BridgeBio Pharma, Inc. (BBIO) Earnings Call Transcript & Summary

Good day, and welcome to the BridgeBio Pharma to discuss topline results from Phase III ATTRibute-CM study conference call. [Operator Instructions] As a reminder, this call is being recorded. I would like to turn the call over to Neil Kumar, Founder and CEO of BridgeBio Pharma. You may begin.

Thank you, operator, and thank you all for the time. I'm sorry I don't have better news to share with everyone today. But Part A of our ATTRibute-CM trial did not meet its primary end point. Although the drug, as we'll discuss, seemed to have performed as we would have hoped, the placebo arm of the trial, especially on 6-minute walk test, simply did not decline as compared with historical clinical trials or the natural history studies we reviewed. There was no real ability for our drug to show any signal against the measured baseline in the study to date. Before I continue, I'd like to acknowledge the disappointment many of you are feeling right now, for the patients and physicians we serve and for the investors we seek to perform for. I'm hurting as is this team today as well. I'll break my comments up into 2 categories: number one, the search for answers regarding the seemingly puzzling finding regarding 6-minute walk; and number two, the case for continuing this trial where we'll explore the drug's effect on all other parameters, excepting 6-minute walk, which looks very positive to our eye, to the trial steering committee's eye and to the DSMB that reviewed the unblinded data and recommended that we keep the trial going. I'm going to refer to slide numbers as we move through this deck for those of you following along with the documents. We'll start on Slide 3. Slide 3 lays out the disappointing data related to the 6-minute walk test on acoramidis. Patients declined around minus 9 meters of mean or minus 5 at median over 12 months. On placebo, those respective numbers were minus 7 and minus 6 meters. What's remarkable, at least to me, is that the placebo performed almost 70-plus percent better than patients did on the drug arm of the ATTR-ACT study. We would have needed something like a 15- to 20-meter plus benefit or so to meet a p-value given these data. Moving to Slide 4. We present the context for this wildly divergent placebo arm. As you can see, the placebo of our trial closely mirrors historical healthy age-matched adult volunteer declines. The very best this company has seen in terms of decline for ATTR-CM patients on placebo over 12 months for a 6-minute walk is a little under minus 40 meters in all of the studies that we've reviewed, as you can see the major studies pointed out here on Slide 4. Turning to Slide 5. We think about -- we start to think about what really accounts for this deviation in placebo performance. The variant subpopulation analyses present themselves in the face of such a surprising result. We don't have time to move through all of these. I'm sure we can move through some of them in Q&A. But so far, we found no answers. Here's one snapshot that further reinforces the strangeness of the 6-minute walk result. Even for Class III heart failure patients, as you can see here, there's barely a decline in mean 6-minute walk distance at 12 months. When we search elsewhere, as you can see it on Slide 6, easy answers are not to be found. We enrolled a very similar cohort as compared with ATTR-ACT as we've talked about before. The mean 6-meter walk baselines were essentially on top of each other at 360 and 352. The percent of Class II and III patients in each trial were 89% and 92%. We observed a slight shift to Class II patients, but this effect does not account for the discrepant result as we'll touch on later, in fact, just the opposite. We did have fewer variant patients where we did observe a minus 2-meter change as opposed to minus 40 in placebo. But even modeling a 25% variant contribution to our trial, we still miss our primary endpoint handily. As discussed, we had more ex U.S. patients than U.S. patients, but we did not see a difference in effect there when reviewing the [ fluoroscopes ]. Turning to the next slide and drilling into our trial itself, we also see that the study was relatively well balanced. A slight imbalance in Class III patients on treatment does not account for the lack of effect seen here either. Moving to Slide 8. We summarize and expand on some of the points I've made thus far. The lack of benefit on 6-minute walk and our placebo outperformance are not easily captured by the proportion of Class II versus III patients in our trial, U.S. versus ex U.S. patients in our trial or wild-type versus variant participants in our trial. We're exploring quite a few current hypotheses as to why we might have seen this type of placebo response, and coupled with expert input on 6-minute walk test, current hypotheses include context and training bias as well as an evolution in standard of care and diagnostic approach. The first category suggests that because patients', one, new small molecule stabilizers work; two, had a 2/3 chance of being on a small molecule stabilizer that was active in our trial since they were randomized 2:1; three, wanted to believe that they were getting better; and four, were engaged in a rigorous testing infrastructure that prompted the best performance that they outperformed even on placebo over this time frame. The second category suggests that patients are simply, despite the class and BNP levels, healthier on track -- healthier than they were on a track and that the deterioration associated with the disease is likely to come later in this trial. Okay. So I'd like to turn into the remainder of the deck to the case for believing in acoramidis. Starting on Slide 9, and as we will walk through, on almost every other measure resolved in this trial, we see promise for this drug. The enthusiasm has been mirrored by the steering committee of our trial and supported by the independent data monitoring committee. As you can see on Slide 9, acoramidis essentially flatlined BNP levels while placebo increased about 25% on the study. On Slide 10, on quality of life. We see that the drug had a nominal p-value. Although our expectation going in was that this endpoint and this questionnaire was going to be the most sensitive to variants coming from the COVID era. Slide 11 shows that biochemically the drug was performing the way we wanted it to perform. It seems to be having the desired effect immediately and with immediate and sustained elevation of serum TTR in the 45% range, especially impressive considering the greater wild-type patient population in the trial. And finally, on Slide 12, we find a couple of points of promise in the safety tables. First, it should be said that to protect the integrity of Part B, our access to AE data from Part A excludes AEs leading to CV hospitalization, excepting events with an outcome of death. Furthermore, AEs leading to death account for a fraction, about 50% if you look at ATTR-ACT at 30 months, of total death and are collected to ensure the safety of our population. Looking at the event rate first, about 32 deaths, which equates to about 5% approximately mortality rate, we see this tracks closely with the ATTR-ACT in terms of event rate, which had about a 10% total mortality at this time, 12 months in. We might assume that the contribution to this mortality from AEs is about 5% given their 1:1 split in their trial at 30 months. Further, looking at the numbers in the table based on a 2:1 randomization, we can calculate a 27% reduction in AE-driven mortality, which is encouraging. Slide 13 summarizes our findings. No improvement on 6-minute walk. All other measures positive. DSMB has recommended that we proceed after an unblinded review of the data. Slide 14 takes us outside of the trial where BridgeBio is set to deliver on the promise, not only in Part B but other efforts with the current cash on hand. Although we've experienced a very real setback today, we have the runway into 2024 to build back better. In the first half of 2023, we anticipate ADH1 and acoramidis Phase III data plus the realization in 2022 of multiple proof-of-concept events including infigratinib for achondroplasia, our programs in LGMD2i, dystrophic epidermolysis bullosa and congenital adrenal hyperplasia. Finally, on Slide 15, although today did not meet our expectations nor yours, what does end all was, will continue to meet our expectation is the courage and dedication of the physicians and patients that are on our trial and in the ATTR-CM community. We hope that together we can play a role in furnishing a better future. With that, I'll turn it over to the operator and take any and all questions.

[Operator Instructions] Our first question comes from Salim Syed with Mizuho.

Sorry to hear the news, and Neil, thanks for all that transparency, and it was super helpful. I guess if we kind of zoom forward here, I appreciate the analysis on why we didn't see a difference on 6-minute walk, and I guess, we'll wait for more color there. But if we are looking towards the case here now for mortality benefit, I think you noted in your prepared remarks, and I think it's also in the press release that there was about a 5% in each arm of adverse events. Is that similar to the overall mortality as well, if that's a number you guys have? And I guess, since these -- I mean, this does seem to be tracking with how we would think wild-type patients are doing. But I guess these looked very similar. I know it's 12 months, but -- what are -- I mean, how should we be thinking about the risk here for mortality given these are tracking similar, we are seeing more wild-type? Is 30 months long enough here? Just things around that? Or do you think you need to go beyond 30 months in order to have something beyond tafamidis here when we start to compare the 2?

Yes, good question, Salim. Thanks for your comments. I guess what I'll say at the outset is we are actually -- I mean since the primary endpoint here on Part B is -- or includes mortality, where -- we're blinded to it and are maintaining a straight bind around that. So we don't know what the total rate looks like. That's why we had the DSMB have a look. All we can see is AE-driven mortality. And that, as I mentioned, in ATTR-ACT was about half of overall. So if we're seeing 5%-or-so mortality rates right now in the AE table, it stands to reason that we're tracking somewhere in that what ATTR-ACT saw to date, right, that 10% or so after 12 months. The real question is, obviously, the amount of mortality increased in ATTR-ACT by 30 months. And that event rate needs to increase analogously in our trial so that we have a baseline against which to show signal. There's no reason to believe that it's not so far. I mean one thing that's obviously -- that we've looked at is if you look at the -- if you were just looking at the blinded 6-minute walk data, you'd have seen about 25% of the overall signal at 12 months. But if you look at many of the other parameters here, KCCQ, BNP, mortality and what we can see, et cetera, you're a little closer. You're still not fully at ATTR-ACT levels of deterioration. So the -- so I guess the prediction would be that 30 months would be a long enough time to capture the patient baseline going down. But it's hard for me to predict that right now, to be honest. I guess let's just elaborate on that. The best I can do is looking at natural history and ATTR-ACT. Obviously, we've sampled a slightly different patient population here. But how different that is? Is it going to be just healthy for a long period of time? If that's true because of some standard of care change or some way that we're relabeling what a Class II or Class III patient looks like, then all bets are off.

If I could follow up with the tafamidis drop and then -- what sort of looks is the DSMB taking in terms of -- like is futility analysis even something we should be wondering about now? Like every 3 months, like there's an update from the DSMB. Or how long will this go from -- what sort of looks do they take? Now that we have tafamidis drop and the risk of the trial has increased, right?

Yes. So there's about 3% tafamidis drop-in. Just as a reminder, there was no tafamidis drop-in on Part A. But it's a good question. So the DSMB is reviewing this on an ongoing basis, obviously. But the overall ability to look and say, is the event rate high enough to discriminate on mortality? That's going to come very close to the end of the trial because these trials enroll like a hockey stick. And so we're going to see a lot of patients coming in right toward the end. And as you remember from the ATTR-ACT study, you saw some -- really the statistically significant separation. You started to see separation after 18 months, and you started to see large separation in those latter months. So I would expect that -- I mean, first of all, there's no formal futility analysis built in. And secondly, that event rate is going to pick up pretty close towards the end of the trial. But yes, the DSMB is reviewing the data on an ongoing basis. And that's why we did ask them to take an unblinded look at the data just to make sure that it's good to reason that we continue this trial.

Our next question comes from Anupam Rama with JPMorgan.

Sorry about the disappointing outcome. What additional analyses are you guys thinking about doing here on 6-minute walk distance to better understand what happened here? And also -- yes, I guess that's my first question.

Yes, it's a good question, Anupam. I mean we're going to continue to refine. I mean these data are very recent. So we'll continue to refine our analysis around any hypothesis that we might have around subpopulation performance. I think the critical thing will be looking side-by-side to understand whether or not we can learn anything at that level of detail. There is not a whole lot of other hypotheses that we have right now. Obviously, we'll be digging into some of the context and training bias questions, but that is not easily resolvable from our dataset alone. We'll have to look at some external data as well. And then the other thing we'll just be trying to do a little bit more work on who these patients are and how standard of care might have changed over the course since ATTR-ACT started to this trial. We can see some of that right now, but not a tremendous amount. So we'll just have to dig in and do some more qualitative research. Certainly, there's nothing quantitative right now from the data associated with the patient populations that jumps out.

And because of the sort of COVID-19 pandemic, were there any trial conduct issues that could have resulted in something like this?

No, it's a great question, but we didn't see any trial conduct issues. In fact, I think the trial was well conducted to date. So I don't think that, that's the explanation nor is it immediately apparent to me that COVID, although it obviously does have effect on some of the other issues that we're measuring in the trial, questionnaire and hospitalization, it's really unclear as to what type of effect it has on 6-minute walk. Honestly, there aren't enough studies that have read out in -- or using 6-minute walk throughout the era of COVID to understand that.

Our next question comes from Paul Choi with Goldman Sachs.

Sorry for the disappointing news. My first question, Neil and team, is just on how you're thinking about your next step of regulatory discussions and interactions given that you missed here on the primary endpoint. Can you maybe just walk us through the path in terms of next steps in terms of your discussions with the regulators? And then secondly, how you would think about your filing strategy or regulatory strategy with the 30-month data endpoint coming up in another year or so?

Yes. Thanks, Paul. I appreciate the question. In terms of regulatory interactions in the near term, obviously, there's nothing to file here. So there's nothing specific to do with the agency or European authorities in the near term or -- and we have checked in with all of the relevant parties. Our goal here, as I discussed, is to finish the trial and against the same end points of CV hospitalization and mortality, ultimately file on the basis of that data should that data be positive at the 30-month endpoint, which would be about 15 months from now.

Okay. Great. And then as a follow-up, as you think about the 30-month data here. My second question to you is with regard to the potential continuation for compounding factors here, how are you thinking about any other trial review or execution procedures or checks or quality controls just to make sure to what degree you can check on this -- these quality control factors can be implemented to hopefully drive a positive outcome?

Yes. I mean I think quality control to me lies in two buckets. Number one is actual trial conduct. And as I mentioned, we looked at a lot of that, and I think we're -- we feel pretty comfortable that the trial is being conducted at a high level of fidelity. The second is obviously the ongoing event rate because you need to have a baseline against which you can resolve signal. And that I discussed with Salim's question as well, which is that the best we have is effectively allowing our DMC to have a look on it on an ongoing basis and to understand that the blended event rates, or obviously, the blinded event rates continue to track close to ATTR-ACT over time.

Our next question comes from Mani Foroohar with SVB Leerink.

I'd like to add my condolence and I'm going to touch on a tough update, somebody else's. I guess I have a little bit of a different trial conduct question. There's nothing you can do about the enrollment at this point, et cetera. The bullet is kind of out the gun in that regard. One of the questions I've already gotten a few e-mails from clients about is, does this result potentially increase the risk of patient dropout, increase tafamidis drop-in, and the things that you and the rest of the Bridge team or all legacy Eidos team can do to try and keep that at a reasonable minimum?

Yes. Great question. Obviously, that -- when we first got the result, since kind of step 1 was trying to understand for ourselves do we still believe in this drug and its promise against the hard endpoints of Part B? And once we answered that question, the related next question was, do the physicians that are conducting the trial believe that because they're going to be a big part of keeping the patients on the trial? And we've had a couple of PI calls since we got the data, and they've all been resoundingly positive, to be honest. People can check that out for themselves. But from our co-chairs, Dan and Julian, all the way down, we've heard a lot of enthusiasm for finishing this trial. And that's really what's going to drive people staying on trial. At the end of the day, the median patient is about 20 months in to this trial right now and access to tafamidis has been challenging for a great many of our patients, which I think is why we see only a 3% drop-in on Part B to date. So I think from that standpoint, lots of enthusiasm for finishing this experiment, at least from the physicians. And we'll be continuing to monitor the dropout rate over time there. We'll -- I can report back to people on it.

Our next question comes from Ellie Merle with UBS.

Can you just elaborate a bit on some of the kind of like working hypothesis for, I guess, why the placebo arm performed the way it did? I mean you mentioned that a couple of in your prepared remarks like that context bias, training bias, some of the evolution and diagnosis. Maybe just if you could sort of elaborate on sort of what you know so far and what could have potentially happened? Because, obviously, some of these could impact mortality data at 30 months and some might not have an effect such as a training bias and just kind of elaborate on what we know, and I guess, what we might learn in the coming months as you have more time with the data.

Yes, great question. Thanks for that. Yes, I would say that there is kind of 2 categories of hypotheses. The first relates specifically to the 6-minute walk test, and that's training and context bias, I mean, very simply because this wasn't a primary endpoint for Pfizer, they probably weren't super focused on this test or not quite as focused as we were. And so in the context where patients knew small molecule stabilizers were working, they wanted to be getting better and they had a 2/3 chance of being on a drug, and we made it a big deal. I mean this test was a big deal, and we used a lot of rigor for that 12-month visit. There could have been just sort of an added performance effort associated with the test that we haven't seen or we didn't see in ATTR-ACT. So that's one ongoing hypothesis that we're trying to study. The second is, obviously, does -- would affect mortality or could potentially affect mortality, which is that this is simply a different population of patients given the fact that we're diagnosing using the technician scan and that's different, obviously, than the Pfizer. And given the fact that standard of care might have changed, I mean, it could be that some of the concomitant meds are slightly different now, that we're picking these patients up in a way that allows us to care for them more effectively. It's not immediately apparent from, as I said, the quantitative descriptors, things like class of heart failure and BNP levels, et cetera, but that could be the case. And if that is the case, then it could well affect mortality rates going forward. We will continue to do research on that.

Got it. And maybe just if you could elaborate, I guess, on the NT-proBNP, I mean, you saw some improvements there. I mean, I guess, what we know so far from other studies and what this could mean for mortality if at all?

Yes. So for the BNP, we saw, I think, basically a flatlining on drug versus about a 24%, 25% increase on placebo. So obviously, that to our eye looks like pretty good performance as compared to what's been seen before in this space. The ability to use BNP as a predictor for mortality is not, quite honestly, well established in this space. It's a reasonable univariate predictor as is serum TTR so that gives us heart. But yes, it's the best I could say. It's based on natural history studies and others, it's a reasonable predictor of mortality.

Our next question comes from Dane Leone with Raymond James.

I guess for me, the key question here is probably the 2-part, Part A problem. The first part of this is just disappointment from investor expectations on a 12-month endpoint, which is -- it is what it is. But I think the more important question now is from the drug development perspective. One thing that stands out to me looking at the data today that we didn't have before is this looks like a wildly different patient baseline population than what was in ATTR-ACT. And at the end of the day, this is fixed in terms of the patients you've enrolled. So when we get to month 30, I guess, how do you think about dissecting that data to eventually making a commercial argument where this drug could fit into the treatment continuum within the context of tafamidis? And I guess maybe a pointed question on that. Were you surprised in terms of the low rate of variant patient enrollment into the study, especially given you guys were more anchored ex U.S. to try and avoid overlap with, obviously, tafamidis utilization within the U.S.?

Yes. Great questions, Dane. Thanks. Maybe I'll start with the variant. I wasn't particularly surprised by that just given that I think -- I mean, ex U.S. wouldn't necessarily bias you toward more, probably less variants, especially the geographies we were trialing in. Maybe you think about polyneuropathy as being -- having a founder effect in some of the European countries where we were open. But on the cardiomyopathy side, I think the slightly lower variant rate made sense given the distribution of sites that we had. I think you asked a good question on patient population. It is fixed. That's why I mentioned earlier that we were sort of looking at some of the other variables to try to understand how this -- sort of different this is from the ATTR-ACT study. It's certainly very different in terms of a 6-minute walk performance. It appears to be less different with respect to things like KCCQ decline, mortality, BNP and the like. And so I think we feel still good about the general evolution of the patient population's performance and our ability to resolve signal within it. Commercially, I guess the only comment I would make is, and we'll have to see when we put all the data together when we get there. But this is reflective of the patients that are being diagnosed today with ATTR-CM, which it does appear is slightly different than the baseline of patients that were first enrolled in the ATTR-ACT study. So yes, that does reflect an evolution of standard of care and diagnosis.

Our next question comes from Greg Harrison with Bank of America.

So what impact, if any, do these results have on your other programs? I know you're well capitalized. But should investors anticipate any reprioritization within your pipeline to either preserve capital or to accelerate other programs that could provide an early proof of concept?

Yes, it's a great question. We remain focused on the diversified prosecution of our pipeline, certainly with an eye towards the 4 key value drivers. Outside of this, we've got ADH1 where we have a EOP2 meeting coming in January this year, and then hopefully, initiating our Phase III shortly after that. For infigratinib for achondroplasia, we should have data on our Phase II some time mid this year. And for congenital adrenal hyperplasia, we should have proof-of-concept data late this year. So all of those will continue to move forward. We haven't yet made any decisions on further prioritization within the pipeline. Certainly, we'll look to leanly prosecute the pipeline as we move forward and as we have in the past to make the most of the cash on the balance sheet right now. But yes, no further decisions have been made.

Okay. And then just one follow-up. Can you kind of set a hurdle or a bar as far as what do you think you would need to show on the mortality endpoint, both for potential approval and also to be competitive within the market?

Yes. Well, great question. I think on mortality, what we'll be looking to do is a 15% relative risk reduction versus what Pfizer showed, which was taking 42% or 30% or so. That would be the bar again, would be outperformance versus tafamidis. Yes, we continue to believe with serum TTR levels of 45-plus percent and a flatlining of BNP that the drug is doing what we wanted it to do if the path or mechanism of disease continues to be what it is, that would be our expectation, and that, I think, would lead to optimal commercial performance in the marketplace.

Our next question comes from Ram Selvaraju with H.C. Wainwright.

Just firstly, can you refresh our memories as to what you would have predicted the placebo 6-month walking distance rate of decline to have been if this study had mimicked other trials that you've obviously reported on before? And then the second question also very quick is, from a regulatory perspective, have you already had discussions where they receive feedback from regulators such that if in 15 months' time, performance on these hard endpoints that you've discussed earlier is positive, there would be sort of a formal or semiformal route to approval for this drug?

Yes. Maybe I'll address the second first. Yes, is the quick answer to the outperformance obviously against CV hospitalization and mortality, that would be a road to approval for this drug. The first question that you asked is around our expectations around 6-minute walk declination over 12 months. Well, we pointed that out in Slide 4, I think, of our presentation. The ATTR-ACT cohort, which we were reasonably mimicking, as discussed in the -- in my prepared comments, declined 56 meters over 12 months. Natural history studies that we've seen are in that minus 45 range or so. I mean you can go back and look at -- and then there's an Alnylam placebo arm as well that was in that minus 40-plus range as well with a reasonably similar patient population. And then there have been some IIT studies, all of which have reflected something between minus 40 and minus 50. So pretty much everything that we've reviewed as a company, certainly, prior to this trial suggested that minus 40 would be the best that patients would do, and we were expecting something based on our characteristics -- baseline characteristics of minus 50-plus range.

So just as a follow-up to that, I mean, you've talked about how there doesn't appear to have been an execution issue with this trial. I also do not think that -- I'm no statistician, but it seems as though the difference between what was reported and what you would have predicted is so significant that this clearly doesn't appear to be a statistical analysis problem either. But do you think at this juncture that it is possible to conclusively exclude any issues with data capture or data entry as it pertains specifically to this endpoint?

That's a good question. I think we feel pretty confident in our data capture here and the conduct of the trial. We'll continue to do work on that. Maybe I'll ask Jonathan to...

Yes, we had a very high percentage of people completing the 12-month visit. So not a lot of missing data. And we had pre-agreed missing data imputation methodologies to account for missing data with the regulatory authorities. So the result you see is pretty reliable.

The other thing I might mention is if you look at the standard deviations, they're similar across the 2 arms, about 60 meters on and off drug, and the distributions look reasonable. They're not wildly off in terms of the shape. So yes, I don't think that there was -- I don't believe as we look in site by site that there'll be one or two sites that set this off or any outlier data set.

Yes. Maybe worth adding that we did have a core laboratory that reviewed everybody's 6-minute walk course setup and execution methodology site by site and had to be certified and trained according to the standard methodology.

There are no further questions. I'd like to turn the call back over to Neil Kumar for any closing remarks.

Great. Well, thank you all for the time, and we look forward to being in touch with any and all who want to be in touch with us over the next couple of days to answer further questions. Thank you all for your support.

This concludes the program. You may now disconnect. Everyone, have a great day.