Louis Fries
Analyst · Kevin DeGeeter with Ladenburg
Thank you, James. I would like to shift the focus now to our RSV F Vaccine candidate and in particular our activities over the last year with regard to the older adult target population. And so if you could advance the slide 31, I'd like to just take a moment to step back and remind you were now speaking of our baculovirus SF9 insects cell derived recombinant RSV F nanoparticle. That vaccine has previously been shown to induce broadly neutralizing antibodies specific for sites 1, 2, and 4 on the RSV F protein in both animals and much more importantly in humans. It has been shown to induce protection against challenge in several animal models including cotton rats which have been predictive in previous clinical programs addressing RSV as well as baboons. The vaccine has shown efficacy in a Phase 2 trial in older adults that we'll discuss momentarily. It has shown the capacity to reduce serologic evidence of RSV infection in women of childbearing age by approximately 50% in two trials, the results of which confirmed each other and is currently in active investigation in Phase 3 in pregnant women where we've already shown efficient transposable transfer in our Phase 2 studies. I'd like to remind you briefly of the details of our proceeding Phase 2 trial in the older adult population. We did a Phase 2 trial E201 in 1600 older adults 60 years of age and older in the 2014-15 season that was a randomized, observer blind, placebo controlled trial which used 135 mcg unadjuvanted dose of the RSV F vaccine. Those subjects were followed for a year for safety, immunogenicity and efficacy endpoints. In brief, the study showed a very well-tolerated vaccine with a benign safety profile. It induced roughly 5-fold increases in palivizumab competitive antibodies and anti-F antibodies in this older adult population. There was a 1.8% placebo attack rate of RSV moderate to severe lower respiratory tract disease that I'll call msLRTD subsequently in the 4.9% placebo group attack rate for all acute to Siebel group attack rate for all acute respiratory disease due to RSV ARD. We showed a vaccine efficacy of 41% against RSV ARD which was a prespecified endpoint in that trial and a 64% efficacy against RSV msLRTD which was a post hoc analysis. Based on that data we embarked on the Phase 3 trial. On the next slide, Slide 32 I'll just remind you of the Kaplan-Meier curves from that preceding Phase 2 trial which showed a quite convincing separation with the first case begin recorded in October 31, 2014 and persistent and widening protection against RSV ARD that was observable all the way through our last case on April 15, 2015 and gave a significant result in this analysis with a p value of 0.039. So armed with that information, we proceeded to Phase 3 and let's go right now to Slide 34. The E301 design included the same target population, older adults at least 60 years old and again it was a randomized, observer blind placebo-controlled trial that was expanded to 60 U.S. sites from 10 sites in the E201 trial. The design was simple. It was 1:1 randomization. Subjects received either the active vaccine at the same 135 mcg on adjuvanted dose or the formulation buffer as a placebo. Those subjects who had not had the 2015-16 influenza vaccine or I IV and that made up about 45% of the population, were offered co-administration on day zero and about three quarters of them accepted that. Randomization was stratified on whether the subject had previously had influenza vaccine, whether they had congestive heart failure and/or chronic obstructive pulmonary disease or COPD their place of residence whether they lived in the community independently or whether they lived in some form of congregate facility and their age, whether they were less than or greater than 75 years of age. They had a standard safety follow-up through a year, serology samples taken at multiple time points and most importantly had active and passive surveillance for acute respiratory disease through May 1, 2016. They were followed to detect RSV ARD or all acute respiratory disease due to RSV, that was any one symptom with RSV detected by reverse transcriptase polymerase chain reaction or RSV msLRTD which required at least three of five lower respiratory tract symptoms and again RSV detection by RT-PCR. Enrollment spanned the 9th of November to the 12th of December 2015 and the last day 182 visit occurred on June 13, 2016. Moving to slide 35 we see the demographics of the study population. We see that the two treatment groups were very well balanced. The mean age in both groups was just a hair under 70 years. About three quarters of the population were 60 to 75 years old and 23.5% was over 75. Our target for the over 75 group was actually 25%. So we hit that very close. 42% were males, 57-58% females and the population was roughly 88% white, roughly 10% black. About 7.5% had either congestive heart failure or COPD or both, 98 or rather 99.9% of the population were independently living in the community. Only 0.1% to 0.2% were either in assisted living or long-term care. As we said, about 55% of subjects had had flu vaccine before they entered the trial, 45% had not had flu vaccine and of that 45% three quarters or about 36% got flu vaccine on day zero along with the RSV F vaccine. On the next slide, Slide 36 we review the top line safety data. I'm not going to discuss safety in any detail because the vaccine was profoundly benign. You can see in this case that there was less than 2% difference in the incidence of all adverse events across the population. The only difference that rose above 2% was in local solicited adverse events, so a little bit of soreness or pain at the injection site, perhaps a little swelling at the injection site that was typically mild and occurred about 3% more often the vaccine than placebos. All the other categories are essentially comparable between the placebo and the vaccine group, so this is a very benign safety profile. On the next slide, Slide 37, we come to the primary endpoint which was RSV msLRTD and look at the evaluation of efficacy. In the upper panel the per protocol efficacy is described and you can see that there were 26 cases in the placebo group and 28 in the vaccine group giving an actual efficacy estimate of negative 7.9% that was essentially no different than the zero and you see the p value is very large and the confidence limits around the efficacy estimate span minus 84 to plus 37% so not distinguishable from zero. We look at the efficacy and some of the randomization strata below and see none of these are significant, none of them are differentiable from zero. The very large negative percentage in the group with COPD or CHF is conditioned by the fact that you have a very small number of cases there. So that huge negative efficacy is really not very meaningful. I would direct your attention to the two rows that outline subjects who received influenza vaccine or I IV on day zero versus those who could not and I would point out that you can see a negative impact, although neither of these results are significant. There's an obvious negative impact on efficacy against RSV msLRTD in subjects who did receive influenza vaccine on day zero. Looking down to the next panel down you see the intention to treat analysis and you'll note that the numbers there are virtually identical with the per protocol analysis. There's only a 47 subject difference between the intention to treat and the per protocol population and the numbers are essentially identical in contrast with the E201 data where we saw 64% efficacy with this endpoint. I would ask you to know two things though, first of all look at the breadth of the confidence interval around the Phase 2 estimate. It spans 1 to 87% and that broadly overlaps the efficacy estimate confidence interval for the current study 301. The other thing I'd point out is that the rate of this endpoint in the placebo group in E201 was 1.75% where as in both the ITT and per protocol populations in 301 it was very much smaller fourfold smaller at 0.44%. Let's move to Slide 38 now and this examines the secondary endpoint all RSV acute respiratory disease. Again, looking at the per protocol data, here you can see the picture is a little brighter. We actually have 12.5%, 12.6% point estimate of efficacy. However, that is not significantly different from zero and you can see that the confidence balance broadly overlaps 0 from minus 14 to 33%. A similar pattern when you look at some of the substrata of the population, again a negative impact of concurrent receipt of influenza vaccine on day zero. Again, no difference in the analysis between the per protocol efficacy population and the intent to treat efficacy population and again the same difference or the same relationship to the E201 results, lower efficacy but broadly overlapping confidence intervals around the estimate for efficacy and notably a 2.5 almost threefold reduction in the rate of endpoints in the placebo group 4.9% in E201 versus 1.98% in the 301 trial. Next Slide 39 looks at a number of healthcare utilization end points that we prospectively studied in this trial. These look at point we prospectively studied in this trial. These look at medically attended events that were related to RSV acute respiratory disease. And you note that there is an impact on these events not surprisingly because they are correlated with RSV acute respiratory disease, but with the exception of emergency room visits you see a reduction in the frequency of physician called, physician office visits, urgent care visits, any outpatient intervention or hospitalization for RSV acute respiratory disease. The next slide, Slide 40 you see some further explorations in this direction. One of the problems that we've noted in carrying out these studies is that often when elderly subjects present to the hospital with acute disease it may be RSV disease, they present to another hospital, not one that is affiliated with a particular clinical site and at that point they pass out of the jurisdiction, the IRB that's approved the trial and into the jurisdiction of another IRB. So we can't send site personnel into that hospital and get a swap. So we can't confirm RSV associated hospitalization in many cases and therefore there's only two, here observed in this very large trial, but both are in the placebo group. So we stepped back and we examined all-cause hospitalizations either for any cardiorespiratory diagnosis or more specifically any respiratory diagnosis and then looked at any hospitalization and from the least specific of these any hospitalization where there's a 3.6% impact of vaccination if you step to a more specific category any cardiorespiratory hospitalization you see a 10.8% impact and with any respiratory hospitalization a 22.5% impact. So very similar to the magnitude of impact we've seen on RSV acute respiratory disease. Very interesting and something that we discovered post hoc actually when we were examining in detail the impact of this vaccine on safety events is the incidence rate of all-cause serious adverse events due to COPD exacerbations. These are patients with pre-existing chronic obstructive pulmonary disease who get transiently hospitalized because the disease gets worse and they may have hypoxemia, they may have lots of sputum production, increased wheezing, increased difficulty breathing which required them to be hospitalized briefly. And those subjects and those events were actually fairly markedly reduced in the vaccine group. Now I think we have to view that with some caution because of the post hoc observation, but it is certainly tantalizing especially because RSV is known to be associated with deteriorations in chronic obstructive pulmonary disease. The next slide, Slide 41 is the first slide that looks at immunogenicity. This is preliminary data for anti-F and for palivizumab competitive antibodies drawn from a representative subset of subjects as you can imagine is the 12,000 subject trial with all those serologic endpoints I pointed out. So the testing here will go on for quite a long while. There are scores of thousands of tests to be done. However, what we can see is that the RSV F vaccine in this case caused about a 6.1 fold increase in the geometric mean ELISA units for anti-F IgG and for palivizumab competitive antibodies about a 4.47 fold increase in the concentration of palivizumab competitive antibodies. And if you'll think back, I described the increases in E201 is approximately fivefold in both those parameters. So the vaccine is behaving about the same with regard to these measurements and in the placebo group you see full rises of essentially one. So this confirms that this is a vaccine specific event. On the next slide, Slide 42 when you look at microneutralization titers these are complete set from the study population. Again you see basically nothing happening for either RSV A or RSV B microneutralization in the placebo group whereas in the RSV F vaccine group we see 1.6 fold rises in geometric mean tighter for RSV A and 1.8 full rise for RSV B. These are again very close to the values that we observed in E201. So the next Slide 43 to summarize, we had a safety profile that was consistent with Phase 2. We had attack rates of both all-acute respiratory disease due to RSV and moderate to severe lower respiratory tract disease that were markedly lower than expected. In terms of efficacy, we did which failed to show the efficacy that we set out to demonstrate we did find a trend suggesting efficacy against acute respiratory disease and that was supported by a number of observations in terms of healthcare utilization and respiratory hospitalization. We did notice that efficacy against both moderate to severe lower respiratory tract disease and acute respiratory disease was improved in the absence of influenza vaccine administration and that is perhaps no not too surprising. Influenza vaccine often plays the bully on the block with other vaccines when they are administered concurrently and so that could be a coherent picture. Our immunogenicity results were broadly consistent with Phase 2. The microneutralization increases were modest, but they were fairly robust PCA and anti-F responses. We have ongoing analyses to look and see if any of these measures provide a correlative risk and just a couple days ago we got our first encouraging results from the - that, that PCA in particular may correlate with a reduction of risk. The curves are shallow and the work is ongoing to understand that, but it's a tempting bit of data. Let's move on now to the E202 results in the next slide and the next slide please? As you will recall E202 was a so-called rollover trial that was designed to examine both the need and the benefit of annual re-immunization. We did that by enrolling older adult subjects who had graduated from E201 and provided kind of an unique opportunity to evaluate our ability to enhance the vaccine response a year after initial dosing to look at the duration of vaccine efficacy and the ability to maintain and recover efficacy one year after the initial dose and the safety of repeated dosing which we always have to keep in mind. We enrolled 1329 of the original Phase 2 participants and we randomized them into four different cohorts. Otherwise the trial design was similar to the Phase 2 and 3 trials. So as you recall, in E201 we had a vaccine and a placebo group of equal size and we split them again randomly 1:1 to receive vaccine or placebo. So in E202 we had a group that had received vaccine twice. We had a group that received vaccine in the first year, but not the second and so we can look at residual protection. We had a group that had received placebo in the first year and vaccine in the second and so this would essentially be a confirmatory check on 301 and then we had a placebo and placebo group to serve as a control. Moving to the next slide, Slide 46 we found in terms of safety that there was a minor increase in mild injection site pain in people who had received a second active dose. So were being boosted in the second year, but otherwise there were no obvious differences in safety. Our surveillance data in E202 demonstrated that the rates of all acute respiratory disease which were 2.4% and msLRTD which was 0.3% or low in 2015-16, so those basically confirmed the performance in E301. And then in terms of efficacy, we saw really no evidence of residual protection of an initial dose when you look to the second year that didn’t terrible surprised us. In older adults antibody responses do decline fairly rapidly. We saw no efficacy of a single dose in the second year, so that was a confirmation of the E301 results. But tantalizingly we saw the group that received vaccine in both years had 75% and 100% efficacy against ARD and msLRTD respectively. They weren’t statistically significant. The numbers are small. But they do stick out and catch your eye. There is there is clearly different efficacy in the individuals who had received two doses. The immunogenicity analyses of that study are ongoing. They are going to be complicated. The performance of the first dose appears to confirm E301. The second dose actually quantitatively looks less but we may have tools to understand that a bit better and Dr. Glenn will speak to those subsequently. So let's move on now to Slide 48, obviously we were disappointed with the E301 results and we wanted to understand what appeared to be a fairly striking divergence with E201. We looked at the trial design differences and there were some based on our increasing experience in time with doing efficacy trials in this population, but nothing that we could think of that would explain the divergences. So we came up with a couple of hypotheses. We wanted to know if the conduct in E301 was flawed. Did we mess up in some way? Obviously we have made it a much larger trial went from 1600 to 12,000 subjects and was there flawed conduct on that basis? We wanted to know if there were differences between the test articles. There had been some maturation of the process and we wanted to know what the test articles in E301 and E201. And the last thing we wanted to think about was the impact of the unusually low attack rate that we saw in E301. So, on Slide 49 we first looked at the performance of the individual sites and where subjects reporting were illnesses that were reported being swabbed and were they being swabbed in a timely manner and point of fact they were. We found that 83% of all reported illnesses received swabs and more than 98% of those were within the five-day time limit carrying out the swabbing. I should note that this is and when doing large studies like this in older adult population this kind of performance is really good. This is industry standard. I reflect back on for example, the older adult trials Sanofi carried out for their high dose flu vaccine. They were able to recover swabs, diagnostic swabs on 79% of their older adult population. Similar, somewhat lower numbers in GSKs older adult trials. So point of fact this is very - this is quite satisfactory performance in terms of diagnostic testing of the subjects. We also looked at both the ascertainment of illnesses overall and the ability to recover swabs in both the 50 newly trained sites and contrasted them with the 10 experienced sites from E201 and found very minor differences, certainly nothing to explain the difference in attack rate that we saw in this study. So then we turned our focus to the RT-PCR lab and we had actually anticipated that we want to know this data. So we carried out on several occasions blind - first of line we had carried out an extensive validation protocol of the lab before starting this study. That was a protocol that was provided to the FDA and we adopted every comment they had on it, so we carried out an extensive validation. And then several times during the course of the trial we administered externally prepared blinded proficiency panels including RSV and other viruses at the beginning and the end of the study and the lab performed 100% in terms of the sensitivity and specificity on those panels even when the RSV was spiked in at the very limit of detection of the asset. So we are very satisfied with the performance of the lab and that was validated because the lab detected whole bunches of other viruses, coronaviruses, rhinoviruses and captured influenza rates and at times that exactly reflected national surveillance. So the lab was performing well. We looked at the integrity of randomization. We validated that against this pharmacy dispensing records at the sites and also by looking at the post vaccination antibody responses we found a very, very low rate of randomization error of 0.4% that could not have explained the result. So the study appears to have been performed well. We also looked at manufacturing changes and we reviewed an extensive data package that was generated and submitted to FDA that established comparability between the two processes. We looked at comparative animal immunogenicity of limiting doses and we looked at post vaccination immune responses and all of these were consistent with the Phase 2 trial. So there was no obvious evidence of trial misconduct or product issues to explain the divergence, certainly nothing that's occurred to date. We continue to examine the product to make sure there are no issues. On the next slide, Slide 50 we also looked at the potential influence of low attack rate and the influence of low attack rate really there's no history to understand the influence of low attack rate in RSV F trials in adults or RSV vaccine trials in adults, but there simply are no large RSV vaccine trials in adults other this one. But the influenza literature does give us some guidance there and in the influenza literature the modulation of detectable efficacy by attack rate is very well recognized in the history of influenza clinical trials. And there are multiple instances of studies using the same vaccine products, the same investigation, the same trial methods and the same study populations that can show 40% to 50% decrements in observed efficacy in sequential years when the attack rates are low. Interest in that effect, the influenza strain changes, but that effect is independent of influenza strain match and you can see the effect of low attack rates depressing observed efficacy even when the strain match in the low attack rate year is better than in the high attack rate year. This effect has been recognized and analyzed in the epidemiologic literature and there's a quote that I won't read from a paper by [indiscernible] and Halloran, but basically that recognizes the fact that efficacy estimates simply may not be comparable unless there is an explicit control and similarity in baseline transmission. We know that in 2015-16 it was a relatively low transmission year in adults that's validated not only in our data set by in data sets by other groups who have been doing either systematic surveillance in adults or in the just by anecdote in people who are RSV interested clinicians who have been following transmission in adults. Our trials of RSV vaccines in older adults are really the only other non-influenza large-scale trials in older adults that address an annually recurring respiratory virus for which prior infection modifies but doesn't eliminate the susceptibility of those. So we think there's a good chance that the low attack rate in 2015-16 may have modulated our ability to find efficacy, possibly because the background resistance in the population was high. We're still trying to understand that. It's an area of new understanding for RSV, but there is an influenza precedent. The next slide, Slide 51, so that brings us really to whether we can establish a unified view from 201 and 301 and 202. This is a hypothesis, but it's become our way of looking at these data. In the Phase 2 trial in older adults E201 we saw significant efficacy against ARD but with wide confidence intervals. The trial was not really powered to detect that level of efficacy. The original power design was targeting say 70%. So the wide confidence interval is not terribly surprising. We also saw 64% efficacy against msLRTD that was noted post hoc. The circumstances, the RSV attack rate was relatively strong, almost 5% right in the middle of the historical range published for example in the studies by Ann Falsey in Rochester. Retrospectively we saw an impact of influenza vaccine co-administration there as well. Moving to the Phase 3 we had a valid trial shows efficacy that's not significantly different from zero for both ARD and msLRTD, but again broad confidence bound because of the low attack rate really and in fact there was this confidence bounce on retrospective examination pretty broadly overlapped by 50% to 60% the confidence bounce in E201. The conditioning circumstances again, low attack rate and repetition of the I IV effect. Last, but not least, there is the Phase 2 rollover trial in older adults where there is suggestion of efficacy that kind of sticks out in people who have got two sequential doses. True, they were separated by a year, but we did something different to the immune system in those people, perhaps than we did in E201 or E301. So our synthesis of this data is as follows: we believe the 201 and the 301 results may be driven by the same underlying efficacy that is the really drawn from the same statistical distribution of results. We got lucky in 201, in 301 we were facing a low attack rate year and we got unlucky. If that's true then the weighted average estimate of efficacy might be 19%-20% against ARD. We know the co-administration of I IV and a low RSV attack rate in 301 both conspired to make success difficult to achieve. So we believe the vaccine can have efficacy and we think it's shown us the efficacy, but it needs enhancement of the immune response in older adults and one of the things we can take away from E202 is that it may suggest a way to achieve this. We may have to hit them twice in a regimen that is designed to make that palatable for use in the real world. Alternatively, adjuvants may help us and my colleague Dr. Glenn is going to present some suggestive data later that indicates we may have a path forward. In the meantime, the next, so next slide, so going forward we're going to continue development of RSV F. We know that immunogenicity is more vigorous in younger adult populations including pregnant women and especially children and adults with low baseline titers coming in. We do believe there are several avenues for enhanced efficacy and immunogenicity that we can explore in older adults and we're going to do that. We do have some concern that our current suite of immunologic endpoints may not adequately capture protective responses and we have work underway that again Dr. Glenn is going to speak about to enhance our analytical tools. So returning to older adults though we're looking at dosing and formulation approaches that can overcome the heterogeneity and background immunity for attack rates. So, on the next slide, as you've heard we are going back to Phase 2 in the older adult population with a trial that we call E205. We anticipate starting our trial in the first quarter of 2017 and having topline data in Q3 that will be done in the Southern hemisphere to avoid the confounding of the RSV seasons. We want to ascertain whether adjuvantation or two-dose primary regimen can alter both the quantity and importantly the quality of immune response in older adults. And we also obviously evaluate the safety of those revised regimens to see if they are still suitable for use. The endpoints will be safety, but also RSV specific immune responses by microneuts, by anti-FIgG, by power vision competitive antibodies, but also crucially looking at antibody avidity assessments especially for the site to target peptide and perhaps others and also to look at T cell responses. That will be a trial of 300 healthy older adults, a randomized observer blind design with both 1 and 2-dose regimens with and without aluminum phosphate and with and without our proprietary Matrix-M adjuvant. And at that point I will conclude and hand the presentation over to our President of Research and Development, Dr. Glenn.
