Welcome to the BONESUPPORT Press Conference 2021. For the first part of this call, all participants will be in listen-only mode, and afterwards, there will be a question and answer session. Today, I am pleased to present CEO Emil Billbäck, CFO Håkan Johansson, and CMO Michael Diefenbeck. Please begin your meeting.
Yeah. Thank you very much, operator. This is Emil Billbäck. I would like to welcome everyone to this information session, where the purpose is to quickly cover the FORTIFY study, the readout that we announced briefly last night, and the path forward. As the operator mentioned, I'm also joined here today by Dr. Michael Diefenbeck, who's our Chief Medical Officer and Executive Vice President, and Håkan Johansson, our CFO. I will march through the presentation fairly quickly and then open up for questions and try to provide as much clarification on this topic as we can. If you help me, operator, please and go to the next slide. The FORTIFY study was defined in 2015 and 2016. At that time, the only path to market for CERAMENT G, a bone graft substitute containing antibiotic, would be a PMA route.
The study was hence constructed to show superiority of CERAMENT G versus purely standard of care. The study design has been selected with that specific and only purpose. I'm not going to go through a lot of details on how the study was constructed because I think there's good information in the public domain on this, and you have it also in the presentation in front of you. What I will mention, though, is that the primary endpoint was a composite efficacy score which weighs together three independent effect parameters and arrive at a final number. Next slide, please. What I do want to spend a little bit of time is on just to describe the patient that has gone into this study. What we mentioned in the design is that this is Gustilo- Anderson II, IIIA, and IIIB.
I can imagine that's a little bit cryptic, but basically what we're talking about here is trauma that have been exposed to such force that the bone has penetrated the soft tissue and the skin. That's an open fracture. The open fractures are the most difficult fractures in trauma events that can happen, and they're categorized on the so-called Gustilo-Anderson scale or categorization. The patients that were included in FORTIFY, 1/3 of them was in category II, 1/3 in category IIIA, and 1/3 in category IIIB. These injuries are very severe, and they also present a very high variability in bone healing infection. The infection can occur quickly after the accident, or it can occur after a few months, even with strong intervention.
Mainly the reasons for ending up with an open fracture is a traffic accident, but there are also patients included from construction workers, for example. There are ballistic injuries, people that have been shot simply or had a gun wound. There are also soldiers, for example, U.S. Army soldiers that have been included. 80% of the patients in FORTIFY are male, and 20% are female. In the control group of the patients that received CERAMENT G, the average age was 38 years, and the youngest patient was 17. We go to the next slide. This is not a real bone. This is a plastic bone. It just displays how the surgical technique was conducted with a metal rod or a metal nail that was forced down through the intramedullary canal to stabilize and immobilize the fracture.
This is not a very common method used in Europe, but the standard of care in the study was defined based on what is used many times in the U.S. surgery. We go to the next slide, please. What are the circumstances specifically influencing FORTIFY? Well, I already spoke about the very high variability with trauma patients. That's why studies made on trauma patients has to have big populations, and they're usually very expensive studies and with certain difficulties to recruit, which we also have reported on in FORTIFY with delays that were announced at a quite early stage of the study. Also, what you see normally with trauma studies is that patients. It's not a chronic condition. It's something that happens, and once the patient has fully recovered, usually they walk away.
They walk away from the situation, they walk away from healthcare, and they walk away from studies that they might be part of. This is a common phenomenon. Unfortunately, the dropout rate in the FORTIFY study has been exceptionally high. In the end, only 143 patients qualified for the primary composite endpoint. The way that the study was constructed, even though patients were followed at certain time intervals, it's the 12 months result that matters. You have to be part of the study for the whole way with all the parameters included. If you withdraw on one of the parameters, you are excluded regardless of how well you have performed on other parameters. 143 study objects. If you compare it to other studies, you could say, "Well, that's quite a high number, actually." No, it's not.
If you have subjects with very high variability, you need a very large patient population to be able to draw significance and make conclusions. This is the problem we have seen with FORTIFY, that during the COVID pandemic, there has been an exceptionally high dropout. A dropout of the nature where patients have decided just not to show up at the hospital visit that they have been requested to come to. Next slide, please. What has then come out in the results? Well, unfortunately, the FORTIFY results are inconclusive. This is very disappointing for, of course, the management team, the board of directors. I can imagine the entire organization at BONESUPPORT, and I can imagine also at quite a few of the shareholders.
We did expect, based on study results we have seen from many other parts of the world with other patients, that this would come out in positive favor of CERAMENT. It didn't. The main reason for this is the very high dropout rate, which was 29%. FDA and the general community has a guideline that below 5% dropout rate is a well-controlled trial with a minimal bias. FDA has a limit at 15%. FDA would not like a study with a higher dropout rate than 15% to go into a PMA application. There are some extreme cases where slightly higher numbers, even up to 20%, has been accepted. Unfortunately, we are at 29%, which means that the entire study is labeled as inconclusive. None of the parameters can actually be used due to that fact. The composite endpoint is inconclusive. The secondary endpoints are inconclusive.
The only solid readout is on the safety parameter for CERAMENT G. That came out very well. The study was intended to show superiority in efficacy and safety. The safety one was met, but it could not be shown that CERAMENT was better. I would like to emphasize one thing very strongly, and that is this study does not show that CERAMENT G is equal to standard of care. That is not what the study is designed to do, and no such conclusion can be read out of this. If that is what we wanted to be tested, the study would have to be constructed in a different way with a completely different design. I think that's important to keep in mind. The study could not show that CERAMENT G is better than standard of care, but that's the only thing that could be read out at this point.
In our press release from last night, we were prudent to show what the primary endpoint success was both for CERAMENT and the control group. 64% with CERAMENT and 66% with the control group. Of course, it would've been nice to see CERAMENT have a few percentage point or quite a lot of percentage point higher than the standard of care. This is actually completely irrelevant given how the study was constructed. Only when these two numbers are different enough to provide statistical significance does it really make any difference. The standard deviation with these patients enrolled makes this number also inconclusive because it only takes a few patients in either direction to completely skew the results. Here we see now the difficulty and the challenge to make a study with few participants, few final objects completing the study, and a very high variability in individual results.
Next slide, please. Concluding up my presentation, what are the next steps? Of course, as I said, we're very disappointed to see this. We always knew that there's a risk of doing trauma studies, and especially when you have a high dropout rate and come closer and closer to where your statistical significance is impossible to show and then eventually pass that line. The study results, though, from FORTIFY does not at all influence the Breakthrough Device designation that has been given both in bone infections and in trauma. The results from FORTIFY does also not influence at all the de novo application for CERAMENT G in the indication bone infection. We have a timeline for submitting additional data to FDA on bone infections, where in all de novos, you have three major themes or three major categories.
You have safety data that has to be in place, you need efficacy data, and you need a control group. With our submission on bone infections, it was the control group that needed to be strengthened, something that we have worked upon since March of this year. We're pleased to announce that we will meet the deadline as communicated back then. The team has worked hard, and we have progressed. When it comes to the indication infection prevention, meaning using a bone graft substitute with antibiotic, even though there's no sign of infection, as you would have for trauma, for example. We will now need to reconnect with FDA, take a look at the results, and also discuss what potential pathways are there to the U.S. market for CERAMENT G for the indication trauma.
We will be open-minded, and we will consider different options, such as piggybacking potentially on a future approval on CERAMENT G for bone infections. Will there be a separate de novo application for the indication trauma? We don't know. We cannot answer that question today, and we prefer to have a discussion with FDA and then give a clearer notice to the market instead of being speculative. We are, as I said, disappointed, but we're not beaten. We are determined to bring CERAMENT G to the U.S. market, and we will continue on that path forward and with that vision. With this, I conclude my presentation, and let's open up for questions that you can address either in verbal directly to me and the team here or in written, then we will also bring them up.
Thank you. If you would like to ask an audio question, please press zero one on your telephone keypad. If you wish to withdraw your question, you may do so by pressing zero two to cancel. There will now be a brief pause while questions are being registered. The first question comes from the line of Erik Cassel from ABG Sundal Collier. Please go ahead. Your line is open.
Hi. Morning, everyone. I have a couple of questions, mostly on clarification. First off, can you give any more color on the individual component endpoints? Was there a statistically non-significant trend towards benefit in CERAMENT G?
Sorry, I didn't hear the first part of the question. Was there a statistical significance on what?
Statistically non-significant trend towards benefit with CERAMENT G. Basically, the point estimate, was it significantly better with CERAMENT G than the control group?
When you look at the detailed results and looking at age groups, looking at different Gustilo-Anderson categorization, and dividing it down for some of the analysis we have done, you can see clear trends in different directions. The unfortunate part with this is that once you go down in the subgroup and you already from the top down have too few patients, you end up in speculations because there's no statistical significance. We have our policy, I think this is very well supported in the dialogue with FDA, that we don't give data that we cannot prove and validate. I prefer not to talk about some of the trends that we have seen in the material as that could create false impressions.
Okay. Understandable. Given the anticipated dropout rate was 15%-20% in the study, which would have yielded somewhere between 160-170 patients, 143 is not that far off. I mean, 17 patients. I mean, the increase in sample variance should not weigh on the significance a lot here, in my view. Could you perhaps clarify this a bit? If it was a close call on any of the secondary parameters?
There was a close call on some of the secondary parameters. What made us anyhow come to this conclusion is that with a dropout rate of 29%, the FDA would, with very strong certainty, conclude the study as inconclusive. The primary endpoint as defined back in 2007 together with FDA is a composite, which of course can have one of the parameters that makes up the composite be in strong favor of either the treatment or the control group. It could then be derailed by other parameters of the composite endpoint. That is a big risk you take when you construct a study with composite endpoints.
Okay, I understand. What magnitude of benefit on the secondary component endpoints on deep infections and reoperation respectively, was the study powered to detect?
We have to look that up, and I'm not sure we released that data as the agreement with FDA was that it was the primary endpoint that should be successful, and that how the study was defined. There are some of the secondary endpoint that are closer to statistical significance. Where I probably disagree a little bit with you is that even though 143 looks to be a decent number, there is a very high variability with those patients that are included. The standard deviation is borderline extreme. There's also this component of who are the people that drop out. Unfortunately, this becomes a bit speculative because once the person drops out, they are excluded, of course, from the final primary endpoints. There's a big difference here between CERAMENT G that has an 18% loss to follow-up, and the control group of 9% loss to follow-up.
Loss to follow-up is, of course, where the patient decides, "I'm not going to go to the hospital." If it's fear of COVID or if it's because the patient has healed or other reasons. We can only speculate in that, because when we cannot reach the patient anymore, we cannot ask them for why they have made such a difficult decision. Unfortunately, when you have this dropout rate, there becomes a lot of uncertainties, and we don't share more data because we're just afraid it's going to lead to speculations rather than clarity.
Okay. Thank you, Emil. Just the last one from me. The U.S. case doesn't seem to be over. The optimum way forward is obviously de novo for trauma and infection. How likely do you think it is that you have to collect any additional data for this to come through? Do you think you can use the EU data for this de novo application? If you have to collect more data, how much cash burn do you expect that would result in?
I think those are very good questions, Erik, of course. I think those are questions that we should jointly try to answer over time. I can clearly tell that you're thinking you're in the right direction. De novo is definitely the pathway which you would find retrospectively now is more suitable for a combination device like this. It wasn't available, as I said, when this study was constructed. I think what we want to do now is, of course, to get together with FDA first and make sure that we have a shared view on what the next steps could be. I'll promise we will be very clear in communicating to the market how we see that going forward.
We should bear in mind, though, that the burden of data, if it's clinical data, retrospective data mining, is significantly less with a de novo than it is with a PMA. We're talking about different dimensions of cost, of course. I would like to come back to that once I can give statements which we can stand behind fully.
Okay. Thank you very much, Emil. I'll jump back.
Thank you, Erik.
Thank you. The next question comes from the line of Stian Westerby from [Analysgalggen] . Please go ahead. Your line is open.
Yes, good morning. Yes, trying to straighten out the patient numbers. 143 of 201 patients were evaluated, meaning that 58 patients were not evaluated. This number, 58 patients, is not corresponding to the loss to follow-up, which I calculate to be some 27 patients. Am I getting the numbers right here? If that is so, what is making up the difference between 58 and 27?
Thank you, Stian. The total number of dropout patients is 58. That means the dropout is defined as people or sorry, patients that have been enrolled and included that then doesn't fulfill the follow-ups and the parameters that are required for the 12 months. Loss to follow-up is one of the reasons for dropout. You're correct on your numbers, and you can say that in addition to loss to follow-up, there are other reasons. There are patients that have died during this period. There are also patients that have had potential comorbidities that have made them excluded from the study or other things that have happened in their life that they haven't continued.
Why we decided to publish not how many people that have died, because we found that to be irrelevant, but how many that are lost to follow-up, is patients that have seemingly gone through successfully the parameters tracked in the study, but for no reason that we have been able to determine, decided to walk away. That measure stood out. There's a lot of different reasons for dropout, but that one stood out in this study with a very high number for the CERAMENT group, and also quite a high number for the control group. The loss to follow-up is a subcategory under the total dropout.
Okay. Second question, if I may. The historical infection rates where we've sometimes, I think we've heard numbers like 50%, at least in the Group III Gustilo-Anderson, they cannot really have been the case in this study. It is enough to fail on one of the components in the composite endpoint in order to not reach the endpoint, I guess. Infection rates appears to have been pretty low in this study. Is there any reason to speculate in a study not reflecting the general infection rate in U.S. centers?
I think that's absolutely a valid question. What we have seen in Gustilo-Anderson Category III is infection rates up to 52%, but it ranges basically from 30%-52%. With lower categories, of course, the infection rate drops even further. You are correct. Every clinical study is a microcosmos with certain procedures and processes that have to be followed. If you walk into a hospital and you sit down as an observer and you look at how patients are being treated with trauma, you will come out with a certain infection rate as an observational study. If you have an intervention and you say, "This is how you need to treat the patients, and we will be standing here monitoring every step," many times there is a slightly better outcome. Infection rates decrease because there's more meticulous protocols to be followed.
We have seen in this study slightly lower infection rates in the control group than have been in other reference studies. That could have influenced also, but again, it could be speculative to make statements on that as the difference between the groups were inconclusive.
Okay. Finally, one last question. We are experiencing extraordinary times in healthcare, and you've not really picked up the COVID item in your study. In what way could the pandemic have provoked the loss to follow-up? It's difficult for me to see why there would be a difference between the group, but would you care to expand a little bit on how COVID could have influenced the study?
Yes. It's absolutely undeniable that COVID has influenced the study and the follow-up, and especially the loss to follow-up. Given also that you see that loss to follow-up is significantly higher with the treatment group, somehow, we know from other studies that when patients recover, they're less likely to be compliant. If you have an environmental factor with concerns about infection, fear about the COVID pandemic, I think it's fair to say that it has a contributive effect on people, unfortunately, decided to self-isolate and not continue the study that otherwise was the aim for following them up with this injury. We followed over time both the recruitment of the study as well as the follow-up. Not all data has been disclosed to us. The very high dropout rate was also a big surprise for us when we opened the envelopes yesterday.
We can see that both follow-up and recruitment rate in the study in general was strongly influenced by COVID. I think also if you look at the announcements that are being made around the world for med tech, biotech, and pharma companies, we are not alone with this observation if judging how many clinical studies that are being paused or postponed or modified during these very special times.
Okay. Thank you. I'll get back in line.
Thank you. Just a reminder that if you would like to ask an audio question, please press zero one. We have no further questions, so I will pass back to the speakers.
Thank you, everyone, for joining this call. Hopefully, we have been able to provide a bit more clarity and detail on the FORTIFY study that came out inconclusive. Again, what that means is that the study cannot be used for the purpose it was intended, no conclusions in either direction can be given. We thank you for your participation, we also encourage those of you that would like to learn more about the company to contact us directly, we will try to be as open and frank as we can also about our way forward. Once we have had further discussions with FDA, we have a clear direction, we will make sure we communicate it in a transparent way as we always do. Thank you very much.
Thank you for attending. You may now disconnect your line.