Idorsia Ltd (IDIA) Earnings Call Transcript & Summary

Dear ladies and gentlemen, welcome to the Idorsia conference call. At our customers' request, this conference will be recorded. [Operator Instructions] May I now hand you over to Andrew Weiss, who will lead you through this conference. Please go ahead, sir.

Thank you, Kai. Good morning, good afternoon, everyone. This is Andrew Weiss, Head of Investor Relations and Corporate Communications here at Idorsia. I welcome you to today's call to discuss the second set of Phase III results for daridorexant in insomnia. In the call with me today are our CEO, Jean-Paul Clozel; and our Global Head of Clinical Development, Guy Braunstein. Both are here to walk you through our presentation and prepared remarks, which will then be followed up with a Q&A session. In that session, Martine Clozel will be able to also take questions. Next slide, please. I do need to remind everybody that we will be making forward-looking statements. Therefore, you have been properly warned about the risks and opportunities of being invested in Idorsia. Next slide, please. So hence, I have the pleasure to hand over the call now to Guy. Please take over.

Thank you, Andrew, and good morning, good afternoon, everybody. This is a presentation of the second pivotal study of daridorexant, and it is a follow-up discussion, as I want to remind you that the first study was -- results were shown in April, 2 months ago, a bit more than 2 months ago. And I want to refer you also to the presentation that we delivered at that time. Of course, as we now have the 2 studies, it will be interesting to look at the data side-by-side, which will be done in the course of this presentation, which is leading me to the main conclusion that we draw from the second study, that we now have 2 positive clinical trials of daridorexant in insomnia and also the remarkable consistency of the results in the -- between the 2 studies. Next slide, please. We must be now on Slide 4. So this is a summary slide of the overall program that we have. We completed a while back, a Phase II program with 2 Phase II studies, one in adults and one in elderly subjects. That was important because we could define the dose level that we wanted to study in the Phase III program. And it was also interesting to note from this Phase II program that the same dose could be administered to adult and elderly patients. As a consequence, we could set up 2 pivotal studies looking at 3 different dose levels. The first study, called 301, that we reported in April looked at the dose of 25 milligrams and 50 milligrams. And the second study that we're having to describe today looked at the dose of 10 and 25 milligrams of daridorexant. In both studies, we included adult and elderly subjects. The studies had a 3-month duration and included patients with moderate and severe insomnia. The efficacy parameters, safety parameters were identical between the 2 studies. We had objective and subjective sleep parameters collected by polysomnography. I described them in details in the previous presentation, and I would like again to refer you to that presentation for details, and we can discuss that in the Q&A if you need. But I also want to mention that we had a range of subjective end points assessing the night, using what we call the LDQ (sic) [ SDQ ], the sleep diary questionnaire, which allows the patient to report on their feeling about their night, and in particular, the total sleep time that they perceive [ they had slept ]. We also had an [ opportunity ] to this program, an assessment by the patients themselves of the daytime functioning using an instrument that is called the IDSIQ, Insomnia Daytime Symptoms and Impacts Questionnaire. This is a questionnaire that we developed at Idorsia and validated according to the FDA guidance. And again, I will refer you to the previous presentation for details, and we can discuss this instrument during the Q&A. As mentioned, we have now replication of the 2 confirmatory studies. The safety assessment included, of course, adverse events and a number of other parameters like vital signs, biochemistry and hematology. But more important and more specific to insomnia products, we also collected, using a specific tool that I will come back on later on, the [ next modern regulatory cycle ]. And we also collected the withdrawal and physical dependence and rebound of insomnia after withdrawal of the drug during a 1-week observation. In addition to this program, of course, we have a clinical pharmacology program, which is not totally completed, and we will report later on, on this program in a separate call. The next slide, Slide 5, is showing the design of the trial, which you may remember is identical to the Phase III 301 study that we reported earlier on. Just to remind you, there was a screening period of up to a month. The second part of the screening period was actually a single-line placebo running phase. And during that placebo running phase, we collected subjective assessments using the sleep diary questionnaire, assessment of sleep as well as polysomnography, which is represented at visit 3 by the 2 bubbles. So we collected the polysomnographic criteria over 2 nights, consecutive nights, allowing a clear assessment of the baseline. Following this placebo run-in phase, the patients were then randomized at visit 4 into 3 groups: placebo, 10 milligrams and 25 milligrams of daridorexant. This was given to patients for 3 months. And we had at 1 month and at 3 months assessment of polysomnography again through 2 consecutive nights. During this whole treatment period, we continued to collect daily assessments of sleep as well as the daytime functioning using the IDSIQ questionnaire that I mentioned before. After the 3 months, when the patient had completed 3 months, they entered then into a single-blind placebo run-out phase of 1 week, during which we continued to collect daily assessments of sleep as well as the daytime functioning, and we also included one placebo night to look at the withdrawal symptoms and rebounds during this period. Finally, at the end of this single-blind placebo period, the patient entered into an extension study, if they wanted. The next slide is showing the study objectives. The primary objective was to evaluate the efficacy at 25 and 10 milligrams of daridorexant on objective sleep parameters collected by the polysomnography in patients with moderate and severe insomnia. The secondary objective was to evaluate the efficacy of 25- and 10-milligram daridorexant in objective sleep parameters collected with the SDQ, the sleep diary questionnaire, as well as the daytime functioning collected with IDSIQ questionnaire in patients with moderate and severe insomnia. And the third objective, the safety objective was, of course, very important to assess safety and tolerability of the daridorexant, not only during the treatment phase but also upon treatment discontinuation to look at rebounds and withdrawal. On the next slide, Slide 7, we have now the description of the primary and secondary end points and how they were analyzed. 2 primary end points were defined. These were nighttime primary end points, looking at the wakening after sleep onset by polysomnography -- we call that WASO -- and also the latency to persistence sleep, LPS, also collected by polysomnography. We had 2 secondary end points, one dealing with the night aspect, which is the subjective total sleep time recorded with the sleep diary questionnaire. And we had the sleepiness score of the IDSIQ questionnaire, which is assessing how patients function during the day. These subjective end points, subjective total sleep time and the sleepiness score were recorded every day on average over the treatment period and every week. The end point was the last week just preceding the polysomnography nights. We analyzed these end points according to the 10-milligram versus placebo dose as well as the 25-milligram versus placebo, and we analyzed these end points at month 1 and month 3. If you make the calculation, you will realize that here, we had 16 comparisons versus placebo. And of course, we wanted to have a study-wide Type 1 error controlled at 0.05, and this is important in the context of making labeling plans. There has been a lot of questions when I presented that slide 2 months ago, a similar slide for the 301 study. There has been a lot of questions about the structure of these 16 comparisons. And so far, we were quiet about it, but it's probably time now to disclose and discuss more fully how this was organized. This is shown on the next slide, Slide 8, which depict graphically how we organized the 16 comparisons. So I will try to explain that in as simple terms as I can. You see here the 16-page hypothesis comparison, H1, 2, 3, 4, 5, 6, 7, 8 on the left and then 9, 10 11, 12, 13, 14, 15 and 16. The H1 and H2 were WASO comparisons and LPS comparison, respectively. The green diagram shows the high dose versus placebo. For that particular study, this was 25 milligrams placebo at month 1. The blue following the green shows the high-dose comparison, 25-milligram versus placebo at month 3. And then moving on to the right side of the slide, we have this pink showing the low dose, 10 milligrams placebo at month 1. And the purple part, which shows the low dose, 10 milligrams placebo at month 3. So essentially, what we did was to prioritize 25 milligrams above 10 milligrams and then within a certain dose level to prioritize month 1 to month 3. And therefore, the order of comparison was high dose month 1 followed by high dose month 3, followed by low dose month 1, followed by low dose month 3. And within each of these buckets, high dose month 1 -- or high dose month 3, low dose month 1, low dose month 3, then we organized the comparison of the primary and secondary end points in a certain way. We started with WASO and LPS, followed by H3 sTST, the subjective total sleep time, followed by the IDSIQ questionnaire, H4. And similarly, H5, 6, 7 and 8 for the low dose -- or, the high dose versus placebo at month 3. And then again, H 9, 10, 11, 12 for the low dose versus placebo at month 1. H13, 14, 15 and 16 for the low dose versus placebo at month 3. What I would like to point out as well is that we started with the alpha value Type 1 error of 0.05. And we carry on this alpha level according to how the statistical response is to the hypothesis. So initially, we divided alpha into 2 and give identical weight to WASO and LPS. So therefore, the H1 and H2 are tested statistically at 50% of the alpha starting with 0.05. We divide into 2, so H1 and H2 are tested at 0.025. If we can reject the hypothesis 1, then we can carry on the alpha level to the next level down. And to do that, we divided the alpha level that is remaining. So for example, if you win on H1, we can carry on 0.025, we don't lose any alpha, and we carry on to H5 and H3, similarly. So therefore, we divide the alpha of 0.025 into 2 identical levels, 0.0125, that goes to H3, and similarly the second one goes go H5. If we also win on H2, we proceed in the same way, carry on half of the alpha remaining, which is 0.025. Half of it going to H3 and half of it going to H6. By doing that, if we win on LPS and WASO, the alpha of -- is received from WASO and LPS and therefore, receiving half of each will again have a full alpha at that level. In contrast, assuming that we file on H2, we don't reject H2 LPS, as an example, then we lose the alpha that goes together with this comparison. So the alpha of 0.025, which was associated with H2, is lost. However, as you see from the network of arrows, we can still continue to do the analysis. The H3 can still be tested because half of the alpha from H1 is retained, and then H3 can be tested at half of the alpha that was given to H1. If we wean on H3 this time, we can also continue to H4. But interestingly, we can also continue to H6. Even if we lose, for example, on H2 LPS at month 1, we can still go to H6 and test the LPS at 3 months. So this is the methodology we have used to test the 16 hypotheses. This has been fully discussed with the FDA and agreed by the FDA. The last remark I would like to provide, which is important for this call as well, is that we don't need to win on more than 1 of the 2, H1 and H2, to claim significance because they are tested independently. We are not in a situation of co-primary end points, where you have to win on both of them to claim positive study. Here, by dividing the alpha from the very beginning, we can claim positive if H1 is rejected or if H2 is rejected. And this is important in the context of the current discussion, and this is what allows me also to claim that the 302 study that we discuss today is a positive trial. I hope that was clear. It's a bit technical and complicated, but because I've received questions on many occasions about this control of Type 1 error for the 16 comparisons, I wanted to give you a bit more clarity on how it was done. So the next slide, Slide 9, is now showing very classically the study patient disposition. As you see that we had to screen a large number of patients, more than 3,600, nearly 3,700 to get 924 randomized. 25% of the patients were randomized. 75% didn't pass the criteria that we set forth in terms of randomization. This was primarily due to the fact that the patient recorded not sufficient insomnia on the subjective diary questionnaire and also for the polysomnographic criteria. We wanted to be sure that we had very clear criteria for moderate and severe insomnia and therefore, a number of patients had milder disease and therefore, couldn't be randomized. So now continuing from the randomization, what we see, which is really very nice is that we have a very high completion rate, 93%, very few patients dropped out. And then the follow-up was the placebo run-out. Also, we have nearly complete data on the withdrawal during the run-out, withdrawal of treatment. And then nearly half of the patients that were randomized actually could continue into the blind and extension. The blind and extension, which is called the 303 study, is still ongoing, and I'm not in a position to report the data from that study now, but this will be the purpose of follow-up conversations. I'm sure you would like to see the results of the study, and that will be shown on the next slide, Slide 10. But first, I would like to describe how the slide is constructed. On the left, you see the results of the first pivotal study, similar to what we presented at the previous teleconference. On the right, you see the results of the new study. On the left, you see the results in green of the 25 milligram versus placebo. And as a reminder, we see that there was significant improvement on both LPS and WASO at month 1 as well as LPS and WASO at month 3. And you'll remember as well that the 25 milligram versus placebo was significantly improved for LPS and WASO. It was positive on these 2 end points as well as at 3 months for LPS and WASO. This is known, and you have seen that before. The new study provides very similar results. The 25 milligrams versus placebo shows numerical improvement of LPS, which was of borderline significance, the same at 1 month and 3 months. And for WASO, it was highly significant at 1 month and 3 months. Interestingly, the 10 milligram versus placebo showed numerical improvement on all these parameters actually not reaching statistical significance. And when we look at the numbers, which I'm not allowed to give you because we would like to preserve the scientific integrity of publications, but I can tell you that the 25 milligram was numerically always superior to the 10 milligrams. So we see an effect at 1 month. It was sustained at 3 months. And there is a high level of consistency. You may question why we don't have numerical improvements, but no statical improvement of LPS. I can't give you again the numbers, but I can tell you that numerically, they're actually very close to the 25-milligram data that we had in the first study. The next slide is showing -- Slide 11 is showing the results for the subjective TST. And as we see here, on the left, we have the positive data for 50 milligram versus placebo at month 1 and month 3 for subjective total sleep time and similarly for 25 milligrams versus placebo. In the new study, the 302 study, we see on 25 milligrams versus placebo significant improvement at 1 month and at 3 months on sTST. And again, for the 10 milligrams versus placebo, there was numerical improvement of a lower magnitude compared to 25, and it, well, didn't reach statistical significance. At this point, I would like to remind you what I said about the structure of the alpha [ scaling ] function so that you understand that, even if we don't reject the hypothesis for LPS, we could still test the subjective total sleep time, although it was a secondary end point. So again, the effect that we observed at 1 month was sustained at 3 months. The next slide, Slide 11, will show the results of the daytime functioning using the IDSIQ questionnaire and the sleepiness domain.

Guy, you're on Slide 12. Just that you said 11, sorry.

Slide 12. Thank you, Andrew, to remind me that. On Slide 12, we see the IDSIQ sleepiness domain. This is an assessment of the patient daytime functioning. The results were significant for the 50 milligrams versus placebo at 1 month and 3 months and numerically better than placebo at 1 month and 3 months for 25 milligrams. We have very similar results in the second study, with numerical improvement versus placebo and 25 milligrams, actually larger than on 10-milligram and not reaching statistical significance. So again, we see the consistency of the studies and the effect that we have sustained at 1 month -- from 1 month to 3 months. So this is in a nutshell the result of the efficacy data that we collected during these 2 studies. I will now move to the safety data, Slide 13. And this is an overview of the treatment-emergent adverse events. We see that there was no huge difference between – okay, again, showing the first study on the left and the second study on the right. We see that the number of adverse events, we don't see much of a difference between placebo and the active doses of daridorexant. We see no tendency to a dose response between the different doses. We see very, very few patients that had adverse events that lead to premature discontinuation. We see very, very few, a handful number of serious adverse events and also a very small number of adverse events of special interest, which were adverse events adjudicated blindly by an independent adjudication committee. Slide 14, the next slide, is showing a bit more details on the type of adverse events. We see high levels of similarity between the 2 studies. The main adverse event being nasopharyngitis, which I believe is really innocent. We see a few patients with headache. We see very little number of patients with fatigue, dizziness, somnolence, you see the data here, accidental overdose and nausea. So no surprise of these numbers, and they are very much aligned to what we observed in the 301 study with, again, no tendency to an increased number of events from 10 milligram to 25 and then to 50 milligrams. On the next one, Slide 15, we give you a bit more data, maybe that you haven't seen so far, events adjudicated by the Safety Committee, and you see the numbers here. It's very difficult to debate about these numbers because they are very, very small and for some, like narcolepsy-like symptoms related to complex sleep behavior including hallucinations, sleep paralysis, you see a very small number, but I cannot give you the breakdown by treatment group because this would have the potential to unblind the 303 study, because these patients are actually continuing in the 303 study. And therefore, knowing which different group they are would potentially unblind the study and the 303 extension is blinded. And then on the bottom, you see, again, the repetition of the somnolence, same number than before and a number of adverse events, which are very relevant to an insomnia product, in particular, the number of fall that we had. You remember the small number in the 301 study. We have also a very small number in 303. And also depressed mood, again, very small number of patients. So overall, we are very pleased with the safety results. I can now move on to Slide 16, which is my concluding slide, using the same format as we have used in the previous presentation. The tick box indicates that we could reject the hypothesis and demonstrate efficacy on WASO and LPS at 50 and 25 milligrams in the first study, and for WASO in the second study, being close to significance for LPS in the second study, and nothing significant for the 10 milligrams, which is actually something we find very interesting. From a safety perspective, we didn't see a dose-dependent treatment emergent adverse events. We see very low rate of clinically relevant adverse events. And I can add that we didn't see morning hangover effect. This was assessed using a visual analog scale every morning as part of the IDSIQ questionnaire, assessing whether the patient would feel some somnolence resulting from the drug that was given the previous night, and we see rather a tendency to an improvement versus baseline and an improvement on active versus placebo. And we didn't see any sign of rebound, and we didn't see any sign of withdrawal symptoms during the placebo run-out period. Having said that, this completes my presentation, and I will hand over to Jean-Paul.

Thank you, Guy. And again, congratulations for this program. Maybe I can go to the next slide. Just to say, Idorsia took a lot of risk in designing this program because in contrast to what is done usually, we tested 3 doses in Phase III. We really wanted to have a statistical analysis, including 16 comparisons, 4 main primary end points with 2 time points, 1 month and 3 months. And also, we had planned a design where the middle dose, 25 milligrams, would be common to the 2 studies and the data would be in a later possibility -- in a later time could be pooled. And I think that this program really shows a fantastic result because the 2 studies are very consistent, as was mentioned by Guy. And now with this data, I think, really, we have a very good idea of the dose response. But usually, in Phase III, you get -- you never get such precision, not only on the dose response for efficacy, but also for safety because we have seen that 10 milligrams has nearly no effect, but 50 milligrams has outstanding effect without safety penalty. And this information, I think, is key for the authorities or the regulatory authorities. And this is the reason why now we are really working -- and the whole company is working very hard to submit an NDA before the end or around the end of this year, and we are going to meet the FDA this autumn to prepare for this timing. So that's the very important point. Now we have to prepare for the launch. So what -- next slide, so we are on Slide 18. So what is going to be important for a successful launch? First of all, of course, we have very clear data. And when you will see the data, you will understand how, I would say, tracking how 2 studies can give data so close. And more important, I would say, than even the data is the drug, because we have a very unique drug. And again, I would like to say, it is not by chance that we have seen this outstanding effect. It took us -- the project started 22 years ago, and it took us -- this is the third drug that we put in clinics. We wanted to have absolutely a drug with a very [ strong ] set of actions, peak effect around a big dose in the pharmacokinetic [ properties ], and we wanted a short half-life around 5 to 6 hours to allow for getting an effect on sleep onset, one, though, having this morning hangover effect. And of course, there is still this finding about the data, and these are outstanding findings. We still didn't know what is the mechanism of this improvement of daytime functioning. And of course, it is going to be very interesting to go through the data and to try to understand better these findings, which is really striking. As we said, this NDA will be filed before the end of the year. And now the company has to get prepared for the commercial launch. And you will hear more during this year, but we are working very hard in terms of branding, but also in terms of preparing, especially the United States affiliates, in order to be able to launch, in the best conditions, daridorexant. Next slide. It is Slide 19. So I think what is important is that this program has to be put into the context of our visions in Idorsia. Clearly, it shows our ability to innovate because we were the first of the team who had discovered -- daridorexant was the first to really go in clinic with a product blocking orexin receptor. This was almorexant at this time. So we already are in the middle of innovation. But as I have mentioned, what was also very important is not only to have the right concept, but to have the right drug. And it took us, I think, more than 25,000 drugs to be synthesized in order to discover daridorexant, which is absolutely unique in terms of pharmacokinetics and also pharmacodynamic properties. But then, as you have seen, what is also very important in Idorsia is a patient-focused drug development. And this focus on really trying to understand better what is needed for the patient, what is its main issue, what does this offer, is quite unique. We did that with ponesimod for fatigue. We are doing that in -- with daridorexant for the daytime performance. We are going to do that with lucerastat for pain and this pain questionnaire. So I think its one characteristic is that we not only go to the simple or, I would say, usual end points, but we try to really evaluate what is important for the patients. And finally, even if we are now just starting with the group of [ Simon Jose ], this building of this commercial organization, I think we [ know this one ] very well, what is needed for commercial success. And I can tell you, we do not, even if we have very good products, very good drugs -- a very good product and a very good drug, we do not believe that it is going to be selling by itself, and we are getting prepared for this very large effort for this commercial launch because we will need to work very hard to really get the full potential and full commercial potential of this drug. I think that now, if I remember you and [ that ] consistency is also not only within the 2 studies, but also with what we say. This is the first slide, and this is our strategy, and it was explained to the investors who have decided to come with us 3 years ago when the company was created. We told that...

Jean-Paul, we are on Slide 20, right? Sorry.

Slide 20, yes. So we wanted to clearly explain what is our strategic priorities. So number one was -- and if we were giving us a 5-year time scale, first one was to really bring 3 products to the market. And something I really think that with daridorexant we will have the first one. We are now working very hard for the other projects, but they are moving also well. So that's in a good way, [indiscernible]. Number two was clearly to create a commercial organization. This is also under way. And as I mentioned, and we wanted to bring Idorsia to profitability. And I think that with a drug such as daridorexant, it will give us a very good tool to grow our sales and hopefully to reach profitability in a reasonable amount of time. We also wanted to create a pipeline with a sales potential at peak of CHF 5 billion or more. And I think that with all our pipelines, we are also in a good way to have the drugs, which will allow us to reach this goal. And finally, we want to use the state-of-the-art technology. This is -- and you cannot see that, but this is now true with artificial intelligence that we are integrating into drug discovery. This is true, as I've mentioned, into the new wave with our evaluating clinical -- we are making clinical programs. So this is going to be true with our launch, which is going to use a lot of digital technology. So I think we are -- this is background of daridorexant. We are in a good [ slide ] and Idorsia is really in a very good -- a much better position than ever. Thank you. That's my next slide. Andrew?

Thank you, Jean-Paul. Next slide, please. So now we are in the position to be able to take your questions. For this part of the call, Martine will be able to join us. Please remember that this is a presentation on Phase III data. So keep your questions to the topic. Questions of commercialization or on financial numbers, we will be able to address at a later time. Operator, please open the lines.

[Operator Instructions] And the first question received is from Rajan Sharma of Deutsche Bank.

I'm just hoping you could help us understand the somnolence rates between the 25-milligram and 50-milligram doses? Obviously, in the first trial, there was a higher rate associated with 25 milligrams versus 50 milligrams, and that seems to be consistent here. So just kind of to get your thoughts on any reasons for this would be really helpful. And then secondly, just to clarify, do you expect to get the data from the long-term safety study, so study 303, prior to the FDA meeting?

Thank you, Rajan. So Guy, can you take those questions on somnolence rates, 50 and 25 milligrams and your expectations for the 303 data?

Yes. So starting with the 303, yes, we expect the data to be available prior to an interaction with the FDA. That's not important from an efficacy perspective, but it's important from a safety perspective because as part of this 303 study, we have a sufficient amount of patients treated for up to a year. And of course, as we are looking for a chronic indication, we need this data, so that would be available. The second question is about insomnia -- sorry, somnolence, and the lack of dose response. I think it's an interesting observation. It is very, very difficult to comment on very low numbers. The message for me is more that we have very low level of somnolence across all dose levels. It is probably in keeping with the pharmacokinetics of the product and the fact that the half-life is reasonably shorter, ideal for an insomnia product, and therefore, the patient would not feel somnolence during the day and not report somnolence as an adverse event very frequently. So we see very few. And what is also very interesting for us is to see that there is certainly no more patients with somnolence at 50 milligrams than at 25 or 10. And therefore, I think having now shown that 10 milligram doesn't provide much benefit and 50 providing more benefit, it gives us a good chance to maybe achieve the right dose in our labeling.

The next question received is from Graig Suvannavejh of Goldman Sachs.

I was curious -- and I might have missed it, my apologies, but did you look at next-day performance? I know that was one of the key attributes that you highlighted in the Study 301, and just wondering if it was measured, what you saw in next-day performance? So that would be my first question. And then my second question really has to do around what doses -- just to confirm what doses you do plan on filing. And do you expect to get differential claims for both doses? Or do you expect -- what's the thought on kind of what you think the label is going to look like?

Thank you, Graig. So, Guy, I think both questions are for you, the one on next-day performance as well as our filing thoughts.

Yes. On the first question, first, before answering the question, I would like to be sure that we talk about the same thing, which is data and performance. There is a bit of semantics here. Of course, I do not want to be too complicated, but there is also the issue of next morning residual effect versus daytime functioning. So I believe what you are talking about is the daytime functioning using the IDSIQ questionnaire that we have developed and validated. And the question was about the effect that we have observed, right?

Yes. I mean, can you describe, without giving the numbers, is there an effect? How is the effect?

Yes. This questionnaire is made of 3 domains, and I would refer you to the previous presentation where this is described, the presentation that is still on our website. One domain is the sleepiness domain, which has a number of questions within that I also described on the slide that I presented last time. And this is the one that we use as secondary end points. And we demonstrated that on 50 milligrams, there is a very highly significant effect on this domain. We have not disclosed the numbers, but I can tell you that as part of the development of the questionnaire, we also validate what would be clinically perceived by the patient that will be clinically relevant. And on this domain, we can claim that the effects we observed is clinically important for patients. We have not reported on other domains. But I can also reassure you that there is consistency of the other domains. One was called the [ alias ] commission domain. The other one was the mood domain. And of course, there is this overall questionnaire. And without disclosing the data, I can tell you that it is very consistent across the 3 domains and the total score on 50 milligrams. Now turning to 25 milligrams, what we observed is a numerical effect, which is not too distant from the 50 milligrams that it's in between placebo and 50, as you would expect. In the second study, we have actually very, very similar results of the 25 milligram on this questionnaire. So there is total consistency of the 25-milligram effect on the daytime performance. As I didn't mention before, but it's in our press release, we are going to complete the analysis of this program by combining the treatment groups that we can combine for the 2 studies through a pooled analysis, and these are the 25-milligram groups and the placebo groups. And I'm very confident that by pooling the 2 studies together on the different end points, we will have extremely good and precise estimates of the effect of the 25-milligram dose, including on this daytime performance. Now turning to the second question. I would be speculating on what we are going to decide in the next weeks and months. And I can do that. What is more difficult is to speculate on what the FDA will tell us because I am not part of them. But certainly, our intention is to take advantage of the lack of efficacy demonstrated -- of a very small efficacy, I would say, of 10 milligrams, the good efficacy of 25 and the very good efficacy on 50. Saying that is giving maybe a bit of a light on our idea of trying to put out in our submission to go with 25 and 50 milligrams. But of course, this will also depend on the interaction we will have with the FDA later in the year and how they reacted to our data. And the final dose that will be labeled will be known once we have the drug approved.

Okay. If I could just have a follow-up. So I believe Dayvigo is approved in 5 and 10 milligram, and Belsomra is approved at 5 to 20 milligrams. So do you have a perspective at this point in time, given the data that you've seen about how daridorexant compares across the 2 other approved DORA products in terms of efficacy?

Yes. This is, of course, a question that we are, of course, debating and looking. I would maybe tell you 2 things. One is versus lemborexant, which is the most recently approved, if you look at the labeling, the data are presented in such a way that it's very, very difficult to know exactly in terms of minutes by how much the treatment effect is. And this makes the comparison very, very difficult. Having said that, we can have estimates and certainly, on the night assessment we are not inferior at all. But what the second part of my answer is that what we have in our program is actually unique because not only we could see an effect, which I believe is at least as good as during the night, as good as what is seen with lemborexant and suvorexant, we are adding a new dimension in our database, which is the daytime assessment. Nobody has ever developed an instrument that is as comprehensive as the one we have. Nobody has done during drug development, assessment of the daytime performance to the level we did it using validated tools' daily measures. And we are in a unique situation now to be able to differentiate the product, not just in the night, but also adding a new dimension with the daytime assessment. And we do that without impacting the safety of the patient. So I think your question is about the differentiation. Here, we primarily differentiate the product by the information we have on daytime performance. And the safety is very good, which is also in relation to the pharmacokinetic profile of our product. And especially, you will see the data, you have seen the data, you can compare in respect to somnolence and fatigue, for example.

Just if I can add, Guy, thank you. I think it is very important but when you make comparison, and it is true with many drugs, very often you can get a better efficacy with the drug. The problem is to get a better efficacy without impairing the safety or the side effects. And I think that what is extraordinarily -- is not really, I would say, that it is much more active or less active than other drug is that we get an efficacy without side effects or very, very limited side effects. Look at the somnolence rate compared to what is on the label of the drug you have mentioned. So we could always -- it's always possible to get sometimes to make the patient sleep longer. The problem is that they sleep not too long. That is an issue. So if you compare only the quantity of sleep, this is really not [ actual ]. What is important is the quantity of sleep and people being able to wake up in good shape. That is, I would say, that is the most important that the people have to remember.