Regeneron Pharmaceuticals, Inc. (REGN) Earnings Call Transcript & Summary

Welcome to the Regeneron Pharmaceuticals COVID-19 Antibody Program Update Conference Call. My name is Gigi, and I will be your operator for today's call. [Operator Instructions] Please note that this conference is being recorded. I will now turn the call over to Justin Holko, Vice President, Investor Relations. You may begin.

Thank you, Gigi, and welcome to everybody listening to our webcast as we provide an update on the Regeneron COVID-19 Antibody Program. Joining me today are Dr. Len Schleifer, Founder, President and Chief Executive Officer; Dr. George Yancopoulos, Co-Founder, President and Chief Scientific Officer; Dr. David Weinreich, Senior Vice President, Head of Clinical Development; Dr. Christos Kyratsous, Vice President, Infectious Diseases and Viral Vector Technologies; and Hala Mirza, Senior Vice President, Corporate Communications and Citizenship. After our prepared remarks on the accompanying slides, which can be found on our website, we will open the call for Q&A from both the financial community and the media, starting first with the financial community. Before handing the call over to Len, I would like to remind you that remarks made on today's call include forward-looking statements about Regeneron, including those related to Regeneron's business and research and development programs, anticipated milestones and regulatory matters. Each forward-looking statement is subject to risks and uncertainties that could cause actual results and events to differ materially from those projected in the statement. A more complete description of these and other material risks can be found in Regeneron's filings with the United States Securities and Exchange Commission. Regeneron does not undertake any obligation to publicly update any forward-looking statements, whether as a result of new information, future events or otherwise. With that, let me turn the call over to our President and Chief Executive Officer, Dr. Len Schleifer.

Well, thank you, Justin, and thank you to everyone for joining the call. I hope all of you are staying safe and well. We're excited to share with you updates on our COVID-19 Antibody Program and to share important data that we feel contributes meaningfully to the scientific understanding of the virus, which patients are most at risk and where treatment intervention may be most impactful. In addition, we believe our data will have positive implications for the likelihood of success of vaccines that target the same spike protein that Regeneron CoV-2 targets. As you know, we began our work on this program back in January, utilizing our proprietary end-to-end technologies. Our scientists quickly isolated thousands of antibodies against the virus and rapidly advanced 2 potent neutralizing antibodies to form a cocktail to both treat and prevent infection. In a matter of months, we began preclinical testing, manufacturing activities and clinical testing in humans. This summer, we secured a manufacturing and supply agreement with BARDA and the U.S. Department of Defense. And most recently, we announced that we are collaborating with Roche on the development and manufacture of the antibody cocktail, significantly increasing the amount of drug that can be supplied globally. Today's announcements mark the next critical milestone in our quest to create a potential therapeutic for COVID-19. As you will hear from George, infection with SARS-CoV-2 starts a race between the virus and the host immune syndrome. Most of the time, the immune system wins that race but not always and often, not quickly enough. Regeneron CoV-2 antibody cocktail is designed to be a substitute for the natural immune response and help the host win that race. Before handing the call over to George, I want to extend my gratitude to our many colleagues, collaborators, patients, families and investigators who made today possible. Without you, today's announcements would not have been possible. Now I'd like to turn the call over to George, who will tell you how Regeneron is doing in its race to beat this virus. George?

Thank you, Len. Slide 3. So I'll be telling you about our data that shows that our antibody cocktail reduces viral loads and symptoms versus placebo in nonhospitalized patients who are infected with the SARS-CoV-2 virus and in particular, that the great improvements were observed in those patients who are probably most at risk and have most to benefit. That is those who had not yet mounted their own effective immune response prior to the treatment. We're going to refer to these as antibody seronegative patients or high viral load patients at baseline. We already begun sharing these results with regulators, and these results will be used to inform next steps. Slide 4. Thank you. As Len already noted, this has been an exceptional whole company effort in the midst of one of the most challenging health crises in history. And as a New York-based company, we were literally located at the epicenter of this pandemic for many months. And I want to thank all of our people who made incredible sacrifices to do their best to create and bring this new treatment in record time to the many patients and physicians who are so desperate for a way to fight back against this horrible disease. And our ability to rapidly respond was only possible because of the decades we have committed to becoming pioneers in the field of antibody and antibody-like technology, which have resulted in many important new medicines for conditions ranging from blindness to allergic diseases such as asthma and atopic dermatitis to heart disease and cancer, and we hope many more. We first applied these technologies to infectious disease for one of the most deadly such diseases that is Ebola, where our technologies and people rapidly developed an effective antibody cocktail, as shown in a trial carried out in the Congo by the World Health Organization, which showed that our antibody cocktail had a marked survival benefit in patients already infected with the Ebola virus. And this antibody cocktail is on track to be the first approved Ebola treatment with a PDUFA date coming in the next month. Our scientists, based on this experience and track record, felt shortly after the outbreak of the COVID epidemic that it was our duty and our obligation to use our abilities to deliver an antibody cocktail for this disease as rapidly as possible. And as we've already described, our scientists produced an antibody cocktail targeting the critical spike protein that was able, in preclinical animal models, to block infectivity and also to prevent an emergence of viral-resistant mutants. And particularly, we're able to show, in nonhuman primate studies, that this antiviral activity could both be very effective both as a preventative agent and also as a therapeutic for subjects that have already been infected with the SARS-2 virus. I'm sure many of you know a lot of these details. As shown in red in the figure up on the left, this critical spike protein that's on the surface of the virus that's required for interaction with the host ACE-2 receptor is required for fusion and entry into susceptible cells. And as I said, once the infections were first described, our scientists immediately started working on developing this antibody cocktail as schematized in the lower left panel. We developed 2 highly potent noncompeting neutralized antibodies that formed this antibody cocktail that bound to the receptor binding domain of the spike protein shown here in blue. It blocked the binding of the spike protein to the ACE receptor and thus, blocked viral entry, and then enhance the viral complexes to be recognized and cleared by the immune system. And as we showed, as I already mentioned, very impressive effects in preclinical and nonhuman primate studies. This antibody cocktail is already in a broad ongoing clinical development program that's already enrolled over 2,000 patients to date. These studies include both studies in so-called hospitalized patients, patients who are the sickest, many of whom require ventilatory support, both in our own sponsor study and also a major collaborative study with the United Kingdom National Health Service known as the RECOVERY study (sic) [ RECOVERY Trial ] in hospitalized patients. We're also doing prevention or prophylactic studies. We're also studying extensively in normal human volunteers. But today, I'll be talking about our outpatient study, known as the 2067 study, a master protocol to assess the safety, tolerability and efficacy of our antibody cocktail in nonhospitalized patients with COVID-19. We have already enrolled more than 900 patients in this study and the results that we'll be talking to you about include this first cohort of 275 patients, whose analysis we'll describe today. It's part of an ongoing seamless Phase I/II/III trial in this outpatient study. The primary objectives of this study was to establish the safety, antiviral activity as well as potential clinical benefits of this cocktail. And the initial analysis was intended to be a descriptive one. Since so little was really known about the detailed time course and natural history in any patient population, let alone this outpatient setting, our strategy was to first explore the time course and natural history and better understand this disease. We had, very importantly, some prospective hypotheses going in, most importantly, that many patients would recover on their own based on their own individual immune response and that any antivirus effect or associated clinical benefit would most likely be greatest in those patients who had not yet mounted a strong immune response. We hope that these data, if robust, could support an EUA, an Emergency Use Authorization, and provide the criteria for a subsequent formal analysis plan in our next larger cohort of patients. We already have the patients enrolled for such a replication effort, and thus will allow rapid replication of these initial descriptive results. Very importantly, these results that we'll tell you about conform -- confirm many important aspects of the COVID-19 outpatient disease course, and importantly, our initial hypothesis that these patients consist of 2 very different populations: those who have already mounted and began to mount an effective immune response that is an antibody, their own endogenous antibody response against the virus; and those who have yet to do so. The patients, as I'll tell you, who've already mounted their own immune response, have substantially lower viral load at baseline. That is the immune system does work, on average, for most patients. The immune response resulted in substantially lower viral load in the patients who have initiated this immune response, and these patients resolve their symptoms quickly. What we see is that medical complications in our study populations are uncommon, occur in 10% of the patients treated with standard care, but they [ over ] only occur as we would predict in these very patients that we think we have the opportunity to help the most, those who have yet to mount their own immune response. So as I said, we have this prospective hypotheses regarding patients who are mounting their own immune response to the virus. So before treatment, we used serological analyses to divide the patients into those who are serology antibody positive. That is they had measurable endogenous antibodies that their own immune system was making to COVID-19. As opposed to those who are seroantibody negative. That is they were not yet making any measurable antibodies. And what we'll show you in a lot of following analyses that, as expected, the seroantibody-positive patients, because they were already fighting the virus on their own, had much lower viral levels when we initiated their treatment as compared to the seroantibody-negative patients, and they rapidly achieved viral loads approaching the lower limit of quantitation even without any treatment. In contrast, those patients who, at baseline, were seroantibody-negative, had significantly higher viral loads at baseline and cleared the virus much more slowly on their own in the absence of treatment. And it was in these patients that we had the much more obvious treatment benefit, as we would have expected. Next slide. So this slide shows in tabular form -- and you'll have all the data. It'll be uploaded for you guys to look at all the details. But it basically points out that, at baseline, we could easily divide the patients approximately in half into those who had already mounted an immune response. And they had 4 logs, higher levels of virus at baseline if they did mount their immune response as compared to those who mounted immune response. I know it's probably a lot easier to see this on the next slide, shown in graphical format, which shows either on a log plot on the left or I think using a graph and description many of us are much more familiar with, just raw viral load on the right, showing you that in blue, the seronegative patients, those who have not yet mounted an immune response, had much higher levels of viral load at baseline than those who have already started mounting their own immune response. Next slide. This shows just the flip side of that analysis. So basically, if you're asking, who are the patients who you might think are at highest risk, that is those patients who are at baseline before we treat them, have the highest baseline viral load. I just bring your attention to that, highest viral load patients, those who had more than 10^7 copies per milliliter of the virus. As you can see, the seronegative, about 60% of those patients, had such high viral levels, whereas you can see, about 5% of the seropositive patients. So the good news is that the immune system works. It works for most patients. It dramatically lowers your viral titers. However, for patients who have not yet mounted their own immune response, they can have very high viral levels on the level of 1 million or 10 million viral copies per ml, and these are really our target population. Very importantly, to link viral load with, at least in this population, symptomatology, what this slide shows is that these so-called seroantibody-positive patients, those who had already started clearing the virus and have low viral loads, also had a strong tend for faster alleviation of their symptoms than the seronegative patients. This is in the absence of treatment. So this is just an analysis just of the placebo patients, breaking them down by their baseline serology or baseline virology status. And what you see again is if you've already mounted your own immune response without regard to treatment, you're going to clear your symptoms about twice as fast as those patients who have not yet begun their own immune response and have not yet started to clear the virus. So once again, this just confirms this hypothesis that the patients who have the most benefit are the so-called seroantibody-negative patients who have not yet mounted their own immune response. And what we really want to do is turn them into patients who already started to effectively fight the virus. Next slide. So as I said, the prospective hypothesis was to use the serology to divide patients to those who are seropositive versus seronegative and compare how these populations responded to treatment with the notion that it was the seronegative population that would reveal the biggest treatment benefit because they didn't have their own immune response. Next slide. This slide summarizes a lot of the data that we'll be going through in the following slides. The basic point are that in the overall population, we were able to show that our treatment could actually reduce viral levels but that this was basically driven by the so-called seronegative population that I've already described to you. The easiest numbers to probably understand is that at day 7, patients who are starting with high viral loads, you could see 2 log reductions with either the high or the low dose corresponding to about 99% viral reductions compared to the placebo-treated patients. In these patients, who are most at risk who have the highest virus loads at baseline, so those patients benefited the most from treatment in terms of reducing their virus loads by 99% compared to the placebo-treated patients. Similarly, these higher baseline viral patients also had increasingly larger benefits for symptom alleviation, about halving the amount of time that they needed to alleviate their symptoms 5 or 7 days compared to 13 days. I should mention very importantly that both our high and low dose showed similar activity in both the viral reduction and the symptom alleviation. I've already told you that there are very small number of medically attended visits. That is people who needed to, for example, go to the emergency room or be hospitalized. As we already told you, outpatients, in general, we know do well. Very importantly, 10 of the 12 medically attended visits occurred in the patients who are seronegative baseline, consistent with the notion that these are the highest-risk patients. This rate was tenfold than would be predicted based on the overall population, consistent with the notion that these patients are at higher risk. And for these very small number of events, we had positive trends related to treatment as we'll also get into. Very important, the pharmacokinetics of our monoclonal antibodies was linear, consistent with what we've seen with other antibodies with very long half-lives of 1 to 2 months. And this is consistent with our notion that we don't have to artificially introduce mutations or other changes, artificial changes, in the antibodies to try to extend their half-life because our antibodies already have these very long half-lives on their own. And I'll tell you, the safety and tolerability profiles were also very comforting. So briefly to remind you the study design, we had 3 arms in our study, placebo-treated on top of standard of care as well as a low dose and a high dose. And we very frequently assess patients for the first couple of weeks literally every other day, for their viral load as well as for symptomatology and other measures. Next slide. The key descriptive endpoints for both virological endpoints that I've already started mentioning as well as clinical endpoints such as alleviation of symptoms as well as we look for trends in these very small numbers of medically attended visits. Next slide. As you can see from the baseline demographics, the populations were very well balanced across the treatment arms. And very importantly, we had good representation of what are standardly underrepresented populations for such clinical trials. And a high percentage of our patients had additional risk factors. And I've already alluded to the fact that the most key discriminating characteristic at baseline was the serology status, which so highly correlated with the viral load, that as the patients who are already mounting their immune response were already clearing the virus with very high nominal statistical significance as opposed to the patients who had not yet mounted their own immune response. So getting to the virological outcomes. In terms of looking at the time-weighted average, in the seronegative population, as you can see, we hit nominal statistical significance with the high dose. We approach nominal significance with the low dose, and in a pooled analysis of both doses, we also hit nominal statistical significance. This was also seen in the overall population, but as I'll tell you, this was almost entirely driven by the so-called seronegative population. Next slide. And this slide gets sort of at the crux of this very important observation and realization that confirm this prospective hypothesis that the patients who are already making their immune response were clearing the virus and had low viral titers. And so as we've indicated, the placebo patients who were doing this had high antibody levels, had low viral titers. And those who were not mounting responses had high baseline titers. And you can see very nicely here that as you correlate the baseline viral titers, you see something that's very interesting. You see in blue that the placebo patients start dropping more slowly and do not get, even at a week, do not start approaching the lower limit of quantitation. Whereas the treatment effect becomes accentuated in the very patients who have the highest baseline viral titers and who presumably need an enhanced immune response as could be provided by the therapeutic cocktail, as you can see here. So you see the greatest treatment benefit for those patients who are presumably at greatest risk will have the greatest viral loads at baseline and who do the poorest job of clearing the viral load on their own. As you can see, even these high viral load at baseline patients, you see 2 on the right, 2 log differences compared to placebo in terms of viral loads at day 7 with ps of 001s for both treatment groups. This is a log representation, which I know a lot of the technical people like, but I think a lot of us like to just see the raw values. This shows the raw values. And once again, it makes exactly the same point but perhaps using numbers and calculations that are easier to understand. So as you can see, once again, patients with the highest viral load, those are the patients who have the lowest antibody responses, endogenous responses on their own. They do the poorest job of clearing the virus. They derive the greatest benefit from the treatment. And as you can see, by day 7, you have 99% reductions compared to the placebo group in terms of the viral load for both treatment arms. Next slide. What you've just seen in terms of viral effects also are consistent with what we see with time to alleviation of symptomatology. What we're seeing is we see a very modest effect in the overall population, which is driven by the seronegative population where, once again, you're almost cutting the time in half that it takes to alleviate your symptoms. Next slide shows it in graphical representation. So in the bottom left is the overall population where you see the treatment arms have a very modest effect compared to the placebo arm in blue. This is exaggerated as shown in the seronegative population at the top. And you can see in the bottom right, there's essentially almost no detectable effect in the seropositive population. Once again, if you break these out, not by serology status at baseline but by viral load at baseline, you get exactly -- this just highlights the seronegative population [ seen ]. Both treatment arms shifted to the left. That means that symptoms are being alleviated faster in those patients than in the placebo group in the seronegative population. But the next slide did the same sort of analysis, the symptoms that I showed you for viral clearance based on baseline viral load. So you can see at the top left, when you're looking at patients who have, on average, relatively minimal viral load levels at baseline, they see a very modest treatment benefit in terms of treatment alleviation. When you look in the bottom right, the patients who have the highest viral load at baseline, those are the patients that we showed you had the biggest treatment effect in terms of viral load reduction due to our antibody cocktail. Similarly, they seem to derive the greatest benefit in terms of speeding up symptom alleviation. You can see both the low dose in red and the high dose in green curves are shifted to the left about -- speeding up by about twice the rate of symptom alleviation compared to the placebo group, once again, very consistent with what we saw with viral load clearance that those patients either don't have their own immune response at baseline or have the highest viral loads at baseline because they haven't mounted their own immune response, are deriving the greatest benefit from the antibody cocktail in terms of both the ability of the cocktail to reduce virus -- enhance viral clearance and/or to improve symptomatology. Next slide. So this addresses the issue of the small number of medically attended visits. Next slide, which shows that the positive trends are, once again, going with the treatment effect. Very importantly to summarize, on the next slide, almost all of the medically attended visits, 10 out of the 12, occurred in the seronegative population consistent with the notion that these are the people who are having the biggest trouble clearing the virus, alleviating their symptoms but also are at highest risk of needing additional medical attention, which we can hopefully avoid by treating them with the cocktail. So the cocktail seems to address these people, at least in this initial descriptive analysis, in terms of the people with the highest viral load, the people who are not mounting their own immune response, in terms of the cocktail, is greatly accelerating the clearance of virus in these people. It's accelerating their symptom alleviation, and it's -- at least trends in avoiding their need to seek further medical attention. So the next couple of slides, let's go to the half-life and the pharmacokinetics. As I noted, half-lives of 1 to 2 months for the various antibodies and the various dose groups, once again highlighting the point that we do not have to artificially engineer or introduce mutations in these antibodies to get these incredibly long half-life. Next slide. These slides make an incredible point that we were trying to achieve certain target populations with the low dose and the high dose at the end of the 1-month cycle. And you can look at these slides in detail, but just a single patient in a low dose, a single patient in a high dose did not achieve levels above the target concentrations that we were going for. And based on the shape and the look of these curves, these people are going to likely be above the target concentrations for quite a while past the first month, which obviously has important implications. Next slide. The safety slides, once again, you can peruse them, but they're very comforting in that there was no indication of concerns in terms of either treatment emergence or serious adverse events or infusion reactions that they were exacerbated by treatment. So in summary, very importantly, this initial descriptive analysis, we believe, confirmed our prospective hypothesis that the people that we wanted to target were the people who are not yet mounting their own immune response. And in these patients, the so-called seroantibody-negative patients, will also correspond to the patients who had very high viral loads at baseline. These were the patients who are both probably at highest risk and could derive the greatest benefit by enhancing their endogenous immune response with this exogenously provided therapeutic antibody cocktail. We've gone over all the numbers in this and as I said, impressive consistency between what we've seen and which patients benefited and the effects of serological status and baseline viral levels in terms of whether you look at the ability of the cocktail to benefit viral clearance, symptomatic alleviation and, in very small numbers, the medically attended visits. So in summary, I think the key takeaways are what we've already told you. We believe the data indicates that the cocktail reduces viral loads and symptoms versus placebos in nonhospitalized patients who are infected, already infective and symptomatic with the SARS-CoV-2 virus. The greatest improvements were observed in the patients who had not yet mounted their own effective immune response prior to treatment. And this could be exemplified by those who are either antibody seronegative or had high viral loads at baseline that was incredibly high statistical correlation between these 2 measures of patients who had not yet mounted their own effective immune response. And as we said, we've already begun to share these results with regulators. And we're certainly building on this data to determine all of our next steps going forward.

Thanks, George. And just for everybody who's listening to the call, receiving several questions. We will post these slides later this evening. We'll get those posted as soon as possible. We know there's a lot of interest in the amount of data that George and the team have generated and presented today. We're now ready to open the call for Q&A. We have several callers in the queue, so we're going to limit each caller to 1 question. We will begin with callers from the financial community. Gigi, we are ready for our first question.

[Operator Instructions] Our first question comes from the line of Geoffrey Porges from SVB Leerink.

Congratulations on shedding so much light on our understanding of the disease and on the encouraging results. So my question is about the assay. Is the seropositive assay -- or could you give us a sense of the format, the availability? Is it a commercial assay? And it sounds as though you would recommend everybody who's symptomatically screened with both seropositivity and viral load and then directing them to treatment. Is that a fair summary of how you see the data?

Yes. I think you bring up great points, and you bring up, obviously, a great societal challenge. We are hopefully on our way to providing a therapeutic intervention that could greatly impact patients. How we, as society, choose to deploy and employ this potential therapeutic solution is going to depend on having the right diagnostic tools available and particularly, in point-of-care fashion. We used very specific serological testing here that is not necessarily what will have to be used in the field to determine seropositivity. There are a great number of companies that are actually working on point-of-care diagnostic testing for either. Remember, you don't have to necessarily look for serological status only. You can also just simply use high viral load because we showed the incredible correlation. So this is going to be a major societal imperative. If the decision is going to be made to deploy such a therapeutic solution in the patients who might benefit the most and need it the most, we're going to have to solve the problem of using the right point of care, diagnostic tools, either for serology or high viral load and many such as our partners at Roche, Genentech are developing exactly such tools that can be used in concert with this potential therapeutic intervention.

Our next question comes from the line of Yaron Werber from Cowen.

Also congrats on the really nice data set, and thanks for getting it out so quickly. So I have a question. You're seeing a resolution of high viral load. Are you -- George, are you seeing a resolution or an improvement in time to resolution of symptoms or in the severity of symptoms? And then secondly, there was no dose response. Any thoughts on that, why that's the case?

Well, I believe, if I understood your question, we were able to see the correlations with either baseline serology status or baseline viral status, which were so highly correlated, that those patients who had -- who were either seroantibody-negative or who had high viral loads at baseline, those were the very patients that we could demonstrate the best treatment benefit for because they were neither clearing their virus as well nor were they alleviating their symptoms as well. So the endogenous immune system in the seroantibody-positive patients was doing a good job to clear virus and to relieve symptoms and so it's the remaining patients who are seroantibody-negative or had high viral load, which were really overlapping same population of patients who benefited the most for both measures. And that was one thing that was very important in terms of the consistency of the data, which I think is one thing that very much is gratifying to us, and we find very comforting. The robustness of the data and the consistency is that everything goes similarly. It's not like we saw some sort of one trend or one thing going with viral clearance and different thing with symptomatology or medically attended visits. It was all the same. The benefit seemed to be the greatest in the patients who needed it the most. You also raised this important question about the low dose and the high dose. Once again, that was actually very gratifying to us. We, for sure, wanted to provide a very high dose to make sure that we weren't leaving any potential treatment benefit on the table. But what the data suggests that the low dose and the high dose are indicating that we are very close to saturating the potential benefit that you can see here from an antibody therapy because both doses seem to have a maximum benefit compared to each other. The importance of that, of course, is this allows for the possibility that many times as many patients could be treated because the low dose is looking like it could be almost maximally effective. Of course, this is just the first analysis from the first set of 270 patients -- 75 (sic) [ 275 ] patients. We're going to look at many more patients, and we're going to continue to evolve and get details on the story. But I think it's a very promising finding right now vis-a-vis the fact that it looks like we may have saturated the therapeutic benefit effect, and we can achieve it with a low dose, and thus, we can treat many times as many patients.

Our next question comes from the line of Terence Flynn from Goldman Sachs.

Congrats on all the progress on this front. The question I had was, I just wonder, George, does this increase your confidence in seeing an impact in the hospitalized setting and/or the prophylaxis setting, the 2 other areas where you guys are developing this for? And then any changes you're contemplating on either of those trials in terms of patient selection, given what you saw here in this group?

Yes. I think those are excellent questions, and they're really important. The fact that we're using the same doses in the hospitalized patients and the fact that we were able to saturate the effect here gives us a lot of hope that we are going to be at doses that could benefit those patients as well. Another critically important observation that we are now revealing to you that we are intending to put into the deck but we didn't have time at the last minute, is that I think many of us believed early on, which is why we did the world's first large Phase III study using an interleukin-6 inhibitor, that the patients who were hospitalized and who were suffering the most and had the biggest respiratory problems, that these problems were not due to high viral load but due to the body's response to the high viral load, the so-called cytokine storm problem. As we all know, those studies did not -- in our hands and in other people's hands, did not show a dramatic benefit. But what we and the rest of the world have seemingly learned since then is that the hospitalized patients seem to actually be a subset of the patients who are not doing as well at clearing the virus. And they actually have the highest viral levels, consistent with the notion that the virus is still playing a very important role in these hospitalized patients. So what we will probably release in some form, but what I can tell you right now is that, on average and very surprisingly to us but has now also been reported by others in the literature as the hospitalized patients on average have tenfold higher levels of virus at baseline entry into our study than the outpatient study patients. Consistent with the notion that maybe these patients are sicker because they have more virus, that their body is doing worse at dealing with the virus and that there's a chance that a treatment that actually has already been shown to reduce viral load could potentially benefit the patients. So we think that this data very much, both the viral data and the appreciation of how important virus is and the fact that we can lower it and that the doses are both [ maximally ] able to do it, give great hope for the hospitalized patient study. But it also, likewise, it gives great hope for the prevention prophylaxis or household contact study. In there, we actually -- our preclinical studies has shown that you need much lower doses to prevent infection than to treat it. We're using half the dose of the low dose in our prevention studies. The fact that twice the dose was able to have a maximum treatment effect bodes very well that half the dose is going to have a long-term prophylactic benefit in prevention in our household contact study. That, together with the exceedingly favorable half-life of these antibodies, that suggest that it may be in the therapeutic or preventative range for a long term, so very positive implications. And I should also say that I think the positive implications go beyond just our work. I think that this -- our study, of course, has positive implications for all directed antibody therapeutic efforts but also for vaccine efforts. Remember, our antibody or antibody cocktail targets the receptor binding domain of the spike protein. The whole point of vaccines is to try to induce an endogenous antibody response that will similarly target the spike protein in the receptor binding domain. So we think that our findings have implications for our studies, for target antibody therapeutics in general and for vaccines, even more generally, putting aside the fact that we all know that vaccines, because of their mechanism of action, trying to induce a host immune response, are more prone to, of course, immune sort of side effects and concerns and autoimmune problems that people have long been concerned with vaccines. But in terms of the efficacy, I think that these studies speak positively, as I said, not only to our future studies but to other antibody therapies and to the efficacy of potential vaccine therapies as well.

Thanks for the question, Terence. So we're going to take one more question from the financial community. And then we're going to switch over to callers from the media. Just a word. Following this call, the IR team as well as David and Christos will be able to speak with analysts and investors to answer additional questions.

Our next question comes from the line of Chris Raymond from Piper Sandler.

Congrats from us as well. So just make it quick here. George, I was wondering if I could maybe drill down a little bit on the last question. So I guess, we're assuming that there's going to be fewer hospitalized patients that are seronegative. And given your data, you're showing the best impact in those -- on patients that are seronegative. And I guess, first, is that assumption correct? And if so, is there a protocol to test for seropositivity in hospitalized patients?

So we're going to have to, at some point, just because I think the world needs to have access and be aware of this data, that data in a more detailed manner. But what we can tell you is you are right that they are a somewhat lower percentage. However, there are still substantial numbers of patients, almost 30% of patients, are still seronegative in the hospitalized study. Moreover, the patients who are seroantibody-positive, they themselves have higher levels of antibody, suggesting that their immune systems are just not doing -- that you can get antibodies, but some people may not get the best kind of antibodies, the sort of antibodies that are in our cocktail. So the notion is some people will not be maybe mounting a good immune response, for example, those people who might, for example, have low interferon levels or so forth. And other people may be mounting antibody response, but they may not be managing the right kind of antibody responses. So I think that this gives us a lot of hope that endowing these people with an exogenous provided excellent immune solution here could really help lower the higher viral loads that we are seeing in the hospitalized patients, whether they're seronegative or seropositive with still high antibody levels. And I got internally in the room another point that I forgot to make but about sort of the point, not about hospitalized but about the medically attended visits in general. And this point I should have made during the presentation. The fact that almost all the medically attended visits, 10 out of 12, were seen in the seronegative population, even though they comprise only about 40% of the population, shows the importance of the endogenous immune system because if these events would have occurred randomly, if there were 10 in the seronegatives and there were 60% of patients with seropositive, you would have expected 15 to 20 events turning to seropositives. The fact that you saw only 2 suggest that, for most of these people, their endogenous system, their endogenous immune system is controlling it. And that's why you didn't see 15 to 20 medically attended visits in the seropositives. You only saw 2. Moreover, I think this speaks to the question that you just asked about the hospitalized patients. It's the rare seropositive patients who, for whatever reason, their antibodies aren't doing a good job of clearing the virus, maybe because they're not making the best of antibodies, will potentially benefit from a really good therapeutic solution.