Good morning, and thank you for joining Guggenheim's 2025 healthcare innovations conference. I am Debjit, one of the therapeutic analysts, and my privilege to welcome our next presenting company, Centessa Pharmaceuticals. Joining us from Centessa are Saurabh Saha, CEO, and Mario Accardi, President of the Orexin Program. Thank you, gentlemen, for making time for us.
Thank you.
Thank you.
You're not just a narcolepsy company. There's a lot going on behind it, so I'll let you do a quick introduction before you get into the Q&A.
Yeah, so thank you, Debjit, for the invitation and opportunity to speak. Very exciting time for Centessa. It's been a very exciting week for the company. We're very happy to have announced our data from all three indications: NT1, NT2, and IH, showing clinically meaningful and effectiveness with ORX750, our first and our most advanced molecule in clinical studies to date. We are very excited about the class in general. This is one of the most well-validated targets in a very long time in neuroscience. Now the clinical data is showing that this molecule, in fact, is effective in narcolepsy and idiopathic hypersomnia. We're the first company to show that. More importantly, I think to your point, is that the narcolepsy patient population, the rare hypersomnias, are—it's a devastating, devastating set of conditions. Hopefully, we have an opportunity to help these patients.
Beyond just the rare hypersomnias, now, as of last week and with the data that we showed from our company, this now portends well that in patients and diseases where normal orexin signaling may be happening, but there's a comorbidity of excessive daytime sleepiness and potentially other comorbidities that are associated with orexin dysfunction, there may be large populations of patients in other neurological indications that may benefit. I think the data that you've seen in NT2 and the very early in our first dose cohort in idiopathic hypersomnia portends well that this class of drug is likely to be effective much beyond the rare hypersomnias.
Today morning, we had the Vibrance readout. Any high-level thoughts on the data?
Mario and I were both in investor meetings—thank you for that—this morning, so we did not get a chance to look at the data in any detail or get to hear the presentation or read the transcript. From what we have heard, it is exactly what we had expected and exactly what we have been telling all of you for a year or so now. We are very excited about where we are now positioned relative to the others as they have disclosed their data with time today and earlier with the other competitor. We feel excited. What it does, I think most of all, show is I think there were doubts initially on Centessa's development plan of how to develop these orexin agonists. It was kind of unusual and unorthodox, unconventional for us to do real-time drug development of these orexin agonists.
What I mean specifically by that is to be able to look at the data after each cohort, after each dose, after each indication, and be able to adjust your dose accordingly based on MWT and ESS data as well as the PK data. That was absolutely key. Not picking doses based on day one data, which we did exceptionally well, and we've disclosed that already, but not just based on day one data, but looking at data as time progresses and then being able to dose adjust up. That's precisely what we did. That, plus the combination of the beautiful chemistry and research work that Mario has done on generating ORX750, is a very differentiated molecule from a chemistry standpoint and from a PK standpoint. We can talk about details, but we are seeing good dose-responsive data. We're seeing dose-responsive data.
We're seeing good exposures that are linear with dose. We saw that in our healthy volunteer studies. We see that in our patient studies in NT1 and NT2, having done multiple dose cohorts. We see that correlate with effect size as well. This is the type of drug development we took a while to get to this point by selecting the right molecule and having very, very careful screening funnel and picking a molecule that would be dose linear in terms of exposure and response.
Got it. Let's talk about the definition of best in class. Now that you have competitor data, both in NT1 and NT2, how convinced are you Centessa has the best in class molecule?
Yeah, I'll start with what's the flavor of the day. NT2, I think it's clear based on the data that we've disclosed in MWT and ESS that that is perhaps the best data ever reported in that indication period. Greater than 10 change of baseline in MWT. ESS drop of 16 to 8, roughly. Very, very significant in terms of the P values. Despite being smaller size in cohorts, you don't need to go to much larger numbers and do large studies if your molecule behaves the way it should and is dose proportional in terms of effect size. You get a strong effect size. That's how you get low P values in very small subjects, number of subjects. That's NT2. ESS and MWT, hopefully setting the benchmark going forward, but that's our initial dose cohorts.
Those are the first two we tested at 2 mg and 4 mg. We're not done. As we're speaking, we're dose escalating. Nothing's preventing us from dose escalating in terms of dose-limiting kind of AEs or any kind of issues there. The beauty of it is that our lowest doses are as good or better than anything else anyone else has reported to date, and we keep going up. That's NT2. In NT1, the story is the same. It's more of the same, where we're greater than 20 change from baseline. Half our patients at the 1.5 mg dose are greater than 30 on mean MWT. Our ESS is 5, which is one of the lowest reported. Massive change from baseline there. Very static across the board. Dose responsive across the two doses we've tested. Generally tolerable and safe.
When we say best in class, the data, I think, speaks for itself.
Got it. There are two aspects for best in class. One, obviously, efficacy, which you talked about. Talk about the tolerability aspect of it. What are you seeing in the clinic? To Saurabh's point, nothing prevents you from dose escalating higher.
Yeah, that's right, Debjit. It's a very, very well-tolerated molecule. There are no AEs that we've seen that define the therapeutic index of the drug. We have ample headroom to continue to dose higher, dose escalate, and really harness the full power of this biology with ORX750. We feel very, very confident around the dose escalation strategy. All the AEs were very mild and across mild and moderate. We had, obviously, polyuria, insomnia, with probably the most true prevalence and true rates of AEs, which have been seen across the class. They're not dose limiting in any way. It was very interesting because at the 1.5 mg NT1 group, we also had no visual disturbances, which I think speaks just to the PK profile, most likely, of the drug.
The flat Cmax to AUC, which is allowing a very, I think, very, very good and well-tolerated profile. Yeah, right now, dose escalating. Based on the fact that we only had one discontinuation in the clinic due to polyuria, a patient that had a pre-existing condition due to urinary urgency, we feel great about continuing the dose escalation and really cementing our position as best in class.
As you have your investor discussions today, the controversies that we hear, durability, what does it look like from week two to week four or week eight? In terms of this being a once-daily dosing or BID? Any thoughts or any clarifications on those?
Yeah, to answer your first question, there were a lot of questions around our design. The primary analysis was done in the first four weeks of the six-week study, looking at two-week MWT on drug and on placebo. It gives us a very high statistical power because of the crossover study design. If you look at external data sets, including the data set that we all seen this morning, once you hit two weeks, and it's probably between the first dose and two weeks, we do not really know because we do not have efficacy measures between those two time points. Once you hit two weeks, you really see robust effect. You see that on ESS. You see that on weekly cataplexy rate in NT1 and other external data sets. You even see it across both NT1 and NT2.
That's why it's so important to include two-week analysis in your dose selection for potential registration studies. That's why we're looking at the two weeks. Now, we have also evolved our study design. We are now transitioning to a four-week parallel design. That's really just to bring the study more in line with what a potential registration would look like in terms of a parallel study design. Two weeks, I think, has now been demonstrated and de-risked as the key time point in order to really allow decision-making for prolonged and chronic studies.
Let's talk about the weekly cataplexy rates. That was a pretty remarkable outcome from the NT1 patients. Do you attribute this to the durability throughout the day, or is there something else going on there?
ORX750 has been designed to have an ideal QD PK profile with very, very good metabolic stability and allowing blood concentrations to really go out up until the late part of the day. That is the beautiful thing about PK. We do not have a biphasic PK profile with a sharp Cmax to AUC profile. It does enable QD dosing. I think that has been demonstrated by, you are completely right, by the very low incidence rate ratio in our weekly cataplexy rate, which is 0.13. Incredibly low. That does suggest excellent symptom control on cataplexy all the way out to the later part of the day. To Saurabh's point around the ESS, the fact that we are hitting an ESS of 5 down from, what is it, 18 or so, really gives us confidence that we are also seeing symptom control in the afternoon.
Because on the ESS, there are some questions around how patients are doing in the afternoon. Now, to answer your earlier question also about where does this fit with split dosing, Debjit, let's put it this way. You can move from a QD drug to split dose. You can, let's say, divide the dose into, let's say you have a 2 mg QD, and you're going in with split dose of 1 + 1 to further give patients flexibility. Flexibility to help them better modulate duration of action and symptom control. Maybe they have a late evening commitment, social commitment, or they have a different schedule that varies from day to day. With that second dose, they're in control of their wakefulness needs. They're in control of their symptom control. That is what we want to address as Centessa.
We really want to cater for 100% of the patient population. It's very different from a BID drug. You cannot go from BID to QD. BID is a necessity. That's, I think, one more aspect in which we're looking to differentiate.
Would you like to have a BID drug for all three, or this would be more in NT2 and IH?
I think split dose means flexibility. I think it would make sense for any indication where you need the diurnal symptom control. We do have a QD drug. We've shown it with our phase II data. We have a QD drug. By looking and exploring split dosing in our CRYSTAL study, it's just an additional toolkit that we can use here to provide additional flexibility.
Will the BID dose help you address any of the early tolerability issues, or tolerability issues fade after the first couple of weeks?
We actually still did not do the full analysis to look at the onset, offset, and symptoms. We have seen from external data sets that you are completely right. A lot of these AEs tend to wane over time and are even transient in nature. We just have not completed the full analysis. We cannot even say that there is still a dose response on many of these AEs. What I would share is that, again, there are no AEs today that define the therapeutic index of the drug. It is extremely well tolerated. In exploring split dosing, obviously, we are also looking at tolerability. Not because we have to. It will be really an interesting data point to understand, does split dosing, does the reduction of Cmax, does that contribute to a reduction of untargeted AEs? That is absolutely something we are looking at, not out of necessity.
It's really just about exploring, cementing our best-in-class position here as we move into selecting doses for registration of program.
Can we talk about what normative wakefulness means? Is that a different threshold in NT1 versus NT2 and IH?
The wakefulness threshold is these minimum blood or brain exposures that you need to hit in terms of drug concentration to allow patients to have full symptom control. In this case, basically, stay awake. What we've noticed is that the wakefulness threshold is definitely different between NT1 and the non-orexin loss indications, which is why you need more horsepower. You need more dose as you move from NT1 to narcolepsy type 2 and probably also idiopathic hypersomnia. That is why it's important to have a molecule that allows you to go higher in dose with confidence, that allows you to have completely linear PK, so you can get linear and linear increase in exposures, and hopefully also pharmacodynamic effect. That wakefulness threshold, based also on what others have reported, seems to be somewhere between 2x-3x between NT2 and NT1.
Of course, we're still refining this as we move up and continue the dose escalation and gather, obviously, a lot more data. I mean, as Saurabh said, we've got more than 50 patients being dosed as we speak today in higher dose cohorts.
Just to add, in our NT1 and NT2, our ESSs have achieved normative wakefulness, as you saw, below 10.
What I was trying to get at is, do you need to get to 20+ minutes in NT2 and IH? You clearly need to get there in NT1.
Go ahead. The key is the change from baseline. Because the variability in baseline and response in NT2s is so profound, that is the metric that regulatory bodies look at. If you look at the current standard of care, it's a good comparator as well. It depends on patient to patient. If they start at 18 and you get a change of baseline of 1, that's not going to get you there. If you get a change of baseline much higher, then obviously, you'll easily cross through the 20-minute threshold. It depends on where that individual patient is. That delta, that change is the horsepower that your drug is providing to provide that effect size. You're feeling much better from wherever you started. That's what's important. The ESS has to correlate.
Obviously, you want to get down to less than 10, which is what we did with very low P values.
That's why the registration endpoint is based on the delta, the change from baseline, especially in NT2, where we know there is some variability.
Let's talk about the difference in diagnostic criteria between the U.S. and Europe. In Europe, you're doing a spinal tap to actually make sure the patient is actually NT2 or not NT1. We do not do that in the U.S. We saw a competitor data readout with very low MWT baselines, almost like an NT1. How do we know these patients are truly NT2 and not pre-cataplectic NT1 patients?
I mean, it's an interesting observation. If they're truly NT1, then you'd expect really good data. I mean, right? Because at higher doses in NT1, you should have blowout MWT data.
Yeah. I can't comment about the others, Debjit. Obviously, in our case, our team has a tremendous amount of experience with running studies in these indications. I think we put a lot of extra care in patient selection. We've got a very robust inclusion criteria to give the best possible likelihood of recruiting real NT2 patients. You're completely right, there is a difference between the U.S. and the EU in terms of diagnosis. We also partner with a lot of PIs that have a lot of experience and provide a lot of guidance and training to a lot of the sleep centers to be able to really refine that.
It is in our best interest. This is what we have done. It is Centessa with our inclusion and exclusion criteria for NT1 and NT2 to be as rigorous as possible, using as many orthogonal measures as we possibly can to understand whether the patient is NT1 or NT2. Scrutinizing every patient because it does not serve us well in a phase II setting to get confounded data and then go into phase III and have picked doses on that confounded data. Whatever doses we are picking for NT1 and NT2, especially NT2, we want to make sure those are accurate doses for that patient population, which truly is NT2.
Centessa is the first company to show a clinically meaningful improvement in IH. What is a clinically meaningful improvement in IH where I think the baselines are probably higher than an NT2 from the Vibrant study?
Yeah. If you look at the American Academy of Sleep Medicine, a reduction of two points on the ESS is deemed to be clinically meaningful. A reduction on the MWT, or rather an increase in the MWT with two points, is also thought to be clinically meaningful. We included the MWT as an inclusion criteria in the IH study exactly to avoid recruiting patients that have a very, very high baseline and would be, I think, more awake than asleep because of that variability in IH. We looked at multiple measures, including ESS and including the idiopathic hypersomnia severity scale. As you know, it is unlikely that the MWT would be the key registration endpoint for IH given precedent. We showed stat-sig and clinically meaningful effect on various efficacy measures beyond the MWT.
At 2 mg.
At our very first dose of 2 mg. I think we feel we're very excited about continuing the dose escalation in IH and restoring normative wakefulness, which, as we know, on the ESS, it's below 10. There's only one drug that's approved for IH today. There's a huge unmet clinical need and, I think, a huge opportunity to do better.
Can we talk about the tolerability profile as you escalate from the lower end of the spectrum to, I don't know what the therapeutic window is for you as yet, but does it scale?
To be seen in the sense that, as I mentioned earlier, we really haven't seen much of a dose response on these AEs, such as the polyuria, insomnia. I think it just suggests that we have ample headroom to continue dosing higher. The molecule has been specifically designed to give us a wide therapeutic index, whether it's in the orexin loss in narcolepsy type 1 indication or in the non-orexin loss.
Where do you think orexin agonists fit in commercially? In NT1, it's a very clear case. This becomes the backbone therapy. What happens in NT2 and IH patients?
Yeah. We strongly believe in the monotherapy approach for orexin agonists across both NT1, NT2, and also IH. In NT1, orexin agonists have shown profound impact on wakefulness and cataplexy, as we've shown, which are the two key registration endpoints for NT1. The only unanswered question is disturbed nighttime sleep. There is a lot of biological rationale why the lack of orexin is the cause for disturbed nighttime sleep. I think it's really about not if, but rather when, i.e., which orexin agonist is going to help with disturbed nighttime sleep because it's fundamentally a PK question. Can you manage to achieve a PK profile, whether with the inherent properties of the drug or through formulation options to allow you to have a residual tone at night and help consolidate sleep?
By the way, we're looking at PSG and disturbed nighttime sleep as part of our crystal study. In NT2 and IH, these are very much diurnal symptoms, similar symptoms across the board. We're still talking about excessive daytime sleepiness. We're highly likely that orexin agonists will, again, be a monotherapy here. The only question mark is on the sleep inertia and idiopathic hypersomnia. Will orexins truly help there? I think there's a lot of evidence, again, from how the inherent mechanism works. Especially with 750, we've designed a drug that has a rapid onset of action. We have a Tmax for about two hours. That should hopefully fundamentally help with sleep inertia. I think the most important thing, Debjit, when I think about differentiation between orexin agonists and that's how we're really looking to differentiate versus the others, is duration of action.
If you have a drug which can be any orexin agonist, which just does not have the right firepower to keep symptom control throughout the entire duration of the day, what's going to happen is that patients are going to go back on polypharmacy. They're going to go back on stimulants in the later parts of the afternoon. They may need to resort to anticataplexy meds in the later parts of the evening. We want to avoid that. We want to give patients the opportunity to go on orexin agonists beyond monotherapy and give that extended symptom control. To really cater for 100% of the patient population, we're doing that by dosing higher. We're doing that with split dosing.
I would add that the oxybates, let's take that for example, is that the tolerability challenge is there. I think the vast majority of patients do not actually take the oxybates. We want to be able to offer something for that population.
How confident are you with your EPI data? It seems to be much bigger than some of the others are. We're putting out numbers there.
Yeah. I mean, we feel we've done quite a bit of work in this space. Based on the numbers that we've put out there, we feel very confident. I think others have done the same. This is a very substantial market. It's important to note, as I mentioned earlier, that the current drugs that people take, mostly in polypharmacy setting, do not do an adequate control in terms of providing what could be a functional cure. The tolerability concerns are really substantial. A number of patients today are on generic medicines. They're not on the branded medicines. Those EPI numbers do not necessarily take that into account if you do not look at that entire population that's not taking the drugs that are branded today for the narcolepsy indications.
The total addressable population is actually much larger than what the total revenue numbers that people are generating today with the narcolepsy drugs indicate.
I think we can go on for another hour. I just want to finish with the regulatory side. The company intends to start the registration program early next year. FDA discussions and when do you have the next 50 patient data or 50+ patient data?
Yeah. As a company, we don't comment on regulatory interactions. Certainly, you would expect we'd be having those discussions if we're planning on registration studies in Q1 of next year. We will provide an update, as we did last week, in Q1 on data that's being generated as we speak today.
Appreciate the time, gentlemen. Thank you so much and good luck.
Thanks.