Bill Sheridan
Analyst · JPMorgan. You may begin
Thanks, Jon. I'm very excited discuss the results of our interim analysis. I'd like to start by thanking the patients, site staff, investigators, and their teams here at BioCryst for their commitment and dedication with this important research. In recent days, we've had the opportunity to work through the analyses with our principal investigator, Dr. Emel Aygören, who is joining us on this call and discuss the findings with other leading HAE experts and their patients' advisors. The feedback has been very positive. Before we get into the results, let's go over the design of this placebo controlled clinical trial. The trial has been conducted in sites in several European countries Australia and Canada. Despite two trial was designed to test several doses in two stages, with study drug dose for 28 days. In Part 1, we wanted to look at 350 milligram once daily and in Part 2 which is currently enrolling, we are looking at 250 milligrams and 125 milligrams daily. The key elements of any HAE study is the qualifying rates with angioedema attacks, for this study we selected a minimum range of two attacks per month or 0.45 per week and we expected to see an average qualifying rate about twice that. Another key element is have the data collected on our test symptomses reviewed as in our previous studies, we had an expert panel of three independent HAE physicians adjudicate the attack diary. Also as in our previous studies subject head to head access to specific medicines to treat HAE attacks hugely either a C1 inhibitor given intravenously or a CAT event given subcutaneously. The most common management strategy for HAE patient in the countries where we conducted the trial is on demand treatment with acute attacks. Other prophylaxis treatments such as Amgen or C1 inhibitor IV were not allowed during the study. The main goal of Phase 2 clinical research is to understand dose response and for the full study, we will be able to look at several doses for efficacy and safety profile what sort of drug levels we get and the degree of target enzyme inefficient with those doses. With that information, we will be in good shape to select dose levels for Phase 3 trial. We look forward to those analyses once the study is completed. With the interim analysis of Part 1, it was very important to see if we could establish this drug works to prevent HAE attacks and then estimate the magnitude of benefit. That information and the pharmacokinetics and pharmacodynamic measurements could help us understand if any changes might be helpful for the rest of the study. Turning now to the analysis results, the first question relates to the demographic and disease characteristics of the subject with hereditary angioedema who enrolled in the trial. We had 28 subjects randomized and treated 14H for 7353 and placebo and this defines the intent-to-treat population. Although subjects have to have a medical record documenting mostly one inhibitor levels to support the diagnosis of Type 1 or Type 2 HAE and indeed full 28 subjects satisfied that requirement, we could not confirm low levels by Central Reference Laboratory testing for one subject on each arm. In addition, two subjects in the 7353 group which I will describe in more detail later discontinued study drugs early. So that leaves 11 active and 13 placebo for a total of 24 subjects with Type 1 or Type 2 HAE completing 28 days of treatment for the third protocol assessments of efficacy. We are very pleased to see the compliance with once-daily study medicine with near perfect with 98% and 99% compliance in the 7353 and placebo groups. Subject to demographic characteristics, we are generally well-balanced across the two groups with the exception that more subjects on active drugs 11 out of 14 had prior androgen exposure compared to six out of 14 on placebo. Talking about eight subjects had alanine aminotransferase level or ALT above the upper limit of normal at baseline prior to any study drug exposure. This is likely a reflection of their prior end use for the treatment of HAE. As in all cases, these baseline elevations were seeing exclusively in subjects with that treatment history. The average qualifying attack rate was about one per week as we expected and very similar in each arm. This high rate and the low C1 inhibitor level with an average level of less than 20% of the normal mean are both good indicators that the population study in this trial represents a severely affected group with patients with HAE. The statistical analysis plan laid out a series of pre-planned analyses of efficacy, safety, drug levels and PD effects on the target enzymes as indicated in the top box on this slide. One of our goals in this analysis was to help us think about selection of the primary endpoint for a Phase 3 trial. So we looked at the clinical disease outcome with interest namely angioedema attacks in a number of ways including attack rates, counts, and proportion of subjects with Type 3. Similarly we knew the steady state trough levels of this drug are not reached for about six to seven days, so we pre-specified two analysis periods weeks two to four can be effective dosing periods and weeks one to four the entire dosing period. As already noted, we also have third protocol and intent-to-treat population, so we analyze efficacy using both of those definitions. The attack rate and proportionate attack free analyses that we present are based on adjudicated attacks. The results presented are based -- we also reviewed analyses of all patient reported attacks to check the consistency of the set, with or without attacks adjudication. The usual clinical characteristics and time course of abdominal and peripheral angioedema attacks are different. So we also wanted to see what efficacy looked like if attacks were categorized that way. An important point of interim analysis of Phase 2 study is to learn as much as possible from the data. We did in fact see a difference based on anatomical locations. So we also looked closely at the abdominal attacks and sorted them into those that also had peripheral symptoms and those that did not, as indicated in the lower box on the slide. This classification was done with blinded data and we conduct a post-hoc analysis with number of attacks in each of those categories. Now turning to the efficacy results, we were very pleased to see a strong and highly statistically significant treatment benefit and this is just a great slide to spend a couple of minutes on. On this slide, we provide the details of attack rates for the effective dosing period for both the PP and ITT population. For all the right analyses is an analysis of an covariance model with qualifying attack rate as the covariate. For the per-protocol analysis, the placebo attack rate was 0.915 per week and on BCX7353 was 0.343 per week, a reduction of 63% with a P value of 0.006. Looking at the intent-to-treat analysis, the results are very similar with a reduction in attack rate in favor of 7353 with 52%. The strength of this effect is reinforced by the small sample size in this interim analysis. These results are displayed graphically on Slide 9. Together with the results, the entire dosing period in each panel, the left peer is placebo and 7353 weeks two to four and the right peer is weeks one to four. The protocol analyses are on the left side of the slide and ITT analyses are on the right side. As you can see, the benefit from 7353 in reducing attack rates is very consistent across each analysis duration and each analysis population whether per-protocol or ITT. Sensitivity analyses were also conducted. The first approach we took was to set the small amount of missing diary data as a failure i.e. as an attack. The second approach was to impute the attack rate for those missing days using the baseline attack rate. Neither of those sensitivity analyses affected the magnitude of benefit indicating a statistically robust treatment effect. These analyses confirm that 7353 is a safety that's been preventing angioedema attacks in patients with HAE who are severely affected by their disease. The second step pre-specified analyses was to look at attack rates by anatomical location divided into peripherals and abdominal. In these panels we have also indicated the overall results for weeks two to four in the left hand peer bars in each panel that were shown on the previous slide. Per-protocol analysis is on the left and ITT is on the right, as is obvious the benefit from 7353 is much greater in peripheral attacks compared to abdominal attacks, for example with an 88% reduction for peripheral attacks compared to 24% for abdominal attacks in the per-protocol analysis. This is a rather curious difference. The third phase of pre-specified analysis was to look at proportionate subjects of Type 3 and we displayed here the results for all attacks and also by anatomical location divided into peripheral and abdominal and exactly the same sort of format as the last slide. Again as it's obvious the benefit from 7353 is much greater in peripheral attacks compared to abdominal attacks, 81.8% of subjects for pre-peripheral attacks by per-protocol and 78.6% by ITT. Compared to placebo, there was a 51% improvement in the percentage of subjects peripheral attacks compared to 7% improvement for abdominal attacks in the per-protocol analysis. These observations with a parent differential efficacy based on location of attacks together with the blinded adjudicated notes indicating that many of the abdominal attacks appear to be atypical prompted a post-hoc analysis of attack counts and characteristics to try to understand if this observation was really true and why it not be so. Patients who provided an angioedema attack diary and asked to record the anatomical side of the HAE symptoms as well as check off what the symptoms were each time they had an attack. That patient diary data allowed us to split up the attacks that had any component of abdominal symptoms into two groups, attacks with no other part of the body affected i.e. abdominal only attacks and attacks that these include other body sites in terms of mixed attacks. The counts of attacks with peripheral, mixed size and abdominal only symptoms show striking pattern here which has the same direction and similar magnitude in both the per-protocol and ITT population. For example for per-protocol there were two, two, and seven attacks in 7353 treated subjects for fewer peripheral mixed and fewer abdominal attacks respectively compared to 22, 12, and 2 for placebo. The 7353 subjects had counts of peripheral on mixed attack categories, 91% and 83% less than placebo and in both cases there is no doubt that the attacks are unequivocally angioedema because the patient can see the event of swelling or specific pink greenness on the skin that heralds an attack. In contrast, the counts of attacks with symptoms were confined to the abdomen showed the exactness reverse pattern with for the per-protocol analysis seven on 7353 and two on placebo, this reversal makes no sense at all. There is no reason that a kallikrein inhibitor circulating plasma would not work in all locations in the body. We also noted the atypical nature of many of the attacks for example symptoms lasting several days not changing after a single dose of HAE medicine and not interfering with continued study drug administration. That really isn't compatible with an abdominal attack of angioedema. The likely explanation is quite simple and that is with patients that are so used to abdominal pain or nausea or vomiting being part of their disease that they assume that that's the case whenever those symptoms happen and record those symptoms as an attack and some subjects may not have been able to distinguish drug related GI upswift from early signs of an attack. We then looked at the counter events of adverse events for abdominal pain, nausea, and vomiting. That fibrillation is in the left hand column on Slide 15 and shows only one event in each category on 7353 and none on placebo. We would have expected to see more GI adverse event in the active arm based on the event rates recorded in the Phase I study in healthy subjects at this dose. If we tabulate the same symptoms recorded on the HAE attack diary page for all the attacks that had any abdominal symptom and using the same classification into abdominal only versus mix as on the previous slide, we have for example, one and six events of abdominal pain of 7353 and placebo in mixed attacks and the reverse pattern in the abdominal only attacks seven for 7353 and three for placebo. The excess in the 7353 arm for symptom count in the abdominal only group are more in line with drugs related to AE than with angioedema attacks. To understand the potential impact of any misclassification of abdominal pain, nausea, and vomiting as an HAE attack instead of a study drug related adverse event, we can look at all of the days that patients in the ITT population recorded any symptoms of HAE in their diary which is represented on the slide. I compare that to what happened when we strip out days in both the 7353 and placebo groups on which there were only one of those symptoms. Each row in this table represents one study subject and each column one study day, one through 29, the top half are 7353 subjects and the bottom half are placebo. Each color square indicate days with symptoms recorded. The white squares indicate no symptoms on that day. This snapshot gives visual sense to the overall benefit in the 7353 subject compared to placebo. These two slides are a good example of the pictures being worth a thousand words stripping out the days in both the 7353 and placebo groups on which there were only the three GI symptoms from both groups makes clear that this likely degree of benefit that we can see with 7353 with now 9 of 14 subjects free of symptoms for the entire study compared to one placebo subject. I would now like to turn to the pharmacokinetics and pharmacodynamic results. We had a very good PK profile in our Phase 1 trial in healthy subjects and we expected to see similar drug exposure in HAE patients as you can see we did. The left hand panel shows the serial blood drug levels on a clinic visit day at steady state and the inset shows serial trough levels. This interim PK analysis has summary PK parameters that are very similar to the healthy subjects with no evidence of having hereditary angioedema effects either absorption or clearance of 7353. The trough levels in other words 24-hours after the previous dose far exceeded the proposed target range for efficacy with values from 11 to 32 times with 50% inhibitory concentration on plasma kallikrein contact activation assay. At the same time point, we also measured kallikrein inhibition; in this assay kallikrein enzyme velocity measured in the post-dosing sample is compared to that in baseline pre-dosing samples. We made sure to activate this plasma to the contact activator to ensure that all of the pre-kallikrein present is compared to kallikrein. Again the results are very similar to those we saw in our Phase 1 study in healthy subjects at the 350 milligram once daily dose level. Turning now to a summary of the safety profile that we saw in the interim analysis. There were no serious adverse events and no severe drug related adverse events. As mentioned previously, two subjects discontinued 7353 study drugs early; one subject discontinued because of persistence of the pre-existing liver disorder. The other subject who discontinued study drug had intercurrent Gastroenteritis and developed abnormal liver function, these abnormalities result. The treatment emerging adverse events seen at least twice in the interim analysis are listed and included the common cold, diarrhea, flatulence, and fatigue. There were no clinically significant clinical examination findings and no clinically significant abnormalities of renal function test and electrolytes, hematology panels or year analysis. One subject, the completed treatment was noted to have an ILT level more than three times the upper limit of normal on the day 29 clinic visit. This subject had a complex medical history with colitis, fatty liver with baseline LSTs above the upper limit of normal and had been treated for more than 20 years with androgens with hereditary angioedema up until three years prior to study. In this population based on abnormalities of liver function are common, we will continue to carefully monitor all safety parameters as our clinical programs continues and look forward to seeing safety profile of 7353 in full study analysis. In summary, the interim analysis achieved all of its objectives. First and most important we saw a large treatment effect that was clinically meaningful, statistically significant, and statistically robust with 64% reduction in overall attack rate compared to placebo in the per-protocol analysis with a P value of 0.006 and 52% in the ITT analysis with a P value of 0.035. These results aren't particularly impressive given the relatively small Phase 2 interim analysis sample size. When we looked at angioedema attacks they were unequivocal and characterized by cripple swelling or the cutaneous ignosure of the disease including those that also had abdominal symptoms the reduction in attack rate was even more striking. Our analyses strongly suggest that some GI symptoms that are being recorded on the HAE diary when subjects have only those types of GI symptoms and not other symptoms with angioedema are in fact related to study drugs. Second, the observed safety and tolerability profile supports continued clinical trial of 7353. Third the drug levels that we achieved with once daily oral dosing of 350 milligrams greatly exceeded the proposed target range of efficacy supporting the study of lower dose levels. To help us think about those dose levels we review the available information on trough levels in dose ranging and Phase 3 trials of another kallikrein inhibitor namely CSL830 subcutaneous form of C1 inhibitor and compare those to trough levels of their drug. To do that, the levels of C1 inhibitor were compared to multiples of its 50% effective inhibitory concentration for kallikrein i.e. with EC50 and we did the same thing for BCX7353 levels using EC50. The left panel is the CSL830 data and the middle panel is 7353 plotted on the same scale. For 7353, we showed both the interim results of 350 milligram in this analysis and the levels of the indicated doses in the Phase 1 healthy subject trial. We have indicated our proposed effective target range of four to eight times the EC50 in the blue bar in each panel. The average level for C1 inhibitor in CSL Phase 3 study were approximately five to six times the EC50. The average baseline level without twice weekly CSL830 was about three times this EC50. So in other words twice weekly subcutaneous dosing of that drug on study increased the levels of that kallikrein inhibitor by two to three EC50 units. With the measured trough levels of 7353 in the APeX-1 interim analysis over 11 to 32 times with EC50, clearly we have plenty of room to lower the dose. As you know, we are already starting 250 milligrams and 125 milligram dose levels in APeX-1 Part 2. The right hand panel updates our previous stimulations of the proportionate subject that we expect to exceed different target levels of trough concentrations. What we do see is to add stimulations for target level of two times EC50 and we can see that with the dose of 62.5 milligrams once a day we would expect more than 50% of subject to exceed that target trough level and might expect to see benefit based on the CSL data. So with this said the interim analyses and reflections on the emerging dose response data from CSL, our confidence in the efficacy potential of lower doses of 7353 has increased. We have decided to expand the evaluation of the 125 milligrams and 250 milligrams dose cohorts and also evaluate some subjects at 62.5 milligrams. We are doing this by adding an additional double-blind placebo controlled Part 3 element to the APeX-1 trial. This path add six subjects each to the 125 milligrams and 250 milligram cohorts and six subjects at 62.5 milligram QD and two additional placebo subjects. At this stage, we have fully enrolled Part 1 and Part 2 is currently recruiting with eight randomized and six on screening. We very much look forward to seeing the results of Part 2 expected in the second quarter and then completing the trial. I could not be more pleased we had such strong results in this interim analysis of the APeX-1 trial. In all of our interactions with patients and expert physicians we are constantly reminded how this almost seriously impacts patients and that there is a strong medical need when all prophylaxis treatment. I would very much like to thank Dr. Emel Aygören, the principal investigator for the trial who helped us in fully announces you have seen today and encapsulate a special trip for several days from Frankfurt to the United States away from her duty clinic in order to help us go through the data. Emel is Head of the Centre for Hereditary Angioedema and a specialist in Internal Medicine and Hemostaseology in the Department of Child and Adolescent Medicine at Goethe University Hospital in Frankfurt, Germany. This clinic is very famous and cares for one of the largest cohort of patients with hereditary angioedema anywhere in the world. I would now like to invite Dr. Aygören to provide her perspectives on the interim analysis findings.
Emel Aygören-Pürsün: Well thank you for the introduction. I want to draw your attention to the fact that the disorder that we are dealing with is a lifelong disorder that is very severe in many cases and is connected with episodic swelling of various body parts and one of the biggest limitations of that disease is that the tests are unpredictable and maybe even life threatening. The scientific literature has shown that in the patient population quality of life is significantly decreased and that many patients are hindered in their academic and professional lives. So when we look at that setting, we can only estimate what any reduction of the attack rates needs to patients especially if we look at the severely affected patients with a high attack frequency. And so the result of the interim analysis of our study are really striking given that we've had severe population in that study and that we have seen significant and pronounced therapeutic or rather prophylactic effect that was achieved by as a great class by an early nanostep [ph] medication given once daily. So from the patient perspective a further success of that prospect will have begun to grow with the pursuit really have made a benefit. For some patients may be even that's changing.