Good day. Welcome to the Tenax Therapeutics phase III LEVEL Top Line Results Conference Call. All participants will be in a listen-only mode. Should you need assistance, please signal a conference specialist by pressing the star key followed by zero. After today's presentation, there will be an opportunity to ask questions. To ask a question, you may press star, then one on a touch-tone phone. To withdraw your question, please press star, then two. Please note, this event is being recorded. I would now like to turn the conference over to Chris Giordano, President and Chief Executive Officer. Please go ahead.
Good morning, everyone. Thank you, Betsy. Thank you everyone for joining us. I'm Chris Giordano, President and Chief Executive Officer of Tenax Therapeutics. This morning, we released top line results from LEVEL, our first phase III trial of oral levosimendan in patients with PH-HFpEF. The primary endpoint did not reach statistical significance. The difference between groups in six-minute walk distance at week 12 was 3.5 m with a P value of 0.63. The evidence we will share with you today impresses us. Levosimendan's impact on key objective biological targets directly translates benefit from pulmonary pressure and BNP lowering to functional improvement in the more advanced patients we treated. We believe we have a drug for the many patients who need it, and there are millions of them around the world. Our phase II HELP results found a new target population for this unique drug.
LEVEL was always intended to validate the efficacy shown in that study, more precisely measure the treatment effect, and establish the right patients who benefit in this very heterogeneous HFpEF population. The trial did not meet its primary objective. When it comes to sizing and locating the treatment effect in these patients, the trial has achieved very important goals. We received data only last week. In beginning to dissect it, we see evidence of a very responsive patient population in LEVEL, and we also see the reason their response is obscured in measurement of the primary endpoint. We see clearly where one inclusion criterion let this drug down in this study. What I will tell you today is this: our therapy works very effectively in a large, common, easily identifiable population of PH-HFpEF patients, and they need it.
We are not in a situation where the drug and the placebo groups are indistinguishable in our data, or a situation where drug exposure or adherence was inadequate in the trial, at least not in the full analysis set we're looking at. On the contrary, our results define the population of high responders coherently, and the adjustment we need to make in our phase III program is crystal clear. LEVEL was a well-conducted trial. It was executed on time, randomizing 241 patients in about 24 months, with a reassuring 2% rate of missingness on the primary endpoint. Very high patient retention and therapy adherence, and no site-specific problems or six-minute walk test performance anomalies we're going to point to as a culprit behind the primary miss. We have invaluable information from LEVEL, and we intend to use it.
Here's how we will use our time this morning. Dr. Stuart Rich, our CMO, will take you through the complete top-line data set, the primary and secondary endpoints, the safety profile, the biomarker and hemodynamic results that yield important information about the efficacy of levosimendan in a well-defined PH-HFpEF population. The subgroup analyses, which will guide the direction of our program from here so that the drug's effect is not obscured a second time. We will hear from Dr. Sanjiv Shah, Professor of Medicine and Director of the HFpEF program at Northwestern University and LEVEL's Principal Investigator. I will close by discussing where our phase III program goes from here. We will open the line where Stuart, Dr. Shah, and I will be joined by Doug Randall, our Chief Business Officer, and Tom Staab, our Chief Financial Officer.
Before we go further, we will be making forward-looking statements on this call, including statements about our clinical data, regulatory plans, our future trial designs, and our business operations, all of which are based on management's current expectations. Actual results may differ materially from those indicated on this call due to inherent risks and uncertainties described on this slide and in the company's SEC filings. Please read them accordingly. Although we may voluntarily do so, we undertake no commitment to update or revise forward-looking statements except as required by law. Before Stuart walks through the results, let me tell you how we are reading the evidence we have so you have a frame for what he's about to show you. Let's start with what this trial confirmed, because it is much more than people may assume from the headline. LEVEL started out with several key questions still outstanding.
Is chronic oral levosimendan safe in the population tested? The answer is yes. Do we have a clear and robust set of evidence of a dose that is biologically active with NT-proBNP and pulmonary pressures both responding in the way this mechanism predicts they should? Yes. The phase II efficacy results from HELP were unprecedented in Group 2 patients, but were they reliable? Would we be able to validate the treatment effect we saw in those 36 advanced PH-HFpEF patients? Again, yes. Can the sites and the clinical operations team execute? Yes. The reason those answers did not convert to a win on the primary is apparent. It actually jumps off the page at us in the first hours of our review of the data. Tenax did not anticipate the impact disease burden would have on the treatment response.
In patients whose baseline walk is below the median of 333 m, the result is not eight meters or three and a half meters. The placebo-corrected improvement below the median is 26.3 m, with a 95% confidence interval of six to 47 m and a nominal P value of 0.0112. Above that line, healthier placebo patients out-walked healthier levosimendan patients by 17.6 m. It's not hard to find patients with limited exercise function in HFpEF practices in clinical trial sites. A patient's track record on this standard assessment is one of several markers of advancing disease their cardiologists follow closely over time. In the results we have reviewed thus far, these markers of severity align in predicting exercise improvement levels from levosimendan, and we are moving forward knowing the six-minute walk distance at baseline, in particular, highlights patient responsiveness to our mechanism of action.
We know much more clearly the right patients to enroll. LEVEL has taught us how impactful this particular baseline characteristic is on drug response in these patients. We set the ceiling for the baseline walk at 450 m, which has been the ceiling for all of the successful PAH trials. Capping the walk at 450 allowed less advanced patients to be enrolled. Our median baseline walk came in about 50 meters higher than in HELP, and roughly half our population, based on that measure of severity and several others associated with it in our pre-specified analyses, ended up being patients healthy enough that the placebo arm improved on its own. In this one aspect, the study design permitted the inclusion of patients with moderate disease. The learning lies right there. We did not fail to find a treatment effect in LEVEL. We diluted one.
Stuart will show you exactly how. Stuart?
Thank you, Chris, and good morning. LEVEL was the largest randomized controlled trial completed to date in PH-HFpEF. As Chris summarized, this study has already taught us several important things nobody knew a month ago. I'm going to take you through all of it, including the pre-specified subgroup analyses that elucidate the route we intend to take to bring levosimendan to the many patients who will benefit. Let me begin with trial design. The LEVEL study was a phase III double-blind randomized placebo-controlled trial. 41 sites in the U.S. and Canada randomized a patient. 241 patients were randomized 1 to 1. Dosing was an oral capsule, 1 milligram twice daily for 4 weeks, and then 1 milligram 3x daily through week 12. The primary endpoint was change in six-minute walk distance at week 12.
Patients completing the double-blind period were eligible for an open label extension of up to two years, which is ongoing. Every patient underwent a right heart catheterization and had to meet qualifying hemodynamic criteria, either at rest, with passive leg raise, or with bicycle exercise. That was our enrichment strategy, and it came directly from the hemodynamic profile of the patients who responded to intravenous levosimendan in HELP. While we enriched on hemodynamics, based on that evidence, we had no basis to enrich the minimum or maximum baseline six-minute walk distance, which I will come back to. Baseline characteristics were broadly balanced. As you see on this slide, mean age 69, 71% female, BMI 33, 72% were on an SGLT2 inhibitor, 28% on a GLP-1 receptor agonist, 60% on an MRA, and 87% on diuretics.
These are the drugs that have shown benefit in patients with HFpEF and are considered the standard of care. 45% were New York Heart Association functional Class 2, and 54% were functional Class 3. This was a contemporary, well-treated HFpEF population, which is what we set out to enroll. 60% qualified on provoked hemodynamics rather than at rest. As to the baseline six-minute walk distance, 319.5 m with a standard deviation of 85. The median is 333 m. I'm going to come back to the 333 number many times. The primary endpoint. Patients on levosimendan improved their six-minute walk distance by least squares mean of 14 meters at week 12. Patients on placebo improved by 10.4 m. The treatment difference was 3.5 m with a standard error of 7.3 and p-value of 0.63.
Using the mean, patients on levosimendan improved by 17.7 m compared to 9.8 m for patients on placebo, with a treatment difference of eight meters. The key secondary endpoint, KCCQ total symptom score, did not reach statistical significance. Levosimendan improved 6.6 points and placebo 6.5 points with a treatment difference of 0.1 points, standard error of 2.2. New York Heart Association functional class improvement was 24.1% on drug and 22.5% on placebo. Adjudicated clinical worsening events were evenly split with three in each arm. On to safety. The oral form of levosimendan demonstrated a favorable safety profile in this population. Adverse events occurred in 86% of levosimendan patients and 72% of placebo patients. Treatment-related adverse events were 38% against 17%. Discontinuations due to an adverse event were 8.3% against 1.7%, and dose reductions were 15% against 4%.
The following effects are characteristic in previous studies with oral levosimendan: headache, palpitations, and hypotension. They were largely mild to moderate, and they were manageable, and they were handled mostly by reducing the dose rather than stopping the drug. Serious adverse events, 10.8% against 10.7%. Adjudicated clinical worsening events, 2.5% in each arm. On arrhythmia, among patients with no preexisting evidence of atrial fibrillation or sustained ventricular tachycardia, there were no new events detected during the randomized phase with repeat Holter monitoring. There were two deaths during the double-blind phase, both in the levosimendan arm, both unwitnessed, both determined to be unrelated to the study drug by the patient's physician. To remind you, the annualized mortality rate reported in HFpEF populations of this age and severity is on the order of 15%. Those are the top-line results and allow me to show you exploratory endpoints and pre-specified subgroup analyses.
These are the findings that give us confidence that there is a clear path forward to a successful Phase III program and approval. The biologic thesis behind this program is that levosimendan's potassium ATP channel activation mechanism reduces the chronic constriction of the vessels in the splanchnic circulation. This constriction leads to volume overload. Dilating the splanchnic circulation will lower the patient's central and pulmonary venous pressures. This reduces strain on the heart and lowers pressure in the lungs. Those patients whose pressures are going down would be expected to demonstrate this through improved exercise. New evidence in this trial not only proves that thesis, it demonstrates an impressive and what we expect replicable treatment effect. NT-proBNP is a highly reproducible measure of cardiac wall stress. This is called strain on the heart, I just mentioned.
Levels of this protein were markedly elevated at baseline in our overall population as expected. We observed a 49% reduction in the patients treated with levosimendan compared to placebo, which was robust. This biomarker is not a subjective measure. It correlates with the severity of heart disease and pulmonary retention in a multitude of studies and is trusted by physicians for diagnosis and treatment. A treatment effect of this magnitude has been associated with improvement in outcomes and is a reliable indicator of treatment efficacy. A treatment effect of the magnitude of levosimendan in LEVEL is larger than any I have seen in a large HFpEF trial. What about pressures in the lungs, as I just mentioned? We measured with echocardiography the right ventricular systolic pressure, or RVSP, also sometimes reported as PASP, pulmonary artery systolic pressure.
Just realize this is the pulmonary artery pressure that you are used to seeing in pulmonary hypertension trials. We observed that the PA pressure was reduced with levosimendan by 3.6 millimeters of mercury, whereas the placebo patients had an increase of a half a millimeter for a treatment difference of 3.5 millimeters of mercury, which was also robust and with a nominal p-value of 0.0045. In patients below the 333 m median at baseline walk, the PA pressure dropped 4.9 millimeters, nominal p-value 0.009. The pulmonary vasodilators that have been approved for Group 1 PH demonstrate the similar reductions in PA pressure in those patients. Lowering pulmonary artery pressure is the most important biologic objective in a therapy for pulmonary hypertension. Population studies in patients with pulmonary hypertension have showed a reduction in the pulmonary artery pressure is associated with a survival benefit.
A paper published just this week by JoAnn Lindenfeld and others demonstrated that a three-millimeter mercury reduction in the PA systolic pressure delivers a 20% reduction in heart failure hospitalizations. The fact that levosimendan was able to markedly reduce BNP and reduce the pulmonary artery pressure with this drug in this population is a critically important finding. When we look to have a better understanding of which patients benefited most from levosimendan, we first look at the baseline six-minute walk distance as a measure of disease burden. The lower the walk distance, the higher the disease burden. Our first step was to look at those patients whose walks were above and below the median, which was 333 m. In those with a six-minute walk below the median, in blue on this slide, the treatment effect was 26.3 m with a 95% confidence interval and nominal P value of 0.0112.
When we looked at the patients with a baseline walk above the median baseline, in red, the final six-minute walk distance in the levosimendan patients showed a treatment difference of negative 17.6 m. Here's what's happening. The drug is helping sicker patients more. What we became aware of in this population is that patients who were less sick and on background therapy did not show treatment benefit from three milligrams daily of levosimendan. Those with a higher burden of disease were very responsive, despite being well-treated on the latest recommended medical therapies for HFpEF, with a very meaningful clinical response. The six-minute walk test was reliable in conveying to us the patients with Group 2 PH who respond best to this drug. When we looked by age, in the oldest tertile above 74, the difference was 37.6 m, 95% confidence interval, and nominal P-value of 0.0013.
Look at the confidence interval there in the older patients in this trial. In patients at or above the median age of 71, the difference was 27.1 m, 95% confidence interval, nominal P-value 0.0021. This compared with patients below the median age of 71, the difference was negative 19.2 m, 95% confidence interval, nominal P-value 0.0972. This finding about age persists whether we look at median, tertiles, or quartiles. We also look at natriuretic peptide as a biomarker of disease burden. The higher the BNP, the more severe the patient. Patients who came in above the median baseline NT-proBNP improved 16.7 m against placebo. That is the same gradient, again, arriving from a completely different direction, from a blood test rather than a walk test or a birthday. Thus, these three markers of disease burden, exercise capacity, age, and natriuretic peptide, point us to our responder population.
I remind you, these are nominal P values. They are not adjusted for multiple comparisons and the multivariable work that is still running, they do provide answers to questions that were specified in advance rather than derived from data review afterward. There is a practical finding sitting inside this that I don't want anyone to miss. The six-minute walk distance turns out to be both our endpoint and our best tool for selecting patients. We do not need to go and find a new biomarker to identify who responds. We need to take this measurement at screening and use it to decide who gets into the trial. We relied on change in six-minute walk as a solid primary endpoint in our studies, and these results confirm its trustworthiness. The biologic measures of improvement parallel the clinical efficacy benefits, the results bearing this out.
In summary, we believe that levosimendan has now proven itself as a safe and effective therapy in patients with PH-HFpEF who have more disease burden. In those patients who had lower baseline walks, NT-proBNP was higher to start with, and it came down dramatically on therapy. This very impressive clinical treatment effect demonstrated here is of a scale we trust would satisfy regulatory authorities for product approval. There was not enough HFpEF literature to rely on when refining the upper range of this criterion. We also had to rely on PH literature. Remember, in any HFpEF study, many of patients have PH that may not be described. It's very difficult to tease out in these databases how much the pulmonary hypertension versus the HFpEF is driving walk response. We just don't have a trove of successful Group 2 trials to guide us.
On the other hand, in PAH, FDA guidance suggests drug developers use a range of 165 to 450 m, that guidance draws on libraries of successful PAH trial databases. In Group 2 PH, we now have a very clear indication of where to set the ceiling. The Tenax team is exploring now whether an amendment to the ongoing LEVEL study would generate a data package that, when submitted to authorities, will bring this therapy to patients. Keep in mind that LEVEL-2 was sized based on the safety database requirements and is already powered above 90%, even with the 55% standard deviation and an assumed dropout rate more than four times what we saw in the shorter trial. LEVEL randomized 241 subjects. LEVEL-2 currently aims to enroll more than double that number.
We are optimistic about our ability to modify this trial once regulators have reviewed our LEVEL results and heard our proposal so that we can generate a successful data package and get this drug to needy patients. With that, I am very pleased to introduce Dr. Sanjiv Shah. Sanjiv, we would value your thoughts about the meaning of these results.
Thank you, Stuart. It's really a pleasure to be here and discuss my thoughts on the trial result as the academic principal investigator of the trial. I've been taking care of these patients for the past 25 years, and in the past 20 years, directing the Northwestern HFpEF program. As you are well aware, this is the dominant form of heart failure now. We think that based on the current data, 60% of patients in the United States alone have a preserved ejection fraction greater than or equal to 50% at the time of diagnosis. Of those, we think that 75%-80% have pulmonary hypertension. It's just extremely common, and there are no approved therapies for PH-HFpEF.
Of course, for the past two decades and longer, we've been looking for drugs that can improve this patient population, and it's been littered with neutral trials or even some trials that harm the patient. It's been a really challenging patient population. We have had successes recently with SGLT2 inhibitors, GLP-1 receptor agonists, non-steroidal MRAs, and that's been a great advance. Some may say, "Well, maybe we don't need to do anything more for HFpEF," but clearly we do. Even in these trials, even on patients with all of these medications, they're still quite debilitated, and they still have high morbidity and mortality. Those who develop pulmonary hypertension, and the more severe the pulmonary hypertension is, are the ones that really suffer and the ones that need the most help.
Now, in the past, we have tried with pulmonary vasodilators to see if we could help these patients with PH-HFpEF like we've helped patients with pulmonary arterial hypertension. It's been a universal no, not an improvement at all. What's consistent about all of these trials is that NT-proBNP didn't come down. In fact, it either stayed the same or went slightly up on the treatment. This is the case for drugs like bolen relaxin in the recent trial that was published last year, PDE5 inhibitors with endothelin receptor agonists, sGC antagonists, and sGC stimulators. There's a clear signal there that there is not an improvement of congestion. There's not an improvement of wall stress. Those trials were neutral and didn't improve the patients.
We look at the recently published CADENCE trial, which was in patients with a pulmonary vascular resistance of greater than four, which is quite high and is the minority of patients with PH-HFpEF, but we do see those patients. They're quite sick. The CADENCE trial did reduce PVR and seemed to improve other markers, including NT-proBNP came down, but not nearly as much as we see here with levosimendan, which is a much easier-to-administer oral medication that seems to be better tolerated. While I felt that six-minute walk distance, if we could show it with this number of patients in a wide proportion of patients with PH-HFpEF, sort of a broadly defined patient population, we would have a major advance. Sometimes we get lucky, and it's easy, and it's simple.
Here, what we saw was what we often see in HFpEF, that it's not simple, that it is a heterogeneous patient population, what we see here is that sicker patients really are the ones who benefit. That's the most important finding because those are the ones that we need to treat. If we have a patient who has a six-minute walk distance of, let's say, 375, 400 meters, 425 meters with PH-HFpEF, they're doing pretty well, and we have treatments for those patients. We have SGLT2 inhibitors. We have GLP-1 receptor agonists, non-steroidal MRAs, those are not the patients that are having the events like heart failure hospitalizations and higher mortality.
The ones that we want to treat are the ones with the lower six-minute walk distance with the higher NT-proBNPs in the sense that those are the ones that are at highest risk for heart failure, hospitalization, and death. Those are the ones that are costing the healthcare system so much and are really the ones where the disease of HFpEF and pulmonary hypertension is driving the patient's journey. In those patients in particular is where we found the benefit in the LEVEL trial, those with the lower six-minute walk distance, those with the higher NT-proBNP. I think that these findings, while they're not a slam dunk and an easy win for us that will quickly move us forward on the path, I think that what they show is actually better than we would have expected.
Based on these data, I feel like we have evidence here that levosimendan will reduce heart failure hospitalizations in PH-HFpEF. I say that because of the reduction in NT-proBNP. There's sort of a linear relationship across all trials, not just the big large outcomes trials, but all trials that have looked at this. That linear relationship suggests that the reduction in NT-proBNP that we see here, which by far is the biggest reduction in NT-proBNP in any PH-HFpEF or HFpEF trial ever done, will reduce heart failure events dramatically. Of course, that needs to be proven. We need to do that trial. To me, that is really compelling. You couple that with the fact that those with a six-minute walk distance at baseline less than 333 m did quite well, 26-m improvement compared to placebo.
I think it gives us the whole package there and tells us that this drug really has a future. I'm still a big believer in the drug, and I've done many of these trials in HFpEF, and I take care of many of these patients. We see that in the open label extension, that patients feel well and are doing better. I'm still quite positive about the drug and think that we really have a lot of work cut out for us. I'm very positive and hopeful for our patients based on the results that we've seen here today. Thank you.
Thank you very much, Professor Shah. I know that in a few minutes, we'll have questions, and we'll probably be directing a few of them to you. We appreciate your support, we are also reminding everyone that there will be data coming out in the future in publications and in presentations, we're not going to steal too much of your thunder here today. We're looking forward to hearing more about that linear relationship between BNP and heart failure events, we'll wait for the late breaker, et cetera. All right. With the last few minutes of the call, we want to do three things: tell you what we know now, tell you what we are going to do about it, and tell you what we cannot answer yet today. What we know.
We have a drug candidate that lowers filling pressures and pulmonary pressures in this disease at a dose we no longer have questions about, with a safety profile consistent with over 25 years of clinical use and study. We have an efficacy signal above 26 m in the patients who aren't walking near the normal range, and we have several ways to identify those patients before they enroll. The question that kicked off the PH-HFpEF program at Tenax in 2018 was, what might levosimendan do in a totally new group of heart failure patients without a drug and with high unmet need? Numerous health findings led us straight into phase III. In LEVEL, the question so many of you asked was whether oral levosimendan at the 3-milligram daily dose would transfer those benefits. That question is now retired.
What is left is a trial design problem, a new challenge our team will overcome to deliver this drug to patients. Trial design is the kind of problem we know we can solve, our aim will remain shortening the time patients have to wait. I want to be clear about that 26-m signal because I worry it will get lost beneath the headline. 26 m is not a rounding error we are trying to talk our way into. It showed up in the patients who have the most to gain. It moved with two independent biological markers that dramatically improved in this population, it is about the same magnitude HELP produced in about the same kind of patient. In the patient with highest indications of disease burden and LEVEL, we saw effects in the 30-40-m range. The RVSP and BNP improvements are profound.
Those improvements are evident across all quartiles of levosimendan patients we studied, they are undetectable in the placebo patients. We have pressure tested these observations already with the data we have so far, I can tell you our pathway is very clear. What are our next steps here? Tenax will request a Type C meeting with FDA to present these findings and our recommendations. We are seeking, in parallel, scientific consultation from EMA. We will go into these meetings with the complete dataset, including the multivariable analyses we don't yet have, and with our scientific advisory board and regulatory advisors aligned behind a recommendation. Our statistic experts are working on the right strategy to construct criteria for an impressive LEVEL-2 based on several disease severity factors they and the medical team are analyzing now.
Their goal is to create a correlated set of factors, preserving as much of the population as possible on the way to a positive result. They will avoid overfitting by using modeling techniques that repeatedly test our rich new LEVEL datasets through these filters. Tenax's statistics leaders inform us we have strong evidence, undergirded by impact on NT-proBNP that is clearly translating to exercise improvement for designing a successful next trial. The predefined analysis reveals demonstrable levosimendan effects in those with higher disease burden at the start of the study. The pre-specified analyses consistently show this positive effect. This is not exploratory. I should pause for a second here to thank the incredible data management team at Medpace and the biostatistics team at Calian, who've worked tirelessly to get us where we are this morning.
Dr. Shah plans to present the LEVEL results at ESC, where the community will next learn quite a bit more about LEVEL, he and the steering committee intend to continue presenting and publishing their findings. The one thing I will commit to this morning, something that doesn't require a validated staff output or a regulator's input, we are not intending to enroll another patient into this program whose baseline six-minute walk distance sits in the range that diluted LEVEL. In the COPD, heart failure, and PAH literature, there is an association between a walk of less than 350 m and increased mortality. In ATTR cardiac amyloidosis, a walk less than 350 carries a 2.2-fold higher mortality risk. There is plenty to support lowering our ceiling to something in this range.
We're finalizing the exact threshold quickly based on our new LEVEL data, when that number is fixed in the coming days, the steering committee will communicate it to every active LEVEL-2 site. What can I not tell you today? We have yet to decide whether the right vehicle is simply an amendment to LEVEL-2, which is enrolling now, or an amendment and the addition of another trial built from the ground up for this population. Several options are under consideration. We may limit the walk ceiling, or we may create a combined disease severity score to avoid enrolling patients who are a few years ahead of the point in disease progression where levosimendan is going to help them. Amending LEVEL-2 preserves what we have already built.
Initiated sites around the globe, a highly functioning group of national lead investigators, a running supply chain delivering IT, training, ethics, and site approvals and randomizations. A continually building footprint is active across four regions. A trial population built on the findings of LEVEL could enroll at a reasonable rate, and we would not be starting from scratch on site activation. These are key decisions that we're going to make with the FDA rather than ahead of them and not on today's call a few days after we were unblinded. We believe the agency will see that oral levosimendan does what we said it would do at a dose we know is driving big improvements in patients who heart failure physicians clearly recognize in their practices and who our sites will now identify before they ever enroll.
With an effect size larger than the one we powered LEVEL to detect. The FDA wants therapies for the millions of Americans suffering from PH-HFpEF. The data we have shown you this morning tells us what needs to get done. We intend to develop and ultimately deliver an authorized oral levosimendan to these patients because this drug has proven itself to us in our data. To everyone at Tenax, thank you and keep going. To the investigators and the coordinators, thank you. We are very grateful to the patients making this research possible. Thank you for listening, and Betsy, we will take questions now.
We will now begin the question and answer session. To ask a question, you may press star then one on your touchtone phone. If you are using a speakerphone, please pick up your handset before pressing the keys. If at any time your question has been addressed and you would like to withdraw your question, please press star then two. At this time, we will pause momentarily to assemble our roster. The first question today comes from Olivia Saunders with Cantor. Please go ahead.
Hi. Good morning, guys. Thank you for the questions. I wanted to ask, what is your best explanation for why those higher baseline patients, AKA the third and fourth quartile on walk score, are seeing such a pronounced worsening effect versus placebo? Is that something that you also saw on KCCQ scores, or is that just a six-minute walk distance phenomenon? Then also wanted to ask about how each arm did in that lower disease burden patient population and whether those patients on TNX-103 actually saw a decline versus baseline six-minute walk, or whether the placebo patients maybe just improved disproportionately, or maybe it was some combination of both. Thank you.
That's a great question, Olivia. A great set of questions. I'll hand over to Stuart. Let me just mention to everyone a couple of things. The Tenax team was unblinded by our stats team last week. We have been pretty much heads down through today with our effort to get the data where you saw it so that it's in front of you and everything is clear. There are going to be detailed questions we can't answer. There are also going to be really important conceptual conversation topics like the one you just brought up that fit in that category. If we can't be certain here today that we can quickly lay our hands on accurate data, we may punt on a few of those things.
Second, again, please remember, that sort of analysis is a great question for the SAB to answer in upcoming presentations and published topics. It's an open one. I'll hand it off to Stuart here, and then maybe if Sanjiv has comments on those topics. It's KCCQ, it's the patients with the highest walk at baseline who seem to go the wrong direction.
Thanks, Chris. Olivia, that's a lot of questions, and let me kind of go through one at a time, and I may have to ask you to repeat some of them. You have to imagine, the one thing that I focused on right from the get-go was why the ones who were less sick and received levosimendan did so poorly. Because we have convincing data that they were taking the drug, and we have convincing data the drug had a biologic effect on their BNP. In this population, where these patients have congestion, we talk about the preload effect on the heart and lungs, which means the blood volume coming to the heart and lungs in these patients. As you know, we talked about our mechanism being to reduce that blood volume.
There is some literature suggesting that in these patients, if you reduce the blood volume too much, they actually develop worsening symptoms, and it's a term preload insufficiency, which is relatively new and being talked about a lot in the literature. To come up with something that makes sense, and I'm not saying this is actually going to pan out, is that the drug worked too well in these patients. It dropped their preload to a point that they didn't have enough filling. Again, if you don't have enough filling, that will limit your ability to walk. That fits the drug's mechanism. It fits the disease and may turn out to be the answer to the question. As we've already alluded to, our drug would not then be the ideal treatment for these patients at this time.
If they get worse over time, then it's a different story. Okay. Next question, Olivia.
Thanks, Stuart. Very helpful. Yeah, the other question was just around KCCQ scores, whether you saw maybe more pronounced benefit on KCCQ scores in those patients that do have the higher baseline six-minute walk or higher disease burden, however you all are categorizing it.
Here's what we'll tell you about KCCQ at this point, along the lines of the topic you just raised. One of the things that we can do in looking at the patients in each group on each treatment over time and their responses, which right there you know you're talking about a multivariable analysis, right? Is also look at their KCCQ and see if in those patients that Stuart just described, who we may discovered here were not preload dependent, et cetera, that their KCCQ expresses what their walk expresses. We know that the patients in that group who walk high at baseline. Well, you know what? Let me ask Sanjiv. Sanjiv, are you still there, and do you have a theory on this?
Yeah. First of all, they're great questions, and we still have a lot of work to do to analyze all of this. I would say a few things. First of all, I think what Stuart said is very possible. The patients who walk far at baseline and have lower filling pressures are different type of patients, and they may sort of depend a bit on this sort of splanchnic phase of constriction during exercise to help them get to that walk distance. You remove that or you lessen that, and they can't do that. That's something physiologic that we'll have to figure out and take a look at. Also, when you look at the quartiles, the 95% confidence interval at the higher end of walk distance, levosimendan versus placebo, I think crosses unity.
We're not as confident at that higher range as we are at the lower range, where it is very much the confidence intervals don't pass and you see a very statistically significant improvement. I think the last point that you made is sort of like, is the issue that these patients are really worsening, or is the issue that the placebo is getting better? From my analysis of the data, that last is seemingly what's going on. It's not that levosimendan is harming these patients in some way. It's just that these placebo patients, when they're already doing quite well, they kind of did a little bit better over time. Is that a fluke or something? I don't know.
In the ones with the lower six-minute walk distance, those are the ones where it really does seem like there's a clear improvement on levosimendan. I think there's still a lot that we have to unpack here. I'm hoping that by the end of the month, when I present it at ESC, we'll have a lot more of these answers.
Great. Thank you.
Thank you. Appreciate all the thoughts.
Okay. Betsy, next?
The next question comes from David Risinger with Leerink Partners. Please go ahead.
Yes, thanks very much. I guess, Dr. Shah, to follow up. Regarding slide 13, do we need to wait for ESC to understand the absolute figures behind the calculation so that we can assess to what degree the above median and below median performance was driven by the placebo responses? That's my first question. I have another one, please.
We'll try to go to 13. No? All right. I think people will remember it. It's the blue and the red walk bar graph. Dr. Shah can address it in a second. Let me address it in one way, Dave. It's a great question. Let me just share a little bit of the observed data over time that sits behind this bar graph, right? What we're looking at is 12-week data. Let me give you the walk distance in observed data. This is raw data, not statistically adjusted, et cetera, but these are just the numbers that we have at four weeks, eight weeks, and 12 weeks. In the levo patients below the median baseline at entry into the study, the improvements are 19, 21, and 27 m. Right? They're on two milligrams. They walk 19 further.
They go up to three, they go to 21 and 27. Do they stay there? Do they go up? Do they go down after week 12? Your guess is as good as mine. We, of course, have the OLE data. What about the placebo patients who walked less than the baseline? Their walks are -1, 4.8, and -3. A tight clustering of unimpressive numbers. The same trends are available in different analyses, and that's part of the data that we're looking at initially here that points to the same conclusions that that's our divider. Professor Shah, we're not showing all the data behind this now because we're still getting through it, but what would you expect one will see, I think?
I think that's right. It's not that the patients who benefited Let's look at the lowest quartile of baseline six-minute walk distance. I think the question that we just heard is it that the levosimendan patients are truly doing better, or is it that the placebo patients in that group are just doing a lot worse and we're preventing worsening? It looks like the levosimendan-treated patients are actually doing better, which is really kind of amazing to see because a lot of times when we have really sick patient populations like ATTR cardiomyopathy, we're just sort of preventing or slowing worsening. Here, I think we're seeing an improvement. Again, we've got to really dive into those data and tease it out. That's what Chris had just mentioned. That's what we're seeing. Placebo remains flat and the levosimendan group improves over those 12 weeks.
Thank you. Then just to follow up. With respect to LEVEL-2, how many patients have been dosed in LEVEL-2 to date? What do you anticipate FDA and EMA's receptivity will be to a protocol amendment? Stuart, could you remind us about FDA's historical advice on a required P value to file with a single successful trial? Thank you.
Great question, David. We'll answer two of the three. As in the past, we've not disclosed information like where we are with enrollment. As far as what do I anticipate the FDA's susceptibility is to our arguments. One is wise not to do that. We believe that when they see our data, we'll be in good shape. Stuart, why don't you take on the last question there about the previous discussions on P value, et cetera.
David, the guidance that we have in agreement with FDA for LEVEL for the program was they would agree to a single trial with a PO1 endpoint or two trials at PO5. As you know, we sought the latter for several reasons for our program, and a lot of it had to do with the resources, because as you know, when we had a big raise, we just started the second trial right away. I think the question on mind is what the FDA will now say regarding our program, and it's too early for us to comment. We've been talking internally about it. We have some very prestigious FDA advisors on this, and so we're confident that we will come to them with a very compelling program.
I'll give you one more piece of information that came from stats on the weekend. Just looking at based on the new data we have, looking at the powering calculations of LEVEL-2, with the dropout rate being higher than we anticipate it will be with the standard deviation of even 55 and a treatment effect like 25, we're well above 95% power with the size of that trial. All right. Thank you, David. Next question.
The next question comes from Yasir Rehmani with Piper Sandler. Please go ahead.
Good morning, team. Thank you so much for all the color. A few questions for you, team. Maybe I'll go one by one. The first one is, do you know if the baseline six-minute walk test correlated whether patients were enrolled via exercise or qualified via rest?
Okay, we'll go one by one that way. Here's what we'll tell you. We have started looking at this issue that has been discussed a lot in the last couple of months of whether the position of the patient that qualified them for the trial might have biased us, et cetera. What we have seen so far tells us that our criteria are fine. We got a lot more digging there to do, but it's a good question.
Thank you, Chris. The second one, I think Dr. Sanjiv noted, right? When we look across the subgroups that levosimendan had a profound improvement and the improvements were not due to placebo, but just drug effect when you broke down the six-minute walk test based on quartiles. I guess that raises the question, would you have on hand what the non-placebo adjusted six-minute walk distance benefits were in that post-hoc subgroup analysis that you broke down among quartiles?
No, we're not sharing that at this time yet. I think we just need more time for that. It's a good question.
Okay. Maybe the last one is, team, I know you're working diligently to figure out what the new cutoff is going to be for the six-minute walk test, but have you had a chance to look at LEVEL-2, where you are in regards to total number enrolled and what their median looks like, and if you did pick 333, what % of the patients would be in that? I don't know if you've done that exercise yet or not.
No. I think your question is a lot like Dr. Risinger's, right? Where are you going to do the cut to get the impact you want, et cetera? The data's there. We will not have a problem with quantity of reliable data, but we will also not rush into it. I'll remind you guys that in LEVEL, early in the trial, we decided we would amend the protocol to require at least 90 days on therapy with GLP-1s or SGLT2s. We implemented that through an immediate action so that it stopped happening, and it worked. Months later, sites continued enrolling and they basically stuck to the new requirement in the protocol. We can do the same thing here on the walk distance.
In other words, it can be a very rapidly implemented change that will show up in a future amendment, but the investigators listen to the steering committee on a topic like that. We're in good shape.
Got it. Thank you, Chris.
Operator, next? Thanks.
The next.
The next question comes from Seamus Fernandez with Guggenheim Securities. Please go ahead.
Great. Thanks for the question. Just hoping you guys could give us a little bit more color on what you hope to show at ESC in particular, not in terms of the data, but the subset analyses. It seems to me like we'll need to know a lot more information with regard to the pre-specified analysis above and below in terms of the baseline drug characteristics and the patients that were on the different potential drugs that could impact performance. Obviously, we got a little bit more information in terms of the placebo performance as it relates to the below 333. I guess the question is, did the patients on placebo improve in the upper bound group?
I'm sure everybody's very interested in all of that, but it seems like baseline characteristics in terms of the patients on drug or perhaps even mid-trial, although only 12 weeks drop in, could also play a role here. Just also wondering when you sort of run the statistics against multiple comparisons, things like that, I know that that's challenging given the fact that these were not statistically significant results, but it seems like it's necessary to add in statistical evaluations of multiple comparisons here just because there are factors that come into play. When we see that P value of 0.01, how statistically valid is that pre-specified analysis? Thanks.
Okay, good. Thanks, Seamus. I'll hand to Sanjiv in a second. I'll say just very quickly, though, there's sort of two things you do when you get your top-line results. We've got the full analysis set. You immediately start asking for the ad hoc data you want next. You think and you wait. A lot of the things you've asked about are in the top-line data set. What we've really shown here today are pre-specified analyses. There's a lot there already, but you still have to think about it, right? Two weeks between now and ESC is quick. Sanjiv, can you respond to some of those topics and at least give folks an idea of what they might see at that time based on what you and your co-authors want to do in that presentation?
Yeah, well, I just want to temper expectations. I think we get eight minutes total for the entire presentation. A lot of it is going to be what you see here today. We just don't have enough time to go through everything, but the papers will have it. We are well underway of writing papers on the primary results and NT-proBNP. I suspect that we'll do a deep dive in the baseline six-minute walk distance categories. I think in the next few months, hopefully you'll start seeing those publications come out. We got the results too close to the meeting to do a simultaneous publication. I think given the interest that journals will be picking this up to publish it.
I think that the multiple comparisons question is fair, and that's the reason why we pre-specified the number of subgroup analyses we are doing for the primary endpoint, which is in the statistical analysis plans. We can go ahead and control those subgroup analyses for multiple comparisons. If we do that, just based on what we've seen so far in terms of the interaction P value for that baseline six-minute walk distance, I think it'll stand up to that. That's something that we and the statisticians have to calculate still.
Thank you so much, Professor Shah. Is there another questioner? Operator, have you got any other questions? I'm not hearing from Betsy. Operator, are you there? I wonder if we ran out of time here, folks.
Yeah. It'd be.
Megan? Folks, it looks like we've lost our operator, and that may make it hard for us to hear questions. Just give us a moment here to try to keep connected. Okay, looks like an IT issue is being worked through at the conference center, so it may just take another few seconds, folks. We know we'll also be talking with two or three of you later today, so we can answer questions there. All right, folks. It's a Monday just after nine A.M. I know how busy you are. Our operators aren't able to get on, so I think we've lost the ability to talk with you here. Sanjiv, thank you so much for your being available this morning. I know you're heading out to an academic meeting.
We will look forward to talking to you folks in the future, and thank you very much for joining the call today.
Thank you.