Relmada Therapeutics, Inc. (RLMD) Earnings Call Transcript & Summary

Greetings. Welcome to the Relmada Therapeutics results of registrational Phase III RELIANCE I trial for REL-1017 as an adjunctive treatment for major depressive disorder. [Operator Instructions] Please note, this conference is being recorded. And I will now turn the conference over to your host Timothy McCarthy of LifeSci Advisors, you may begin.

Thank you, Shomali, and thank you all for joining us this afternoon. With me on today's call are Relmada's Chief Executive Officer; Sergio Traversa; Maged Shenouda, Relmada's Chief Financial Officer; Dr. Paolo Manfredi, Relmada's Chief Scientific Officer; and RELIANCE Program Principal Investigator, Dr. Maurizio Fava, the Chairman of Harvard's Psychiatry Department. Please note that certain information discussed on the call today is covered under the safe harbor provisions of the Private Securities Litigation Reform Act. We caution listeners that during this call, Relmada's management team will be making forward-looking statements. Actual results could differ materially from those stated or implied by those forward-looking statements due to risks and uncertainties associated with the company's business. These forward-looking statements are qualified by the cautionary statements contained in Relmada's press release issued today and the company's SEC filings, including in the annual report on Form 10-K for the year ended December 31, 2021, and subsequent filings. This conference call also contains time-sensitive information that is accurate only as of the date of this live broadcast, Wednesday, December 7, 2022. Relmada undertakes no obligation to revise or update any forward-looking statements to reflect events or circumstances after the date of this conference call. Now I would like to turn the call over to Sergio. Sergio?

Thank you, Tim, as always, and good afternoon to everyone, and thanks for joining. Earlier this afternoon, we announced the top line results from RELIANCE I our first of 2, Phase III [ assist ] two-arm placebo-controlled pivotal studies evaluating REL-1017 at 25-milligram as a potential adjunctive treatment for MDD. RELIANCE I evaluated the use of REL-1017 in addition to a standard antidepressant for patients who have had inadequate response to at least 1 and up to 3 standard antidepressant therapies in the current depression episode. In this trial, REL-1017 was administered for 28 days. I'll give a brief summary, then I believe Dr. Maurizio Fava has been kind enough to accept the offer to present the data in detail. And we also have a chance to ask any question that he will try to answer. RELIANCE I did not achieve the primary endpoint of a statistically significant improvement in depression symptoms compared to placebo as measured by the Montgomery-Asberg Depression Rating Scale or MADRS on day 28. As was observed in the monotherapy study of REL-1017 RELIANCE III, RELIANCE I exhibits similar and implausible results in certain study sites that also affected the monotherapy study, where placebo dramatically outperformed REL-1017. In this study, the REL-1017 treatment arm, 113 patients showed a MADRS reduction of 15.1 points at day 28 and versus 12.9 points for the placebo arm on 114 patients, which is a clinically meaningful difference of 2.2 points on the MADRS as well as a statistically significant difference in the response rate with a response rate of 27.2% on placebo versus 39.8% in the REL-1017 arm. As was observed in the monotherapy study of REL-1017 RELIANCE III implausible results were observed in 2 of the same high enrolling RELIANCE I study site, where placebo dramatically outperformed REL-1017. While the patient population in RELIANCE I was different than RELIANCE III in that subject enrolled should already have been diagnosed with the depression and did not respond adequately to at least 1 up to 3 courses of antidepressant therapy, a limited number of the same high enrolling sites had an implausible rapid and sustained placebo response that outperformed REL-1017. To better understand the implausible results in RELIANCE I, post-hoc exploratory analogy was conducted excluding the 2 high enrolling sites. Data demonstrated that excluding the 2 high enrolling sites in the study that showed implausible placebo response, the REL-1017 treatment arm showed a 4.1 point difference versus placebo with a p-value of 0.02. A second post-hoc confirmatory analysis using the well-established band-pass method that includes only site with plausible placebo response showed a robust difference between REL-1017 and placebo. While the overall top line RELIANCE I results are disappointing, we believe these post-hoc analysis show a clear efficacy signal for REL-1017. We know that this efficacy signal was also observed with the post-hoc band-pass analogy in RELIANCE III, our recently completed monotherapy MDD trial of REL-1017. We understand though that these post-hoc exploratory analogies are not sufficient alone to be used in an FDA submission. I would like to highlight that at the [ indeed in ] RELIANCE III, REL-1017 again, demonstrated very favorable tolerability and safety in RELIANCE I confirming the results of Phase I and Phase II study with no opioid-like effects, no withdrawal effects and no psychotomimetic effects. Looking ahead, as previously discussed on our recent third quarter earnings calls, we applied several protocol and operational changes and make certain improvement to how they currently involving in Reliance II, the second ongoing Phase III 2-arm, placebo-controlled pivotal study, evaluating REL-1017 as a potential adjunctive treatment for MDD is being conducted. We expect top line result from this study next year. As a reminder, one, we have finalized and determined how to best execute on these announcement, we'll provide a firmer timeline for top line results from RELIANCE II and potential additional studies. We anticipate data from the long-term open-label safety study in the first half of 2023, and they will be part of the planned NDA filing package. In closing, we remain highly confident in the potential REL-1017 to be a safe and effective new therapy for the treatment of MDD and that the issue faced in RELIANCE I and III are related with execution of the studies and in particular site and patient selection. It is important to note that we have the financial flexibility to continue advancing REL-1017 in the clinic due to the strong balance sheet with cash, cash equivalent and short-term investment of approximately $184 million at the end of the third quarter of 2022. I will now pass the call to Dr. Maurizio Fava, who will discuss more details about RELIANCE I results. Maurizio, line is yours now.

Thank you, Sergio. So I'm going to kind of recap a little bit the design. I think as mentioned, it was a parallel comparison design of 28 days in which we were [ preparing ] 25 milligrams of REL-1017 against placebo in patients who had certain characteristics response [ to antidepressant ] treatment. The primary endpoint was a change in MADRS. And the secondary endpoints were change in CGI-S, change in MADRS score at day 7, remission response rates at day 28. There was a follow-up assessment at the end of the 28 days. The key inclusion criteria for the study as opposed to the 303 that was a monotherapy, patients had to be stable for these 6 weeks on an antidepressant treatment with a duration of the episode of 8 to 36 months -- 8 weeks to 36 months. And with a HAM-D score of at least 19 or 24 and 24 on the MADRS. The demographics of the sample. This was -- this involved 227 patients randomized, 114 to placebo, 113 to REL-1017. The mean age was quite typical of an adjunctive trial, approximately 43 years. The BMI was on the lower side given the restricted range of BMI that we had required and of course, more women than men were randomized and the percentage of patients with either Hispanic or African American -- about 82 patients were either Hispanic or African American out of the 227 subjects. As Sergio has mentioned, the difference in terms of change was not statistically significant at day 28. And in fact, as you can see, in the first 14 days, the 2 curves are pretty overlapping and then you can start to see a separation that at day 28 was for the active 15.1 points and 12.9 points on placebo, again, a 2.2 point difference on the MADRS, which is considered clinically meaningful, although not statistically significant because of the standard deviation in the trial due to the variability of a few sites. The good news is that among the secondary outcome measures, there was a statistically significantly greater response rate on active of almost 40% versus 27%. So almost a 13% difference in response rates, which was statistically significant and a 9% difference in remission rate, which was naught. When we removed the 2 in high enrolling centers, and I want to point out that these were exactly the same centers that with no plausible results in the 303. If you remember, we ran the trials in parallel, both 303 and 301. And therefore, sites could participate in both. And these 2 sites had nonplausible placebo responses in 303 and they had the same issue in 301. So when we remove those 2 sites, the statistical significance was obtained at the level of [ 0.02 ] with a difference of over 4 points versus placebo. And removing only about 40 patients out of the 227 from these 2 sites. I've been myself a big supporter of the band-pass. I'll show you the results. But from a safety perspective, there were no serious treatment-related adverse events in RELIANCE I. As you can see, that when you look at the percentages, 11% versus 8% for headaches, upper respiratory 7%, so 5%, nausea 7% versus 4%, diarrhea actually higher on placebo, constipation higher on placebo and dizziness 6% versus 1.8%, very well-tolerated treatment across the active. Again, noise related to QTc prolongation, no evidence of increasing [ abnormality ] or dissociative events or abuse potential. So from a safety perspective, we confirm what we saw in Phase II by what we saw in 303 and now in 301. The well-characterized band-pass showed an even greater difference in terms of -- drug-placebo difference when we look at all the sites that had a plausible placebo response. But even just removing 2 and the same 2 that also were implicated in affecting the results of 303 statistical significance. So overall, the study did not meet the primary outcome measure prespecified. But the signal was there in terms of response rate and removing the 2 sites, which we determined in 303 to be this abnormal patterns of improvement on placebo. We achieved statistical significance. So I'd like to thank all of you for your attention. I'm happy to take any questions as the principal investigator for 301.

[Operator Instructions] And our first question comes from the line of Marc Goodman with SVB. I'm sorry, our first question comes from the line of Uy Ear with Mizuho.

So I guess my -- a few questions. My first question is, can you elaborate on the factors that contributed to the high placebo response at these 2 sites? As previously, I think in RELIANCE III, part of the issue was patients who were not truly the depressed but have one-off situation of depression. So just wondering what the factors here are? And secondly, are these 2 sites the highest enrolling sites? Or are there other sites that have higher patient numbers? And I guess my last question is what are the protocols and operational changes are you planning for RELIANCE II? And do you need to consult with the FDA or have you already consulted with the FDA?

Thank you, I think Maurizio is -- a question he can answer.

So the -- so these 2 sites were among the top enrolling sites. As you can see, the enrolled 40 subjects out to the 227. So they were what we consider high enrollers. The question is, if you notice, our placebo change of about 12 points was, I would say, within the normal range of what we see in retrospective studies. So if you remove those 2 sites, it goes -- let me go back and show you if you remove the 2 sites, the placebo change at the day 28 becomes again, 12.8, so still around 12, which is kind of in the normal range. But the difference is that in those 2 sites, you have this paradoxical 10-point difference in the reverse direction for active. And so when you remove them, you suddenly have a greater 4.1 difference. So while in 303, there was a general tendency across all sites to have higher placebo response. In this study, the placebo, I would say was, all things considered, in a normal range, but we did have these 2 sites with this very paradoxical pattern of the placebo outperforming by 10 points or more the drug. And that's exactly what they did in the [ standing ]. So in terms of how do we interpret that, there may be issues of rating because it's very odd that -- at the same site underperforms when all the other sites show a benefit of the drug over placebo. These are the only 2 sites that show a marked worsening of the drug or yes. So I would say in terms of the explanation, something very unusual has happened at these 2 sites. And they're not part of 302 which is a good thing there. And -- but we're setting up a series of refinements in terms of subject selection. And also, we're going to tighten up the quality of assessments to make sure that the ratings are conducted properly. So we're putting in place -- because in the case of these 2 sites, there may be an issue of unplausible ratings rather than subject selection.

Our next question comes from the line of Marc Goodman with SVB Securities.

This is [indiscernible] on the call for Marc. I have a question regarding the trial design, given the study was 90% [ over] with 2 points of difference in MADRS score and end up being 2.2 points. Can you provide more color on the lack of it here? And I know you mentioned the standard deviation issues, but any additional color would be helpful?

Yes. So the -- if I understand correctly, you're saying a difference of 2.2 with a larger sample size could be statistically significant. As you remember, we were planning -- we -- based on the feedbacks of the DMC, we terminated enrollment about 2/3 of the way in the trial with 227 subjects randomized. So if I understand correctly, you're asking, had we gone further in the study to that 2.2 have become statistically significant. Is that your question?

Yes.

Yes. So you're correct. You're correct. It could have. It could have. But the -- we conducted the analysis with the smaller -- the interim analysis with a smaller sample size. And at that point, the curves were too close to encourage us to continue. Had we done the interim analysis with a little bit more subjects probably the recommendation set keeps growing. But it's one of those things where -- you never know post-hoc, it's kind of easier to say, "well, had we gone further with the trial, we might have achieved statistical significance." We clearly achieved it even with a smaller sample size with response rates. So that was very good. So -- but again, unfortunately, those 2 centers really hurt us.

And if I may Maurizio? The -- because of these 2 sites with this placebo outperforming by 9,10 points, the drug, the variability just increased dramatically. And that's kind of -- was not -- I don't believe it was anticipated in the statistical plan.

Got it. Can you also talk about onset of action and the treatment effect over time during the 4-week treatment period? I know you mentioned there's like no separation during the first 2 weeks. Can you maybe compare the data for both arms versus the prior Phase II study?

Yes so -- go ahead Maurizio.

So the separation here was at day 14. But so we didn't see it as early as in the Phase II but that could be because you have to realize that by day 7, we had an average of 11 points of placebo. So there was -- even though the placebo response overall was in the normal range the [indiscernible] of the response was quite significant. And so that might have prevented as our ability to detect a signal earlier. Does that address your question?

Got it. Yes.

Yes. In the Phase II, the placebo response at day 7 was, I believe, 8 points, including the study was an in-clinic that usually has a higher placebo [indiscernible]. So in 301 and 303 this placebo response has been very, very high, very quickly that is not very difficult to explain since these patients were supposed to have taken an antidepressant for 2 months and not to respond to an antidepressant and all of a sudden, they take a placebo and they get all better very quickly in a few days. So it's difficult to explain, but I don't -- we don't have the real explanation.

In clinical trials, expectations modulate the response rate. So in general, one of the reasons why we see higher placebo response rates in Phase III than in Phase II is because the sites and the patients have greater expectations of the treatment. So it's [ methadone ] at this point, is a well-known among investigators entity. There's a lot of interest, a lot of excitement. And that so we do see that expectations can modify. And in fact, the [indiscernible] because of the Phase II study results start to have a rapid effect. So all these things tend to contribute to more rapid placebo response and more robust placebo response. But as I said, at the end of the day, and at day 28, a change of about 12 points is not -- it's kind of in the range, if you wish.

Our next question comes from the line of Andrew Tsai with Jefferies.

I guess my first one is a bigger picture. Will you likely need to start another Phase III on top of RELIANCE II because at the end of the day, we do need 2 successful studies, right?

Well, I think it depends on the interpretation of the FDA, whether the Phase II trial can be considered a positive trial? But again, we will require discussions with the FDA, whether or not the Phase II would be considered sufficient, supportive if we only had 1 that is RELIANCE II positive. Can we use both, I think this is -- leave it to the regulatory experts. But certainly, we clearly had a robust positive effect in the Phase II trial. And if we have a robust effect in Phase III in RELIANCE II. And with the band-pass, both 301 and 303, showing a robust effect with the band-pass, one could consider that supportive. So I think we have to look at the big picture here.

Yes. And -- when we compete the 2 -- go ahead, sorry.

Yes. Andrew, If I may. Corporate-wise, it would be very prudent, and we are nicely capitalized, very prudent to start at least more study, maybe not with the very -- exactly the same design, but in terms of spreading the risk it would be beneficial to have more the only 1 ongoing. We haven't decided. We just got the data yesterday. So we really crunch the time to put them out today.

Got it. And so I guess, as a follow-up to that question, when I compare this study at day 28 to the last study, at day 28. I guess the MADRS reduction for 1017 around 15 points, it does look consistent across the 2 studies. Placebo does look a little bit better, and maybe it drove a better delta this time. So as you think about a new study, I know it's still early. Could it make sense to push up a higher dose or would the next subsequent study be focused mostly on controlling the placebo response?

Well, I mean, Andrew, I can give you my -- [ wait ] 2 seconds and I'm not the super expert there, but I think the issue is the placebo response, especially early placebo -- response to controlled placebo especially early and site selection would be -- and the rating would be the focus. But Maurizio and Paolo, they are on the call. So they can talk more about the different dose.

Yes I think with 303 showing with the band-pass a robust effect with 301 showing a robust effect with the band-pass or with removing the 2 sites with highly implausible results and even response rates that are statistically higher with 25 milligrams and 25 milligrams working very well in Phase II, I don't think there is an issue of pushing the dose. I think it's an issue of controlling the placebo. So if there's going to be a study as Sergio mentions, RELIANCE IV the focus there would be on managing the placebo response, designing the study in a way that weeds out the placebo responders.

And very last quick question is for RELIANCE III, how far along are you enrolled out of the target, say, 350 patients?

About 25%.

Our next question comes from the line of Andrea Tan with Goldman Sachs.

Maurizio, maybe a question or a clarification on some comments you made here. Just regarding placebo response. I guess when I look at the overall population at 12.9 versus the 12.6 for the post-hoc analysis. I think you've mentioned that, that's generally in the normal range for a placebo response. So I guess, could you help us understand if it's still much an issue that it's a high placebo response at this type versus REL-1017 maybe underperforming there? And then curious what factors in your opinion might lead to those patients maybe being less responsive to the drug? And is there anything that you can detect from patient baseline characteristics that might kind of inform that these patients would be less responsive?

Thank you. So these are great questions. So the -- yes, the placebo response when you take historical failures can range between 10 and 13 or 14. So 12.6 or 12.9 is kind of in the normal range. However, the variability caused by those 2 sites because the -- was a significant problem. And so practically speaking, when you're trying to eliminate or address this problem. I think that you have to look at several factors. One is subject selection. And clearly, in 301, the subjects were kind of had a more reasonable placebo response. So they were more likely to have a true condition or major depressive disorder because we had discussed 303 that with the pandemic, we have a lot of people who [ get ] criteria for depression, but don't necessarily have a disease they have more stress induced depression, it's self-limited or short-lived. But here is the placebo was not necessarily the big issue. The issue in my mind for those 2 sites is that they perseverated in showing a greater than 10-point difference in favor of placebo and that is very, very odd because you can say, well, that can happen in one trial. But happening in the same site -- 2 sites and happening in the same direction, whereas all the other sites show the opposite trend that is drug improving more than placebo. As a clinical investigator, I have some concerns about the ratings. The ratings conducted at the at those 2 sites. That would be my guess that something happened in the ratings that would explain why in 2 parallel studies, the same problem, the same implausible pattern was reported. I don't know if it addresses your question.

Sure. Yes. I guess -- I mean, I guess, maybe I'm sorry for digging in on this. But I guess I understand that placebo is greater -- 9, 10 points greater than the treatment arm. But if placebo is in the range of what's normal, doesn't that imply the issue might be more on the treatment arm where there's -- that's like what's the most surprising part versus the placebo? Or does it only just matter that the direction is implausible.

Well, both are important. That is -- what I'm trying to say is that nested within any anti-depressant effect, there is a placebo effect. And that's why, in some ways, you typically don't subtract the placebo from the drug. It's an additive effect, 2 points, 3 points, 1 point but there is a placebo effect nested within each active. The fact that you have endpoint in the opposite direction, is very hard to justify because it's as if these patients had no placebo response whatsoever. All of those at site on the drug and no placebo response, but all those at site on placebo had a placebo effect. That kind of doesn't quite make sense. Now you could say, well, if the drug had a nocebo effect, had some sort of a negative effect, well, then it would not be limited to 2 sites, you will see it in other sites. So that's where, in my mind, that's the odd aspect of these 2 sites is that they somehow mysteriously attributed no placebo whatsoever to the drug in both studies and in some way had this paradoxical placebo being 10 points better than drug. I don't know if it addresses your question.

Andrea, if I may just to give you like one example. 1 site had a placebo response of over 20% improvement from baseline in both studies, but 20%, meaning all the patients, to get an average, all the patients on placebo on that site were cured after a week or after 2 weeks so that is a magnitude.

20 MADRS points.

It was 20 MADRS points. So they had a drop of 20 points on MADRS. And somehow, the same site with the same because -- you could say, "hey, these are sites that are very nice to patients. They provide tremendous support." But then you should see some of that placebo benefit are [ a sign ] to drug, right? Because we know but we didn't see it. So normally, I would say, on high placebo response site, you see drug and placebo are looking the same. So having no signal detection because the drug is no better than placebo. But you don't see drug being 10 points lower.

Our next question comes from the line of Jay Olson with Oppenheimer.

This is Matt on for Jay. I really appreciate all the color. So we were wondering if you've received any additional feedback on the new results since you top line and especially if you've received any feedback or commentary on the band-pass method, on that study? And as a corollary, what you would expect a physician to comment or provide feedback about this study and the post-hoc analyses? And another question we had was just based on the modifications that for [ RELIANCE II ] data such as excluding those problematic sites, what placebo response range would you expect in those results? Definitely, really appreciate it.

Sure. So clearly, we're going to try to lower further the placebo response in RELIANCE II and in the eventual study that Sergio has mentioned, with a number of methodological approaches that can -- that are known to further reduce the placebo response. Now the band-pass is a way of basically saying, let's look at only the site that had a plausible response. So at least a 3% improvement and no more than a 33% improvement so -- on placebo. And when you do that on 303, you clearly see a very nice signal. And when you do it in 301, you clearly see a signal, both in the 5-point range. So the critical thing is to try to get the average placebo response rate as much as we can on most of the sites within a range that -- 3% to 33% improvement and no more than that. And we clearly did better on 301 and 303, which we would have expected because it's an [ augmentation ] study. As I said, if we just did not have those 2 sites, we would be now talking about a positive trial.

And our next question comes from the line of Joon Lee with Truist.

This is Les on for Joon. Can you comment if there have been any changes in the percentage of patients rolling over into the open-label extension study? And would we expect the readout from the open-label study to come before RELIANCE II top line?

Paolo, maybe you should take that.

Sorry, the question was about rollover patients into 310. And so mentioned...

I believe you previously mentioned it was about 75% rolling over. I wonder if there's been any changes in the percentage of those patients rolling over?

No. There are still very high numbers rolling over and staying into [ 310 ] being retained there.

Okay. And then on the time line readout, will that come in before the RELIANCE II top line?

Yes, 310 should be ready in the first half of '23. And -- but it is already -- it's an open-label study. It's going very well in terms of safety.

The enrollment is done with long-term safety. So we're not enrolling any more patients.

Got it.

I think there was 600 and some change -- Paolo, do you remember what roughly the number is?

It was around 600, the enrollment.

Yes, around 600. Yes.

Our next question comes from the line of Uy Ear with Mizuho.

Just -- I just want to get a better sense of the raters. If the issue is -- seems to be 1 of the rater rating problem with these 2 sites, can you elaborate on whether these raters are site specific or do they go from place to place? Or is it centralized? Or is it -- I just want to get a sense of why these 2 sites have -- it seems like the raters are reversing things, I guess, in a sense.

So no, we did not use centralized raters. So the raters were site raters. And so traditionally, that has not been a significant issue in clinical trials that is most trials use site raters. But in this case, the good news is that we have the recording. So we can now go back, be post-hoc, but we can do some quality assurance post-hoc on those sites.

Okay. Do you think that the rating issue was -- could also be part of the issues in the 303 study for these 2 sites?

Yes. Yes. I think that because they had the same pattern. So it's possible that they had issues. Again, we're trying -- we're puzzled by the results in those 2 sites. And the most logical explanation -- positive explanation is that issues with ratings.

And we have reached the end of the question-and-answer session. I'll now turn the call back over to Sergio Traversa for closing remarks.

Thank you, and thanks, everyone, for the interest and for participating in the call. Thank you, Maurizio and Paolo for helping out on the -- answering the question on the clinical and scientific side. In the drug development there are setbacks. And this one, this data look a little better than the other one as it was expected. We are not -- didn't get where we want to be yet. But moving forward, it looks promising. The signal is very clear of the efficacy and tolerability and safety over REL-1017. The question is just to improve and we're working on that very hard, improving the execution and the operation of the trial and improving by simplifying protocols and improving site selection, patient selection and rating, as Maurizio has pointed out. And hopefully, the next call on the results of a clinical trial will be different than this one. And I remain very confident that we do have a drug, and we'll continue to work and getting the drug to the market and make it available to patients. So thanks again to everyone, and looking forward to the next call.

And this concludes today's conference, and you may disconnect your lines at this time. Thank you for your participation.