And then...

There will be. Yes.

There will be, yes. And then generalized. So what's the breakdown of generalized versus focal when it's all said and done?

So the intent here is to present all, let's call, 75 patients at AES, of those a little bit over 60 are going to be focal onset seizures and the difference there, let's call 15 for the sake of the argument right now should be generalized.

And so basically, is it fair to assume that the 61 patients that were dosed to date, that's all FOS?

That are generalized patients, those they are not included on the 61. So all those 61 that are here are generalized -- focal, sorry.

61 patients dosed today are local, but they were not included because they haven't crossed an 8-week follow-up?

Correct. That is the only reason.

I see. So there were 37 that had 8 weeks' worth of dosing data and then 24-ish that did not -- that have less than 8 weeks?

So anywhere between 1 week and about 5 weeks.

Got you. Got you. And then subsequent to that, you started adding generalized epilepsy patients, that's up to 75?

And we're well underway there.

That's very clear. Thank you for that. All right. And then, let's get into the sort of the demographics, 81% as we discussed, were on sodium channel blocker, 65 were on -- is it SynGAP vesicle or Keppra?

Yes, yes, that's right, Keppra or brivaracetam, yes.

[indiscernible] modulators. Is this typical of -- does this represent like typical patients in the epilepsy -- focal epilepsy, what's the kind of drug they're on?

Yes, absolutely. On the -- we showed one slide on our claims work here, but it's actually a very extensive analysis, not only of like the type of drugs they're in, the duration they're in, how they cycle, what's the decision triggers and stuff like that. So I can comment a little bit about the U.S. market in detail there. So this is fairly typical, right, of the market. And normally, the vast majority of patients in the U.S. start with Keppra or with Levetiracetam and the vast majority of them fail, right? That drug has the lowest response rates of like any drug that is widely used here. So it's not unexpected in a sense. It's a very easy drug to be given to patients, and that's why it starts there. One of the reasons why it starts there and then they move normally to a sodium channel blocker on that case and they go back to other things, they come back, they add sodium channel blocker. So the steady state, if I may call that way, end up being something like this, a makeup that is similar to this.

I'm not an expert on epilepsy, but just being a biotech analyst, looking at other trials, if I were designing a trial, I would have avoided patients who were on sodium channel blocker because it seems redundant. Like you're working -- you're not a blocker, you're a modulator, okay, that's a new one. You work at the sodium channel and you have patients already on sodium channel. So I would have avoided them and actually try to like not enroll these as an inclusion criteria, but you haven't done that, and this actually represent demographics.

Yes. I think once upon a time, maybe some people still believe that there was this idea that people were responsive to a mechanism, right? And I think for some patients, it might be true for certain mechanisms. But generally, you're trying to stop the seizure. And I think there is one thing that we can say quite pretty confidently. You cannot have one if you properly modulate sodium channels, right? They're fundamental for initiation propagation of seizures in general. What we know as well, right? Yes, they are on all those things. but it's not working, not completely at least, right? You're looking to the median seizure there at the bottom, like this is very high. So this is fairly refractory in terms of refractory seizures, right, not only to a given treatment. And most of the time -- and we made this in another slide, this point, but maybe just to make a little bit more nuanced way, there is a lot of stacking here. And the stacking of different drugs is not necessarily one would call rationale, right? They're very much like empirically driven at that point in time or if the patient in front of them. And oftentimes, it is driven by whatever they can do at the moment. So when you look into a drug that actually you are very confident can stop seizures, as you've seen on the results, I'm sure you're going to go there next. You kind of have to ignore this noise about what is in the background and ask like, can we do that? Can we actually -- and of course, 10 steps ahead that is the easiest one to replace, right? Like if you're thinking about POWER3, for example, like this is a very obvious replacement, there is not a lot of other markets or a lot of drugs in development that I know of that. The easy replacement is 80% of the markets and the drug is going to be better, like I think everyone here in this call knows how to do math, you can just imagine like what kind of markets that can be.

All right. I'm getting e-mail questions, but I'll save all that for the very end because -- and we need to move fast. But just one quick point I want to ask here. Were patients on -- in this slide, were there patients who were on more than 1 sodium channel blocker?

Absolutely. Yes. There were patients in 2 sodium channel blockers, a fair bit actually about half of all the patients on sodium channel blockers were in 2. And there are even a situation patients were on 3 sodium channel blockers. And then we have, obviously, adding the vormatrigine on top of that.

Okay. I mean -- but -- if something works, I mean, if you have one sodium channel blocker that works, is there a point of having other sodium channel blockers or sort of targeting drugs on board?

Yes, I would say absolutely. I think when you talk to like the physicians who are prescribing these drugs, of course, those are real patients. There are real physicians behind us. They are not dumb, right? They are actively trying to help these patients. They're very smart actually. There are limitations on all these other drugs, right? Like we look into vormatrigine, there's certain limitations from an immunological or immune response perspective. There is so much you can do with that or lacosamides that are other limitations or even like older generation drugs like carbamazepine or ascorb, all of them, I think they are trying to do the best with the tools they have at hand. Those tools are suboptimal, but they are really well intended. They are really trying to help these patients. And for the most part, they are helping these patients. So that is -- it's not to say that there's not a better way to do it, but definitely, they're all trying to do the best here.

And one more point. So -- and we will get to this towards the end, but you talk about trying to take off drugs like to avoid AEs. How do you go about picking -- like let's say the patient is on 3 drugs, like one, Keppra, sodium channel blocker or GABA, which one do you take off? And how do you -- it seems like a very imprecise?

Yes. And we can get there. Maybe we'll move a little bit to talk about the -- that's okay. Joon. Since you said you want to manage the time, I'm going to help you manage the time.

All right. I mean, look, this is -- there's no controversy here. I mean you saw this 50% reduction in 60% of the patients, and you see a median 56.3%. And again, you have week 1, week 8 response improving over time? And 22% seizure freedom, that's really the impressive aspect about the drug, which is at the last month. You had like seizure freedom even on top of Synovo being 30% of patients. Anyway, but before I move on, like, have you -- are there any like particular response based -- have you stratified response based on background therapy, did this work better in patients not on sodium channel blocker or what sort of a channel blocker with or without -- how does that efficacy look?

Yes. So we'll go there in a minute. I'm just want to make 2 points about the previous slides, right? So one, when you look into this, these are all patients. So all 37 patients are included here on both panels, right? When you look into the overall median like 56.3%, this is like considering every single patient in the study, it does not remove any of the discontinuations that I'm sure are going to look into that. I think we got questions before is like when you look into the overall media, this number looks bigger, right? Because they point for all of that and what's happening, obviously, we always get a little bit impacted by discontinuations and things like that. The same thing with the more than 50% response on the right-hand side, this is the entire cohort like every single patient counter their response. And you look into those 2 together with what you're looking to the baseline characteristics before, this is very hard, right? These are very hard results to be seen on these treated population. So I just wanted to make that point because I think it's quite important to put in context. The patients are capturing their seizures every single day. And then when you put this all together and having such a robust reduction here, it's very, very clearly a drug and is very clearly derisk of the program as we move forward. But maybe to answer your question about the background, we did a few cuts that we thought was interesting to explain that, right? So we're seeing this very high response rates. We look into patients without any seizures at the last 28 days, we're like asking the question, okay, what is driving this? What is driving this high? And there are a few things that are kind of obvious questions. And if we could go to the next slide, it seems like you have the entire deck there. I think the first question one can ask and has been debated a lot in epilepsy is what is the influence of where they start, right? So they come into this study, they are treated with a lot of antiseizure medications. They have this disease oftentimes for decades, they failed tons of other drugs before and they are seizing a lot. Is the drug working on people who are seizing very little and not working at all on the ones that are seizing a lot at baseline, of course, on the 28 days of baseline or is that is a general impact. And I don't think this is very controversial here, there's always a little bit more numerically response on the lower patients, right, just by definition. I think it's easy to understand. But it is a question, right? Like what happens when you split this group in the middle. No one finds the right like perfect time points, like it's not like if you pick like 8 here and like I split the group on that like when you split by the median, it's nice to see they split like being so homogeneous. If you go to the next one, I think that's where it answers part of your questions, actually. One thing a priority, we did not expect, Joon, was the amount of patients actually being on a stable dose, high dose, right, to give you a little bit of a view. I think when you actually ask a lot of physicians when you look into the claims data, the dose of cenobamate in the U.S. is normally between 200 to 300 milligrams per day. Very few people go higher than that. Here, the mean was actually 300. It had a significant portion of patients with more than 300, 400, even 450 milligrams per day. This is a very unique opportunity that we had because no one so far developed like a drug that had a result after cenobamate became so important in the market. And we know how effective the drug is, right? So one could -- if you asked me this before, I was like, what is your expectation for how well this is going to work on top of cenobamate, I would probably have said not that high, right? This drug is very effective. There's probably a lot of -- there is not a lot of residual like ability to help there. But if you flip to the next one, and then I'm really going to answer your question since I've been lingering here on answering your question. And you got the 3 most common like ASMs used here, right, like sodium channel blockers in general, and that might be 2, might be 3, of course, includes cenobamate as well, Levetiracetam and Brivaracetam, which is a very prevalent one is on this cohort and cenobamate by itself. So it's very effective across the board. One could argue is like if you're going to remove something a priority what would you remove, of course, this the discussion to be held with the PI and with the doctor.

Isn't that an easy question here because cenobamate is the riskiest drug. So wouldn't you remove cenobamate first? Or I don't know. Isn't that?

I think that is a fairly reasonable way to think about that. But as you can see, there are multiple ways, but it may be the most fundamental questions, do you actually need to remove something? If you're talking about POWER2, and the answer is probably not always or most of the time, not yes, then when you go to POWER3, then yes, we're looking for switch to monotherapy. So we're going to have to reduce things and it might be an order of the reduction.

Got it. Marcio, before we move on, I just want to better understand. So there were some patients who discontinued, but these data represent all like 37 patients who were included in the analysis. Now 23% discontinued which is not out of there. We can talk about that. Why don't we actually go there? So this is a paper from Xenon's publication where in the 25 milligrams that they're taking forward, 114 were enrolled, of which 26 did not complete for whatever reason, about 23% discontinued. So it's not out of the ordinary. And they also had to deal with some imputation of data that we're missing based on the discontinuations. I don't know, and this is what they showed, which kind of looks similar to what you show here in terms of the fluctuations here and there. So there is definitely some noise there, but how were the missing data imputed from the patients who discontinued?

Yes. So epilepsy has this interesting -- if you read slightly statistical analysis plans and the reviews, they say there's no imputation, right? But technically, it's not exactly right because you use the rates of the patients at that given point in time to use as the terminal rates or to -- and that is equivalent to imputing the average. So -- and basically, the way it works is, let's say, a patient had 2 weeks and they had whatever 10 seizures, right? So we're going to do like 14 days on that week, then seizures calculate a rate per day multiplied by 28 on that one is pretty easy, it's 20 and that is the -- basically the rates and you compare with the previous 28 days. And that's how you calculate and then you carry that forward. I think some people say that's a last observation carry forward, is really not, right? It's not the last time you observed is the overall rates carrying forward. So all of that is included. It tends to penalize the median. It did here. As you can see, it's probably the best way to look into the overall impact, but what we know from treating physicians, they are not treating the discontinued patients, they are treating the continued patients. So there's always an interest on the overall impacts, which is what you have in this slide, but also what is impact in general when I keep treating these patients.

But Marcio, because the patients gained the most seizure reduction in the first week or 2, as shown here, if you actually have a patient who discontinued at weeks 3 or 4, and you're using -- applying the rate of seizure reduction of week 1 and 2 out to week 8 or so, wouldn't that actually favor greater -- unfairly favor greater seizure reduction?

Yes, I wish, right? So what you are seeing here is the median per week. So in that given week who is having the response and the median for that week, not the overall. So you have on any given week, people who respond really well, people that don't respond so well, right? If those are the ones who drop as it tends to be, right, you actually don't influence as much on the weekly basis, but you influence a lot on the final.

It depends on who drops out. And is it fair to assume that people who discontinue are -- they're self-selecting in a way, they drop off because it doesn't work as well rather than continuing it really well?

Yes. And that is -- and it should have been obvious in the chart but maybe did a poor job on the scale. There is a fairly substantial reduction between the first periods and the second periods, right? By being first periods or second periods, is the first month on drug in the second month on drug. So in a sense, the worst thing is actually towards the end to up for discontinuation to happen or the best thing for the clinical imputation is actually the worst thing towards the beginning.

And is it fair to assume that how you analyze the data here in terms of imputation of discontinuation, patient data from discontinued patients are no different than how the [indiscernible]?

It's actually identical.

So it's an industry standard?

Yes.

Nothing suspicious there. Okay. Good. Good enough. The other thing I want to touch on is, let's say -- yes, I think that we're going to end up spending quite a bit of time here. And I definitely want to leave 5, 10 minutes at the end for questions and also to go over POWER1 and 2 and 3. But 23 patients -- 23% discontinued, which again, is not out of the ordinary versus Xenon's trial and cenobamate trial in hindsight. But you also mentioned in both this like discontinuation or reducing the background medication should improve adherence or -- how are you going to do this -- I know that you're -- maybe you're encouraging investigators to say, you know what, some of the background medications you don't need or not working why bother and just -- how is that going to work in POWER1, 2 and 3 in a blinded trial, investigators don't know if the patient is honored not on very much. So would be more hesitant to peel off the drugs?

Fair. No, absolutely fair. A couple of points there, even before we jump into the POWER1, 2 and 3 studies, right? You're right, like 23% is I would say, similar in general to what we're seeing in other epilepsy studies, nothing to be alarmed there. We do think we can do better, right? And I think that's why we're talking about that. Do we need to do better? Maybe that's the first question. And I think we always have to try to do better for patients, but let's take that aside for a second. And the answer is no, right? So if that was it. This is going to be a competitive drug. This is going to be a drug that's going to get a very significant market share in a competitive setting and can move towards actually being a dominant player when you remove sodium channel blockers or other drugs in the market. So like if nothing else, all things being equal, that is totally fine. I think the observation we wanted to give here is because that option was on the table for the investigators. They're like -- and some of them took and some didn't, right? So we're saying, one they did talk. There are a couple of things that are from a learning perspective, this is a learning study, right? We said from the beginning, we want to learn a number of things. The very first learning is does the efficacy deteriorates? And that is the most important question. I don't get asked in these slides, but you reduce the dose of the background. If the background is what it was keeping things together, for these patients. And I can say quite definitively, no. Those patients responded really well. So that is like you can entirely attribute that response to vormatrigine and that's quite important, right, as we move the drug forward, then the implementation on POWER1, and it might be a little bit too late on POWER1 for an implementation. I'll talk about that in a second. But it doesn't seem to be needed either, right? So a couple of things from a POWER1 perspective, like going pretty well, we're not done yet, but we're pretty happy with where we are. And what you've seen so far if we use like the worst case for placebo, like meaning no one discontinued and -- or the best case in the worst case for drug and like the majority, we are not seeing the levels of discontinuation you're seeing here. So I think we feel that even the training that was reimplemented and really the counseling of the patients and a lot of that was sufficient for -- to minimize some of those, because the majority of the AEs were really fairly quick transient, mild in nature. So is that a big deal when we have like, again, in the grand scheme of things are fairly low, and I know 59% does not look low, but it's comparatively fairly low. Treatment emerging AEs and we can actually manage. I don't think that we are in a territory that it is an issue here. So we are starting with this is not a problem and then we are moving to can we make it better. And I think that that's quite important as we discuss these aspects, right? We're trying to make a problem out of something is not a problem. And -- but just maybe because we said we can do better. But our intent is always to do better. That's why we spend time here.

What percentage of the 23% discontinuation was due to AEs or some other reason?

Yes. So we were very good about following up with the patients and we think that scatters and understanding was there an AE associated with, right? And one could argue true or false, because one patient discontinue and you go and ask them afterwards like I wanted to discontinue and you record that as patient choice, you avoid the terminology of the AE, but you first say you had an AE discontinue. You normally record the AE. So the vast majority of the patients, said, well, I had an AE and I discontinue. And that's what I believe happens in most of these studies, but that's the case here.

I mean, I guess you did some of the work for us and in comparing these drugs -- and yes, I mean, look, 59% or was it 52% PAE drug-related. It doesn't look at that bad relative to other drugs...

Well, I'm going to take an issue with that. Like if 30% or 40% less patients are having treatment emerging AEs, that is a lot better. We can talk about being comparatively everywhere else, but I'll take a problem here because it's not only the rates when they have it, that's how AE sometimes are calculated, but are they having treatment emerging AEs at all? And 60 versus 90 versus 85 is our very different numbers. You're talking about a significant proportion of patients not experiencing a treatment emerging. And for the vision of the drug, for the get to the markets, be competitive, get a good share might not be so important. But in the long run, it is quite important for where we want this drug to be.

And I mean, so when you unveiled the plans for POWER3 yesterday for the firs time in an effort to study vormatrigine as a standalone or monotherapy, right? Same thing standalone monotherapy, while they get washed out of the background therapy. People thought -- I mean you could have a bearish and bullish interpretation of this, right? The bullish is that you want to study this as a monotherapy potentially positioning it as a more frontline therapy ahead of other sodium channel blockers. The fairest argument is that there must be some drug-drug interaction that you're trying to avoid as a monotherapy. When did you when did you decide on the POWER3, tell us a little bit of the history and conversations that went on with the investigators and internally in decision to do POWER3?

Yes, I love the conspiracy theories, right? In the end slides, that is like the GGI profile for all these drugs. And by the way, we combine the drug with all these orders, like that. There is no one is trying to avoid the GGI here on the very concept of GGI. That has nothing to do with that. But moving to what it has to do with, right? If you think about the markets or go back to the market with epilepsy, right? Getting off the gates, we're going to add another drug for refractory patients, right? POWER1 and POWER2 are going to position us to do that. The market is $1 billion to $2 billion. And if there are 5 other drugs that are all going to be $1 billion to $5 billion to $2 billion, like that is not a winner take all market, right? That is like a market that people cycle, there are several patients there. And one could argue that is enough and sufficient and great in a win to actually be in a position to derisk that position. But when we take a step back and ask like where does this drug can go? Then we believe strongly that it can keep moving lower in the adoption curve. And the best way to show that is to actually remove some of those drugs. Now do we need to remove it? I think we answered that before, as no, as a condition to be successful in the market. But should we remove it or should we test it and make sure we can remove it. So we can move faster towards first line and then actually replace all the drugs that they shouldn't be in eventually and have a market that, hopefully, at one point, is going to be 2 or 3 drugs, that are used in general instead of like this, in a sense, madness that happens with all these drugs right now. So this is a conversion to monotherapy, right, where patients got reduced sequentially. To answer your question, what do you take first. They're going to be 1 or 2. You can choose the 1 or the 2 and you're going to take that and you're going to introduce vormatrigine at the same time for these patients with the expectation that they're not going to lose ground and they're going to gain response. Many studies like this were done before. I think some of them are fairly old and some of them use like subclinical dose of other drugs. I think those are all deemed to be unethical to be run because like -- or using subclinical dose for other therapies and things like that. Some of them use like scape rates, during that conversion and so on. I think we're all still having some debate on how to do it. I think our strong preference is really to try to keep these patients as stable as possible, of course, as safe as possible. Some of those escape rates in the past were very aggressive, not towards the patient like 2 or 3x increase in seizures. We don't think that, that's fair to patients at this day and age. But we're going to get back to the design of this once we finalize -- we're going to start this next year. Maybe the other part to be excited is actually the POWER2, one that we are starting clinical now by the end of this quarter, where we had a key question for RADIANT. Going back to the things that we intended to answer if RADIANT, right? So was 30 milligrams effective dose. Can we dose that off the gates like without any titration, without any adjustments there? And I think we checked those boxes, right? Or should all agree we check those boxes. But in a secondary question there was, do we have a reasonable evidence that if we give more drug on board, that the efficacy could be better, one, so from a response rate perspective versus like exposure. And then on the other side, is the side effect profile worse if we increase the drug because if they are, then you got to be very careful. But if they are not, right, if the efficacy is higher and the side effect profile is equal or lower, then you know it's almost you have to go there. Because like you have like a base quite strong, solid foundation of the 30, but you can explore these new highs with the 40 and that's how we picked the 40 here, basically just got the rationale for you there. Now on POWER2, we can more aggressively, I'll call, they were aggressively not like with aggression, but with being incisive here to reduce the background drugs by a bit by 20% or so in the first week. So our levels are fairly safe but can actually keep all these patients responding better and managing them better. And that's the answer to your other question, how do you manage on POWER2. POWER1, too late, but we're pretty confident that all the measures already in place, we're able to manage, we be able to -- we're able to manage, and I think POWER2 is going to be even more effective.

Going back to POWER1. What's -- I got a lot of questions around the timing of POWER1. In some slides, you say you will finalize the POWER1 in 4Q. On other slides, I think you've said you'll report in second half of the year. What's the status there?

Yes. So last mile for POWER1, right? And when you are at the last mile of something, you always like to start asking the question is like, okay, when does it end to get like the actual race, and you got like marked up the time and that's when last patient last visits. But then like when do you get the matte and that requires a lot of other stuff to happen. And when you look into that, it is the end of the year. And being the end of the year we though it was prudent to just like give that at the end of the year view right now, and that's the disclosure we have right now. There is no -- I know there were some conspiracy theories about that as well. and there is nothing there. It's just like a fast-moving study. Every fast-moving study has thousands of pieces that are moving and no -- or very confident it's like we're very happy with the number of cases that are being referred every single day to sites on the medical review, on the quality of the patients, like on everything else, just like we were on RADIANT. And here we go, not only deliver on the 35 plus that we said we would. But may more there. Almost every single patient between 35 and the 60 or your single RADIANT would qualify for POWER1. So we got to ask the question, like we got to be fairly confident to be able to keep them on RADIANT when you actually didn't need that instead of like actually funneling them to POWER1.

So fair to assume that the baseline characteristics in POWER1 will be very similar to what we saw in RADIANT?

Very, very similar. Yes.

All right. And the last question before I take some of the investor questions, EIC -- IEC on August 31. What do you think will be -- what are you going to share or to the extent you can share, what will it be?

Yes, will tell a little bit. So of course, a little bit more about everything we talk today, like that, obviously, a scientific meeting. There is a lot more -- a little bit of the details here and there. We do have a few new analysis that we conducted like, for example, one that we're very excited about is by seizure type. So really showing that, as you know, there are 4 major seizure types that we actually measure. As we go through 3, of them get part of the endpoint. One of them is normally not part of the endpoints, but really like showing that, I think it's quite interesting. It's quite interesting in general, but it's incredibly interesting from the impact you're having like that across the board that seizure type is like what about the ones that are like focal to bilateral, right, generalized type of seizures and without giving too much and saying we're excited about that as well. And a few other things here and there. Of course, we need to be careful scientifically and otherwise where IC had a purpose. AES has a different one. We submitted the full or the expectation to have the full data sets to AES over the weekend, and that's where we expect to be the entire wrap up for all the patients.

Yes. All right. I don't even know where to start with the questions. What evidence do you have that going to 40 is going to lead to better efficacy? What kind of internal -- what exposure data do you have that supports going to 40 milligrams?

Yes. So again, going back several months ago when people would ask, why are we doing Radiant? Then one thing we said is conducting intense PK sampling is easier in a study that you know patients are on drug. It's easier to ask for a patient to stay like a full day, taking samples in another top line study. So that was one. And with that, we can not only understand better the PK of the drug itself in patients, not in health volunteers. And second, we can use that data to fit to a PK modeling to do a lot of exposure response to explore covariant matrices and things like that. So starting with the first one. The very first question is the exposure different, right, in how volunteers in patients with epilepsy. It answers indirectly and directly your question, is there a significant CDI hidden here in plant sites that we actually haven't detected that is influencing the PK of this drug. And I can tell you quite categorically, the answer is no. The exposures are very, very comparable between healthy volunteers. And that is very important because, of course, we know the drug is behaving well. But second is we have safety on 45 in how volunteers, and we can borrow that experience to predict safety. And as you said, about POWER2 and how we select that. And then the second question is the one I answered before, but just to reiterate, when you actually use now the exposure response parts help this modeling. And you asked the question, are higher exposures, and I would say preliminarily because for people familiar with these models, that it takes a while for them to be finalized. Preliminarily, I can say that the base is positive, meaning you give more drug, you see more response for some pains, as it's not something exaggerated. It's not something by model, what you see like no response and then high response, but there is some hints there that we could benefit, number one, and you're not harming number 2, which is quite important. So when we put that all together and the fact that the FDA told us they would kind of like for us to explore higher doses. I think that's seals the deal.

You had a pre-Phase III meeting with the FDA and you brought up the POWER2 dosing up to 40 and they were, okay?

Yes. So we had an interaction with the agency where we basically expose the entire program and we like, this is what we're trying to accomplish here with this drug, right? POWER1, POWER2 ranging from exposure to GGI like Phase I studies, you name it. And I think the feedback at that point in time, we're under the impression that they would want us to go lower to like 10 milligrams. And you might -- if you go back to previous disclosures, you even added that 10 milligrams, like we're so convinced they're going to tell us to go there. They were like, maybe we should go there. And I think the feedback was quite different. So like if you really believe on the profile of the drug that we were describing to us, right? It might be warrant to actually go higher. And I think we took that to heart, and here we are on a point that all scientific boxes are tracked towards going to 40 milligrams because we can expect potentially, of course, we need to run the experiments a higher yes then, like POWER1 should be very solid. We expect POWER1 should be very solid efficacy and safety. But if POWER2 replicates that efficacy on the 20 and 30 milligrams and actually gets higher on 40 gives a lot of flexibility for physicians to use this drug on how to treat their patients.

And there were on average 2 background medications and ASMs. How many -- is there data on how many they failed?

Yes, very substantial number of ASMs they failed. The -- I would say the worst case here was 22 or 23 ASMs patients in general, like 5, 6, 7, 10 was not uncommon on this data set. In general, we would see 4 or 5 ASMs being failed throughout like that. And that is even the ones that I would argue the patients recall and the physicians have on their records. You tend to have a limit for like 10 years of medical records in most patients in the U.S. and the majority of the patients have more than 10 years of epilepsy treatments.

Got it. And then someone was asking, you mentioned something about the interim dropout rate in POWER1 being lower than that in RADIANT. What does it look like on a blind basis so far in terms of dropouts and discontinuations in POWER1?

Always tricky to talk about ongoing studies. So I'll please take with a grain of salt. But I think what we are seeing so far is lower. And it can be for many reasons, right? And it's hard to speculate. I think one of them is for an open-label study as RADIANT was or is, I guess, still ongoing for those other patients, have something like I have a headache, I'm going to call -- you attribute immediately to the drug and you're like, I'm not sure I want to stay with this. I think when you are in a double-blind study, and you have headaches before and dizziness before and somnolence before because these patients lived with all these other drugs, it's not so easily attributable. And then, so there might be one. I think the other one is there were somewhat of our relationship with a few sites where they have higher dropout rates than others. It so happens that they were not sites, had a lot more patients to offer to clinical studies. So they're not really contributing as much or contributed as much for POWER1 proportionally which allows for actually the sites, we didn't have any. We had bigger -- to be a bigger contributor. And then the last one is training. Price is just like going back and on the counseling on the ICF and everything else being better. So I can't possibly disentangle which one of those parameters actually is influencing the most certainly, all of them are.

And how many of the 37 patients actually had their background ASMs reduced?

So 6 of those patients had their background reduce. None of them got completely removed during the 8 weeks period, which is a question I got as well the other or yesterday. The -- it was a reduction at the majority of the time is like 30% to 50%, the reduction they had, and dosing was completely investigator-driven, right? They would think it's a good idea and they would reduce it.

Okay. There are so many questions...

Pick the tough ones.

Somebody is asking, should we still expect essential tremor update in the fall.

Oh gosh, okay. So you really managed to pick a tough one. Yes. So we just updated the disclosure yesterday to talk about like we finalized since we went there and seems there is some general interest. I'll always start by saying we ask everyone not to attribute any value to essential tremor right now to take it out of their models. Having said that, we have completed all the explorations. We wanted to complete the modeling and so on, on the blinded data for Study 1, and Study 2. We believe that we found like some small chains that would quite significantly impact the result of the analysis, but not the integrity of the analysis, which is always what we are aiming for. Here we've made those changes to the statistical analysis plan. In the process of doing that, we enrolled a few more patients. So we wanted to get the studies to rolling before we actually finalized documents. That's all said and done. So in the past, we said the fall because it's really like the kind of the very end of September to like the beginning of October. And someone joked with me yesterday. It goes all the way to November is that we're just talking about and like no, that's no what you're talking about is really that interface between the end of September, beginning of October.

Great. And then do you have Xenon's trial allows cenobamate as a background therapy or not?

Yes. very unclear I think by the public protocols, I believe so. I would imagine so it would be interesting, if not, in the most American way, I can say the word interesting. But the -- I would imagine like cenobamate prescriptions grew about 12, 13 fold since the last study was conducted by then. So there is a lot of those sites that you end up going to have a lot of cenobamate use. as you've seen, was over 30% in our study. So I'd imagine is the same on theirs and anyone else studies.

I think we're past the hour, and we asked a lot of questions. And the questions keep coming in, maybe I can ask this one last question. What was the seizure free [indiscernible] first 28 days?

I'm not going anywhere. Yes. So the seizure free -- and of course, this is stuff that we are holding for our IC and AES. But I'll give a little hint, right? The seizure free rate throughout the study, both first 20 entire study, last 20 was all high double digits. So that is nothing we...

[indiscernible] seizure freedom rate.

Yes.

Interesting. All right. The questions that keep coming in. But I think we can also do another one of these and -- but it's very interesting data. I think the seizure freedom rates efficacy on top of sodium channel blockers, multiple sometimes, says a lot about the drug's profile. It's just about designing a trial design that shows up the true power of this vormatrigine. So we look forward to the presentations on August 31 and further disclosures in 4Q.

Likewise. Thank you so much for the time.

Absolutely. Thanks, everyone, for joining. Goodbye.

Thank you. Bye.