Inventiva S.A. (IVA) Earnings Call Transcript & Summary

Ladies and gentlemen, thank you for standing by, and welcome to the Inventiva's KOL webcast event from AASLD. [Operator Instructions] I must advise you that this conference is being recorded today, and I would now like to turn the conference over to your speaker today, Frédéric Cren. Please go ahead, sir.

Thank you, Petra, and welcome, everybody, to this webcast from AASLD 2020. As you know, for Inventiva, it's been very positive 5 months, started in mid-June with positive data from the NATIVE study with lanifibranor meeting the 2 regulatory endpoint that are needed or required by FDA and EMA for the conditional approval. On the back of that, we're extremely pleased to get breakthrough designation from FDA and also released last week, the key summary from the end of Phase IIb meeting we had with the FDA, which confirm that we need only one single Phase III to get approval. And very importantly, that we're eligible to get accelerated approval based on biopsy. So of course, on the base of the success, it's in the -- very appropriate for us to organize this webcast, and we're absolutely very pleased to welcome tonight 3 very talented speakers that have taken time out of their busy SLV schedule to be with us. So we will -- you will hear from Professor Francque about the NATIVE study. It will be a copy or -- of the presentation he gave at the Presidential Plenary on Sunday. Then Professor Sanyal will give an overview of the NASH feed and especially about Phase III design, and then we hear from Professor Cusi, how he treats NASH patient and how he feels PPAR can play a role in its track. So let's now move -- yes, before we move to trend. There will be a Q&A session at the end. So just please specify to which KOL you would like to ask your questions. So that will be easier for them to answer. So thank you very much. And now we move to Sven for an overview of the NATIVE trial.

Yes. Thank you very much. Can we have the next slide? So as Frédéric said, I will mainly present the results of the NATIVE study, as I have presented them yesterday at the Plenary session at the AASLD. Next slide. So my way to the introduction, the PPARS or the peroxisome proliferator-activated receptors. That's the clot of of nuclear receptors consisting of 3 isotypes, and they are differentially expressed in different tissues and different cell types that you can see from this quite complicated cartoon, but this complexity is important. Now with that said, you have 3 isotypes. You have PPAR alpha, which is mainly, but again far from exclusively expressed in the hepatocytes and in the brown adipose tissue. There's PPAR gamma, which is expressed in stellate cells in the liver and also in the adipose tissue and we have delta, which is expressed also [indiscernible], immune cells and in muscle. So it's a complex scheme of key regulators of metabolistic inflammatory processes and fibrogenesis. And as such, they are deeply involved in the petrogenesis of NASH. And this is important because lanifibranor as a pan-PPAR acts on the 3 isotypes and at different sites, not only in the liver, but also outside the liver, which means that it potentially tackles several key mechanisms of the pathogenesis of NASH at several levels, and this is potentially highly relevant for the efficacy of lanifibranor as you will see. We can move to the next slide. So lanifibranor, as a chemical structure, is an indoor sulfonamide. So it's a different chemical structure compared to some of the other PPAR agonists that are listed on the table. It was tested, of course, in Phase I, and then subsequently, patients with type 2 diabetes and also in scleroderma, and in this disease, the patients were treated up to 48 weeks and the safety profile and tolerability profile in these patients, even after 1 year of treatment, was favorable. We can go to the next slide. Here you can see the design of the NATIVE trial. So patients were included based on qualifying level biopsy, and I will discuss the criteria of inclusion on histology in subsequent slide. And then they were randomized 1:1:1 to receive placebo or lanifibranor 800 milligram once daily or 1,200 milligram once daily. So 3 treatment comps. Treatment period was 24 weeks. And thereafter, there was a repeat biopsy. So for third biopsy studies for the short term treatment. And treatment efficacy was, hence, assessed based on histological endpoints. I will mainly present the results on the full analysis set, which includes all patients that have been randomized and received at least one dose of placebo or drug, and this is 247 patients. In this analysis, looking on the treatment biopsies we'll consider nonresponders. The per protocol population, which are all patients that have had both a baseline and a readable and a treatment biopsy and our other protocol deviation impacting on these efficacy results consisted of 194 patients. Next slide. Now in this trial, we used the SAF scoring system and not the NASH CRN method activities score for the definition of the inclusion criteria and the primary endpoint. Why so? We know, of course, that driving force of this disease [indiscernible] processes of cell damage and information that are key to the progression of the disease. The question is, of course, how best to capture these processes? Well, I think the field is still not perfect in identifying that, but on histology. Those processes of cell damage and the explanation are probably best reflected by the ballooning of the cells and the lobular information. And of course, ketosis is also a very important feature of this disease, but still those does not always equal liver cell damage and inflammation. Now in the NASH CRN network activity score, or NASH, the 3 cardinal features of the disease of the lobular inflammation, the ballooning and steatosis are combined, whereas in the SAF scoring system, activity only refers to the sum of ballooning and lobular inflammation. So in the SAF scoring system, steatosis is assessed separately. Also in the NASH CRN scoring system, lobular inflammation can be scored up to 3 and ballooning up to 2. Whereas in the SAF scoring system for both lobular inflammation and ballooning get an equal rate and a maximum score, 2 each. This means that the SAF-activity score ranges from 0 to 4, and we'll only allow it in a NATIVE trial, patients with the 2 highest scores for activity, mainly 3 and 4, so active disease. Next slide. As I said, this was a trial that lasted for 24 weeks of treatment, and we based the inclusion on the activity to really select patients with highly active disease and also for this rather short-term treatment in a paired biopsy study, the primary endpoint also focused on the activity -- on the reduction of the activity, of course. So the primary endpoint based on that same SAF-activity scoring system and requires an at least 2-point reduction of the SAF-activity score. Besides that, we incorporated classical histological endpoints as secondary endpoints. So you know that in Phase III non-cirrhotic NASH trials, regulatory endpoints on NASH resolution without worsening of fibrosis and also an at least one stage improvement of fibrosis with no worsening of NASH, and these 2 endpoints were hence included as secondary endpoints. We also included the composite endpoint of NASH resolution and concomitant fibrosis improvement and one in the same patient also as a secondary endpoint. And of course, we looked at liver enzymes and biomarkers and also in this population and with this drug. We also looked, of course, at lipid profile and parameters of glycemic control as outcome measures. Next slide. In total, 247 patients were randomized, equally distributed, of course, in the 3 arms of the study. There were low numbers of withdrawal and they were comparable across arms so the patient population was well-equilibrated over the 3 treatment arms. Next slide. So this is of course, the classical table of baseline characteristics, a little bit busy, but what it mainly shows you is that there are no significant differences between the 3 arms. Over 90% of patients are of Caucasian decent and more than 40% of the patients are living with diabetes and diabetes -- presence of diabetes results in a stratification factor. So this is well balanced across arms. Now as I said, we defined the inclusion based on the SAF-activity score, but if you look at the classical NASH CRN activity score, you can see that by this selection, you include over 70% of patients that have enough activity score NASH CRN equal or above 6. So this is highly active NASH. In terms of fibrosis, we excluded the presence of cirrhosis, so F4, but we did not set a priory, a lower limit for fibrosis. Despite the fact that we did not set a lower limit for fibrosis, the selection based on the high activity of the disease with the SAF scoring system resulted in recruiting more than 70% of F2 and F3 patients in the 3 arms. Next slide. So we now move to the results. So first of all, of course, the primary endpoint, you can see that there was a dose-dependent and significant effect of lanifibranor on the primary endpoint of reduction in disease activity and then disease activity was only encompassing lobular inflammation and ballooning in this definition of the primary endpoint. So this primary endpoint was convincingly met, both in full analysis set as input protocol. If we restrict the analysis to F2, F3, which I will show you for other endpoints, you get the same results. And also the presence of diabetes did not impact on the results of patients living with diabetes responded equally well to both doses of lanifibranor as did patients without diabetes. Now in the next slide. If we can have the next slide, you will see summarize the results on the classical endpoints, and this is only the full analysis that's depicted here. As you can see, lanifibranor has a dose-dependent significant effect on NASH resolution without worsening the fibrosis, with quite a large effect size, more than 25% for the highest dose compared to placebo. Also, the endpoint of improvement of at least one stage of fibrosis without worsening of NASH was met. So lanifibranor is the first drug in Phase II that shows efficacy on both these regulatory endpoints. And again, this is after 24 weeks of treatment. Looking at the composite endpoint, which is the combination of natural solution and fibrosis improvement, you also see that the drug is significantly efficacious dose dependently with an almost 25% effect size compared to placebo foreseeable for the 1,200-milligram arm. Next slide. The next slide shows you the same analysis that restricted to patients with F2 and F3, which is the classical target population that we currently enroll in non-cirrhotic Phase III trial. And as I said, you get the same overall picture and the significances and effect sizes are even higher than if you take the whole -- full analysis to the whole population. Next slide. The efficacy on this histological endpoint, as I showed you in the preceding slide, it's also supported by a significant decrease in liver enzymes. So [indiscernible]. And as you can see from the curves if they appear on your screen, you see an early decline already at week 4, and that improvement is maintained throughout the trial. Next slide. The results are also further supported by improvements in other relevant biomarkers of fibrosis, apoptosis and inflammation, like Pro-C3, the TIMP-1 and MMP-2 ratio, for example, or CK18 and by ways of example, you can see here the results for the Pro-C3 and the CT18-M30 fragments. Next slide. It's, of course, a population that is at high risk of cardiovascular disease. There is metabolic syndrome or a lot of components of the metabolic syndrome. So it's important, both from investigative point of view as from a safety point of view, to look at some aspects of cardiovascular risk factors. And as you can see, lanifibranor and -- it is also what you would expect with the mobile action and the impact on 3 PPAR isotypes, it's improved the lipid profile with a significant increase in HDL cholesterol and a decrease in triglycerides and that's what we've seen for the liver enzymes, you see already important improvements at week 4 of treatment, and they are sustained throughout the treatment. Next slide. If we then look for the parameters of glycemic control, we also see significant improvements in fasting glucose, fasting insulin and insulin systems and in HbA1c and this is particularly relevant and also confirmed in the patient group with diabetes. Next slide. The safety is, of course, also a very important point to discuss. Now as you can see from this stable treatment emergent adverse events were numerically more reported in the active treatment arms compared to placebo, but the rates of TEAE is leading to drug withdrawal were very low, lower than 5% in all arms, and there was no difference between active treatment arms or placebo and drug-related TEAEs, so as assessed by Inventiva investigators were even only reported in the placebo arm. The most frequently reported adverse events with active treatments were gastrointestinal symptoms and headache. Lanifibranor is a pan-PPAR with the PPAR gamma component, and as you know, that will lead to an improvement in the function of the adipose tissue. And that improvement in the function of adipose tissue, which is reflected, amongst others, by an increase in adiponectin levels, is probably part of the mechanism that lead to the efficacy that I have showed you. It also implies that a better function of -- especially the subcutaneous adipose tissue, you have better expandability of the adipose tissue. So there will be a shift from the ectopic fat and the visceral fat to the subcutaneous set, and this is reflected in a weight gain that was observed of 2.4 and 2.7 kilograms on the 800 and 1,200 milligrams, respectively. Peripheral edema was reported more frequently on active treatment than on placebo, but with only 2 cases in both treatment arms that were considered as drug-related. And all but one were mild to moderate, and they could be easily managed and did not lead to drug discontinuation. Next slide. So in conclusion for this results, main results of lanifibranor 24-week treatments in patients with non-cirrhotic NASH. First of all, clearly, lanifibranor significantly reduced the activity of the steatohepatitis and again, not only of the 24 weeks of treatment. It's the first drug that reports efficacy on both the endpoints of NASH resolution without worsening of fibrosis and fibrosis improvement of at least one stage without worsening of NASH, and the 2 regulatory end points for conditional approval in non-cirrhotic NASH Phase III trial. The drug is also very significantly positive on the composite endpoint of NASH resolution and fibrosis improvement. As I showed you, the results on histology are supported by improvements that were observed in different enzymes and in relevant biomarkers. Importantly, in the population at high cardiovascular risk, lanifibranor improved lipid profile and glycemic control. Previous data, as I mentioned in one of the first slides on safety and tolerability were confirmed in this population of NASH patients that were treated for 24 weeks. And as you heard also from Frédéric, this is probably one of the considerations that led to a breakthrough therapy designation that lanifibranor has received a few weeks ago. Phase III trial of this compound, of course, we expect in the first half of 2021, but I suppose that will be discussed later on. I also pointed out why we use the SAF scoring system for inclusion in primary endpoint definition, mainly because this reflects high activity of the disease, and that was the primary focus of this trial but as you have seen, we also saw a very significant effects on fibrosis, along with the effect on the activity. And with this, I thank you for your attention.

Thank you, then. And now we'll move to the presentation from Professor Arun Sanyal.

Hi, good evening. This is Arun Sanyal. What I will do today is try to cover sort of some of the concepts that are involved in Phase III development. And of course, frame it in the context of lanifibranor. So the current regulatory pathway for the approval of therapy for NASH is shown on Slide #23. On this slide, what you see is that there's a dichotomous pathway. First, if you do not have cirrhosis, you have pre-cirrhotic stages of NASH, through the accelerated subpart pathway, one can have 2 possible endpoints, resolution of steatohepatitis without worsening of fibrosis or improvement in fibrosis with no worsening of activity. Now based on mechanism of action, the FDA will accept either/or for approval, whereas in the European guide reflection paper, they required combination of both. Now obviously, for lanifibranor, the data from the current trial shows that you can win on both. So this is likely to be the path forward, but then eventually, all of this has to be connected to a long-term study to demonstrate that it translates into less progression to cirrhosis and improvement in clinical outcomes. For people with cirrhosis, it's fibrosis improvement and reduction in melt progression, which are the primary Subpart H endpoints with improved outcomes in the long term. We currently have -- are getting better and better stents of what the actual event rate in patients with compensated stage 3 and stage 4 disease actually are, so that we can start planning not only the Subpart H but also the long-term part of the study. These are data from the NASH CRN. We hope this paper will come out in the next few months, it's under review. But basically, this is the first prospective data set of a fairly good-sized population, covering the entire spectrum of fatty liver disease with almost 1,800 adults followed for 8,000 patient years and you can see that mortality is closely linked to fibrosis stage. Prior to stage -- up to stage 2, the mortality rates are actually very low and do not really -- which is why they were lumped for this analysis and also not dramatically higher than population-based survival figures. It's a little bit different than the data from retrospective meta-analysis published by [indiscernible] and more recently by Taylor, where the outcomes start increasing at stage 2. This may have to do with retrospective analysis, bearing case definitions because in the retrospective study, there is no case definition because you have no pure case definitions. So somewhere -- the truth probably lies somewhere in the middle, and we actually have liver outcomes data now that we can start modeling our trial sample size and power on. Next slide, please. So in developing the Phase III. We have to follow some core principles, which are really the fundamental principles of drug development. And that is -- you've got to have the right targets, which we seem to have. We now need to make sure we have the right population, the population that is most likely to benefit from the treatment. In my personal view, Stage 3 is really the sweet spot because you have enough disease that you're really at risk of progression. There's clearly a step-up in mortality at Stage 3 and then a further step up at Stage 4. So they will benefit from reduction in their fibrosis stage. And also, the fibrosis not too far advanced for it to be malleable, but so some combination of stage 2 and stage 3, but a substantial portion of stage 3 patients. You have to demonstrate right safety and then, of course, the correct efficacy measures. So this is our current paradigm where we take people with active NASH stage 2 and 3, put them in a study with the idea of preventing progression to cirrhosis and eventually out of NASH. If you are patient-centric, you will say the ultimate goal is to reduce outcome, but in the setting of a clinical trial, we already know we have a gating with the Subpart H to show that you have improvement in histology before you get there. So the study populations, again, I think having a larger number of stage 3 patients means that when you get to stage in the long term, the post-Subpart-H, you'll have a higher rate of progression to cirrhosis in the placebo arm. So you will be able to meet that final endpoint of progression to cirrhosis and outcomes easier by having a higher proportion of people with stage 3 disease. And given their increased mortality, it is also more likely to resonate with third-party payers. The exclusion some of -- given that the profile of this drug seems to have many things that appear similar to a PPAR gamma, we want to make sure you don't have people with volume overload congestive heart failure, specific types of cancers that have at least historically been linked to PPAR gamma, although some of this has been debunked now. But I think it's best for safety reasons to exclude that population because it really will not impact the recruitment by doing that. One of the major issues in clinical trials is this placebo response problem and we continue to struggle with this conundrum of what drives this completely variable placebo response from one trial to another. And so one obvious thing is lifestyle and so it probably will be valuable. And in the development of this study, which is going to be a very large study conducted in many regions with very diverse lifestyles and food habits, so we'll have to come up with some common sense guidance, and we look to the Liver Forum and the paper that's been developed to the liver forum on lifestyle recommendations in the context of Phase III trials to try to develop some principles, use those principles to develop this, but I think this will be a very important ancillary document to be provided to all of the sites so that you don't have certain sites, which turn into lifestyle nazis and start absolutely enforcing it with military precision, whereas others are not paying any attention at all, which can create a lot of diversity and noise from center to center in terms of the placebo response. The next thing to remember is that NASH does not occur in a vacuum. There's no point saving the liver if you die of a heart attack, so you have to think about what are the comorbidities in this population and the bar for an effective treatment of NASH is not only that you have to benefit the liver, you need to be at least neutral in terms of the competing threats to life. So you have to look at the population in terms of the background, hypertension, cerebrovascular, cardiovascular, peripheral vascular, diastolic dysfunction, type 2 diabetes, chronic kidney disease, which are all independent threats to life and make sure those are evenly distributed, hopefully through randomization, and you also have to pay attention to the background therapies that patients will be on and which ones could interfere with your treatment response or interpretation of a PPAR gamma response. And these will need to be structured into an estimate-based analytic approach, which defines the population in which you will study, the outcome measure and exactly how you will deal with the development or addition of specific therapeutics that may be confounders once the patient starts treatment. So that's a long conversation, but we can talk about it in the discussion. Moving on to safety. We know that NASH has a major impact on extra hepatic events but this is part of that manuscript that's currently under review, but there's clearly a substantial number of people who have a chronic kidney disease, type 2 diabetes, hypertension, coronary artery disease, et cetera, no surprises here, and so we have to make sure that you don't take people who are likely to die from these diseases before you reach a liver endpoint. So you have to pick your population carefully, follow them with structured endpoints so that you don't get surprised by some random investigator-driven perception that there is one event happening more often than others, particular attention to weight loss, the weight gain and bone density and other common things that are PPAR-related. So in the development of any drug for NASH, very early on, I think it's very important to derisk the development by getting simultaneously, markers and tracking the proportion of people with nuance at hypertension. So I think in the context of the trial that you just heard about, there are probably additional data that can be mined, markers of sympathetic activation, which may link to long-term renal, cardiovascular and hypertensive outcomes, atherogenic risk profiles, markers of arterial stiffness, markers of endothelial injury, markers of diastolic dysfunction, beta cell reserve and eGFR. The more you know, the more you can derisk and present a very solid safety profile because at the end, your approval will be based on benefit versus risk. One of the biggest challenges in these long-term NASH trials is the assessment of liver toxicity, especially since the measures of liver toxicity, which is a rise in liver enzymes and bilirubin already can occur because of the background disease. So there are very well-established guidelines. This is I'm showing you on the next slide, sorry, I lost track of the slide number on the next slide -- I'm at Slide 33 now. There are -- there's guidance from the FDA on the type of parameters that you should follow, and how to follow them, and what the stopping rules are for individual patients. These will be obviously incorporated into the protocol. So on Slide 34, then coming at the end to write efficacy, Slide 35. This is a table, which shows sort of the profile of the competitive market space, where we're looking at the profile of what would constitute a good NASH drug in terms of weight, atherogenic profile, impact on MACE, stability of GFR glycemic control and then, of course, the ability to improve liver disease. So if you look at the extra hepatic profile, you can see none of these are really quite ideal, but clearly, GLP-1s have an advantage because they have been shown to produce survival and GFR benefits and MACE benefits in a high-risk population with type II diabetes. And so that is likely to be an important important threat post-Subpart-H. So the more endpoints we can include and particularly showing the improvement in lipid profile, et cetera, and long-term data to actually show that at least the surrogates for cardiovascular outcomes actually are looking good, will allow lanifibranor to maintain a competitive edge in this space. In terms of histology assessment, the ideal way to do it has to be specific, measurable, attainable, relevant and time-bound. Traditional histology assessment is still the reference standard that regulatory agencies use. We just need to be aware that there is substantial sampling variability. There is variability based on the quality of the biopsy, and there's variability because the subjective nature of assessment by the pathologists, no matter how good they are, and there are emerging best practices. There is no one set of best practices, as shown in Slide 36, that are uniformly available or used, but I think this is something that we will need to pay particular attention to as we develop our Phase lll, and we already have had conversations with the FDA about how to approach this. So we have a pretty good idea of how we will tackle and minimize the biases in the assessment of histology moving forward. So on Slide 37, this summarizes some of these best practices in Phase III with 2 pathologists, establishing their cap on an initial set of training biopsies and then periodically, as the study progresses, to make sure there is no temporal drift in reading tendency and have a specific format for biopsy reads and at the end, you would have a master reading of biopsies. I think this is all sort of tentative. There is probably room for refinement in this, but this is sort of an emerging framework for a lot of Phase III. The role of AI, I think, is still emerging. We are on Slide 37, in case anybody is still behind. And this is, I think, going to be a supportive technology. A lot of people are talking about how it is replacing the pathologists. I think that's completely incorrect. That is not going to happen. The pathologist is still the primary leader, but some of these will provide assistance to the pathologist in evaluating the histology. Obviously, in -- shown in Slide 38, a lot of interest in the different types of AI mechanisms, they ask different questions. So PathAI is focused on reducing the variability in histology assessment, whereas Histoindex and PharmaNet actually can obtain data at collagen fibrinor level and not only can do the same, but also actually provide some evidence, some data on the dynamic nature of fibrosis evolution and resolution as it progresses. So this is clearly an evolving field also, but different companies are approaching it in different ways. So one size definitely does not fit all. So then moving on to the next slide. The overall design has to proceed within this estimate framework. This is the recommendation both the EMA and the FDA, where you have to clearly define trial objectives. And big issue is this estimate concept really deals with post-randomization confounders and how you will deal with that. And some of the key post-randomization confounders we always struggle with are weight loss, use of diabetes medicines, which in long-term studies, you cannot ethically withhold the use of statins, which now there's, at least, some data that it may have some antifibrotic effects. And the impact of cardiac events and the treatments that can occur or cancers or other type of intercurrent events that are not liver-related, not related to the primary objective of the study, but whose treatment may actually impact mortality and other outcomes. So those will all need to be addressed within the context of this design framework, and it already is being -- these are being addressed. Now at the end, you don't get drugs approved just because you met an endpoint. It requires an analysis of benefit and risk. So you have to have clearly greater benefit than risk to be convincing that for a surrogate endpoint that you can get approval. So we showed decreased activity and fibrosis with lanifibranor, both, which makes it very reasonably likely that this will translate into decreased cirrhosis, decreased liver outcomes and decreased mortality. It's a little bit of a promissory card, but this is shown on Slide 40, but that's sort of where we sit. And on the risk side, of course, we have to have a plan in this Phase III about how to mitigate weight gain, what kind of advice we give the patient, how we monitor the patient's CHF. We exclude high-risk populations, we monitor the patients, and we have some active management built into the protocol. And then, of course, obviously, monitoring for hepatotoxicity, but clearly, I think paying attention to the safety as well as efficacy are both going to be important for the long-term success of the planned Phase III program. So there are other alternate paradigms that are in evolution, may not be that relevant for lanifibranor. This is shown in Slide 41 where you can take a large population and directly follow them through outcomes, but this is sort of conceptual. Nobody, to my knowledge, is actually following this, but in conversations with regulatory agencies, we have -- expect they'll be open to reviewing an application that looks like this. Others may take people with F3, F4, who have already got advanced disease and follow them to outcomes, but that also -- these are all emerging paradigms. Nobody to my knowledge is actually doing the study. So I'll finish over here on Slide 43, and I will let Ken take over from here.

Thank you. Thank you, Arun. Great presentation as usual. And thank you Inventiva for having me take part of this interesting program, and all of you who are after a long day here listening to us. I know it's late in the day. So what I'm going to try to do is simply as the only endocrinologist probably in this liver meeting, but with many friends in the audience and in the liver field, expand a little bit the context of the management of these patients to tell you what is going on in the diabetes field, in the primary care field, and how the different drugs that we have to treat diabetes and in particularly -- in particular, PPARs, which I have used extensively all my career, fits into this. And in my view, that lanifibranor will now really shift this class of PPARs into a liver-centric approach. So I'm going to be going now to slide number 46. And this is a very conceptual slide in which it appears that diabetes and fatty liver disease make the worst of each other disease in a way that diabetes, in ways we don't understand, seem to exacerbate disease activity, promote more cirrhosis, even hepatocellular carcinoma and having a fatty liver or NASH seems to make the diabetes more difficult to control. And as Arun showed, maybe linked also to worse cardiovascular disease. So in Slide 47, I want to tell you, you might wonder why is it that my peer to endocrinologists and primary care are just now waking up to this reality, and this slide tries to tell you exactly why. So this is a study that we just got accepted in Diabetes Care, which is the official journal of the American Diabetes Association. If you look at the upper left panel, we screened 561 patients with type 2 diabetes, just walking for the regular checkup, healthy outpatients, and we did a study with elastography to look at steatosis and fibrosis. And as you can see there, 70% of your patient with diabetes have NAFLD and 21% of them, fibrosis. Now if you look at panel B on the right, you will see that only 10% had elevated ALT, which has this rather high cutoff of 40, but this is a cutoff that makes primary care doctors seek additional tests. So most of the primary care doctors and endocrinologists feel that most of the patients are doing just fine, but if you go to the lower left panel, where you can see the fibrosis stage. Again, almost 10% have F3 or F4. And if you consider F2 also targets for future pharmacological therapy, you have 15% of about 30 million patients that should deserve therapy or additional workup. Again, on the right panel, you see that even under these circumstances that minority have elevated liver in the problem is vastly under-diagnosed, and we think this is going to change in the near future. Now one thing that's really important is to put NASH into the context of where we are in diabetes and obesity. And now we choose drugs for diabetes, not based on their glucose lowering but on their ability to reverse cardiovascular disease, heart failure, chronic kidney disease, again, both conditions, much more common in people with NASH, and we plan to integrate this in the selection of our drugs to those that can help us also with NASH. And you can see that there is a very close relationship into the 3 classes of agents. Now Arun mentioned very nicely the data that he -- that has been gathered from target NASH in which cardiovascular disease is a big component, again, not only in those with NASH, but those with fatty liver. And again, this is something that is extremely important in the selection of drugs to treat not only diabetes, but NASH. And that's where we think that PPARs have some advantage. Now the data on pioglitazone, I think, is widely known by this audience. So we don't need to dwell on it. You are very familiar with the PIVENS study, and we tried to harness the advantages of vitamin E and pioglitazone in a diabetes CARE study, but the take-home message from this slide is that an approach based on a PPAR gamma with mild alpha activity has shown benefit in NASH. And again, this has been incorporated into the guidelines. And again, probably with the data from semaglutide. This is going to be updated to include GLP-1 receptor agonist as a valid approach. Now in the interest of time, we are now in Slide 51. I'm going to move on to Slide 52 in which I think that the future of the management of NASH will be pretty much as we have seen it in type 2 diabetes, in which we don't expect a single drug to manage the disease. We don't expect metformin or even any of the newer drugs to be the -- treat all each drug has pros and cons from what works for one patient, well, might be an exclusion criteria that another one cannot use. And I think that the audience has been heavily tilted in the recent years towards anti-fibrotics, and there's a very good rationale. It is fibrosis, what's going to lead to cirrhosis and endpoints. However, the real problem is although -- all the drugs who are tested, have worked in mice. In humans, then pathways that lead to fibrosis are a lot less understood. And my view is that until we understand them better, we might better use what I think could be like a shotgun approach in which either you reduce adipose tissue that drives promotes the state of lipotoxicity and liver damage, and that can be by weight loss or change in the biology of fat as you can achieve with pioglitazone and now lanifibranor, hoping that then these agents will be able to turn off disease activity and prevent disease progression. That was the central hypothesis with pioglitazone, but as you've seen from the NATIVE study, that approach also allows to achieve the best results on fibrosis. So there's much more that we need to learn, but clearly an approach that can turn off disease activity is very exciting. So one key aspect is the importance of having PPAR gamma activity in your compound. So as you see in the upper right, the main target of PPAR gamma, although it's broad, is an adipose tissue where the receptors are tenfold higher than in any other tissue. There are also PPAR gamma receptors on macrophages, TMB cells in the liver and Kupffer cells and stellate cells at, like, collagen, and it is incredibly important that there are other PPAR activities that are very relevant. So if you look, for example, again, to the different family of PPARs that have been tested, it's a very difficult field in which we don't fully understand how these different compounds work. If you look at fenofibrate, a pure PPAR alpha, that does not lead to reduction in liver fat in humans. You look at PPAR gamma alone in rosiglitazone, that lowers fat a little bit, but does not change NASH. The results of a PPAR delta alone like seladelpar have been a bit disappointing. PPAR alpha and delta like elafibranor has been promising, but falling a little bit shorter than what we wished and then we have in the middle, again, lanifibranor, that somehow combines this in a way that led to the results that Sven showed you on the first slide. On the other hand, you have a combination of PPAR alpha and gamma, that, as you know, reached some of the benefits. So this is a complex field, but in which you need to have the right combination of PPAR activity, and you have to have some PPAR gamma activity. And again, to make a long story short, the data that we have with pioglitazone has been very exciting in the sense that it reduces cardiovascular disease and contrary to PPAR gamma agonists like rosiglitazone. And for the lanifibranor program, I think pursuing this avenue is critical because there is a very good chance that it will have similar or even better effects than pioglitazone from the preliminary data that we have from the NATIVE study. So again, remember that most of your patients with NASH are going to die of coronary disease or cardiovascular disease. So keeping this into the overall context of the management of patients is really critical. As a couple of examples, I'm now in Slide 55, and we see the early study PROactive that showed a reduction in fatal and nonfatal stroke and fatal and nonfatal myocardial infarction. Moving on to the next slide. We have the effects on a recent trial, the IRIS trial in people who took the medication. They looked at those who had taken the medication for at least 80% at a time. We're in Slide 56. And again, as shown initially in the PROactive, there was a 40%-plus reduction in stroke and MI. If you looked at acute coronary syndromes, they were reduced in half. And even new onset diabetes. As you know, NASH promotes the progression of individuals from prediabetes to type 2 diabetes. So here, you have the potential to not only target the liver with PPAR gamma and alpha and delta agonism from lanifibranor, but eventually to replicate findings seen earlier with pioglitazone. And again, I'd like to emphasize this because many times, in many of the NASH trials, we -- all of us fall into maybe a narrow liver-centric approach, but for patients and for payers, there will be a greater bang for the buck if you are really treating multiple targets that are meaningful to your patients. And as you can see, one of the factors that -- or aspects that I've seen most confusing about pioglitazone has been the belief that it causes heart failure. It only exacerbates heart failure if you have pre-existing heart failure. But as you can see in that reference at the bottom, and there are several papers of this kind, pioglitazone improves less diastolic dysfunction in patients with diabetes and does not cause heart failure unless you already had heart failure to start with. And again, as we move on to the next slide, it's important to realize that there have been really no increase in heart failure in basically any of the studies done after PROactive, in part because we learned not to give these medications to patients with pre-existing heart failure or in whom we suspect that they may have heart failure. So that's an important thing. So many times, people ask me, so why isn't a PPAR gamma alpha like pioglitazone used more commonly? Well, some are not aware of the efficacy. I think one of the aspects really important is that it's not being developed as a NASH drug, and it doesn't have an FDA indication. Other times, I find that people just say, I'm not a diabetes doctor, I'm not an endocrinologist. I'm not going to prescribe it. Others can be because of bone laws. I mean there's been a long debate about the bladder cancer, but most studies now have been negative, but if I have to choose, one is a misperception about weight gain and edema. So let me just spend the next 2, 3 slides or -- to explain this a little bit better. Number one, weight gain is typically due to an increase in subcutaneous fat. It's not related to water retention, although there might be about 5% to 8% of patients that might have lower extremity edema and there are many mechanisms that -- which we can talk about, most of the weight gain is true fat that is deposited in the subcutaneous tissue. And what really happens is that subcutaneous tissue becomes the working again. So it's going to be an efficient tissue to store fat if there continues to be overnutrition. So as you can see in Slide 61, one of the aspects that's most important to clarify is that it's not the same to gain weight with pioglitazone as it is from excess caloric intake. Insulin resistance and all the factors that you can see under insulin resistance such as even free fatty acid levels, liver fat, skeletal muscle fat, they all improved, while they all worsen in individuals that have gained weight just from overnutrition. Diabetes risk decreases, as I showed you, by 80%, hyperglycemia improves, while both worsened with excess weight gain from overnutrition, and you can see a laundry list of risk factors for cardiovascular disease that improve while they worsen with weight gain from excess caloric intake. In other words, what really happens is you put the excess calories in the tissue where it should be in the subcutaneous tissue. So you're making that adipose tissue be the same of the adipose tissue of a lean person while sparing other tissues from that excess fat, and I'm going to just skip this slide in the interest of time, but this is from our initial New England Journal paper because the message can be more simply put here. As you can see, if you plot changes in liver fat, with changes in visceral fat with pioglitazone, visceral fat goes down. If you look instead at the green triangles, which were the individuals in the trial that gain away it tends to put both visceral fat and liver fat in the -- increasing over time. If I show you Slide 64, this comes really to life in a patient, a 59-year-old person who, at baseline, had a hemoglobin A1c of 7.3 and a fasting glucose of 184. And you see on the right that after 16 weeks of pioglitazone, fasting glucose improves markedly, A1c normalizes, subcutaneous fat goes up and visceral fat goes down. Although there's an overall weight gain, this clearly explains that the more harmful visceral fat decreases and can be one of the clear factors why PPAR gamma alpha reduces cardiovascular events. Now in Slide 65, people always ask me how can you mitigate the weight gain with pioglitazone, or well, other PPARS. And this study just shows how the effects on glucose and lipids can be similar at lower doses with very minimal weight gain, as you see, 1% or 2% on the right. And this has been tested in the context of diabetes treatment in which we typically reduce the risk of weight gain with pioglitazone by combining it with other agents that have other positive effects on cardiovascular and renal risk. Let me show you in the next 3 slides, some examples. So if you go to the next slide, Slide 66, you can see very nicely that SGLT2 inhibitors that are now rapidly gaining market share given their effects on heart failure, cardiovascular disease and chronic kidney disease, also reduced body weight and reduced liver fat in basically every study, whether controlled or uncontrolled. Now in Slide 67, and I'm just going to show you one example of study from the EMPA-REG, which was one of the landmark studies in diabetes for this class of agents. You see that in the upper left, you see the A1c reduction that is typical with these agents. And you see that when you add it to pioglitazone, you have a net weight loss instead of a small weight gain, from the TZD. The other big group of agents in which we are combining pioglitazone is with GLP-1 receptor agonist, in this example with exenatide or dulaglutide, you look on the left, again, the typical A1c reduction in gray, dark gray with exenatide, in blue with a higher dose of dulaglutide. And then if you look in Slide 68 to the right panel, you see that with both, the GLP-1 leads to a net weight loss. And this is typically the way that we use TZDs in many of our patients. To finish, I know all of you might have heard about the good results with semaglutide in NASH, wanted to share what happens if you combined semaglutide with pioglitazone. Again, in the left panel, the A1c. Typical reduction of A1c in lighter blue with the 0.5 weekly -- 0.5 milligram weekly dose with a darker blue with the 1 milligram weekly dose that's different than the daily dose from the NASH trial. In the middle, you see there's a marked weight gain even when added to pioglitazone. In the right panel, you see that between 2/3 to 3/4 of patients can achieve A1c at target with no weight gain and no hypoglycemia. So I'm now in Slide 70 to finish. Again, the -- from an endocrine or non-strictly hepatology perspective, we are seeing that this is an epidemic that is going to just get worse. There's a lot of -- not only hepatic but extrahepatic morbidity. And again, PPARs can going to be used safely with proper patient, dose titration or a combination with other agents that may also have very positive effects in NASH. So I think PPARs offer really a very comprehensive approach, not only for the long-term benefit of the liver, but also for prevention of cardiovascular disease and also prevention of diabetes. That's an aspect that is a little bit overlooked, but it is critical, at least as an endocrinologist. So with that, thank you for your time, and I think we're going to be heading into questions. Correct?

Yes. This is correct. I think the operator can tell us how to ask questions.

[Operator Instructions] Your first question comes from the line from Etzer Darout from Guggenheim Securities.

Great. Just a couple of questions. I guess, maybe in terms of sort of the trial, on the Phase II trial and how we think about that, I guess -- just wondered how important will the background therapies be in the Phase III? And your thoughts around that, for example, you have a larger proportion of diabetic patients moving to SGLT2 and GLP-1s. And so your thoughts on those therapies or things like placebo effect and weighted, as you mentioned in the presentation. And so just overall, how you think about sort of Phase III? And then I have a second question?

I guess, this is for me or, I guess, suppose so. Well, to be honest, that will be a little bit of a problem, but on the other hand, it will be real life. So there is really no solution other than to hope that with these -- and most likely with these large trials, they aren't -- will be breakeven or allow a certain amount of each of these compounds per arm. The other alternative would be to exclude them, which would have problems of its own but would allow a cleaner approach. However, since these trials in the end are going to be long-term trials, I think that we -- that's really how the field is going to evolve and, again, will be taken into account in the end -- in the final statistical analysis of the study. So it's a real problem, but in reality, it's going to simply be an aspect that will be telling you how the real-world will be treating these patients.

This is Arun. I can maybe add a couple of lines to that. I completely agree with Ken. I think there are some in the methodology ways to deal with that. Well, obviously, if somebody just started a GLP-1 or SGLT2 and is not on stable therapy before they come into the trial, those should be excluded. If somebody has been on a GLP-1 or SGLT2 for an extended period of time and still have a biopsy that meets entry criteria, potentially could be included as -- but some sort of a cap on the total numbers of such patients could be placed. Once patients start therapy, especially for 3-year, 4-year post-Subpart-H, in the Subpart H period, probably these will be prohibited medications. And certainly, that would be probably the best way to go, but in the long term, you're absolutely right. It's not ethical to really withhold these classes of medicines from patients who have indications for them. And then this could be worked into the endpoint and the estimates where you could look at potentially benefit from a liver point of view without the need of additional secondary -- additional therapies for add-on therapies for type 2 diabetes because you're trying to create a profile of overall benefit, right? So those -- but these would, of course, have to be clarified and negotiated out with the FDA ahead of time, but it's not an insolvable problem.

Yes. So let me just clarify. In the short term, where you're just trying to prove the benefit like, in those, you have them on stable doses for 3 or 6 months or if not, they don't get in. I was thinking in the 4-, 5-year and the long term, you have a number of options in diabetes, so you could avoid including them. Although the point we're trying to make with Arun that there are some caveats. It'll depend on how the FDA wants to handle that, but you can do a very clean Phase III to establish the efficacy in the first 18 months of the trial. And then later on, the field will have to adapt the add-on medication to some extent. But again, I think it's something that can be managed.

Got it. And then, I guess, for Frédéric, if you can or Dr. Sanyal, I guess, to the extent that you can comment, we previously talked about sort of the F4 population and sort of the benefits that we've seen with drugs like pioglitazone in that group. And I guess if any additional thoughts on how you'd tackle sort of this subgroup, right, high unmet need as well in terms of sort of the fibrosis population and what plans could be moving forward for that population?

Yes. So on the positive side, we have learned that if you treat the root cause that drives the inflammatory process, which in turn drives the fibrosis. You can actually induce compensation and actually slow down progression of cirrhosis. Good examples are hepatitis B, where even in a decompensated state, if you have effective antiviral therapy and you really knock the virus down, the inflammation goes away and you actually can restore compensation. And then for compensated cirrhotics even lead to fibrosis -- to reversal of -- cirrhosis. So that is probably the best example. There are also people with hepatitis C, where, probably not to the same degree as hep B, but you can see stabilization. And then bariatrics, that's a proportion of people who improve, although these are all different diseases, but the core concept that if you treat the underlying root cause effectively and reduce the inflammatory drivers, injury and inflammatory drivers of fibrosis -- fibrosis doesn't occur in a vacuum, right? Something is telling the stellate cells to make scar tissue. So if you can reduce those drivers with effective therapy, you may not be able to reverse the cirrhosis -- I don't know, maybe you will, maybe I just don't know. There's no basis for me to say either way, but nobody has been able to really do it effectively and consistently so far, but if you slow and stabilize the cirrhosis, so you just live with some bunch of scar tissue, but good liver function and your portal hypertension is manageable, then that's really not a bad place to land at least for the foreseeable future. So the main role of lanifibranor, I see, would be doing that. Now obviously, it has anti-fibrotic benefits that were clearly evident in NATIVE. And so it is possible that we will actually see this translate into fibrosis reversal. You have to remember that clinical outcomes occur not because your fibrosis reversed, it is because your fibrosis progressed. So lack of progression is far more important than fibrosis reversal. I think the FDA has got it completely backwards on that. This is an argument that we have all the time, but I think they are beginning to see it, and they do look at the data in both ways. I think in the Subpart H, they just want to be more convinced that there is an anti-fibrotic effect for people who are claiming an antifibrotic drug, so they want to actually see fibrosis reversal, but they are actually aware that outcomes do happen only when your fibrosis progresses.

Next, confirming the plan moving forward. So we're totally focused now on the F2, F3 Phase III, which we plan to begin first half of next year. But certainly, with the activity we've seen, we're very interested of exploring the potential of the drug in the F4 patients, but that would be in a separate study.

Got it. Congrats again on the progress.

Thank you.

Your next question comes from the line from Lucy Codrington from Jefferies.

Just a couple. Just -- I could be wrong on this, but I was just wondering on the weight gain, this is for Professor Francque, have you looked at whether there actually is any correlation between weight gain and degree of efficacy seen in the trial? Just -- you very eloquently explained how the weight gain is associated with the trial, the drug being efficacious. So I was just wondering if there's -- if you've alluded to that at all. And then I guess this is applicable to all of you as clinicians, what proportion of your patients have concomitant congestive heart failure or would otherwise be inappropriate for lanifibranor treatment based on the current safety profile?

Yes. In terms of the weight gain as far as the analysis has been done, there seems to be no clear difference in the weight gain between responders and nonresponders. There is some difference in the increase in adiponectin that is greater than responders than in nonresponders, but not in terms of the weight gain. In relation to the question of heart failure, at least in this trial, there was one case reported by the investigator in the placebo arm, and that occurred relatively early in the trial after 2 months with the nature of fibrillation and then the decompensation and hospitalization, but that was in the placebo arm. The only other case where the investigator noted cardiac failure was in the 1,200-milligram arms just by the end of the treatment, but in that case, there was no hospitalization and no further investigation and then the symptoms, it was just peripheral edema. So it was probably more cases peripheral edema than a case of heart failure. For the rest, there was no case of heart failure reported in 24-week treatment period.

I would like to add, this is Ken Cusi, at that in the -- our -- even our pioglitazone trials that our 3 studies included together about 110 patients, we never treated with pioglitazone. We never had a case of heart failure, and probably Arun can say about PIVENS, but in the larger diabetes trials, the latest with the IRIS study that included 3,000 people, half of them treated with pioglitazone for about 4 years, there was no increase in heart failure compared to placebo. So there is heart failure, again, sometimes heart failure, subclinical -- can be subclinical and you may miss it, but currently, it is rare with well-selected patients. I know -- Arun, what has your experience been?

Yes, we really didn't see it at all. So -- but I think there's an important distinction to be made that when you have a trial being run by hepatologists, if you get a little pedal edema, it can get reported as a heart failure. And once an investigator puts that on an -- a form, you're stuck with it, which is why in my talk, I was trying to highlight that derisking development means you pay as much attention to safety as you do to efficacy by creating structured case record forms to capture outcomes so that all the criteria -- and then you have an independent adjudication committee that reviews those and actually make sure that the outcomes meet the prespecified criterias and are not ad hoc because a study coordinator decided, who really does not have the medical background to make these decisions, put something on a sheet of paper and a busy clinician is not paying attention just signs whatever the coordinator puts in front of them. That's the reality of clinical trials, particularly in large clinical operations where everybody is really busy. And so this comes back to trial conduct, and I think it will be very important in the context of Phase III to make sure that all of those derisking procedures are in place and rigorously followed so that we will not get burned by something like this. A great example with the FXR agonist is we had -- because everybody is hyped up about pruritus, when you have the investigator-initiated assessment of pruritus, very high rates. When you actually give the patient a questionnaire and let the patient respond, the numbers are about 30% lower. So there is a difference that, depending on -- how you get the information makes a difference. So with those kind of things in place, I completely agree with Ken, that actual florid heart failure -- edema, getting a little spitting around the feet is not heart failure.

That was very clear. I actually wanted to know, though, of -- I guess, of just your NASH patients, F2, F3 NASH patients in general, what proportion have concomitant heart failure that, therefore, wouldn't be able to be treated with lanifibranor rather than the lanifibranor causing heart failure.

I don't know if we have those numbers. There is an association of NASH with heart failure because they have obesity and diabetes, both risk factors, but I would guess 1% or 2%. Arun, you might be the best informed or Sven? What do you think?

I think the numbers are below [indiscernible] they exist, but the numbers are low.

Yes. I agree.

Your next question comes from the line from Lenny Van Steenhuyse from KBC Securities.

Perhaps 2 initial questions for Professor Francque on the trial results. I was wondering if there would be any explanation for more outspoken ALP and GGT reductions in the lower dose versus the higher dose, which is a bit counterintuitive. And secondly, we also saw some reporting of anemia as a side effect. Has that been excluded as a treatment-related side effect? And then I have a follow-up question.