Eli Lilly and Company (LLY) Earnings Call Transcript & Summary

Good morning and welcome to the Eli Lilly's session of Cowen's 41st Annual Healthcare Conference. We're very pleased to have with us today, Dr. Dan Skovronsky, who is Senior Vice President and Chief Scientific Officer. Cowen has been recommending Lilly's stock now since 2014, and it has been a great stock because great companies tend to be great stocks. And great companies are led by great leaders. And we have one of those great leaders here today. So Dan, thank you for your time.

Thank you, Steve.

Maybe we could start out by talking about tirzepatide because of the data that was just issued this morning, the SURPASS-2 data. Maybe you could just take a moment to recap what you think is most important from this data.

Yes. It's another great day for Lilly and a great day for tirzepatide, for sure. Looking at the data for today, I think we named these trials well, the SURPASS series of trials. This one, SURPASS-2, again, exceeded our expectations. We're really pleased with the results, particularly if you look at the efficacy across the doses but really also the lowest dose, the 5-milligram dose, showing superiority to the comparator, which is semaglutide 1-milligram dose on both A1c and weight loss. That's pretty spectacular to think that the lowest dose here could actually deliver best-in-class efficacy. And then we've got -- and by the way, best-in-class tolerability, as far as I can tell. And then we've got 2 more doses above it that just, again, exceeds, surpass, what's ever been thought to be possible in type 2 diabetes both in terms of A1c, and now you're looking across a number of SURPASS trials that have read out, and we're consistently getting about half or more than half of the patients on the highest dose to completely normal A1c, which has just never been contemplated before as a treatment goal. And I think that's the future, 90-plus percent patients getting to current treatment goals of less than 70 -- 7% A1c. So efficacy on glucose lowering, that just allows us to think differently about what might be possible in people with diabetes. And then weight loss, again, that's really incredible to see it. I said the lowest dose was statistically superior to semaglutide 1 milligram. The highest dose had twice the weight loss, double the weight loss. That's incredible. And we've now been seeing across these trials, even trials of 40 weeks duration, like this one, 13% body weight loss in a type 2 diabetes population. So that's not an obesity population where drugs typically could do better. And again, it's only 40 weeks. I think this drug has just performed so well in the Phase III trials. We have a lot of confidence coming out of Phase II, but there's a lot we didn't know about safety and tolerability and longer-term efficacy. It's all playing out as well as we could have hoped.

Can this drug meet Lilly's expectations if the 5-milligram turns out to be the most frequently used drug or dose? And is that your expectation that 5 milligrams would be the most frequently used dose?

Yes. I think it can meet, will meet, will exceed our expectations in just about every way. I think the -- it's hard to know what will be the most frequently used dose over time. I think there was a position out there, and people might have thought looking to Phase II data. Well, this drug is sort of an incretin that's so powerful that we reserve it for patients who failed other things and needed the most -- the most severe patients. I think these data completely eliminate that possibility. The 5-milligram dose is a great first incretin. It delivers efficacy that is best-in-class, with tolerability that is as good or better than other leading incretins. So you do have a starting dose here of 5-milligram that could be appropriate for many, many patients. Now over time, people may see that they desire better A1c control or they may need it as their disease progresses or they may desire or need better weight loss. And so over time, perhaps the patient starts on 5, 6 months, a year or 2 years later, comes back to their physician, then they say, "Okay, now you're ready to step up to the 10 or the 15." That's a really powerful opportunity because doctors could feel comfortable prescribing this if it's approved in the future and then knowing that they've still got dry powder here to help patients as time goes on. On the other hand, there will certainly be many patients who see their physician and because of their obesity or the severity of their A1c measures, they may be asked to go straight to some of the higher doses. So it's great. It's hugely, I think, unusual to have 3 doses all performing well, taken into Phase III and then to be pushing to take them all to the market. But I think in this case, it was a good decision on [ surface ] as well.

Do you think prescribers should -- or do you think they will compare tirzepatide to high-dose GLP-1s?

Well, we compared it to semaglutide 1 milligram in SURPASS-2. I think if you're sort of asking about sema 2 milligrams, that's a cross-trial comparison, so we don't encourage those. And there's so many caveats about cross-trial comparisons. But I think if you look across the weight loss here and the A1c, I feel pretty comfortable with where we are. I think some people wonder, "Oh, are you going to go head to head against that?" I'm not sure I see the need actually. I think the numbers here speak for themselves, even for the 5-milligram dose and even more clearly for the 10 and the 15. But I think there's something more important in that question, Steve, which is that if you look at what Novo's been able to do with semaglutide, they've really maxed out what GLP-1 alone can deliver. So you can just look at the 0.5 and the 1 and the 2, and we're reaching the limits of what fully hitting the GLP-1 mechanism can accomplish in terms of weight loss and A1c. There's diminishing returns. And then you take even the 5-milligram dose of tirzepatide and certainly the 10, the 15, and you can just see quite clearly, there's something different going on here. We're off that curve. This is not a diminishing returns game. And you're seeing the power of the GIP mechanism. And so that's like a last point from this trial package, is that there were questions a couple of years ago. Is this really dual agonism that's important? Or is it just a more powerful GLP-1? And I think the comparison with semaglutide answers that definitively that this is the power of the GIP mechanism. And we've seen what GLP-1 can do. And it's a great drug, but it's reached its limits in GLP-1 alone.

You mentioned the powerful weight-lowering capabilities of tirzepatide. Does Lilly view obesity as a medical condition that should be addressed independently of comorbidities or a risk factor for other medical conditions but should not itself be the target?

Yes. It's a good question, a good point you're raising, Steve, which is that unfortunately, probably obesity is one of the probably last diseases that's still a bit marginalized and not seen as a real disease or a real treatment objective by many in society. I think that's going to change over time. I think the idea that obesity is both a disease unto itself and a risk factor for other diseases is important, so I accept both sides of your dichotomy there. And it should be treated for both reasons. So of course, any person with obesity will easily tell us the perils of obesity in their everyday life. I think the medical community easily tells us the consequences of obesity in terms of other diseases: diabetes obviously, heart failure obviously. Even over the last year, we've seen COVID-19. When I look across the COVID-19 treatment trials, obesity is one of the most important risk factors for death. So it's -- it will be great to have powerful new medicines to treat obesity. I can't wait to see our obesity results. If we're getting 13% or 14% here in sort of 40-week diabetes trials, imagine what we're going to get in much longer trials in obese populations. We're going to reach levels of efficacy, I predict, that haven't been seen before. And just like I'm predicting that tirzepatide will change how we think about treatment goals in type 2 diabetes, tirzepatide will change how we think about treatment of obesity. I think, though, to take your question head on, it's on us, the pharmaceutical companies working with the medical community, to show that reducing -- weight loss also improves morbidity and ultimately mortality from obesity. We'll generate that data, and we'll make those arguments to the payers and prescribers. And ultimately, I think we'll be able to impact one of the biggest causes of illness.

One more question on tirzepatide. And by the way, if you're listening in on the line, you can send questions through the interface on your computer or you can e-mail Mike or I, and we'll pose your question. So Lilly has a dual-acting GIP/GLP in development that supposedly is more potent than tirzepatide. What does Lilly see as the primary clinical advantage of that follow-on?

Yes. Well, I think it's normal for companies to have follow-ons in case the lead molecule gets in trouble. That would have been important a couple of years ago. Now no, I'm not worried about that. So now I see that purely as a next-generation play. More potency doesn't inherently translate to better efficacy in any way. It just translates to lower dosing. And why do we care about lower dosing? 15 milligram is a pretty low dose already once a week. The reason we care about lower dosing is because of the potential for oral bioavailability. So at the same time as we engineer this to be more potent, we engineered in features that could make it orally bioavailable. So it's -- one of our next-generation plays is to make a drug like tirzepatide, this level of unprecedented efficacy, but make it as an oral molecule. Now it will still be an oral peptide with all the complications that we've seen that come into play when administering oral peptide. But perhaps if we can get this kind of efficacy, that could be worth the trade-offs. We just have to wait and see that data. I think, though, there's a broader theme here, which is that the incretin pathway, these hormones produced by your gut in response to food, which are your body's own way of ramping up metabolism and decreasing energy consumption -- decreasing core consumption, increasing insulin sensitivity, this is a great way to develop treatments. And we've seen the GLP-1s play out, I think, now to their conclusion. Now we're seeing what GIP plus GLP can do. But I think there's more to come. And we understand this biology and multiple incretin actions. And we have a triple incretin coming that I'm excited about, and combinations are possible beyond that.

Great. One question about the basal insulin-Fc. This is probably a class of drugs which Wall Street doesn't pay enough attention to. What do you see, if you -- if this answer -- if this question would even be answered. But what do you see as the advantages and limitations of the technology versus what Novo is working on? So they have icodec and they have insulin 965.

Yes. So look, I think the first molecule you mentioned from Novo as well as our insulin, those are weekly insulin. So by different technologies, and there's many ways to extend the time action of a molecule, we've extended the time action of basal insulin to allow potentially for weekly dosing. So think about that for a second, an insulin that you just inject once a week instead of every day. Now for people with type 2 diabetes, particularly, I see that as a huge advance. Think of when we launched Trulicity as a once-a-week incretin and what happened to the incretin market. Not only did it capture -- rapidly capture large market share, but it grew the whole incretin market and transformed how people think about incretins. Once-weekly drug is a huge advantage. Now with insulin, there are lots of concerns about a once-weekly drug. It's hard to do this. First is titration and how do you get the level correct and make sure you don't have hypoglycemia. And so you're looking for a very flat action, a flat profile. That's one thing that could ultimately be compared against different molecules. Another, and we have personal experience in this, is how much of the insulin is distributed to the periphery versus the liver. And that's an important aspect with respect to the safety of insulins that we've encountered in the past. And so those are important things. But ultimately, this is really an opportunity to change the treatment paradigm. So I think rather than focus on how 2 molecules might differ with each other, it's the similarity here, which is, by the way, riding on a tailwind of weekly incretin. So what I mean by that is even today, probably close to 2/3 of patients who are on their first injectable for type 2 diabetes, that first injectable is daily insulin, which is not a good medicine compared to a weekly incretin. Daily insulin causes hypos. It's -- causes weight gain. It doesn't have cardiovascular benefits. It's daily. Incretins are not. So over time, we've been changing that. I think the incretin business can still double from here as we change. So most patients have -- first injectable's an incretin. But imagine that future where most patients, their first injectable's a weekly incretin, their second injection should be a weekly insulin, not a daily insulin. So the changes to the diabetes paradigm that we've created with incretins and weekly incretins, in particular, now open the door for the weekly insulin. So I totally agree with you that this is a huge opportunity. It's not yet fully understood probably by investors or even by the medical community, but I predict we can make it happen.

Okay. Let's move to perhaps the largest of all opportunities, and that's donanemab. Curious, has Lilly met with FDA to discuss the TRAILBLAZER-1 data yet? And if not, is a meeting scheduled to discuss it?

Yes. Thanks, Steve. That's a pretty specific question. So I think across the portfolio, we probably try and avoid getting into the back and forths on FDA meetings with investors. On the other hand, we're committed to maintaining transparency with investors about our thinking and our actions and our expectations, which, of course, are informed by FDA conversations. So we try to sort of limit our comments to those kinds of areas. And you've heard us say some things, and we'll continue to keep investors informed as things might change.

Okay. So when we see the TRAILBLAZER-1 data on March 13, what would you urge investors to focus on first?

Yes. It's -- there's a lot to take in. First of all, I should say there are unique features for the trial. We talked about them. But the design features, you sort of have to really understand them and lock them into your thinking, starting with the idea that we windowed on tau levels. This is not like past Alzheimer's trials have been, where they enroll people regardless of their pathologic stage defined by tau pathology. This is more like a cancer study, where we pick patients of a particular stage of disease and analyze drug effect in there. So that's one difference. The second design difference is the idea here that we're treating them to clearance. So we treat until plaques are gone, and then we stop. There's no reason to stay on drug if the drug target has gone out of the body. Again, this is like not done in Alzheimer's trials, but it's done all the time in other therapeutic areas. Obviously, in infectious disease, you do it, and sometimes in oncology. Third is the unique pharmacology here of the drug. This isn't a design feature. It's [ unfair ] grouping. But this is a drug that specifically binds plaque. So when you look at this data, this is the effect of complete clearance of plaque. It doesn't bind the invisible putative oligomers. Some people think about it, it doesn't bind the [ sideway beta ]. It's just the plaques. And it binds them and clears them more deeply and more rapidly than anyone's ever done before. So you should look at this data and think about it as a test of the amyloid hypothesis specifically around plaque. So what do you get when you clear plaques in a clearly defined tau-windowed population? So what would you look at? The first, and this is how I, well, think about trials. Look, we've disclosed the magnitude of effect. So that's not going to be a surprise. You know what it is on the primary end point. The placebo performance is probably the first thing to look at in any trial. Did the placebo group decline at about the right rate? Or was this a really fast-declining placebo group, and maybe that gave you the effect? Second is the noise. So knowing the effect size, people pester us about what could the P values be. The P value is just determined by the effect size and the noise, the standard deviation of the means, in the treatment group and the placebo group. And so just look at it, is this a noisy trial? Are the standard deviations high? In a noisy trial, anything can happen. In a study where the standard deviations are more typical, it makes you feel better about the result. We've commented on consistency. Consistency is an important way of looking at trials. Sometimes at a superficial level, people just say, tell me the P value of the -- one end point at one time point, and I'll use that to make my decisions. No, I make my decisions based on consistency of data. So I think those are 3 big things. And maybe the fourth is some of the more exploratory end points that are still hugely informative, things like subgroup analyses, is this driven by one or the other, and that could be a compound sometimes. And biomarker effects, are they overall supportive of the hypothesis? So it's a lot to look at. It's a short presentation. But we'll try and be as forthcoming with details to answer all of those questions as possible. Keep in mind, of course, it's a small study. And there's always a risk of overinterpreting a small study. So with the robustness of the conclusions based on the sample size, I think, is an important consideration.

Lilly uses the word consistency when talking about the data set that we will be seeing. Can you just take a moment to define how you're using the word consistency? What does that mean?

Yes. Yes. Thanks, Steve. For us, it's consistency in 2 dimensions. One dimension is time. And you can imagine trials where the lines cross and then cross again. And then by luck or drug effect or whatever, you have a different set at the final time point. That wouldn't be considered consistent over time. The second dimension is consistent across end points. So Alzheimer's trials are notorious for, oh, well we missed on the primary, but look at this other end point. This is really the important one, this one hit. Or we -- well, no trial's ever hit on its primary, by the way. This is the first study that's been successful on it's primary. But you can imagine hitting the primary at the prespecified analysis, and then the secondary is going in the opposite direction. I wouldn't want that drug. That wouldn't be a drug I'd take forward to development. So when you look at consistency, it's a question of, with the totality of the data integrating in your mind, because we don't have great statistical methods, across time and across end points, is this a real drug? Is it likely to reproduce? Or is this a statistical fluke and we just got lucky with a small effect size or no effect size drug? That's the key question to take out of Phase II trials. And I think you've heard from our comments, we believe it's real, and we think this is a drug that's highly likely to reproduce and be an important treatment option for patients. But I think investors will make their own calls when they see the data.

Obviously, you can't tell us how the placebo performed in this trial. But in this type of population, what would you expect the placebo group to do over 18 months?

Yes. I think when we show the data, you'll be able to draw those comparisons. It's not far off, but sort of the modern Alzheimer's trials which have selected patients for amyloid positivity and looked at this sort of MCI/mild AD type of population, those give you a pretty good indication of where placebo groups should go. And of course, we've done something no one's done before. We've windowed out the lowest tau and the highest tau. But I think we still have a slice in sort of the middle of the population, which will make those comparisons easier.

Okay. When you look at the iADRS and the CDR Sum of the Boxes scales, how do they compare? For instance, if you achieve one, would you expect the second to be achieved? Or does one represent a lower bar? So how do these scales compare? Do they move in unison? Do they not?

Yes. So all of the scales in Alzheimer's disease have some correlation with each other, and none of the correlations are 1.0. I don't know the exact correlation between iADRS and CDR Sum of the Boxes. But of course, there is a correlation. On the other hand, different scales can have different levels of noise. And that noise can be within patient on repeated measures. It can be across patients in a trial or it could even be across trials or countries, depending on how things are assessed. Now why did we pick iADRS versus CDR Sum of Boxes as the primary this trial? It's obvious. We looked at all of the data that had been accumulated in Alzheimer's trials up to this point in time. And we said which clinical end point performs the best, is the most reliable in all the different ways I just mentioned, and selected that as the primary. CDR Sum of the Boxes, of course, FDA has said that's a good end point for them. If CDR Sum of the Boxes and all those analyses looked similarly reliable as something like iADRS, of course, we would have favored that one. But it doesn't. And we've published and talked about that in the past. In historical studies, CDR Sum of the Boxes is extremely noisy. It has some significant deficits that iADRS doesn't have. Now what are -- why then is the FDA sort of more favored -- has commented more favorably on CDR Sum of the Boxes? I think with iADRS, and this is true of any composite end point, there's a risk that you could have a benefit on the composite that's driven by one part of the composite looking really good and another part of the composite looking modestly bad. And you wouldn't want a drug saying that improves cognitive testing but decreases function. Even if on the composite, it looks like a net benefit. So of course, for the composite like iADRS, it's important to look at consistency across the different parts. But other than that, I'm quite excited about the potential as biomarker as I was before -- of this end point as I was before we started this trial.

And we only have a few more minutes. But do you have an opinion -- or do you know what FDA thinks of iADRS? Do you have -- do you know what they kind of view the end point as?

Yes. I mean we certainly have a sense for it, and they've made different comments about CDR as a preferred end point. But look, I'm not particularly concerned. I think a lot of this comes out in the strength and reproducibility of the data. Our second trial, I think there's still discussion and flexibility in terms of the population and the end points, et cetera. I think if you've got 2 trials in Alzheimer's disease with positive results on iADRS and consistency across cognition and function and support of secondaries, do I think that's a drug that could get approved by regulators anywhere in the world? Yes. But we're so far from that right now in the field, and there's so much discussion about aducanumab, which has just one trial that was sort of -- well, it was definitely negative before it was positive. And there's discussion about is that enough. So I'm not concerned about having 2 positive studies that hit on iADRS. That would be a great scenario for any company to be in anywhere in the world.

And where does the subcu of donanemab stand? One of the points or differentiation points that Roche points out is that gantenerumab is delivered subcu. So where do you stand with the subcu of donanemab?

Yes. So donanemab has a definite drawback that we've commented on. It generates high levels of antidrug antibodies. And so we dosed it pretty high, 1,400 milligrams every 4 weeks. So that's an infusion. 1,400 milligrams is just too high for subcu. I think for different drugs, you could sort of formulate things at 100 to 150 mg per ml and then there's less absorption for subcu. So you're talking about like a 10 to 20 ml injection to get these kinds of blood levels, which I don't think is desirable for patients. Now we have a next-generation version of donanemab, another antibody against the same species of a-beta, N3pG, that's in clinical development. It's not generating those levels of ADAs. And so therefore, perhaps the dosing could be lower and it could be subcu. But we're not switching horses. This is -- we're going to go all the way with donanemab but then over time, perhaps work on this next molecule. Maybe it could be subcu.

Do you think there's a decent possibility -- and you can define decent in your own mind. But do you think there's a decent possibility that the FDA would view TRAILBLAZER-1 and TRAILBLAZER-2 as complete and fileable and ultimately approvable?

Yes. It depends on the results of TRAILBLAZER-2, of course. So -- but that's our plan, is that these 2 trials in conjunction will serve the basis of a BLA. I'm not putting words in the regulators' mouths. I'm just saying that that's our assumption, and that's what we're working for. Now that doesn't necessarily mean TRAILBLAZER-2 is as it's currently constructed, and we have opportunities to think about that and change that as we enroll the trial. But given that situation, Steve, of course, like what's our #1 priority? It's enrolling TRAILBLAZER-2 and getting that trial done and analyzed as quickly as possible so that we can bring this important medicine to patients. I think I like our odds. I think we could have a really powerful data package if we can have 2 studies that replicate each other, a really powerful science here if we have a well-defined stratum of patients, just like we do in oncology, defining the patients by pathology. By the way, everybody passes through that pathologic stage, so everyone gets a chance to get this drug. And then finally, a really powerful value proposition for patients, because this is a limited-duration dose, this is not a drug for the rest of your life. You take this drug until the pathology is gone. And hopefully, you can get the benefits for the rest of your life, but you don't stay on it. So to me, I like that package of attributes. And so let's go as fast as we can to get to that end result. I hope -- expect we'll enroll the trial this year. Maybe not going up to the end of the year, I hope not. And then it's currently designed as an 18-month trial, and so that puts you -- and we'll get the next readout.

Great. So unfortunately, we're out of time. I do want to ask one last question. If I, in my own mind, summed up what Lilly has said about donanemab, and it's the tone in which it was said. What I see is a very, very high level of confidence, and that on March 13, the Alzheimer's community is going to realize that a substantial advance has been made. What do you think about my summation of Lilly's words and tone?

Well, I'll just say it myself. I have a high level of confidence in donanemab, and I think a substantial contribution to the field has been made, and I'm excited about that. I can't put words into the mouths of the Alzheimer's community. Who knows what the different scientists and people think? Let -- we're not far off. Let's see what they say. And sometimes, new data readouts take time, and sometimes we're surprised by the things that people focus on. But from our own analysis and the small number of experts we've shared the data with, we surely believe a very substantial advance in Alzheimer's research has been accomplished.

Great. Well, congratulations on all the advances that Lilly Research has completed in the last year, and we look forward to many more in future years. And this has been a great discussion and a great session. So thank you, Dan, for your time, and best of luck on the 13th.

Yes. Thanks for all the great questions, Steve. Bye-bye.

Bye-bye.