December 10, 2010 Transcript: Product Development Program for Interventions in Patients with Severe Bleeding Due to Trauma or Other Causes
U.S. FOOD AND DRUG ADMINISTRATION
CENTER FOR BIOLOGICS EVALUATION AND RESEARCH
PUBLIC WORKSHOP: PRODUCT DEVELOPMENT PROGRAM FOR
INTERVENTIONS IN SEVERE BLEEDING DUE TO TRAUMA OR OTHER
Bldg. 10, National Institutes of Health
8800 Rockville Pike, Bethesda, MD, 20894
Friday, December 10, 2010
LIST OF PARTICIPANTS:
JOHN HOLCOMB, M.D.
Chair, the Department of Surgery
Director of the Center for Translational Injury
Research, University of Texas Medical School
ERNEST GENE MOORE, M.D.
Chief, Surgery and Trauma Services
Denver Health Medical Center
DAVID HOYT, M.D.
American College of Surgeons
DALLAS HACK, M.D.
Director, U.S. Army Combat Casualty Care Research
Program, Ft. Detrick
JEFFREY LAWSON, M.D., PH.D.
Duke University Medical School
JERROLD LEVY, M.D.
Emory University School of Medicine
KENNETH MATTOX, M.D.
Baylor College of Medicine
BARBARA TILLEY, PH.D. University of Texas Health Science Center
KEN MANN, PH.D.
University of Vermont
NIGEL KEY, M.D.
Director of Hemophilia Treatment Center
Distinguished Professor of Medicine at the
University Of North Carolina at Chapel Hill
CHARLES WADE, PH.D.
Professor at the Department of Surgery
University of Texas Health Science Center
GRANT BOCHICCHIO, M.D.
RICHARD WEISKOPF, M.D.
MICHELLE BIROS, M.D.
|CLINICAL TRIAL OPTIONS FOR SEVERE BLEEDING INDICATIONS|
|PIVOTAL CLINICAL TRIALS FOR SEVERE BLEEDING IN TRAUMA PATIENTS||5|
|LIMITATIONS OF TRAUMA TRIALS||22|
|PANEL DISCUSSION: CLINICAL TRIAL OPTIONS FOR SEVERE BLEEDING INDICATIONS||30|
|DISCUSSION AND SUMMARY OF OUTCOMES OF WORKSHOP DAY 1 SESSIONS|
|CLINICAL ENDPOINTS-LOCAL AGENTS||104|
|CLINICAL ENDPOINTS-SYSTEMIC AGENTS||115|
|ETHICAL CONSIDERATIONS IN TRAUMA||158|
|CONCLUSION AND ADJOURNMENT||173|
* * * * *
CLINICAL TRIAL OPTIONS FOR SEVERE BLEEDING INDICATIONS
SPEAKER: If you can make your way to your seats we'd like to get started. So welcome back for the second day of our workshop and we're going to start off this morning with session 5 with the Clinical Trial Options for Severe Bleeding Indications. And I'd like for you to know that there's actually one small change to this schedule this morning. At 9:15 we're not going to have a summary of the discussion for trial options, that's going to be later on after the coffee break.
So for now our first speaker this morning is Dr. John Holcomb, chair of the Department of Surgery and director of the Center for Translational Injury Research at the University of Texas Medical School in Houston and he's going to be speaking on Pivotal Clinical Trials for Severe Bleeding in Trauma Patients.
PIVOTAL CLINICAL TRIALS FOR SEVERE BLEEDING IN TRAUMA PATIENTS
DR. HOLCOMB: Okay, here we go. All right, we'll get started here very quickly. You know yesterday we talked a little about this in the -- several times we glanced into this subject. You know, I have -- clearly have a bias and you'll understand my bias as we go through this talk. There are clear differences in animal physiology. Gene Moore brought the point that between animals there's huge differences in physiology in especially the coagulation physiology and then all of those animals are very different than the human system really making it difficult to translate results both positive and negative directly into the human. That obviously has implications for the CFR 50.24 discussion that we had yesterday.
There are unique aspects to trauma pathophysiology not seen in elective surgery if you -- both discussed in the literature and from clinicians, anesthesiologists and surgeons who do these cases. And the challenge as we started talking about this in setting up this program was finding elective surgery procedure that routinely, routinely loses a lot of blood. There are certainly elective surgery procedures that lose lots of blood but the question is that done routinely.
And you know every -- the trauma studies we'll just say upfront are difficult. They happen. You have to enroll patients 24x7. That causes an increase in expense. For many of the trauma studies you have to use the 50.24 mechanism. Everybody wants to use elective surgery as a surrogate, if you will.
I find that from a data driven point of view a little challenging. And I'm going to run over some reasons why that may be challenging and then suggest from my review of the literature where with maybe some possibilities in elective surgery arena. So this paper is annals on surgery in 2003, hepatic resection from metastatic disease. So resections of the liver, 15 -- over 15 years 1,300 cases, only 3 to 4 a year were massively transfused out of 1,300 patients, 4 percent received more than 10 units of blood and fully almost 24 percent only had 1 to 2 units of blood. So while going in and doing hepatic resections at a major cancer center most patients did not lose a lot of blood and very few were massively transfused.
Now you go to Whipple procedure from this year from the same group and mortality is zero so mortality studies usually is extremely low. You can't use mortality for these things and I'm not sure if people would want to, but only 18 percent of the Whipple procedure were transfused at the 1- to 2-unit range. So Whipple, removing the pancreas is not a good elective surgery operation to transition to trauma.
Again in annals of surgery now looking at blood transfusion, blood loss, and 30-day mortality in older patients kind of a grab-bag, if you will, from the VA in a large study of 2,400 patients, the mortality was very low and the units of red cells transfused also was very low at 0 to 2 units fully 200,000 patients had no transfusions and only 22 had a mean of 2.8.
Looking at liver transplantation and we all who work in the ICU understand liver transplantation is a big deal. This study is a prospective randomized study published in Lancet and obviously some of these studies weren't looking at blood transfusion, but they had blood transfusion in them so that's the reason I picked them out. In this group placebo patients bled a total of 5 liters and were transfused 2.4 liters of red cells. So this looks pretty good in this single center study or multi center, mortality very low both at 24-hour and 30-day. Within the other range of this is a single center study published more recently in 80 percent of these patients, 350 transplants over 5 years, 80 percent received no transfusion. These are liver transplants, so large variability in this area.
You can't find papers that -- where they transfuse routinely a lot of blood. But the flip side is you also find many really done. That's certainly our experience in our hospital. We work with the transplant surgeons. The vast majority of liver transplants now get no blood transfusions or 1 to 2 units which is what this paper reflects.
Spine surgery is interesting so single level spine surgery doesn't lose a lot of blood. It's not that big a deal for those again tending in the ICU. Somebody gets multi levels, five or six levels of their spine repaired. A fair amount of blood loss approaching 5 to 800 ccs of blood fairly routinely. This might be an interesting area. Now many of these patients have co-morbidities, but a large group of these patients are young and healthy. And this spine surgery might be an opportunity in elective surgery that routinely loses a lot of blood to investigate. Unfortunately, the numbers here not many spine surgeons routinely do these really large spine operations so numbers may be important.
And then in general orthopedic trauma from last year 187 patients over 3 years, a single center study, so about 60 patients a year with large pelvic acetabular fractures after trauma not emergently. So these are done now several days later so you could get consent either from individual or LAR. And these patients routinely lost 6 to 800 ccs of blood and/or transfused accordingly with transfusion correlating pretty closely in that regression with the amount of blood loss. So the those two opportunities, spine surgery and pelvic, are kind of interesting that literature would support that there is routinely some amount of blood loss and the frequency is somewhat acceptable in a multi center fashion you could do some studies.
The cardiac, as you all know who do cardiac anesthesia, is interesting and there are well -- there're certain individual cases who lose a lot of blood. That pathophysiology extremely different when the patients go on pump, but 73 percent of these patients in annals of thoracic surgery receive 0 to 2 units of blood. And then the -- another study that's saying exactly the same thing, 40 percent with no transfusion and most of them with 2.5. And obviously in that group of patients the cardiac patients multiple co-morbidities and age comes into play.
Now there are -- this is a data from our institution. It's unpublished data from Dr. Zhou who's an anesthesiologist looking at thoracoabdominal aneurysms. The -- our group does -- our hospital, this is Dr. Safi's group, does about 150 of these a year. For those of you all who don't remember, this is basically replacing the aorta either from the root or just distal to left subclavian artery on down the abdominals type 1 through 5. These are big operations.
If you use the deep hypothermia circulatory arrest, they lose about 7 units of red cells. If you just do it without that, they lose about 20 of red cells. So routinely losing large amounts of blood in these open procedures -- Lena Napolitano's group has actually published recently that blood loss goes way down with non open or closed procedures with stenting that's starting to be done for some of these. But this is a group of patients that loses lots of blood. Now a 150 a year is a lot. There is only about four or five centers across the country that do this volume of thoracoabdominal aneurysms. And the amount of blood loss does correlate with mortality and discharge status at the end of the hospitalization.
So there's a wide variety of elective surgery. As we -- as a group of patients, anesthesiologists and surgeons have got more concern with blood loss, blood loss is going way down elective surgery. It does still happen but in general about 10 to 20 percent of these -- or up to 80 percent of these patients don't lose a lot of blood routinely. There's not a lot of massive bleeding going on. There's low mortality and except for these three opportunities there's not routinely a lot of blood loss going on across the country. We've talked about this study a little bit already with massive transfusion.
I just bring this up that in 16 centers across the country fairly recently we found 466 massive transfusion patients. The characteristics are as from a pathophysiology is what we want to find are hypothermic, acidotic and coagulopathic. There is large variability across centers. Now this is both the problem and an opportunity. It's a problem from a study design point of view. If you randomize this though you will have some variability and I think what this means is that there's opportunities for improvement in these patients maybe moving the group up to more this area.
The -- I would bring -- this is -- I put this slide up just to highlight one of the discussions from yesterday that as more plasma and platelets are given to patients what you see is a decrease in blood loss and death from bleeding. What we did not see in this study was an increase in multi organ failure or an increase in death from head injury. And the 24-hour Kaplan-Meier occurs just to reiterate and I think everybody has gotten this point that the action is all here and it's very early.
Bryan Cotton published an interesting paper. Bryan's in the audience and spoke yesterday at his time at Vanderbilt while implementing a massive transfusion protocol this is a before and after design, obviously not as good as a prospective randomize. But in a large group of patients 141 massive transfusions versus 125. This is over 4 years so 2 years 141, 2 years 125, actually lowered multi organ failures. This goes to what Janice Zimmerman was saying yesterday. If you give -- what this suggests is by giving plasma platelets earlier as you avoid crystalloid substitute with plasma is the primary substation fluid that you may in fact lower multi organ failure and improve survival. Lots of papers suggesting that this is true and then just some data from our center and Bryan touched a little bit on this yesterday. This is looking at our damage control laparotomy patients. They are the sickest of the sick.
If you look at our last 1,200 laparotomies over 6 years emission and demographics, physiology were similar and the damage control laparotomy the guys that we -- I believe, they are out and is open in pack and do all really focus on physiology and sacrifice normal anatomy for a couple of days. The crystalloid used in those damage control laparotomy patients is markedly less as is the red cell used plasma and platelets. So by using plasma and platelets earlier actually decrease the amount of blood products given to patients survival is higher and massive transfusion defined as 10 units in 24 hours -- whether it's right or not, but it's just a definition -- actually went down as well.
So there is a lot of bleeding going on around the country and a lot of opportunity to routinely study blood loss in these trauma patients. Snyder has published an extremely important paper. There is maybe some methodology issues and the numbers of patients in this paper is pretty small, but brings up the point of survival bias. The survival bias question basically says are patients living because they are getting plasma and platelets earlier? Are they just living long enough to get plasma and platelets? And the answer is only going to be answered by prospective studies. All the retrospective studies, every one of them, has this bias. Some of them is measured, some of them is not. And it's really -- highlight of the fact as the Alabama group put in their paper that the -- they give red cells at 18 minutes and plasma at 93. So it's impossible to tease out the true answer from these data. But raising the question is extremely important.
Now what are people trying to do. We did a multi center retrospective study. We've just completed the prospective observational study laying the groundwork for a prospective randomized study looking at massive transfusion in trauma centers around the United States. I'll go through a little bit of PROMMTT data of the study just finished about 6 weeks ago. The data by no means one hundred percent clean, but just to -- it's an opportunity to discuss -- the opportunity to do trauma studies in the United States. There was projections based upon John Hess' paper several years ago what we might see and we projected that we would see -- screen 12,000 patients, eligible patients. We thought 1,200 was eligible. It's 1,250. These are patients who received a unit of blood in the ED and the red cells in the ED were eligible. We projected we would see 300 massive transfusions and we saw 296. So not bad from basing our projections on retrospective data and collecting this over little over a year at 10 trauma centers. And what we saw was that there there's high volume centers and low volume centers across here, but averaged about up 4 to 5 massive transfusions per week which gives us some numbers to really do our projections for our prospective randomized study. And as you can see you can incrementally improve our CER enrolment numbers took a little dip here in the October-November timeframe around Thanksgiving and then picked back up again.
The epidemiology or the demographics as you will, the non-massive transfusions were not as sick as the massive transfusion patients replicating what we saw in the retrospective data where the massive transfusion patients presented in the ED hypothermic, coagulopathic and acidotic with a blood pressure, you know, the mean was 99 and heart rate was 111. Now looking at 24-hour mortality and obviously, this study we have these mortality data and minute-to-minute increments. This is just a summary of 24-hour mortality. It's interesting if you look at all eligible patients the 11 percent mortality which is pretty close to that was presented yesterday from the 7A trial.
If you take out the patients who don't receive as much blood, they have a 7 percent mortality and if they get more than 10 units in 24 hours up to 27. So there is a absolute association with transfusion mortality; we think it's pretty important and we're going to work through these observational data and sort out with kind of a 3D calculus, if you will, of injury severity units transfused and mortality and put that together and I think we'll have a data -- from the data we'll have a data driven description of what massive transfusion really is and it's when the change in the slope or the curve occurs rather than some apriority (phonetic) units over time that's fairly arbitrary.
Thirty-day mortality increases as you will expect up to 15 to 39 and just Dr. Wang is one of the statisticians is working with us, put this table together last week. And if you look at this 24-hour and 30-day mortality there is an association with units transfused. It's fairly flat up to about 14 units and bumps up above 15. So maybe the definition of massive transfusion is more in the 15-unit range rather than 6 or 10. The interesting way -- have the opportunity to pull all these together. Now when are plasma and platelets given?
I think this is very interesting. Plasma and platelets at 10 level I trauma centers around the country, platelets are given after 8 units of red cells and plasma started after 3. So these are really kind of laying the groundwork now for what is the standard in the United States. It was interesting that in massive transfusion patients fully 26 percent of patients at 10 well-known level I trauma centers around the country did not receive platelets. Platelets are really an unknown story out there. We know less about platelets than just about anything else.
PROPPR is the prospective randomized study from the DOD through NHLBI and through the ROC consortium where we have gone through PRC and DSMB and are approaching FDA and IRBs to do a prospective randomized study with 24-hour mortalities, our primary endpoint. The -- we've gone back to the PROMMTT data. Now the PROMMTT data, the observational data, and tried to look at enrollment criteria. Who are we going to enroll and how are we going to do this and there's the scores that were discussed yesterday, the ABC score, McLaughlin score looking at sensitivity and specificity with the PROMMTT data that was just collected. We also measured physician gestalt. And one of the participants yesterday talked about you know it when you see it. Well, we tried to measure that in PROMMTT. We had our data collecting people asking the clinicians in the ED if they thought the guy was going to get a massive transfusion. We weren't very good. These are the attending physicians in the emergency department. So that physician gestalt, we know it when see it, not so good.
We did some machine learning score, this is, you know, the black box where you put data in that runs through millions of algorithms and out spits an answer. Looks like it may be better than just about everything. So that question of how you enroll patients who were expected to receive a massive transfusion, I think, is one of the critical research questions.
And I just put Mo Blajchman's study up here titled Blood Still Kills. We must not forget this and nobody does and oftentimes don't mention in these discussions, but you must balance risk and benefit of these products. So to summarize I think that the study -- trauma studies are extremely feasible. Yes, they are expensive and difficult. They actually have, in my review of what's out there in the literature, the largest volume of bleeding patients who bleed routinely, they bleed routinely. So if you go to our blood bank and look at their per physician use of blood in our blood bank 1,000-bed hospital our trauma group are the docs who use the blood routinely. The liver group does use a lot of blood on a small group of patients, but they don't do it routinely.
And just kind of following up what was discussed yesterday, you know, we got to keep after this. We need -- we started -- if you just go back through the studies going back last 20 years of MAST trousers, hypotensive resuscitation, HBOCs, et cetera, you know, these studies are added up, we've learned from each one. We need to continue to learn to really satisfy the societal imperative discussed yesterday the real scourge of death and injury that trauma inflicts upon our society.
So those human studies are extremely important I think that we can very rapidly bring together centers to do trauma studies. We need to use predictive models, work collaboratively, remain nimble, 24-hour end-point and 50.24 for phase II and III studies extremely important concepts. Thank you very much.
LIMITATIONS OF TRAUMA TRIALS
SPEAKER: Our next speaker will be Dr. Gene Moore, chief, Surgery and Trauma Services at the Denver Health Medical Center speaking on Limitations of Trauma Trials.
DR. MOORE: Good morning. Thank you for the opportunity to participate in this timely forum. I guess the first thing I have to do is emphasize potential conflicts as indicated in your packet. I have research funded by Haemonetics both at the basic and clinical level. Well, originally this forum was intended to be a debate between John and me, a path of wisdom of providing evidence at trauma in trauma patients for the use of introduction of new products, but it turned out as John and I compared the notes we actually agreed more than we disagreed. So instead of me discussing the limitations, what I want to do is turn it (inaudible) and talk about how we may optimize clinical trials to make this suitable to test blood products.
I want to go through some fairly simple concepts most of these were covered yesterday. But I want to highlight what I think are some of the goals we should seek in designing clinical trials. First, to investigate a scientific based hypothesis, time therapy to match the window, proof-of-concept in high performance centers and then adequate funding because this is logistically challenging.
Well, first just to cap John's issue should blood product developed in the use of trauma patients be tested in trauma patients. And I agree strongly with him that it's difficult to replicate the complex environment of the injured patient. Even with thoracabdominal aneurysms as you all know these patients are fully systemically heparinized modifying the interpretation of changes in coagulation enormously.
In orthopedic procedures the ones that do have significant blood loss, they tend to be slow and over long time periods and they don't have the confounders listed here that compound the interpretation of the coagulation system. So I think we're left with the inevitable that to license these products for clinical use I think they have to be ultimately verified in the population intended for their use. Well, how can we design trials that will be convincing that are safe and effective. Well, I think that one of the first challenges is to pursue a scientific hypothesis. These data, as eluded to several times, provided evidence that red cell, the RBC of one-to-one had a dramatic improvement in survival and in fact there is widespread adoption of this policy despite the fact there is little scientific evidence in this retrospective study to substantiate that hypothesis.
When we looked into civilian center, a short time period we began to see a little different patterns raising the question of really is one-to-one the answer for post injury coagulopathy. And in fact there have been a myriad of studies you're all familiar with that have suggested that increasing platelets, increasing FFP, increasing cryoprecipitate to RBC has decreased mortality. But the shortfall from all these studies, including our own, is number one, there is no objective verification that the products given have altered the coagulation status of the host. And perhaps more importantly as you've seen in the presentations there is a dearth of information of why these people die. We haven't even separated hemorrhagic shock from head injury. And even more vexing is the fact that some people will die no matter what you do. As a blood banker I worked with BAC (phonetic). When I first started in our institution, would say to me when I'd call for products on the phone, are they dying because they are bleeding or bleeding because they are dying.
I've not seen this several times and I'm not here to suggest that TEG is ultimate answer to this, but I'm suggesting TEG is a move in the right direction to provide us some evidence of whether what we're giving provides the patient a benefit. Rapid thromboelastography as I mentioned with added tissue factor to replicate the extrinsic system and uncitrated blood will provide you a profile well within 10 minutes and this is available in emergency department with acutely injured patients. TEG in fact in not new as many of you know. It was pivotal in Dr. Starzl introducing liver transplant into the clinical arena. They couldn't understand the coagulopathy associated liver transplant until they took this to the lab studied with TEG and they introduced changes into their clinical management.
Thromboelastography is not only quick, but it provides a visual impact of the changes in the coagulation system as illustrated here. Here's a patient with a gunshot wound to the lung and liver requiring emergent operation. And for those who recognize these viscoelastic profiles you see this patient had multiple derangements. Not only can you instantly assess the various defects, but more importantly perhaps is as you introduce products to these patients you can see the profile change and gauge whether you need to use one product versus the other.
One of the most important features of using TEG routinely we believe is the unveiling of the existence of post injury fibrinolysis. And although this has been highlighted in cardiac surgery and liver transplant the issue of whether a fibrinolysis exist in trauma has been debated for decades. But TEG provides unequivocal evidence that it does exist in a certain subset of patients. This is a patient about a month ago who slammed into a telephone pole at 70 miles an hour and the diagram on your far left is what the clot formed in this patient. First of all, there was very weak clot formation, but perhaps more importantly 96 percent of this clot was dissolved within a minute. It wouldn't matter what products you gave to this patient, till you arrested the fibrinolysis, this patient is recalcitrant to any products given. In fact, when we loaded this patient with three sequential doses of Amicar finally arrested the fibrinolysis then she began to accept and improve with products and eventually we rendered her profile normal as you see in far right you see survived this potentially devastating injury.
But we need to fare this subgroup of patients out and if we look at coagulation issues as well. The next is the time the therapy to match the therapeutic window and I won't belabor this. We've talked about this multiple times. Studies have talked about looking at products over 24 hours, but the fact is we give most of this within the first 6 hours. So our window of examination has to be tightened and in the early phase. Additionally, mortality occurs early after those due to death. This is the STO study that Peter Rhee mentioned yesterday. You see here that most of the deaths occurred within 2 hours and in fact two-thirds occurred within 6 hours. So we really need to concentrate in that first 2 hours of therapy. Perhaps within the first hour or even prehospital to impact on the outcome based on controlling hemorrhage.
Similarly in individual trials this is our prehospital blood substitute trial showing when deaths occur. And once again you see here that virtually all the deaths have occurred within the first 2 hours. So again, emphasizing the point that we need to time therapy based on what the therapeutic window is clinically.
Another challenge for us in clinical trials again alluded to multiple times is that we need to test proof of concept hypotheses in high performance centers. These are American College of Surgeons the actual Trauma Outcome studies. These are major level I trauma centers in United States in high volume and what this dramatically illustrates is the marked variation in outcome. These are risk adjusted.
You have centers that -- a cadre of centers that have 0.5 risk of mortality and those on the other side who have four times the mortality risk adjusted. How can you possibly see a signal in therapy when you have such vast variation in outcome? Similarly here's our outcome analysis in the blood substitute trial based on individual institution overall mortality. We have two centers performing a mortality of about 5 percent and two on the other side with 20 percent. Again, how can you mix this in and expect a signal to look at intervention?
And finally, which I think is a fairly well-known problem is adequate funding for research in this arena. Unfortunately, the majority of massively injured patients occur at either off hours at night or on weekends. This is analysis we did a while ago to look at OR efficiency and tried to track when we needed a personnel in the operating room to provide resources for operations. But remarkably and not too surprisingly anyone who works in trauma, the trauma cases typically occur between 8:00 p.m. and 4:00 a.m., a time in which there's not a lot of personnel around to facilitate investigation. So I think if we're going to conduct this kind of work we're going to need assistance. We're going to need scientists in there in the middle of the night helping us work through the management of these patients.
At a time when we and the trauma surgeons are trying to save lives with holes on aortas and multiple drips in the ICU we don't really have time to ponder the implications of the cell based hypothesis of coagulation. So we need some help, but we need it at the time the patient arrives to ultimately change the course of post injury coagulation. So again in sum, I would suggest that we do need to test these products in trauma arena, but we need to develop hypothesis-driven studies. We need to time therapy in that critical early time window. We need a proof of concept studies in high volume centers in which we can look at this at a scientific level and finally we need adequate funding to assist with the logistic nightmares of our research. Thank you very much.
PANEL DISCUSSION: CLINICAL TRIAL OPTIONS FOR SEVERE BLEEDING INDICATIONS
SPEAKER: We're now going to have the panel discussion for this session and it's going to be moderated by Dr. David Hoyt, executive director American College of Surgeons.
DR. HOYT: Good morning and thanks for the opportunity to participate in this conference. Sorry, I couldn't join yesterday, but we have an important assignment this morning with this panel. And it's to really raise some of the questions that have been highlighted by the two previous talks this morning and frame some of the discussion about the uniqueness of clinical trials as we examine new products and how those things might -- how the rules, if you will, for clinical trials may need to be modified.
Sort of tee this up we're going to ask three questions, but I want to just introduce our panel because we have really arguably one of the best panels you could put together to discuss this based on experience. We have Dallas Hack who is currently commander of the U.S. Army Combat Casualty Care Research area and is intimately interested in this and also involved in funding these trials. We have John Holcomb who has, I think, everybody knows his contributions and is currently the facilitator for the proposed PROPPR trial to really examine this question in a prospective fashion.
We have Jeff Lawson who everyone's heard from, but has really contributed tremendously to our understanding of this whole area of coagulopathy. Jerrold Levy from Emory University and involved in clinical trials in this area. Ken Mattox, Ken has really, I think, been responsible in the trauma world for asking really in a precipitous fashion some very critical questions over the last 25 years and one which currently still remains unresolved is this whole question of hypotensive resuscitation, should we resuscitate at all and it was really Ken's contributions that opened that question up once again and then finally, Barbara Tilley who is providing the really the statistical core and organizational core for the proposed PROPPR trial.
So we have a tremendous panel and I want to start off by just then framing these questions and the first one is in three parts but it really -- the question is, is trauma a unique physiologic state? Is it so unique that it's the only thing that we can use?
So let's start off with the first part of this question, as a sub question and ask the panel then to react. And I'd like everybody to react. If you don't want to react then just pass it down the line, but we want to get a kind of a rhythm going here so we can get into the controversies. But with extreme logistic and resource demands of a trauma trial is this realistic?
And Barbara, why don't we start with you?
MS. TILLEY: (Off mic).
DR. HOYT: Got to go the microphone a little bit though.
MS. TILLEY: Sorry. We've heard a lot of discussion over these past several days.
DR. HOYT: So what do you think?
MS. TILLEY: And I think that it is realistic, but I think there have to be modifications, for example, when you're looking at approved products and comparing approved products I'm not sure that some of the same FTA criteria need to be put in place. I think things need to be done quickly. And we also heard about the myriad of devices in surgery from Dr. Lawson the other day. We really have to think of new paradigms to be doing these studies and being able to do them more quickly.
DR. HOYT: Okay. But we just heard from John Holcomb that there is no parallel. I mean a complex aortic reconstruction, a 15-unit spine operation is not the same as somebody with a gunshot wound to the aorta that's exsanguinated in the first 45 minutes. So how can we bridge that gap? I mean, is there a realistic substitute that could get us away from dependency on clinical trials?
DR. MOORE: Well, I think you heard my bias. I believe that categorically that these products need to be tested in the trauma environment. It's not that we can't learn from other arenas but the definitive trial, randomized trial that has to be performed in trauma centers. And I think the other thing we have to remember which we've emphasized but perhaps being emphasized more is that we're dealing with very expensive and potentially toxic products. We have to use these prudently both for the benefit of the patient as well as society and health cost.
DR. HOYT: Ken?
DR. MATTOX: Yes.
DR. MATTOX: With some modification. I think we'll learn a lot of things --
DR. HOYT: Wait, wait, wait. You lost me there. I'm sorry. I can usually follow that but you lost me. Yes, what?
DR. MATTOX: Yes, use it in trauma. Trauma is where it needs to be tested and we'll learn in things in trauma that can be applied to the cardiovascular, to the prostrate and the other injuries. There simply is not enough material to study in these other areas as John pointed out. It doesn't mean they don't exist but the physiologic derangements are not those that we see in the trauma patient.
DR. HOYT: Okay. Dr. Levy.
DR. LEVY: Is a treating bleeding kind of guy. I have two perspectives. I think that the milieu in which we work in is a patient presents in shock academia but also I think we need to get a better handle on the variability of tissue injury because the tissue injury and all of the complexity and the fossil lipid that's generated may be important --
DR. HOYT: How would you do that?
DR. LEVY: I think we need to develop assays and we have a lot of smart people in this room --
DR. HOYT: So a biomarker that would predict the --
DR. LEVY: I'm not saying -- well, biomarkers exactly. There are certain things that we may need to look at measure tissue factor, do things of certain perspective. We got to get a better handle on the variability of the tissue injury which is I think one of the complicating factors of trauma of the variability from the last case we saw. There was hepatic and pulmonary injury. I mean you got you know just countless number of things.
DR. HOYT: Tell me how you -- let's assume that you had a biomarker or some bio assay or some membrane receptor marker that could predict the amount of global tissue injury. How would you use that then in terms of a clinical trial design?
DR. LEVY: I think what we need is some -- we need a better way to standardize the patient populations. The thing about surgical patients is, you know, I can put the reoperative mitral double-valves in one pile. I can stick the reop CABGs, the aortic route so the type A dissection but this is such variability. I think that that's my confusion looking at all the trauma data is the variability of the injury. I'd like to see some way to better quantitate that that might be useful in interpreting and again, I'm an outsider to this. But I think this is somebody who does in vitro and clinical studies and in vivo studies as well as in vitro. I think this is an important think that we need to get a better handle on.
DR. HOYT: One of the problems with that I mean it's a great concept but one of the problems historically has been that maybe some of these markers and I'll take a simple one lactic acidosis may not occur until reprofusion so --
DR. LEVY: Yeah, absolutely. I mean I look at lactate post op in all patients but I think I'm talking about something better than lactate. I'm talking about something that looks at -- and maybe even tissue factor is not the way to look it. Looking at some sort of phospho-lipid some aspect because when we have all this cellular injury you and I know and Jeff knows as a thrombin biologist you stick that in the milieu of haemostatic system and bingo, I mean you get micro particalization other things occurring that creates a sort of yin-yang DIC pro-coagulant consumptive issue. So we got to get a better handle on some of the underlying pathophysiology in the variability of the surgical injury in my opinion.
DR. HOYT: Jeff Lawson.
DR. LAWSON: So I'm going to -- as the non trauma surgeon on the panel I'm going to say that trauma is an extreme example of a continuum. I think it is what you get in that extreme example is a lot of injury a magnitude that is profound but you get a lot of chaos. So it's a -- it's very difficult to study. And I would argue if you could distill it into more definable studiable predictable events the output of your clinical trial well, may be less profound would be more doable and you get to an end point faster.
DR. HOYT: But what about the question of can you only study in a trauma situation something for trauma indications or is the complex vascular operation a surrogate for that?
DR. LAWSON: I would argue that you don't have to invoke a new paradigm of biology. I think it's just an extreme example of a continuum. So it's not like there's a new disease as much as I respect the insight of my colleagues. I think it's just a -- it's like a standard deviation off of a continuum. So I do think you could use surrogate, you know, sort of big whack operations but at a time when you can get informed consent. You can have a study team available. You can collect samples biologic specimens that could be analyzed in meaningful ways with thromboelastography and other things and I think that's the advantage recognizing that maybe it's a less --
DR. HOYT: What operation would you pick?
DR. LAWSON: I probably would pick thoraco-abdominal -- open -- accepting the fact that you use significant -- you give significant anticoagulation and that does not replace or recapitulate being cold in the field for 4 hours when you get scooped up. So I recognize the --
DR. HOYT: The problem with that though is you're talking maybe five centers in the country that would have the volume to be able to justify that. I mean, you know, there are not that many aortic reconstruction.
So Dr. Holcomb, can you use something other than trauma to study or approve a device or a drug in trauma?
DR. HOLCOMB: I thought you said approve that means phase III. I think the answer to that is probably we know. I think in the incremental steps along the way you could use a non trauma patient, yes. We heard yesterday that phase II with CFR 50.24 is possible. I think that's important to additively learn more about optimal studies on in end points before moving into definitive trial. Barbara mentioned CER research. I think much like the ARDSNet trial when we turn the ventilator down, it would been a new molecule, a new device just to turn the dial down. There's huge opportunities for improvement in currently utilized products. There are really so-called low hanging fruit. We just don't know how to use them right now.
And so the CER research regulatory issues I think should be different than new product issues that would really improve the outcome and care of military and civilian trauma patients. The biomarker thing is interesting, but now that's inventing a new test which in itself is a research project. If you remember the recruitment curve in the PROMMTT, it went from 0 to 300 in 54 weeks, at 10 centers. It was essentially a straight line. I'll just reiterate that the -- if we are seeing bleeding, from my review of the literature -- I've not heard anybody say otherwise yet -- the bleeding that occurs most routinely is in trauma patients.
And there is -- as my slide showed yesterday, 20 to 25 trauma surgeons in the United States have had more than 3,000 patients a year with probably 900 massive transfusions a year. So it's -- the population is there we just have to decide to do it or not.
DR. HOYT: Dallas, why don't you comment on that, and then we're going to add part II, and I'll ask you to start that.
DR. HACK: I think that we've jumped to the end and we need -- there are several steps -- before we get there. I tend to agree with Jeff. Trauma indication is one indication of many for these products. It's not the only indication and what we run the risk of is that we don't even get these products approved for other indications if we're insisting that we go to the maximum physiologic state before we do anything. So I think that we need to look across the continuum. There are other studies that have validity for some of these products. Trauma isn't the only indication that we seek.
SPEAKER: David, can I make one last comment?
DR. HOYT: Yeah, can you save out till -- it's your turn again.
SPEAKER: Yes, yes.
DR. HOYT: Okay. Let's just sort of get an audience participation now. And let's just ask for people to raise their hands that feel that for the kind of approval that we are talking about for something in the most extreme kind of bleeding, do people in general think a non-trauma trial could be used to support that? Raise your hands.
Okay. That's a fair number of people that raised their hands. So let's just go through maybe and see what people are talking about. Are you talking about liver surgery? Raise your hands.
Are you talking about spine surgery? Raise your hands.
Are you talking about aortic surgery? Okay. So maybe one sort of summary to this part IA is that at least there's a feeling that some forms of operations in a most severe form probably defined by number of transfusions would at least be supportive data for some of these hypotheses.
Now let's with that switch and go back to Dallas, and ask the panel to address, sort of, this age-old question and I think we've made incredible progress in the last 20 years between the civilian and military community to come together and agree that many of our problems are much more similar than different.
Having said that there's still -- there is concern that the battlefield injuries are different. And Dallas, rather than trying to get into whether that's true or not, let's talk instead about what in your mind and in the panel's mind is unique to a battlefield injury, particularly today? And is it just more tissue injury? Is it more bleeding? Is it logistics and delay in treatment? What are the factors that we really need to consider that are different?
DR. HACK: I think you basically mentioned them. I mean, the massive wounds that our troops are getting is seldom seen in civilian trauma.
DR. HOYT: So what Dr. Levy was saying? I mean that there is a quantifiable type of tissue injury?
DR. HACK: That is one aspect. And then there's all of the others; the issues of logistics, delay in evacuation to definitive treatment and so on. However, the chaos of doing studies in that environment makes them almost impossible to do any kind of valid, randomized type trial. I mean, there -- with all of the other -- and then other ethical issues involved it becomes extremely difficult, I would say, virtually impossible to do a valid type I study in that environment which is why we have been so supportive and working so closely with our major civilian trauma centers to get these kinds of studies done.
DR. HOYT: So but would you accept a big aortic operation is a surrogate to a battlefield injury?
DR. HACK: With the right controls and the right scientific testing and --
DR. HOYT: So possibly.
DR. HACK: -- possibly, yes.
DR. HOYT: Okay, all right.
John, difference between battlefield injuries and civilian, you've clearly worked in both environments and what in your mind really does it come down to?
DR. HOLCOMB: The difference is in the wounding mechanism. I think the physiologic response in every way we've measured -- we've actually looked very hard at this at our research institute when I was in the military. I'm convinced even more so now in care trauma patients in the civilian sector that their physiologic responses are the same. If you get an AK-47 or an IED that blows you up or hits you, your physiologic response, I think, is exactly same as if you get your leg torn off in a guardrail, riding a motorcycle or you get run over by a car.
DR. HOYT: So are there any differences then?
DR. HOLCOMB: The differences are the decreased amount of personnel support that you see on the battlefield, you know, the area that was discussed yesterday and all our extra people hanging around, which makes sense. The medics routinely are taking care of patients in the middle of a firefight that occasionally happens in the civilian environment much less than talked about at meetings. And although the similarities with transport time are much more similar than different as well. There are some extreme examples of transport time in military but most transports are actually fairly short as well. So like probably 85 to 90 percent very similar. There are some differences that relate to logistics and transport in a permissive versus a non-permissive environment.
DR. HOYT: Jeff?
DR. LAWSON: I would say there are three considerations. The first is the degree of -- if you'd excuse the phrase, but the degree of the whack, how hard somebody gets hit. And I think we could probably reproduce an injury that might not be as chaotic, but could easily cause as much cellular tissue destruction when you do a thoracoabdominal biscuit-can operation you split somebody open sideways, that's a huge injury.
So I don't know, you could -- but you could come up with almost a scale to quantify the magnitude of the injury and say, yeah, they had a whack that was big enough.
DR. HOYT: Right. How would you do that because, you know, that's not a new concept. I mean would you do that by doing an MRI and quantitating cellular tissue injury? Do you measure phospholipids as a biomarker? I mean how do you quantitate the whack? I mean, we've had physiologic scores; we've had anatomic scores. We've really been trying to do that for three decades. But we actually haven't been able to come up with something.
DR. LAWSON: I don't have a magic answer to that. I think it would be somewhat of a subjective scale. But that being said, I think, there are two other things that you can't recapitulate, that is unique to trauma. One is the effective anesthetic. So whatever the catecholamine brain response while who just got shot, you know, you clamped him.
If Jerry's got me asleep I never have that response. I don't have that fight-or-flight. I don't have that catecholamine rush. And the other effect of having anesthesia if even if I'm having a bad day, Jerry doesn't let me get in trouble as much as the field will let you get in trouble acidotic , hypoglemic. So there's the magnitude of the how sick you can get and still live is, I think, much more severe in trauma because the anesthesiologist never lets me get that sick.
So those are that what I'd say, but I do think you could come up with a scale of injury that we could all -- we wouldn't be perfect, but we could agree with it. DR. HOYT: Okay.
SPEAKER: I think Jeff's point is very important, you know, in a controlled environment it's totally different. The other thing is that the massive injury that occurs in the non-controlled anaesthetized patient is probably incredibly different. I used to believe that surgery was a potentially good correlate, but John Holcomb is going to talk me out of it. And I kind of starting to believe it now that all of the things that are mentioned.
But I think also the fact is that the presenting with the substrate of shock, acidemia has major effects not only on tissue profusion since I have been interested in certain subsets of distributor shock, but also because of the fact that it makes the hemostatic system quite different. So therefore trying to correct it is very complex.
You know the vasopressin and other response is a major activator of endothelium and why will bodies just get massively degranulated. So there's a lot of interesting things in the awake patient versus in the patient who is under major control with blunting of all these complex pathophysiologic mechanisms. So I'm not sure that it's really extrapolatable.
The other problem with thoracoabdominals, with the interruption selectively aflow to the kidneys and other selective organ system with a spinal cord and everything else is such a unique phenomenon. When we do these major thoracic versus thoracoabdominal procedures we usually use spinal drain to improve CSF flow. So there is a lot of things that really alter this, and I'm not always convinced that it's really now extrapolatable.
DR. HOYT: Ken, you've been part of the discussions over the last 20 years really of, sort of, the similarities, the differences between military and civilian injuries. You've been a cardiac surgeon; you've done a lot of these big operations in your entire career. What's your perspective, in 2010, in terms of what are the differences?
DR. MATTOX: Number one, we have to decrease the static by eliminating the head injury, the person with a spinal cord injury and define very narrowly what we are talking about. We're talking about the big, bad injury. In the civilian practice, what challenges me is the truncal injury and the pelvic fracture. I rarely lose somebody from an extremity injury in the civilian sector.
In the military setting, there's a lot of double amputees and triple amputees and the truncal operations are not the same percentage as we see in the civilian setting. But they still have lactic acidosis, their pHs are low. And in the thoracoabdominal aneurysm, which I have participated in, and the cardiac surgery I rarely see a pH lower than 7.2; even if 7.2, I often start with a pH of 7.0, 6.8. This week I had a patient, I started at 6.4, and his pre-hospital time was relatively short, his emergency room time was 15 minutes.
So I think something that's unique among the military are -- the same in the military and civilian is the lactic acidosis, the low pH. And in the big whack, there is a difference in whether it's extremity or truncal as I look at the military injuries that get back to Walter Reed, go through Landstuhl and the like. But then we have got to eliminate the head injury and the spinal cord injury.
DR. HOYT: Get ready, Barbara. I may have a challenging question for you.
Gene, you brought up a very important point in your talk. And that is that despite, you know, universal access and good pre-hospital care that there is a variability and mortality ascribable to the process of care in a trauma center presumably, in a trauma center not just across hospitals, that in and of itself could overshadow a treatment effect. So how do we deal with that if we are going to try and approve products in clinical trials that use that infrastructure?
DR. MOORE: Well, we have, as you know, a myriad of national outcome databases that track the performance of trauma centers. Frankly, I think if you can invest NIH money in our research project, it ought to be given to centers who can perform better than average. We know those centers. We can eliminate them. But I submit they have to be eliminated to study this.
But just to quickly address your issue about extrapolating from the civilian to the military, I would certainly defer to John and others who have been in both environments. But I would certainly accept extrapolation there much quicker than I would a patient undergoing a thoracoabdominal aneurysm resection under full heparinization. And all of us know at a investigational level, even trace amounts of heparin in murine and rodent models has profound influence on the inflammatory response. And the more we study coagulation, the more we understand the cross talk between inflammation and coagulation. We are just beginning to understand some of these parts. But certainly someone fully heparinized would be difficult to accept as a model for an injured patient.
DR. HOYT: Okay, so Barbara, everybody has talked about the big whack and role of physiologic, you know, lactic acidosis, et cetera in terms of influencing this. But you've got to create a statistical analysis between two groups and control for the big whack. How are you going to do that?
MS. TILLEY: Of course, that's where you go to the randomized trial. But I think that, you know, when I spoke about new paradigms, I think that if we have sufficient historical data in the military and sufficient historical data from some of our major trauma centers that we might be able to design some short-term phase II studies and we would probably have to have different rates for the big whack group versus the trauma patient that you'd see in a more civilian population.
But the other thing is that perhaps we should look to the cancer model where, you know, the goal in cancer is to have every cancer patient in a clinical trial. And I think that, you know, we perhaps need to think that way in trauma that maybe we should have more networks not just of the big trauma centers but networks of places including the military where trauma patients are seen, where we're gathering some data, and we've set up a group of people that can quickly participate in these studies.
DR. HOYT: Great. All right, well I think everybody would certainly agree with that, and that's one message back to the NIH is that the opportunity to expand clinical trial networks would be very useful. Let's finish off question one very quickly with part d which is we really touched on the rest of them. But c and d deal with two very specific issues; are historical controls adequate and can we use data from other countries in justifying trials.
You want to start on that?
MS. TILLEY: Yeah, could I -- yeah, I'll finish that. I think they are an important first step. There is a whole concept of futility where you try to decide if something's worth carrying forward to phase III. And I think that we can use historical data in that setting. But I think that for very important questions you're still going to need the type of evidence that you would get in a phase III trial.
DR. HOYT: Any other comments from the panel about European -- or out-of-country data? Is that useful? Do you believe it? Would you use it as part of a defense for phase II trial?
SPEAKER: You have to -- I think you have to look at how it was collected. If they are well done and adequately controlled studies and good data collection procedures I think that that needs to be supporting data in what -- in the work that then goes on for more definitive indications.
I use this as an opportunity for a plug here. In my comment yesterday morning, it goes back to who is funding these type of trials the NIH is, I mean, basically we have put the majority of the money into these kinds of -- this kind of work. There is completely inadequate funding for trauma trials in my opinion. It is going to take a major investment than yet. What pharmaceutical company, things are going to make enough money to do -- to invest in these. We need to really change the way we are doing things and I need -- we need bright ideas on how to make this stuff happen. There can't just be one funder of this kind of research.
SPEAKER: Not to say that the world is flat, but if you look at the cardiovascular track of approval, they are international networks and that's how they can enroll 40,000 patients in a pivotal coronary trial of some type or cardiovascular intervention. And I think we'd be foolish not to at least consider that. Knowing there are some mitigating circumstances about access and -- but the world there you can collect data internationally and it can be good data.
DR. HOYT: How about historical controls? Any value today?
SPEAKER: Let me say this. I agree with Barbara on historical controls. I think they can give you handsome directions on important questions. We pretty need a control group. On the U.S. data, you know, if there is a bias at the U.S. centers it would be where data must be collected. That clearly is untrue. There's some very outstanding centers around the world that have better results than we do.
If you look at the CRASH-2 trial one of the things that's interesting. I have, kind of, have gone to several transfusion meetings, international transfusion meetings. And many places in the world use still use whole blood and don't use component therapy. And many of the countries in the CRASH-2 trial, I'm absolutely sure resuscitate the patients with fresh whole blood. Wouldn't it be interesting if the survival difference was in the so-called Third World countries with fresh whole blood and the increased mortality was in the Western European countries with component therapy. That's what the -- their data that will support that concept in -- from the military 8000 unit fresh whole blood experience.
DR. HOYT: Okay, role of primate studies.
DR. MATTOX: I'd like to go back to the historic control question just for a moment. We must be very careful. The big elephant in the living room we've talked around for the last 2 days is the increasing reduction of crystalloid use. And the agreement that Ringer's Lactate and normal saline are good way to kill a population and the admittance that we are not dosing those solutions in the same way we dose medicines.
If we -- and we have seen the decreased trend in a lot of bad complications mainly because we've reduced crystalloid, and if we reduced crystalloid use even further, a lot of the coagulation problems, the --
DR. HOYT: So you are saying historical controls, at least from that standpoint, might not be valid?
DR. MATTOX: The historic controls are all in the old paradigm and would have a lot of static and wouldn't be applicable to today -- even today's experiments.
DR. HOYT: Okay, just a couple of questions from the audience to end up this part.
Dallas, difference between a 21-year-old and a 7-year-old, how do you control for that?
DR. HACK: That's really the question I raised yesterday morning. It's very difficult. We get -- a 7-year-old typically will not respond as well to trauma, so probably know what's going on with them sometimes quicker than we will with a young person. I think we need to develop the studies that will help us in that area more than we have. We were trying to -- I think we have one broad-brush across the population. I think we --
DR. HOYT: So age is part of this whack scale then?
DR. HACK: I think so.
DR. HOYT: Okay. And then what about use of primate studies in trauma?
DR. HACK: I think we've got -- I think that they can help us considerably, especially the larger primates. It's very challenging and expensive to do those kinds of studies. But at least in the early phases, you know, of a product, developing a product I think we can learn a lot.
DR. HOYT: Okay. Gene, question about the difference in trauma center effect is whether this is gunshot ones or car accidents, and I think you said this was using basically risk-adjusted data, but could you just explain a little bit again on what your graph showed?
DR. MOORE: Certainly, you, of course, know this much better than me. But the risk-adjusted profile I showed you was for combined trauma, the same scatter exists for both penetrating and blunt. There's enormous variation in the major trauma centers in this country as much as an eight-fold increase in mortality risk-adjusted. It's astounding.
DR. HOYT: So the whole advantage of these newer databases or new analyses is it tries to take the risk adjustment out of the equation and actually then leave you with an explanation of difference that may, in fact, be due to the actual trauma center.
Okay, let's -- we got 10 minutes left. Let's -- I think we've hit on most of the physiologic questions in terms of different surgical situations.
I think just to summarize that, that I think there's consensus that if you can find a group of patients that people think are equally physiologically affected that that would be a good group to study. But it's going to be a group that's hard to identify in large numbers simply because most surgery today does not have that accompanying problem the way it used to, but that could be a useful subgroup.
Let's go on to the informed consent process and talk a little bit about the potential benefit. So how do you actually define that? Do you have to use animal studies to do that? Are preclinical studies really the way to demonstrate that? How much efficacy is sufficient to hit that threshold so that when you go to a exception from informed consent study, what's the database that you need to say that you've achieved that threshold?
Barbara, let's start with you.
MS. TILLEY: I guess, I'm in the equipoise mode, which is that you have to really feel that you have two reasonable -- if you are comparing treatments, two reasonable treatments, and you don't know which one is better or you don't know whether your new treatment's better than placebo.
I mean, it's --
DR. HOYT: But this law or this rule says you really have to have pretty good evidence that there's clinical benefit. So what's the data that you need to suggest that?
MS. TILLEY: Well, I mean, it's very subjective. It can be, you know, now thinking of the different studies that I've worked on, often it's either -- it's animal data that's mechanistic or it's some phase II studies or phase I studies where they've tried it in people.
DR. HOYT: So rats, guinea pigs, and mice, is that enough? Do you have to go to pigs, do you got to go to primates, do you have to have a phase I study that shows at least -- I mean, what -- are there general rules that we can develop?
MS. TILLEY: I don't think there is a general rule. I mean, I think it depends on the risk of the intervention and it depends on the knowledge you have of the mechanism. You know, there isn't one paradigm that fits all.
DR. HOYT: John ¬- or Gene)? I mean you studied along with Steve Gould a blood substitute which, you know, had to, at some point, show justification that there was reasonable chance that this would help people to get it through that kind of process. So what kind of database did you have to have to come forth with that indication?
DR. MOORE: Well, we reviewed the tenets of the 50.24 yesterday and as outlined in number one, it has to be life-threatening that we're testing; number two, we have to have shown potential for a benefit in the host, receiving agent. So again, I agree that that's going to be individualized according to the question asked, but certainly obligatorily, there should be animal data and strong phase I evidence of benefit and risk to life.
Moving to informed consent, waiver of informed consent with phase II,, I think, is going to be very interesting. I'm sure in certain environments it'd be appropriate, but I think data will be rare in which we can justify that.
DR. HOYT: Ken, what would you want you -- you've got to sign off on this as head of the IRB in your institution. And you know, otherwise you're going to have the Houston newspaper saying that you violated the rules of 50.24. What have you got to have in your mind to feel comfortable with that?
DR. MATTOX: Is the treatment we currently have evidence-based and is the new treatment equal to or better in our evaluation and --
DR. HOYT: Component therapy for bleeding versus reconstitution of whole blood.
DR. MATTOX: Right. Now you've narrowly defined a product. All I need to see is that the new treatment has basis, in fact, to alter a test. Let's pick thromboelastogram. I'm willing then to push that forward because I know that --
DR. HOYT: That doesn't necessarily translate to clinical benefit to a patient, does it?
DR. MATTOX: Yes, it can. I think we already have significant evidence that what we have done the last 20 years in using component therapy that's 30 days or 20 days old, has not produced and has -- what we wanted, and has been -- has produced some bad results. And we have a coagulopathy that we didn't see in other people. So we need something better. If someone shows me that I have something equal to or better, I'm happy to evaluate that and have waiver of consent.
DR. HOYT: Would you want animal data to support that?
DR. MATTOX: I really don't care where that data comes from. I had no animal data to help me to start doing damage control and packing with laps or doing a temporary closure. And yet that probably is one of the most significant advances we made during the 1990s, and that data came initially out of South America from clinical data.
DR. HOYT: So you are really saying that there is a bit of a difference between what we might do clinically based on just the fact that we can use empiric data to derive a therapy versus the criteria we've set up for 50.24?
DR. MATTOX: Yes.
DR. HOYT: What about the rest of the panel? Is this too stringent, this requirement that you show benefit?
DR. HACK: I think benefit is somewhat broadly defined. Just participating in the study could be benefit, participation in the clinical trial --
DR. HOYT: Remember this is to hit the threshold for exception from informed consent --
DR. HACK: Right, right.
DR. HOYT: -- which we've said, you know, we were able to get because of these criteria. And so do these criteria need to be then modified? Is it too --
DR. HACK: I think that if you go back to the actual language and work with them, I think that they are fine. The -- it says participation in the trial has a potential to benefit the patient. I think that if you stay with that, that you have -- I mean, why do we do science? We do the science to find answers. If we knew the answers, we wouldn't have to do the science. So therefore, if we have absolute proof that this is going to help somebody in this situation, we wouldn't need to do the study in the first place.
DR. HOYT: Do the others interpret it this way? Because I think this is an important distinction and the broad definition of benefit to the world of science to the community versus the individual benefit to the individual patient is where people get hung up on this.
DR. HOLCOMB: Right, I think the words in the 50.24 don't need to be modified, maybe in some of the interpretation based upon the evolution of scientific data. When those words were written, let's just use transfusion, for example, I don't think people realize that pig platelets are not like human platelets. And you can't make a pig platelet work like a human platelet or fibrinogen or gene set, you know, really kind of delve into the acid base and coagulation proteins and all the animals are very different than humans.
So if you just use that example, I want to use, David, so if you set up a piggy or a mouse or a rat blood bank which has been done, does that translate to people blood banks? They actually are very different and the scientific base is very different. So is there a suggestion in the animal data on safety and toxicity and a suggestion of efficacy. I think you can do that.
And does that satisfy the words and the intent and the spirit of 50.24? Absolutely, and then you move into the phase IIA, IIB which can go into three trials. That interpretation as well allows you to maximize the benefits of entering into phase II trials just like any other kind of elective drug or product. So that you don't put patients in phase III trials at risk of failing after enrolling them because of lack of previous experimental human data.
DR. HOYT: Or perhaps you have some very compelling retrospective data which, you know, and the PROPPR trial is probably the most compelling group of data that's -- that has allowed people to go ahead with it.
DR. HOLCOMB: Right, and the prospective observational data. I think prospective observational data is an important step we haven't talked about. And they are starting to pop up more and more. Those prospective observational data, just putting somebody and recording -- measuring what happens is extremely useful and extremely powerful. And this is a step between recent retrospective data, going back to Ken's comment, and before you go into prospective --
DR. HOYT: Jeff?
DR. LAWSON: I think the concept of free will and choice to participate in medical research is very important. I think there is ample evidence throughout recent history when this goes awry when you don't give someone the choice to participate in medical research. So I think it's critical that we honor that from an ethical standpoint.
I think the two criteria for me to do a trial without informed consent, has to be a really good question. And you have to be sure you can't test it another way. And that's the benchmark, it doesn't matter if you did it in baboons or pigs or mice, good question and there is no other way to do it. And if you are going to meet that criteria then I think it's ethically acceptable.
DR. HOYT: Okay, we are out of time. And I want to thank the panel. I think we did our job in answering these questions substantially. Thank you very much.
DR. VOSTAL: Thank you, Dr. Hoyt, and thank you for the panel members. We are scheduled now for a 20-minute coffee break.
DISCUSSION AND SUMMARY OF OUTCOMES OF WORKSHOP DAY 1 SESSIONS
SPEAKER: (In progress) -- it's going to have the summaries of all the previous sessions that we had. And the moderator for this session will be Dr. Ken Mann, who is a professor of biochemistry at the University of Vermont.
MR. MANN: If you'll take your seats, we can get started, which means that you can then go home after the presentations. The sooner we get started, the sooner we go home.
I'd like to begin the -- this session by first thanking Dr. Vostal and Dr. Lindsey for their efforts in organizing this conference which has proceeded very smoothly, I think, because of the excellent logistics. I'd like to thank them and their staff.
MR. MANN: Now the plan of this activity is that each moderator will give a 15-minute summary of their perception of the section that they moderated. And it'll be followed by about a 15-minute audience participation. The idea here is to end up with a synopsis, if you like, of each segment of the meeting that will ultimately be distilled down by staff into a readable document that will be the outcome summary of this meeting.
I think without that kind of document, the output of these kinds of meetings disappears into the vapor. And so it's important that we end up with a tangible position. And I really encourage the audience to participate in the question-and-answer session. And we'll do that as efficiently as we can.
While I have said 15 minutes and 15 minutes, I'm not running the NFL clock, so it will be more flexible as if we -- if there needs to be. With that, I'll invite Nigel Key to come up and summarize the bleeding issues.
DR. KEY: Okay. Thank you, Ken. Well, again, our charge yesterday, I think, was primarily to look at entry criteria, inclusion criteria, if you like, for clinical trial. And again, I'm not a trauma surgeon, so I'm looking at this from a slightly different angle. But hopefully we captured in about six or seven slides here, the essence of what was discussed yesterday.
So very quickly, we are -- the problem is or the target here is hemorrhage and trauma, major cause of death and that these deaths occur very early. And a sense that the eligibility criteria here should be sensitive enough to really capture the entire target population. And we have the issue of sensitivity versus specificity here in any inclusion criteria.
So the first -- I'm going to just go through the four questions. And there is overlap between these answers and the questions actually that were posed to us. And the first one was what parameters should be measured to characterize severe bleeding for study enrollment? Those are really the key question, I think.
nd we talked about severe bleeding presentation or those at risk of severe bleeding. And I was very interested in the data that John Holcomb showed this morning in terms of the physician gestalt versus some of the other existing scores that are out there. But these are really the patients we want to get at and identify.
And the issue is, is this all done on clinical criteria or should there be a role for clinical biomarkers. And just to editorialize here a little bit from someone who also attended the April meeting at NIH, where to some degree, research questions were being raised, I think one of the themes that's come through is what are we going to do if we start a clinical trial tomorrow there -- versus clearly, of course, there we want to encourage and promote research questions that can be developed to better characterize the coagulopathy to look at instrumentation that might be able to pick that up as a biomarker, if you like.
So it was emphasized that the rapid turnaround time as really a critical feature of whatever is used for a clinical biomarker. We heard again this morning, some issues of discussion about the TEG.
I think, John, I've heard you say at one meeting that about 3 or 5 percent of patients have significant fibrinolysis when they present.
But the way I would look at that is that's probably a continuum. And I think the results of the CRASH-2, whatever one might criticize it using an antifibrinolytic strategy and the positive result there does make us wonder, and as a research question what is the role of hyperfibrinolysis here. Is it more universal than we can really think about or that we have thought about? And it is true that this is one of the areas of, in a coagulation lab, where we have trouble really characterizing fibrinolysis.
So again, from my editorializing, I think that that is something that I think despite the fact that TEG is a 1940s technology that we have the ability to develop TEG. And I think that one of the challenges is going to be issues like standardization of TEG, how do you standardize a machine that takes on anticoagulated whole blood. And of course, I think some of the answers to that might be in terms of the more of those -- the hand-held TEGs and things that can be standardized centrally.
Anyway, pro time -- we do have technology to be able to do that at the bedside. But we have issues in terms of being able to do it because of rules and regulations about point-of-care testing that I personally think could be addressed, but at this point it's a major issue.
I think a very interesting potential strategy in terms of using a trans-apical echo was brought up by Colonel Hack yesterday. He was really telling us what's going on in terms of the physiology of bleeding that might have a role. But these are -- at this point I presume these are one would consider research questions of their own.
And then there is the question about near-infrared spectroscopy. I think the conclusion was that at least in the study that being done it didn't add much to the clinical impressions. I think pretty well established and is really the lactic acidosis and base deficit, it shows up in most of the scores. I think it's probably already in routine clinical practice. And the issue about lack of validation as a predictor of bleeding, I think, is really more in terms of some of the TEGs and INRs.
Okay. So the second question, the triggers for the need for massive transfusion, whatever those might be, what should they be, be used as a criteria for defining severe bleeding or perhaps the question should be or risk of severe bleeding. So we talked about the pre-hospital versus the in-hospital setting and the data that are collected at the point at which the patient is initially encountered.
And there was perhaps my interpretation a little bit of uncertainty about the systolic blood pressure and heart rate thresholds. They vary a little bit by study and by scoring system. We heard yesterday and again today about the age effect of the substrate, is a 21-year-old serviceman the same as a 70-year-old person who is knocked down by a bus in terms of these parameters.
The CRASH-2 had a cut off of 90 millimeters, and if we got this right, 50 percent received any transfusion and the ROC trial was discussed by Lena yesterday has a cutoff of 70. So I think this is probably an area for consensus that can be discussed for the development of a clinical trial. And I don't want to lay -- you know, be too dogmatic about that, but clearly I think there is some data that are developing in terms of being able to come to consensus there.
And really essentially we're looking at simple tools in terms of the blood pressure, and heart rate, et cetera which are generally, but not always reflective of the physiology during bleeding. Again, we heard that other things may impact the blood pressure and it's not perfect in that respect as reflecting bleeding. So, the final question I think was if any of the existing scores helpful in deciding severe bleeding and Lena reviewed these yesterday, the TASH score and the McLaughlin score include laboratory results in addition to clinical observations and such things as hemoglobin or hematocrit base deficit were included in both of those scores. And she reviewed, for example, that a TASH score of 16 had a 35 to 50 percent likelihood of massive transfusion. The ABC scores really entirely clinical other than the abdominal -- intra-abdominal fluid assessment. And really I think it was on the panel there was some discussion about this that the fact these existing scores are promising. They really just incorporate in a validated -- semi-validated way, depending on which score you're looking at, some of the things we've already talked about, blood pressure based deficit, et cetera.
And I think it's really that the last sentence there or the last phrase could be that these could be further supplemented with clinical biomarkers as a research question going forward, once we really validate those. It may be possible to hone in more then we can do right now in terms of the ability to pick up the ones at risk of massive transfusion.
Oh, I'm sorry, there's this other question, should the above be considered either independently or composite criterias and I think we've really discussed that already, so I won't belabor this one.
So from the existing data I think we'd agree there's probably not a single criterion be it blood pressure anything -- systolic blood pressure anything else to define severe bleeding or severe bleeding risk. And we already are using some combination of -- if you consider lactate based deficit as a clinical biomarker and physiological criteria, they may be reflecting some of the same physiology. But that's okay.
And really modifying existing trauma scales to fill the gaps, I think this is sort of going forward maybe useful. So again, we have the question what would we do tomorrow versus what might we do in 2 or 3 years when we have a better understanding of the physiology. I think Jerry Levy brought up this morning the understanding the effect of tissue injury and the extent of tissue injury on coagulation and developing biomarkers. And I think those are things for the future, but perhaps more better discussed as they were in the April NIH meeting and at this point in terms of getting a new trial going we have what we have. I'll stop at that point.
SPEAKER: (Off mic).
MR. MANN: Use one of the two standing microphones, we also have a mobile microphone in the back, if you don't feel move to stand up. We would ask each individual to identify themselves, Dr. Golding.
DR. GOLDIN: I get from this that the take-home message that you are imparting is that the entry criteria on still -- there's still a lot of uncertainty there, and the uncertainty is very clear from the studies that were presented. So using a variety of entry criteria the number of people that ended up with severe bleeding as defined if you can define it as getting massive transfusion was quite a low number. So, it was in the different trials it varied but it was somewhere varying from 10 percent to 30 percent. So with that uncertainty, what Amy Shapiro said, I think is at this time the realistic approach to clinical trial design and that is you have to build in to your clinical trial design that uncertainty factor and you have to have very large trials. Until we have better data to -- with -- that predict with greater certainty which patients are going -- which individuals are going to have severe bleeding. Do you think that's a reasonable take-home message from this discussion?
SPEAKER: In general, yes, I would probably agree with that. I think that's the most conservative conclusion from this and that the likelihood of -- one hates, you know, we hate to go through clinical trials and not get a result in terms of really addressing the question but I think you have to conclude that if we can't pick out these patients with greater certainty then numbers become the issue.
MR. MANN: Would you -- Nigel, would you care to discuss the whole issue of the suggestion, I think, was made by Dr. Levy, about using apical filling as a definitive marker of blood availability.
DR. KEY: You know, it's not to --
MR. MANN: I don't know enough about it, so I'm asking for opinions.
DR. KEY: No, and I would have to punt on that one Ken. It's not something that I do. I'm not an intensivist on a regular basis. I --perhaps Jerry or someone else would like to discuss it and whether that could be incorporated into the clinical trial or at least how would one get the data from an initial study to incorporate. So, I -- no, it's not something that I've used a lot in my practice. I'm a mere country doctor, a coagulation hematologist. So --
MR. MANN: I've only used it on the receiving end.
MR. MANN: Anybody want to comment on the potential utility of feasibility, portability, et cetera of these -- that sort of technique.
SPEAKER: I think I would comment on the potential. I think the potential's there. I think its pretty -- there's not many trauma surgeons doing -- measuring the intra-cardiac filling in the ED with the portable ultrasounds. Certainly, up in the ICU it's becoming more common in the ICU setting, but I would say that's a potential.
MR. MANN: Well, it strikes me that in -- both the April conference and this present conference we're not only sort of delineating where the state-of-the-art is but I think we're also trying to encourage enterprise that would perhaps fill in some of these gaps and there are many bright engineers from many and scientists from various walks of life here that potentially would be stimulated to begin to look at such things.
Thank you Nigel.
MR. KEY: Okay.
MR. MANN: We'll go on to the next presentation, Charles Wade on animal models. And obviously, this would (inaudible) of discussion, I hope. It's interesting the -- and I'm a biochemist, I don't touch anything that's warm. And it's always struck me in my 40-years career in science that the less we know, the more we argue. And we've seen some of that in these presentations.
MR. WADE: Okay. We were asked to address some specific questions and the first one was, what is the appropriate type of trauma model to be used, that is, animal model to look at general indications for cessation of severe bleeding and trauma. And the first one was that the animal model should demonstrate the clinical hemostatic challenge for which the drug device biologic is used. One of the issues here, was that there are wide range of models out there that fit different products and in selecting the right model to evaluate a product in because sometimes they can be unsuccessful due to the fact that the wrong model is employed and that is also based on species when we're referring to model.
So there were questions about species differentiations between clotting, et cetera and understanding those differences are very important in the model selection. So, the key factors in selecting the animal model are the relevance of the hemostatic challenge that's the characteristics of the coagulopathy, and/or the vascular defect. And we're going to make a little bit of distinction between those two.
Proving or validate or reproduce the ability of the model and that is consistency of that model which involves a lot of initial development, and as I mentioned yesterday, always including that sham group or the control group in the analysis of your experiment. Potential lethality of the model; most of the stuff that we've been -- the products we've been discussing today are dealing with mortality as an endpoint. And this is because right now that's what the primary clinical endpoint is. There are models that can be used to asses, for example, acute lung injury, LI, renal injury, et cetera. But at present most of the clinical studies are directed at survivability and so we're saying that maybe the animal models should also encompass that endpoint.
Ability to demonstrate the hemostatic efficacy, that's been the theme that's being repeated here is that you have to select a means or a measurement that shows that what you've given is done mechanistically what it was supposed to do and then has an effect on outcome because there's been discussion here of the disassociation of survivability with changes in hemostasis. So, you may change the hemostasis or the hemostatic profile but yet the patient still may die. So, we have to kind of learn to tie these two things together. The ability to apply the hemostatic product as potentially indicated, I'll just make a point that you'd be surprised how many papers I review that somebody is given a drug or a procedure before they do the injury model. So, I mean, you've got to think through what is the process, the timing issues that mimic the clinical environment. And that's one of the things that I think that the clinicians interacting with the basic scientist and having a good understanding of the trauma environment are very important in the design of the models and the applications of the models.
So, finally, pre-hospital considerations are distinct from intra-operative considerations. Well, this is something we wanted to bring up that our products that are developed for field use in the pre-hospital setting, this is basically to control large vessel bleeding. They can be something as simple as a tourniquet or the ultimate goal for intra-abdominal or intra-thoracic is a fix or flat type approach which we're still -- a lot of people are working on versus the coagulopathy that maybe occur intra-operatively. I -- you get control of the big surgical bleeding but you still have massive bleeding from everywhere associated with coagulopathy. So, there are distinctions to be made between these two environments and the continuum between those two points.
Finally, the hemostatic effect should not be considered in isolation from the general systemic effects. One of the things that's been very interesting is that we've looked at products that have been used for resuscitation, IE fluids specifically. And all of sudden, we said, okay, wait, we fill the tank up, but all of a sudden we start looking at the immunological response, it's drastically dependent upon what fluid you use. We started looking to hemostatic response, the hemostatic changes are dramatically related to the fluids you use, and then the corollary to that is that when you look at a hemostatic thing you should be looking at what happens with volume, what happens with immunology. And you have to keep the whole system in mind as was mentioned, this is -- as Dr. Holcomb mentioned in his talk yesterday and the first introduction, is that trauma is a systemic, you know, disruption, and you have to understand what you're doing in terms of manipulation and the impact on other systems. You can't look at a system in isolation.
So those are the comments in reference to question 1 . The next question 2 was, what animal species are preferred for modeling in trauma and pre-clinic evaluation? What are the advantages of small animal models, rabbits, rats and the large animal models, pigs, primates, et cetera?
We feel there's a place for the small animals in initial screening and mechanistic studies. You've got to understand, similarly understand the mechanisms which you may not get good elucidation and the expense, et cetera in looking at large animals. You can do those in the smaller animals with a preference for the rabbit, okay. And basically the reason we brought up the rabbit, it's the ability to do repeated blood sampling and look at trends in hemostasis, et cetera, over time in that model, see things that you cannot do in a mouse and you cannot do in a rat. Specifically, most of the assays that we're using are developed for humans and have not been miniaturized for use in small animal models. So, in order to do those serial sampling, we felt that the rabbit was the appropriate model.
Swine for the pivotal pre-clinical efficacy and safety studies. There's a rich environment out there, I know Dr. Moore brought up the use of the sheep and we just do not have enough data yet in that particular model to recommend it. Maybe in time somebody will make an investment into it and develop appropriate models. In both -- in all these we must understand the basic difference between humans and the animal model and define them well. So, that has to be taken into consideration.
Are two species required for pre-clinical -- to evaluate a product, and I would say that more than one species is preferred. If you're carrying it forward for a community consultation protocol, I think, that's imperative that you have more than one species. The other aspect of this is that that the more that the papers come out in the open, literature and we're communicating on across a wide range of species and we see consistent results that adds to the validity of the product. And that includes different models as well.
When should a non-human primate be used? Okay. When the product is not compatible with non-primate species? There was talk about platelets not being compatible in pigs and rabbits, et cetera, that would be something to do the pivotal preclinical efficacy in safety studies in the non-human primate. To address specific safety issues, those are issues that were possibly identified in from clinical data that cannot be mimicked in the lower species. And potentially if there is a clinical trial that cannot be conducted and the reason I bring this one up, we put this one on is Dr. Pusateri and I both have experiences where there are combinations and situations in which the pivotal clinical trial is impossible, but you need that safety data in order to move a product forward and having it is a major issue that especially when we're looking at regulatory.
Finally, we were asked to come up with a series of citations when we produced this summary, we're going to get the full length citations in here. But basically we broke them down into a series of things. We've looked at general reviews of hemorrhage models, general reviews of coagulopathic models, discussions of the effects of the model to model variations and efficacy, external uses of papers in the groin model that is that particular model that's been developed, consensus model paper by Dr. Kheirabadi that's coming out. Safety papers for external use. Once again, Dr. Kheirabadi is doing extensive work on that demonstrating that external use can impact the systemic system as well. That is and through isolated endothelial damage and also the possibility of dissemination via the vasculature.
Internal use, there's a variety of models. These are the most recent models that have been produced, but there is a wealth in it of literature in this area where we felt these particular models mimic the clinical situation, specifically the hemodilution or the hemostatic changes which were -- are relatively new for this area of research.
Several papers on the new polytrauma model. This is the work that was done with the -- is funded by the Army across centers. One of the things I want to point out about this particular series is the consistency of this model across centers. The idea was to bring in three different laboratories to show that we could do the product consistently and repeatable between centers, and this would allow comparisons of results.
One of the biggest problems we have with the experimental models is all of us tweak them slightly so that the comparing lab to lab is very difficult. And this was to build a consistent animal model we were able to demonstrate that you could do it across through labs.
Topically applied internal use hemostasis, the early work by Dr. Guggart (phonetic) and Kheirabadi again. They spent a lot of time developing these models. And finally, the rabbit models, both for the issue of -- on coagulopathy but also to look at the other side of the spectrum. And that is what are the adverse events of some of these products and their potential to be -- produce a hyper situation. Thank you very much.
MR. MANN: Shall we've some discussion?
MR. SHIRE: I'm Marty Schreiber from Portland. And Charlie, it's a very nice summary. One issue that we haven't really discussed, I think, is the issue of the short-term models versus long-term survival, and I think it's worth some discussion. You know typically, we're doing studies that are either 2 hours, 4 hours, maybe 8 hours in duration, and we're seeing if animals survived for those periods of time. That's not going to tell us much about long-term survival.
In those models we're using inflammatory parameters, cytokines, neutrophil activity, RT-PCR, lung and other tissues to sort of give us an idea if the animals would develop organ failure if they survived.
But clearly, for the ultimate studies, the survival model is clearly the way to go. The primary survival models, is it's an orderly increasing complexity --
MR. MANN: I think it's being recorded -- yes, okay, good.
MR. SHIRE: Everything has to be done using sterile techniques. It's incredibly expensive.You need an animal ICU, et cetera, et cetera. So I think it's worthy, Charlie, to comment somewhat on the issue of if we're going to do pivotal trials involving either devices or drugs, do we need to do the pivotal trials in survival models?
MR. WADE: I would say the answer is yes. Marty, and as you know we've done long-term survival models both for internal and external dressings as well as looking at the issues of doing an ICU-type model. I want to make a point here that once again, this is where the clinical scenario or the clinical endpoint becomes paramount in the design of the animal experiment. So, for example, if the tenor is to change to 24-hours survival endpoint in the clinical model, I would say that we would do the same thing in the experimental models and push those 8-hour experiments out to 24 with maintenance, and I think that'd be a viable endpoint.
There are very few institutions just so people are aware who can do the long-term outcomes and provide the necessary veterinary and experimental support to manage these. And fortunately I've worked in two institutions in which we have full up ICU support with board certified veterinary intensivists to support us. And as well as having imaging capability, et cetera, and they're very few institutions that have that. I know Marty has done some of this work as well.
The other one is the limitations of these models in that environment. Some of these animal models are very -- for example, the rabbit are very difficult to support in an ICU-type environment because you can't use some of the drugs the same way you would in a patient. So there are variety of issues that have to be addressed as we go forward in this particular area, but I do support this issue of long-term survival.
And the other one is as we move forward we've talked a lot about the bow wave effect. It's that if we get people that survived 24 hour we put them in the ICU. We've created a new paradigm in the ICU with a patient that has to be looked at with greater acuity and then we have to correct that and so on down the line. If we're able to do that in the animal model we may be able to see some of that bow wave we carry out, including long-term recovery. So doing, you know, 6-month recovery, say, in a mini pig of a procedure is a viable option I think in some of these things if we have the right facilities available and I think it should be done in certain areas. DR. GOLDING: Something that you mentioned I want to --
MR. MANN: Dr. Golding.
DR. GOLDING: I wonder if you could make a few additional comments about the endothelial cell damage as mentioned by you in some of the talks. Are they blood or serum markers for that? I know that CD40 ligand, for example, is released, but is there anything that we've learnt from the animals in terms of endothelial damage and biomarkers that is predictive of outcome and that could be used possibly and even in humans?
MR. WADE: This is an active area of research in our laboratory and both the animal models and you know, in the clinic as well. There are some markers that we are looking at syndecans, for example, is an indication of changes -- around proteins. There is, you know, we're just starting into this area right now.
MR. FEUERSTEIN: Giora Feuerstein, and I'd like to make a comment today. I know that --
MR. MANN: Go ahead, would you identify yourself?
MR. FEUERSTEIN: Excuse me?
MR. MANN: Identify yourself, please.
MR. FEUERSTEIN: I'm Giora Feuerstein, former head of Translational Medicine Discovery in Wyeth Research and Pfizer. So I would like to make a brief comment and I know it was touched upon in reference to biomarker. But I think the great value of animal model, especially the non-human primate, would be to really put the scientific annularity of understanding what coagulopathy, for example, means as it's probably not the same throughout the situation at large. And I think all that we have done right now, listening to the therapeutic coagulation factor in trauma center, we're probably converting coagulopathy from one state to another state of coagulopathy and expecting at some point to have a better coagulopathy as far as survival, but we don't understand how the relationship between factor really pan out over time. So the mechanistic underpinning of the therapeutic is probably only an animal option, and in that regard the non-human primate is probably the best way to go.
MR. WADE: As I said that there is a wide range of components and differences between the various species. And I think we can learn progressively along the way in which directions we should refine those experiments to get to the primate. And the primates are unique in that. The other specific issue about the primate model is that in my experience, it's limited. I do have work with rhesus and with baboons is to do a trauma model is very expensive, number one, very heavily reviewed. And the final point is that when you do these experiments, you've got to maximize the return on them. That means bringing in together a very extensive team to do it, and once again it's a cost issue. So there are a number of primate centers across the United States that are -- do provide vital opportunities to conduct this type of -- these types of studies, but at present to go after a mechanistic component, I don't think that that would be viable in terms of funding.
MR. MANN: I want to take the chair's prerogative first to comment on an issue that has come up several times with respect to the relevance of animal models. And that is that there are -- there is no substitute for human study in any of the enterprises that we're talking about. But I've studied coagulation in animals going back to C. Lamb (phonetic), Fries (phonetic) and the fundamentals of the reactions are the same. The proteins are not inter-convertible, but the fundamental reaction systems and the nature of the proteins are identified one to the other place. What you did not mention in your presentation on mice, would you care to comment on mice?
MR. WADE: I group those together with rats, you know, this being the small rodents, so I mean the mice have a very -- due to the fact that you've got the knockouts and that you've manipulated strains, et cetera, there is a lot of characteristics that can be employed from those models to look at mechanisms, and they do offer a viable model to go into mechanistic investigations. We use them ourselves in our lab and I think that that is an option. One point though I want to make is once again it has to do with miniaturization of assay techniques is that that you go through a lot of these animals. And it has to be increased because of the variance when you're doing single time point analysis. So there is a lot of considerations that have to be taken into when you go down that route. They are a very interesting approach and I think they do have a place in the mechanistic side of things.
SPEAKER: I really do, I think that since you -- we really don't know the interplay between inflammation, coagulation and vascular cell biology. The mouse gives you a reasonable way to segregate those different components that may be brought into the area of various points in trauma and recovery from trauma.
MR. WADE: Okay. And I want to go back and Ken is involved in a study that is looking at a systemic approach, in the systems biology approach. And I think that this is going to help answer a lot of things in that they go all the way the -- the data that's been generated in mice, rats, larger animals, as well as in humans and to integrate across all these databases and all this different type of data is going to be, I think, where we get the key is out of the systems biology approach.
MR. MANN: There is no substitute for human beings in studies of human beings?
MR. WADE: They have to be an integral part of the designs.
MR. MANN: Other questions, comments?
MR. KHEIRABADI: Just a comment --
MR. MANN: Please identify yourself.
MR. KHEIRABADI: Bijan Kheirabadi from ISR, San Antonio. I just wanted to --
MR. MANN: Excuse me, I just wanted to comment. We're recording everything. So when we put together the document, it's important we know the author of each contribution.
MR. KHEIRABADI: I was just going to make a point here. I think we have put -- quite emphasis on the animal models and animal studies, but there are sometimes when a vitro test can be also quite revealing. Recently for instance, we found out just as human endothelial cell culture can be a very good way of finding the toxicity of some of these hemostatic dressings that we didn't know in the past because generally they are tested on fibroblast, which are much more resistant cells. But when in some -- when we managed to incubate some of these with just endothelial cell, we found the toxicity which caused basically necrosis of the cells. So I think this along with animal studies maybe some in vitro studies might be also useful in terms of showing the safety of these products.
MR. WADE: And I'm well aware of that work, Bijan, and I agree with you that that we have to go down at some cases to that level to understand -- fully understand the mechanism to go for a reduction oriented. And I think that Ken's work, you know, demonstrates that reduction approach. So you both have to work down and up from the animal model.
MR. McPHEE: Mart McPhee from STB. I would agree completely with what Bijan said and they're also ex vivo models that can be used. But a more general point, I think, is what I was seeing in the panel yesterday was a really excellent discussion of animal models for testing products, but not really for developing them. The reality is that for developing a product you -- is a very long process and the bulk of the work gets done in models that have to have high throughput, low variability, low inter-group variation. They have to be practical from a perspective of cost and scale and those are not the same kind of considerations for when you produce a product that you think is good and now you've got to test it into a definitive or near definitive kind of test.
And I just think we need to be clear about the distinction, really I felt this was a very excellent discussion about testing and evaluating products to get whether we should put them into the ultimate experiment in people. But development actually -- man -- development animal models and other models like Bijan said, I don't think really we have gone into that area here at this conference.
MR. WADE: As I mentioned yesterday when we opened this session is that the point of this particular exercise was to look at that -- that the animal model is a window to move into the clinical environment in these particular kind of situations. The point here is, is that that as I said yesterday the -- you have to have input from the clinical side of what the endpoints are, be it the change in hemostasis, that you are interested in or what the condition is as well as what the endpoint is. Because you want to design the animal model for these -- this efficacy testing to mimic as much as you can that particular environment. I agree with you that there are a lot other steps along the way in terms of dressing, initial mechanisms, et cetera, in developing a product. But when we get into the model that's going to be used to lead to subsequent clinical trials, I think that you have to make some changes and also the magnitude of the interventions and the amount of data that you collect change drastically when you're moving up the line.
MR. McPHEE: And I don't think we disagree, in fact, it's really key to have these models in line when you're working on your development one. I guess I just wanted to be quite clear because as you interface with the FDA, if it comes into their mind that development -- these are good models for development then that would be a problem when they're looking at the early data because they do want to see the whole thing.
MR. WADE: Yeah.
MR. McPHEE: I don't know what the terminology that could be used would be to just make a distinction but I wouldn't want us to get you know called -- caught up in a --
MR. WADE: You know having been involved in both, you know, product development both of drugs and devices the amount of energy you expend, I mean, so for example doing three species for toxicology, teratology, et cetera, and thousands of animals to report to the FDA at doses five times what we would use clinically, et cetera, that is a different ball game than looking at this particular endpoint before you ever go into those types of studies and there is a whole series of studies below that, that must be done before you get to the efficacy study. MR. MANN: Dr. Holcomb.
DR. HOLCOMB: Holcomb from Houston, along -- a little bit along the same lines. Do you want to discuss the indications and maybe contraindications along the developmental testing process, controlled versus uncontrolled human trials.
MR. WADE: Okay. What Dr. Holcomb is referring to is that that the clinical scenario is a -- which goes back to our last point, for example, I want to understand a product, let's say a volume expander. I can do that in a controlled, very consistent hemorrhage model by just withdrawing blood, re-infusing the product and looking at the physiological inflammation, et cetera, response to infusion of that product. But it does not mimic the clinical scenario. The clinical scenario is an uncontrolled bleed.
So you have to develop models that go beyond that and this is for example, we talked about a perforation of an artery and letting it bleed out and then resuscitating or cutting a liver, or cutting a spleen.
Uncontrolled hemorrhage is what the clinical environment is. It is not a just a fixed volume where we open a stopcock and then close the stopcock. We can learn a lot about the physiological responses and safety issues associated with the product by doing those experiments as a lead into doing the uncontrolled model.
Uncontrolled models have a wide variance. Okay. They go up almost in an exponential if -- the severe you go the variance goes up exponentially. So if I want to develop a major damage to say a liver injury, the variance and when I cut that or injure that liver it goes up phenomenally. So I'm taking where I was -- in a controlled model using 10 animals, I should be using 30 animals per group, 40 animals per group in a liver injury model because of variance in the model. So that's a different approach, but it is an essential step along the way.
MR. MANN: If there are no further questions we'll go on to the next session.
MR. WADE: Thank you.
MR. MANN: Grant Bochicchio.
Out on a coffee break?
DR. BOCHICCHIO: Pardon?
MR. MANN: Out on a coffee break?
This is clinical endpoints local agents.
PANEL DISCUSSION: ENDPOINTS FOR CLINICAL EVALUATION OF PRODUCTS FOR BLEEDING INTERVENTIONS
CLINICAL ENDPOINTS-LOCAL AGENTS
DR. BOCHICCHIO: First, I wanted to thank the panel for -- yesterday, for their contribution to the discussion. I think one thing was -- an important thing I wanted to bring up was in one of our attempts or one of our goals doing this presentation was trying to actually bridge a little bit between what we know in past bleeding trials in the clinical arena. And what I was trying to do is facilitate that type of thinking into a trauma situation because having been a principal investigator at multiple trials in the elective arena whether it's liver or severe or mild-to-moderate bleeding in soft tissue that there were clear endpoints that have -- the investigators have been asked to perform in addition to what the FDA has at least determined to be important endpoints as far as bleeding and hemostasis. So that's what I was trying to get our panel to kind of grip into or at least try to focus on in that way.
We are going to kind of go individually through the questions. The first question was what local and/or systemic factor should be taken into account to determine the severity of blood loss in a localized anatomic site or considerations include anatomic site, patient factors of coagulopathy, rate of bleeding, arterial versus venous, single versus multiple sites, discussion of objective bleeding score, need for blood transfusion, and how many units. And we looked at anatomic factors, and we looked at this discussion of external versus internal.
The panel supported that challenges in external and internal bleeding control are different, specifically when we're talking about, you know, whether or not the patient would be treated in the field where the groin is different than a patient that is treated in the operating room. The panel did not discuss definitions of internal and external. Some of that is self-evident or clear, but again we did not get to that. In terms of body cavity where we separated it into chest, abdomen and pelvis or solid organ; in terms of the specific organs liver, spleen, kidney and lung, some panel members suggested that there is no difference in consideration of bleeding in different body cavities and/or solid organs. I do caution however in these different body cavities you can also have complications and safety issues, i.e. if you're putting a pad near the lung, and it causes the lung to be lacerated or persistent air leaks that can be a problem as well as in the abdomen if it facilitates a higher rate of adhesions that causes inflammatory response and it could lead to fistulas. Those are real complications that need to be discussed and/or considered.
Named versus non-named vessels or artery versus vein, the panel noted that traumatic wounds often have mixed arterial or venous bleeding so they can tie -- kind of really -- we're not able to separate that. I think the purpose of this discussion was that some trials in the past have been on non-named vessels, mild-to-moderate bleeding in soft tissues, and the question is whether or not products would be specific to large named vessels, i.e. aorta versus mesenteric vein. Some of the big players that we see that are noted and they have severe bleeding versus multiple small branches off a major vein, is that different? And also whether or not an artery versus a vein is different in regards to if you place a dressing down on a vein, would you like to just leave it there and be safe by doing that versus an artery where a high pulsatile vessel or organ would require you to just use it as a temporizing measure. And because of the most likelihood that's going to create pseudoaneurysms, you would have to remove that at that time and only use it as a temporizing measure and that's why that discussion was held.
Next, looking at patient factors, question of coagulopathic versus non-coagulopathic and we know that coagulopathy due to cancer versus trauma versus pharmaceuticals are different. Our prior speakers had talked about how inducing coagulopathy with heparin is different than the traumatic coagulopathy that is created by trauma.
And the question is you really cannot just, you know, extrapolate data from one group to the other. The panel felt that those patients are different although we heard some conflicting opinions earlier from other panels and then on how to deal with that. You know the history of thrombo-emboli, DVT, PE, cancer. A lot of trials for bleeding are in cancer patients. So what happens when you place a device that contains thrombogenic material, i.e. thrombin or fibrinogen in a cancer patient who is already prothrombotic, are you increasing the risk of thrombosis, DVT, PE in that patient population. And those are real factors.
So you know, immediately you may have a benefit for the patient where you stop the bleeding, well now you are doing Whipple and you tear the mesenteric vein at the end of the case and you apply your dressing while that patient is prothrombotic you may induce or increase her chances of developing a PE or a thrombus in that patient because those things are real phenomena.
Next, the question of emergency situation trauma versus elective. Trauma patients are clearly more heterogeneous, multiple versus random bleeding sites. Trauma patients will have multiple bleeding sites. And the elective patient population is really one single bleeding site that you pretty much have a good idea that you're going to get into that.
I use the abdominoperineal resection, where if you do a complete mesorectal excision the bleeding from the pelvis in the bony area is pretty common and, you know, that's going to happen. So after you remove your specimen at that time it's easy to do a trial where you apply a hemostat into that area versus a trauma patient who has, you know, ten different gunshot wounds in ten different organs and you're trying to figure out which one you're going to treat first.
Field versus hospital, civilian pre-hospital, military field, data may be poolable. And I think the panel was very -- was in agreement that treating medics in theater and training medics in the domestic population is something that's comparable. That can be extrapolated. And I think that's extremely an important conclusion from the panel.
We did not discuss prior hemostatic agent exposure. That's one thing that I had up there initially because of Dr. Lawson's work on thrombin exposure and should you -- if you know a patient has had prior thrombin exposure like in a cardiac patient should you -- is it safe or is it reasonable to -- you're giving them secondary or tertiary hits of thrombin exposure, i.e., multiple -- takes back from a damage control patient should you -- they be exposed to multiple thrombin exposures or fibrinogen exposures based on risk. And that's something that needs to be further discussed.
Rate of bleeding, the consensus was that rate of bleeding is quite important. However, as other panel struggled with, there was no consensus on objective measures of bleeding or method of assessment. So as surgeons, you know, we all sit here and say we know severe bleeding when we see it, we know catastrophic bleeding when we see it.
But from an FDA standpoint, how do they get their hands around that to say, yeah, we're going to approve a product that has an indication for severe or catastrophic bleeding. And that's something that we continue to struggle with. And we are challenged to really address that because that's a real need and something that we were hoping to try to get an answer for in this meeting.
Continued clinical meaningful endpoints, at the anatomic point of care. You know, I tried to push this timely at 100 percent hemostasis. And what we mean about 100 percent hemostasis is how often do you apply that product and it works. So if you apply it 10 times does it work 8 times on 10 or does it work 10 times out of 10?
And I think that's important because as you study these different products, what is an acceptable rate of efficacy, is it 80 percent, is it 100 percent. And I think that's something that we did not get a clear consensus on. And I did try to bring in the -- again, the elective hemostatic trials, endpoints -- the predetermined endpoints of 3 minutes versus 5 minutes was that acceptable to the panel.
It seemed like the consensus was 3 minutes was fine. Some panel members thought that it should be a continuous assessment. But in a trauma setting as you keep on peeking and taking the dressing off and looking, are you really disrupting the clot or you're biasing or impacting the study endpoint by actually doing that.
So I think in the elective trial world we really feel that, you know, 3 minutes actually, at a time to assess it, is really an objective way of looking at that. And it works pretty well.
Again, talk about safety, adverse events, functional healing, we did not discuss that. But again, safety is extremely important where you may initially have a benefit, but you know, you may pay later for that complication.
Finally, what covariates may impact clinical trial design? And covariates are very important. I was going to kind of briefly mention them since we did not get to them. Conditions of use, open field versus controlled sterile conditions again it is a -- you know, different products may be good for temporary use in the field, but may be contraindicated for in the hospital. Some products may be good for temporary use; you can use them in a damage control patient, but have to take them out in 24 to 48 hours.
Next question is how long can they be implanted, for 24 hours, 30 days, what -- you know, what is the indication there; elective versus emergency situations. We did not talk about compressible versus non-compressible hemostats as non-compressible hemostats are coming down the pipe, how are they different, how would they be judged differently.
Single exposure versus multiple exposures even being, you know, multiple device placements, safety efficacy, you know, you're giving them a large thrombin exposure, a fibrinogen exposure, how do we measure the safety in that in our patient population. Is it different to be placing a topical hemostat on the aorta or the cava versus a grade V liver injury, how is that different.
Ease of placement and removal, well you -- something may work and if you're using it as a temporizing measure, but when you try to take it off it tears, you know, tears the aorta in half. Is that something we really want to be using? If you use something that we had mentioned yesterday like a powder that you place into the area and it stops bleeding, but then you have to go back and repair it, does that cause or increase your difficulty in trying to repair that later on if the powder causes inflammatory reactions or really clouds the field.
User training, we talked about, you know, there are differences between training a surgeon versus a non-surgeon. That's very important if we are going to be studying devices in the field whether it's in theater or domestic population.
Again, we talked about prior thrombin exposure, a history of DVT. And type and category of products is something that I think that will have to be discussed further as, you know, how are they different, liquid versus powder versus foam, compressible versus non-compressible, is it a biologic thrombin versus fibrinogen product -- project versus a non-biologic, a chitosan or an HM-chitosan derivative and how do they all different in their efficacy and their endpoints.
And finally, what are the standards to compare it against. If you're studying a sugar molecule or a plant molecule, what is the adequate comparator to that? If you're studying a lyophilized thrombin fibrinogen product what is the comparable arm to that, is it Surgicel, is it something else and really what is the real comparator here? And that's the end of our summary. Be happy to entertain questions.
MR. MANN: It could be that you convinced everybody in the audience or could be just getting tired, I don't know so. If there are no further questions, we'll move on.
DR. BOCHICCHIO: Thank you.
MR. MANN: Thank you. The next presentation is by Richard Weiskopf, Clinical Endpoints-Systemic Agents.
ENDPOINTS FOR CLINICAL EVALUATION OF PRODUCTS FOR BLEEDING INTERVENTIONS-SYSTEMIC AGENTS
CLINICAL ENDPOINTS-SYSTEMIC AGENTS
DR. WEISKOPF: Thank you. I introduced a session on endpoints yesterday by talking a little bit about the importance of endpoints in that it represents the critical translation of the desired indication to the investigation. That is, it takes the indication and says, this is how we're going to do it.
The endpoint then drives a protocol and then these are interactive and the -- it all -- in the end it also drives the entire development program. And I think the lack of clarity of endpoints in the past has in part led to the dearth of products we have in this arena.
And asking the panel to answer the questions with which we were charged, which we will all get to in just a few moments, I asked the panel to consider how close should the endpoint be to the agent's pharmacodynamic and pharmacokinetic properties. And I think you'll see that by and large the panel did a very good job in addressing that.
I also asked them to consider the following criteria that an endpoint must not only be scientifically valid, it must also be clinically meaningful both qualitatively and quantitatively. It must be logistically feasible. And it must be resource feasible for -- not only for fiscal issues and for personnel, but also in terms of absolute and relative.
By that, I mean, there are absolute values beyond which some sponsors or investigators may not be willing to go, but relative that many sponsors have competing programs and that the endpoint must be constructed for a program that makes it doable and able to compete with other programs within the sponsors' portfolio.
I note that we must not constrain these endpoints and protocols to the point where they adversely affect product development. In the past day-and-a-half, much of the focus of this forum has been on trauma.
However, I note the title of the program is severe bleeding due to trauma and other causes and that at times I feel that the discussion has focused on trauma not only to the exclusion of other causes, but to the almost elimination of other causes looking at other -- the other causes of -- and other incidents of trauma of bleeding rather and saying, well, these are not adequate models for trauma.
That may be, but in their own right they are valid areas for investigation and that when looking at endpoints must -- one must consider that as well. So with that as a prelude I'll now go into what the panel discussed yesterday and what they concluded.
In discussing whether hemostasis itself is an adequate endpoint, the panel by and large agreed that it is the best primary endpoint. After all that is the pharmacologic action of the agent that is to be studied. However, most thought that it was difficult to apply and challenging to apply. So here is an opportunity for investigators and sponsors, I think, to be more creative and to develop direct measures of hemostasis if, in fact, their agent is an agent that acts by promoting hemostasis.
The issue of bleeding always suggests an issue of hemodynamic instability at least severe bleeding does. But the panelists by and large were not in favor of this endpoint because criteria could not be applied uniformly. And more than that, routine care uniformly changes hemodynamic stability and instability and would make evaluation of such an endpoint quite difficult.
For many studies transfusion has been used as a surrogate, if you will, for hemostasis. All in the panel thought that this was clinically relevant, but there were varied opinions about the magnitude of reduction required. Now remember, we're not talking solely about massive transfusion, we're talking about any degree of transfusion.
The panel all recognized that the risk of transfusion is on a unit-by-unit basis, but that variability -- clinical variability and trial variability makes an endpoint of looking at only one unit of transfusion perhaps inadequate unless the sample size were extremely large and that at least one member thought that we ought to be looking at more like two to three or four units. Massive transfusion was also thought to be likely acceptable, but again, is not necessarily the focus of all severe bleeding.
All agree that mortality is an acceptable endpoint, efficacy endpoint that is. And most thought -- all except one thought that it ought to be somewhere within the first 24 hours. The opinions varied in range from 6 hours to 24 hours with one of the panel members thinking that we ought to be looking at something like 28 to 30 day mortality. All agreed that something in the vicinity of a 5 percent to 10 percent reduction in mortality is clinically meaningful.
As a -- not on the slide, but I would say that if we move and the panel strongly suggested that we should move to mortality that is in the first 24 hours that does not negate the issue of tracking mortality for a longer period of time, but that this could be captured as a safety endpoint rather than as an efficacy endpoint. And this issue directly address what I charged the panel with and that is looking at an endpoint that is pharmacologically -- pharmacokinetically akin to the agent's action.
Interestingly enough, although morbidity has been endpoints in the past, the panel thought that this was not a good endpoint, not a primary -- not good a primary endpoint but perhaps acceptable for secondary endpoints. And the panel thought that composite endpoints similarly were not good that perhaps they would be -- could be manipulated in some way and that they are nevertheless acceptable as secondary endpoints. This stands in contrast to some other areas of product development, where composite endpoints have in fact become relatively standard.
As to surrogates, I think the panel -- in some ways all the other measurements -- all the other endpoints except for hemostasis is a surrogate for hemostasis, but further down the road other surrogates such as length of stay, length of ICU -- length of hospital stay that is, length of ICU stay, multiple organ failure, time to another intervention, there was a lack of enthusiasm in the panel for these surrogates.
And I think that represents the summary of what the panel discussed yesterday.
MR. MANN: Thank you Richard. Dr Wade.
MR. WADE: Charlie Wade, Houston. Rich, I'd like clarification on this 5 to 10 percent delta. I have a little problem with that because we enroll patients that have a blood pressure less than 90, and we had about an 85 percent survival rate, in contrast if I take somebody who requires massive transfusion I get a 60 percent survival rate. So it's easier to demonstrate a difference, and I just feel it's cautionary to put a percentage up there depending on the patient population that we are defining.
DR. WEISKOPF: Well, in some respects you're right Charlie that any percent reduction is a number that is to a certain degree pulled out of the air it's what people think is clinically relevant in reduction. Interestingly, enough although some people have been excited by the CRASH-II study despite it's -- shall we say, less-than-stellar design, that reduction was all of 1.5 percent and most people apparently in the panel at least are not very excited by that.
It is always a challenge in clinical trial design to select a population that has an endpoint that can be moved sufficiently to make a not only a statistically significant difference but a clinically significant difference. If you have a -- if you're looking at mortality and you have an extremely low mortality it's going to be very difficult to have a trial that affects that in a favorable way. I think the panel was talking about a relative change in mortality rather than an absolute reduction.
MS. TILLEY: With regard to the last few --
MR. MANN: Please identify yourself please identify yourself for the record.
MS. TILLEY: Oh, Barbara Tilley from Houston.
MR. MANN: Thank you Barbara.
MS. TIILEY: With regard to your last slide, I agree that those are not good primary outcomes but in the framework of comparative effectiveness research where one of the considerations is the cost of the treatment and when you're comparing existing modalities for treatment, I think those variables like length of stay, organ failure et cetera might be a very interesting subset for secondary analysis with respect to cost.
DR. WEISKOPF: Well, I did say that they are -- that they are reasonable things to look at for secondary endpoints. We were charged primarily with looking at primary endpoints, for secondary endpoints they provide supporting evidence to support a primary endpoint especially if that primary endpoint just barely hits statistical significance or is not a very strong clinically important endpoint. And it also allows a sponsor to potentially market on those claims if they're allowed to put those secondary endpoints in the label and of course, if it's an economic endpoint such as length of ICU stay or length of hospital stay that may help them market as well.
MR. DAKOTA: Joe Dakota (phonetic) Integreon (phonetic), good morning. I have a similar question for secondary endpoints. You mentioned transfusion avoidance, but was there any discussion about total volume requirement in the future?
DR. WEISKOPF: No, I think the transfusion discussion was centered on primary endpoints. I think everything the panel discussed and that was displayed here today, relates transfusion to a primary endpoint not a secondary endpoint.
MS. LISTER: Shirles Lister (phonetic), Puget Sound Blood Center in Seattle. I am a transfusion-medicine person with a significant interest in platelet transfusion, and it's been very interesting for me to sit in on this conference. One of the things I've been struck by is that the majority of people here were very interested in evaluating their product in the trauma situation. That's clearly a multi-factorial situation where you may have a product like say a platelet transfusion that may be effective in that situation but it may not be effective if they're acidotic, cold all those other kinds of things.
So I think it's important that the product that you're evaluating as I think you've discussed is assessed in a situation where it's likely to be effective. So I can give a platelet transfusion to a hemophilic patient and that's not going to stop his bleeding, so I think it's important that for some of the situations a controlled bleeding event like a vascular situation be used as a model for does this product effective because I just -- I think it's extremely difficult in a trauma situation you may miss a very effective product in the complex situation that you're talking about.
DR. WEISKOPF: Thank you for pointing that out. I tried to make a similar comment in introducing this by saying that not all important bleeding is a result of trauma and that I would hope that the discussions here today would not discourage development of products that are not necessarily intended solely for trauma. And your points are very well taken with respect to that, just as other circumstances are not necessarily an adequate model for trauma. Trauma is not necessarily an adequate model for other circumstances.
MR. MANN: Dr. Golding.
DR. GOLDING: You know, you mentioned that the discussion about the mortality rate and what would be significant here, so you mentioned two numbers, one the --
DR. WEISKOPF: I'm sorry, so we are having a bit trouble hearing you.
DR. GOLDING: So you -- we're talking about the mortality rate that would be acceptable, the difference in mortality between the treatment and the control. So, the numbers you mentioned the trial the 1.5 percent where they were, if I remember correctly, 20,000 individuals and then you mentioned that some of the panel members are talking about the 5 to 10 percent difference. So I think there's some practical statistical considerations, and what I heard from John Holcomb that if you look across the United States in one year people that are eligible for massive transfusion or receive massive transfusion is somewhere around 900. So, you're talking about several odd differences in terms of numbers of patients, so that's also a consideration.
The other consideration is that because of the variability of an uncertainty in these trials, you need a robust number in order to show a difference and on the one hand, on the other hand it becomes very difficult if your treatment effect is modest to show any difference because of the heterogeneity of the patients. We talked to Holcomb, and others mentioned differences, incentives that were quite remarkable, so there are a lot of problems associated with figuring out what the right number should be.
DR. WEISKOPF: Your points are very well taken and it's one of things that has among others that has led me to believe that mortality ought not be -- this is my own personal view not necessarily that of the panels -- that mortality is not necessarily the right and appropriate endpoint for some of these trials that as you point out there are mortality with current care trauma has decreased greatly. And it's showing a substantial difference in mortality, good enough to be called clinically important makes for a difficult trial with a substantial end. MR. MANN: Dr. Bochicchio.
DR. BOCHICCHIO: One question which I also posed to the FDA panel here is as we look at data from Europe, you know, the CRASH-II trial is an example of -- we're looking at the endpoints that these studies are reporting. And having audited a lot of these European or these large trials and specifically the CRASH-II when we were invited to participate in that there was no funding in that trial, it was kind of participate. And my big issue with this and these trials is that none of these trials are audited or QA'd, so these data gets dumped and sent.
And, so when we accept these, you know, question these primary endpoints from large -- these fairly large trials, it's never been QA'd by anybody and if you actually QA some of these trial which I've done in the past there's a 30 to 40 percent error rate in their quality of how they enter that data. So how does FDA or how does when you look at that as -- did they really get their primary endpoint in CRASH-II, or is that actually transcription errors because no one ever talks about that, and I think this is a significant problem when these trials get done elsewhere and we're trying to talk about it in these forums.
DR. WEISKOPF: Well, you raise some of the important issues -- one or two of the very important deficits of the CRASH-II trial, the first -- but let me dissuade you from the idea that this was an European trial. If you go to the appendix and add up where the trial site with the number of patients that were enrolled, approximately one percent came from Western Europe or North America. There's a substantial question as to whether the results from those -- that trial even had it been done correctly would apply to current care in Western Europe or the U.S.
And with respect to auditing of the data you're absolutely correct. If you read the methodology of that trial there's no guarantee what so ever that those data are in fact the actual data that occurred. Steve, Dr. Gull (phonetic).
MR. GULL: Steve Gull, Chicago. The day and half of discussions have been stimulating but I'd like to express some concerns, I'm an optimist but I have some concerns and they are based on having, you know, as many people here know having lived through a 720-patient trial in trauma under waiver of consent and I think a lot of what we've discussed in this conference relates to what I'll call variability. It sort of picks up on the theme Dr. Schlichter (phonetic) made. So we may have products that work, we're trying to generate the truth, factual information and we're going into a very noisy chaotic setting.
This morning with the panel that Dave Hoyt led which I thought was terrific, we heard about variability that I've lived through. Gene Moore talked about variability in outcome at sites. That's a reality. We heard if we take the Lawson whack effect, we heard about the variability patient to patient which we live through.
And I want to add another element of variability that we encountered, which will always occur in trauma trials which is the inability to have as good compliance with the protocol as one would like in a stable setting. So, we had an incidents of 17 percent protocol violations for a variety of reasons. Now, some of these I think we need our statisticians to help us out. So, one of our desires with protocol violations was to do the analysis not on ITT or MITT but on per protocol population. If our goal is to generate the truth, patients enrolled to meet the per protocol eligibility give us the best opportunity to asses the performance in the product in the setting in the patient for which it's intended.
The regulatory concern is should the ITT represent the real world and we need that for safety information. I understand both sides of the argument, but we need some help here if we want to move forward, because with all of this variability, in a setting where a patient has defects in an anatomy, in oxygen transport, in their metabolic situation and their coagulation status, we want to evaluate a coagulation profile.
I frankly am concerned that despite the day and a half, we're actually going to find a way to move forward that complies with the regs, complies with good scientific principles, generates information that will lead to new products. So I don't want to sound like a pessimist. I'm very hopeful that this audience and with the accumulating experience of trials like ours and others will lead the path forward. But I want to express a bit of caution about having a panel opine on the question such as what's an acceptable mortality endpoint because Dr. Golding just raised the practical issues. So it's not a direct question, I'm sort of opening a forum here. I'd like to hear some more conversation about those topics.
DR. WEISKOPF: I certainly understand your concern Steve. And I for one have never been one who's been in favor of ITT analysis, and in many cases it doesn't make any sense, whatsoever, if a patient never received an intervention why they ought to be considered to be part of the efficacy analysis.
I'm more in favor, personally, of a modified intention-to-treat analysis, but when, of course, the details as to what represents an appropriate modification, this is not the forum where we examine the statistical approaches to these analyses. Perhaps that could be a good focus for a forum or a workshop at another time. But you raise valid points worthy of discussion.
MR. MANN: It strikes me that the minimum thing you can get from a trial is knowledge. And if you take the CRASH trial, for example, and you use it as an endpoint whether or not there was a delineation (phonetic) of primary fibrinolysis in those patients, you will be able to segregate whether or not the drug basically was effective in the environment used. I mean -- DR. WEISKOPF: Well, let me back up that, and go to the entire basis of the CRASH trial.
MR. MANN: Yeah.
DR. WEISKOPF: Of course, you are right that the idea of doing any of this is to gain information. But the information must be valid information. All that information is valuable. The CRASH trial if -- it's premise, the entire premise of it --
MR. MANN: I'm not defending the CRASH trial.
DR. WEISKOPF: Okay. I'm sorry, I misunderstood --
MR. MANN: No.
DR. WEISKOPF: I misinterpreted you --
MR. MANN: What I'm saying is that in the fraction of individuals that are treated, let's say with tranexamic acid, who happen to be suffering from fibrinolysis as a contributor to their coagulopathy. If you could segregate that subset you would know whether or not the drug worked at the biochemical level in the individuals. I mean given the heterogeneity of the input population --
DR. WEISKOPF: Yeah.
MR. MANN: You can drown out the response --
DR. WEISKOPF: Absolutely, and you are talking about the noise to valid signal issue.
MR. MANN: Yeah.
DR. WEISKOPF: And since the presence of fibrinolysis was neither an inclusion criteria nor assessed in the trial we cannot have the answer from that trial about that issue, but your general point is correct.
MR. MANN: But you could know that retrospective to the trial from the stored samples?
DR. WEISKOPF: Excuse me?
MR. MANN: You could know that answer retrospective to the trial in stored samples?
DR. WEISKOPF: I don't know if they have stored samples.
MR. MANN: Well, I am just saying that --
DR. WEISKOPF: Yeah, I understand.
MR. MANN: The design basically would allow you to de-convolute that subsequently.
MS. DOLHONDE: Deborah Dolhonde (phonetic), Houston. And I would like to address the previous comments about the need for a biostatistical and research design expertise to join the team. And I would like to frame this in terms of thanking the planning committee, the presenters and the panelists for what I consider a truly trans-disciplinary opportunity to address this critically important public health issue.
And I just wanted to extend an ear as an epidemiologist and a member of a biostatistics epidemiology and research design large group that our ears are here, and that our interest and experience in research design and opportunities to come up with possibly new paradigms and things that can address the chaos that's been described, the high variability, the potential for confounding, the potential for misclassification all of those things have a rich body of expertise and methodology that can be brought to bear.
And I do appreciate this opportunity to learn about this area and want to assure folks that these trans-disciplinary teams that you've illustrated here by bringing together trauma surgeons, transfusion medicine experts, ethicists, biostatisticians et cetera that with these networks that were described previously by Dr. Tilley with bringing together large networks where highly stratified patient populations can be considered in reasonable, validated methodologically rigorous studies it can be done.
And thank you again for the opportunity to be here.
DR. DIMICHELE: Donna DiMichele, Division of Blood -- I'm the deputy director of the Division of Blood and I've -- I would also actually that comment was actually a good segue way into what I was going to say since we are here representing the NIH and HLBI.
First of all, we also would like to congratulate the FDA and the planners on a very, very informative meeting. And I think our message is that, I don't know what's gone on before, but certainly the Division of Blood specifically with its workshop that was held by -- earlier this year on Trauma Induced Coagulopathy, certainly has an interest in the basic science that would support and help define some of the issues that are quite important to clinical trials that would ensue in trauma and treatment and the development of new treatments.
We also, of course, have traditionally supported large clinical trials, but in this time of austerity the issue of leveraging and leveraging resources is extremely important. And I just wanted to remind everyone that a lot of what's been discussed here with respect to endpoints is extremely -- they are all extremely important whether we are talking about clinical endpoints, surrogate endpoints, biostatistical models for clinical trial design and there are mechanisms to actually look at this from a research perspective. The R34 mechanism, which is a pilot trauma mechanism, does allow investigators to come together and to actually plan trials and to hone inclusion criteria, study design, statistical methods, and everything from clinical endpoints to surrogate endpoints that would allow for very cost effective clinical trials.
And the last thing I just wanted to say is that, you know, certainly the NIH has long supported SBIR mechanisms which are very complimentary to maybe some of the mechanisms in DOD, the DARPA mechanisms et cetera for developing preclinical analytical models, assays, biomarkers and point-of-care devices that might also facilitate this. Thank you.
DR. WEISKOPF: Thank you for your comments, Donna. But I'm not sure we adequately addressed Dr. Gull's comments earlier. I know you aren't particularly -- that does not -- was not your point, and I'm just following up on his comments again and that whereas most studies, most large-scale studies or studies for registration are conducted in a relatively controlled environment, in fact so controlled that many people say they are almost inappropriate to look at how the drug or pharmaceutical we use when it gets into practice should it get into practice. That is not the situation in trauma and leads to, as Dr. Gull indicated some severe protocol violations including not only patients getting the wrong -- being part of the wrong arm of the trial rather than the trial, the arm to which they had been randomized. And I'm not sure that the statistical world has been appropriately addressing the issues of statistics as it is in an everyday world in the trauma situation and is reflected in trauma trials.
DR. DIMICHELE: That maybe I think -- but I think that the issue of and you know, I'm not going to speak to that. I think that the issue though is that, you know, 75 percent of what we support or investigate or initiate a trial and therefore investigators coming together, putting their heads together and looking at these problems actually will have mechanisms to do the basic and potentially early clinical research that might then lead to the pre-licensure clinical trials --
DR. WEISKOPF: Sure.
DR. DIMICHELE: You know that the FDA will mandate, just something to think about.
DR. WEISKOPF: Yeah, but I -- I can't help but think that when a sponsor looks at the publication from Dr. Gull's study that it just presents another potentially daunting hurdle to conduct trials in this very chaotic environment.
MR. MANN: That has been made very clear. Dr. Kheirabadi.
DR. KHEIRABADI: Yes, Kheirabadi from San Antonio. I had just a quick question. If we are accepting the first 24 hours mortality as one of the primary endpoint, should that be differentiated for the death due to uncontrolled hemorrhage or overall death? And I'm -- that's all.
DR. WEISKOPF: Okay. See you are asking whether it ought to be specific cause mortality versus all-cause mortality. There is a substantial debate going on about that and the panel did not address that directly. So I don't want to speak for the panel about that, but you raise an important issue.
MR. MANN: Thank you very much.
MR. MANN: Now at this point, we are going to insert Dr. Holcomb to discuss the session five that you heard this morning, Clinical Trial Options for Severe Bleeding Indications. And then we will go on to Dr. Biros and the ethics discussion.
DR. HOLCOMB: All right. Dr. Mann, thank you very much. The -- so we put these slides together shortly after our discussion this morning. And I just kind of went through the bullet points and summarized the comments from the two talks and also from the panel.
The trials, trauma trials are feasible; we've talked about this conundrum that maybe boxing ourselves into a corner a little bit where the alternative pathophysiology of trauma patients versus elective surgical patients. We do think that use of existing blood products as a baseline before investigating new products kind of a CBER (phonetic) approach, before launching off into new products is a good idea.
The predictive models using either machine learning or just simple scores are here. They are better than physician gestalt; we should use those to enroll patients. We should consider the use of thromboelastography in the ER or some of the new devices are coming down the pike, even pre-hospital, to help us sort out which patients are coagulopathic and which part of that coagulation profile is abnormal therefore, specifically treat that segment rather than come throw in the kitchen sink of the patients, which is what we are doing now.
I think there is a lot more data that needs to be collected there. Large primate studies with the caveats that Charlie Wade discussed may help new data analysis for some adjustment techniques. We just heard about that comment. And it really is applying the, you know, we have ivory tower statistics to the chaos of trauma patients at 2:00 o'clock in the morning, and how to balance those two out I think are very important or will generate a high quality data within which we can make a logical decision so we can help patients.
Obviously timing, everybody said that, is extraordinarily important. Proof-of-concept studies in high volume and high quality -- Gene More's comment about high quality centers is extremely important. Again, I will go back -- the reason I think there is great variability in trauma centers is because our data supporting what we do are incredibly poor; thus the need and really imperative to improve the quality of our data.
Multi-center studies collaborative, multi-agency high volume extremely important concepts along with the developing networks much like it had been done with cancer. I think cancer is a great model to replicate. Those trauma centers do exist in the United States; the systems of care both we are looking at quality of centers and then distribution of guidelines and research data throughout ATLS et cetera do exist already; we just need to generate the quality data.
Then, of course, in funding there is no lack of initiative and drive and inventiveness and desire and when frankly the regulations don't need to be changed, funding really can drive all these as we have discussed already.
The elective surgery question, I think is fascinating, it's a fascinating question. Everybody, everybody wants to be able to use elective surgery to move this field forward. It would be extremely pleasant to be able to do that. It's just the elective surgery only brings in some unique aspects that may not reflect what goes in trauma patients. How to generate that pathway? (Inaudible) talking about in the back of the room. Somebody needs to put down on paper a pathway.
And then, so right now, we kind of have this blank page, needs to put down a pathway so that we can do constructive criticism of that pathway and really see what the data say. Just because we want something doesn't mean it's true; just because we want to use elective surgery doesn't mean it's right. And just because we have a bias that there is something different about trauma patients doesn't mean it's right that we shouldn't use elective surgery. Somebody is got to come up with a pathway.
There are cases out there that are vastly different than the trauma patient for all the reasons we have talked about. The intrinsic pathophysiology and then also the anesthesiologist at the start of an elective case keeping that patient relatively physiologically normal for 95 percent of the case is extremely important.
Trauma obviously is a continuum. There are great physiologic differences between patients that really affect the outcome here. Massive bleeding is not really seen often in these elective surgery patients and the type of bleeding is very different. In multilevel spine surgery, unless you get into one of the large vessels in a place you're not supposed to be, doesn't bleed rapidly -- but it bleeds a lot, and it bleeds consistently over a long operation. There is not really arterial (inaudible) in that kind of case.
So we need to take in the type of bleeding which was discussed by Grant, early as well. Mortality very low in these elective surgery patients and very low, even in the biggest operations, exceedingly low. And so mortality won't be an endpoint. You'll have to use units transfused or estimated blood loss as an endpoint and those obviously have the issues discussed earlier.
The -- I think do think elective surgery can be informative. But it really needs to be considered very carefully on how informative it would be. The military and civilian setting, we've discussed this before -- I absolutely feel like the results taken both from the military setting and what we have found over there and from the civilian setting, the results can go both ways.
There are some differences in wounding agents; there is some difference in wounding patterns. And Dr. Mattox brought up extremity wounds versus truncal wounds, but that basic physiology is remarkably consistent.
There is a -- it was a very good discussion of a biomarker tissue injury severity score or scale and the issues within -- you know, from a meeting like this where we are trying to figure out how to move forward, you hate to have one of the conclusions is to develop a new biomarker or a new test which, of course, has its own research pathway as well. It's something that people have talked about for a long time but nobody has done.
And then the ethical issues, I think we will have a whole separate panel on this. It is -- so I am not going to specifically address those other than saying that I think one of the major opportunities is we kind of lead into the next panel is this potential for phase II studies with 50.24(a). I strongly believe that one of the reasons other studies have failed or have been inconclusive is we didn't have the opportunity to enter into phase II studies in the population of interest venturous with community consultation, so that we could learn the operational issues that we learned with PolyHeme and how to -- you know, how best to analyze the data, what's the right population, what's the right endpoint, that's the way other products are developed.
We need to do the same thing in trauma and not go from -- right from a pig into a phase III definitive trial, 50.24. I also think the observational studies have been extremely important. That's -- it's popping more and more in the literature and those observation studies in lieu of prospective randomized studies, so really allowing you to get high quality data and help design the phase III trial.
Again going to that cancer model, both with the philosophically the endpoints with the adapted trials with the Bayesian analysis in the cancer area, I think, there's lots of lessons to be learned there in getting away from a 0.5 - 28 day mortality endpoint is extremely important concept. And that's where we are bringing in the experience of biostatistics epidimeology support into this trauma milieu is extremely important.
That the funding issues again can't just be DOD it's a societal imperative; it needs to be addressed in relation to its impact on society. And the 50.24 a lot of discussion more and more and more, and I do think the regulation actually is pretty good. I think if you'd read those words carefully, the regulation was well constructed. And how those regulation is interpreted needs to reflect this changing epidemiology and changing knowledge of the animal models out there.
DR. SCHREIBER: Martin Schreiber from Portland, and John a very nice summary. I think it's important that we take a moment and reflect on those trauma studies that we've done and we've talked about a little bit. But I want to talk a little bit about the ROC studies. Both the hypertonic saline study for head injury and for shock were stopped early for futility. And I think there is a little bit of a problem with that because one of the issues is okay, maybe the hypertonic saline is not superior to normal saline. However, if it's equivalent that's still important because being able to infuse a 250 cc bag of fluid and have that being equivalent to infusing one or two liters of normal saline is important in some settings particularly the military setting and it's also actually important in the field.
If you can get by with a small amount of fluid then I think that there is certain logistical advantages. So I think it was an error actually to -- in midstream change that study to a superiority study as opposed to a non-inferiority study. I don't think we have talked about very much. And I think it's worth some comment. Should these studies, you know, should we look at more non-inferiority studies? I understand it takes more patients from a statistical standpoint, but I think it would have been important to finish that study and show that the hypertonic saline was equivalent to normal saline.
DR. HOLCOMB: I would just reiterate Dr. Schreiber's comment that non-inferiority design in the initial ROC studies that was started. The non-inferiority design was placed into that study design on the request of military. You know, if you are a pre-hospital medic carrying a 250 cc bag versus two or three liters of saline, it's a very important difference especially between multiple casualties, non inferiority would have been an important outcome of that study.
MR. MANN: Dr. Golding.
DR. GOLDING: So, you know, in considering the 50.24 regulation and in the prospect of benefit it goes on to talk about a risk-benefit calculation. And when we look at it and I think most people who are experts in the field would say that you have to consider safety very heavily in terms of the risk benefit. I don't think we have spoken a lot about safety at this meeting. But the one point I'd like to make about experience with elective surgery and HPOX (phonetic) is that we did learn a lot from elective surgery studies regarding safety. And that went into considerations of the risk/benefits for allowing studies to go ahead under 50.24s. So at a minimum you can learn a lot about safety, I think, in elective surgery studies before you can get into a 50.24 study and that is very informative.
DR. HOLCOMB: Dr. Golding, I think I agree with you. There are things to be learned and, you know, it's -- you are right, so there's the safety part, there is the efficacy part in bringing in the phase II and then the phase III. So we kind of jumped around that question quite a bit, obviously, probably it's worth its own separate conference.
MR. MANN: Dr. Weiskopf.
DR. WEISKOPF: Richard Weiskopf; I would like to follow up on Dr. Schreiber's comments. And I think it's very -- an issue that is not perhaps paid sufficient attention to and that is when the investigative team or the sponsor whoever may be involved in the trial appoints a DSMB, the DSMB obviously has several critical roles. And many of us have been involved in trials either from appointing DSMBs, or being parts of DSMBs, or being investigators where trials may have been stopped early for reasons that are perhaps not what many people would have agreed to stopping a trial. And in part I have seen this done when members of DCMB may not be sufficiently versed in the issues regarding the trial./
And so it's just a word of caution or advice, if you will, that those responsible for appointing DSMBs perhaps pay a great deal more attention to that issue than that I have seen done in the past that these are -- has a potential for breaking a trial when it ought not be broken.
MR. MANN: Dr. Hoots.
DR. HOOTS: Keith Hoots, NHLBI. I want to comment on the phase II. I'm no statistician, but I read -- there was an article in Blood, I think it was the last of October this year, well they looked at the NCI data on phase IIs and forming phase IIIs. And I agree with your statement based on at least what they found with cancer trials which was first, as we said earlier during the panel discussion, the problematic use of historical controls even in cancer where the variability in terms of progress may not have been as profound as here, but also the fact that phase IIs protect potentially good products if they are randomized against false positive expectations, but reducing the likelihood of false positivity.
And I think that's a really important thing when we are talking about very expensive development to market products like you have in trauma or in cancers. So I just wanted to reemphasize that based on I think some really interesting and good data.
DR. HOLCOMB: I just, you know, from a financial, logistical study design, how to do the randomization Steve, you know, should both products be on the reg at the same time or will you learn that from a phase II versus the phase III. And then, you know, these products you are exposing patients to these projects. The phase II may be a smaller study than the phase III maybe expose fewer patients to wherever this product is even it's a CER study in a phase II design versus a phase III in 100 patients or 150, 200 patients, you see absolutely no difference, then you wouldn't go to a phase III.
DR. HOOTS: Right.
DR. HOLCOMB: It's -- I think it's extremely interesting concept that needs to be explored. Maybe a difficult pathway, but it should be explored.
MR. MANN: Dr. Gold.
DR. GOLD: Steve Gold from Chicago again. I have a question and a comment. I want to pursue Martin Schreiber's comment. This is a formal question for the regulators and statistician. I don't know if you can do a non inferiority trial under 50.24. If you set out with the premise that my test therapy is non-inferior to the control, how does that offer the potential for benefit? MR. MANN: Sara.
SPEAKER: You may be preempting the next presentation there.
DR. GOLDKIND: So we have actually approved 50.24 studies with non-inferiority design. And in those cases, it's -- there are other factors not specifically let's say efficacy that we are concerned about, perhaps administration of the product in the field, the product that's -- that they are testing against is one that's approved and indicated for that condition but it might be an IV product and we want to see if we can ascertain, we can approve a product that has an IM administration, or it might be a product that has significant safety concerns. So there are opportunities for doing ethically and scientifically designing non-inferiority trials under 50.24.
DR. HOLCOMB: Well, I'm glad to hear that. And I'm in favor of doing that. But to be consistent with the reg -- the safety I understand, Sara, but a different form of infusion, IV or IM trying to find out if a smaller volume is as effective, raises the question are we ever going to do that in the civilian setting for use in the military population. So there are a whole host of other questions. I'm not trying to attack it. I just want to make sure that we are being consistent with the reg. They are important questions.
SPEAKER: Steve didn't say, okay, so you need to say okay at this point.
Mr. GULL: My comment, let's go back to the phase II and the elective surgery, John, because I think there is a theme evolving here, which is in phase IIs and in elective surgery, we should be able to get information about the effectiveness of these products. Now, we go into the phase III in trauma we need to show in a pivotal trial safety and efficacy.
And the other question that has been put forward in different ways in the last couple of days is there any mechanism to have a large scale trial, a pivotal trial where you have the established efficacy, but really looking at safety. I think the answer is no, but again the regulators may want to comment on that. That may be one way forward to avoid missing the ability to approving a product. I don't know if I have expressed it clearly or not.
DR. HOLCOMB: Yeah. I think the answer to your question is a little complicated and convoluted but there seems to be a way to mix and match, all right, elective surgery and trauma. The trauma -- if I was thinking out loud about a phase II trauma study. With 50.24, you would want to learn lots of things. Should we put both products on the regs or should we do alternate day product on the reg because of the unique variability inherent in 5,000 medics across the country doing the randomization and treating at the same time?
That's probably an important variable to sort out in a phase II study. That would be, you know, so you would learn that variable and then many more variables besides the endpoints and patient outcome to bring into the phase III. Right, I just think there's so many unique variables to these trauma studies that go beyond the physiologic endpoint of mortality or blood loss at time of hemostasis that we ought to learn those in the iterative phase II studies before going to phase III. And then potentially design it so that your IIb rolls into phase III, you know, where the shortfalls of the -- talking to the FDA and the DSMB.
MR. MANN: Dr. Lawson.
DR. LAWSON: I Just would like to follow up on that because I think that's an emerging theme about, you know, instead of trying to do a grand slam clinical trial in an incredibly complex environment to take pieces of insight you can gain from different environments and be able to sum them --
DR. HOLCOMB: Okay.
DR. LAWSON: Will it be the efficacy in a well designed, thoughtful animal experiment? I mean there's nothing wrong with that if you can demonstrate your product has a biologic response and you can correlate that with a controlled surgical environment or you can make the injury and have your product respond to that. So you have been able to demonstrate the biologic utility of this and then maybe in an independent mechanism demonstrate safety in another clinical environment and to be able them to piece them together instead of trying to do a 10,000 patient trauma trial in the middle of the night where you have so much heterogeneity you may have an effective product but you just lose it in the noise.
DR. HOLCOMB: I think those examples of trying to go for that home run fairly recently didn't work. That's because of --
MR. MANN: Mike.
MR. FITZPATRICK: Mike Fitzpatrick (phonetic), in following up on that that the biggest issue, I think, in making the leap from the clinical to the trauma is the thing that the panel talked about, anesthesia. So removing from a controlled situation, under very careful monitoring, looking at efficacy, looking at safety then looking at efficacy. Now we leap into trauma thinking the product is effective and safe at the doses that we think we should use. So in a totally different patient population so you really have to start over, in my mind, at escalating doses looking at safety in that population before you can even try to look at efficacy in the phase II.
MR. MANN: Thank you, John.
DR. HOLCOMB: Thank you.
We have already started a bit of a ethics discussion, Ethical Considerations in Trauma, Michelle Biros.
ETHICAL CONSIDERATIONS OF CLINICAL TRIALS WITH TRAUMA
DR. BIROS: Thank you. Again, I wanted to thank the organizers for this outstanding forum, and I've been privileged to be part of it.
Our session yesterday began with presentation by Sara Goldkind, who was able to refresh our memories about the various aspects of 50.24. Especially relevant, I believe, were some of the themes that came out not only in her presentation but also in subsequent conversations that we had.
Well, many of us have considered this to be a community consultation requirement; it's actually an exception from consent requirement and we have to remember that community consultation is just one of the many patient safeguards that need to be fulfilled in order to apply this particular set of regulations appropriately.
Sara emphasized that when the FDA evaluates these studies community consultation is part of it, but many of the studies that fail to go forward are basically forwarded by the study design and so we need to pay particular attention to make sure the study design we propose is indeed consistent with an exception from informed consent study.
Our panel was given -- we had a very interesting panel. It included individuals who have done exception from informed consent studies, individuals who have reviewed such studies. And we also had a community member who was personally impacted by an exception from an informed consent study.
And I think that individual gave us a perspective to remind us how serious this is when we do studies in which a patient is unable to speak on their own behalf and also demonstrated how perception of disrespect can occur if these regulations aren't followed to the letter. So we were fortunate to have that particular perspective which I think is unique and again gave us a lot of food for thought.
Our panel discussed four questions that were tasked to us. And the first had to do with what happens in the pre-hospital situation when a paramedic or a field emergency physician or a trauma surgeon once has a patient who is eligible for a study, but the patient is unable to speak on their own behalf to agree to enrollment. So what is it that we can communicate to individuals present at the scene?
Well, there was quite a bit of discussion about this, and the comments ranged from nothing to more than that. And again it was very circumstantial. It depends quite a bit on what exactly is happening, what the circumstances are in terms of the physical environment, the safety element for the medics in the field, the chaos of the setting, the ability to identify who should be directly spoken to. But what we do think should be possible again, if it's feasible in terms of a therapeutic window, it would be possible and perhaps desirable to give the family notice of the fact that an enrolment is about to occur and provide them the opportunity to object as soon as possible. This may not be at the field but it may be a little bit later.
We have to recognize that this would not be informed decision making because this is a circumstance that doesn't provide the opportunity to sit down and actually educate people about the risks and benefits of the trial. There just isn't time. What we do have to remember though is this is a complex situation when an individual is enrolled in a project or an experiment in which they have no say. So we have to try to respect the feelings that the family would have and as demonstrated by one of our panelists these can cause serious after-effects that are not necessarily the most reasonable or desirable.
We also talked a little bit about clinical trials in the military setting in which exception from informed consent might be considered. Now, this is an interesting topic and certainly relevant especially in the context of this workshop on severe bleeding and trauma and other causes. But when you think about the way the military is structured there is a rank-based structure that creates a power differential and that may make it very difficult if not impossible to complete some of the requirements for the 50.24 regulations. For instance, it would be difficult to have a two-way communication as was required by community consultation if a commanding officer is talking to his troop. Chances are they wouldn't necessarily voice contrary opinions.
So this type of circumstance might raise concerns related to undue coercion or undue influence. There is also logistical issues related to doing any research in the combat setting, but in particular a circumstance in which exception from an informed consent would be employed. There is danger to study personnel. There is quite a bit of lack of personnel and the scarce resources in these circumstances. It would be difficult in these circumstances to conduct a research to the level of scientific integrity that we would expect in more controlled environments and that would provide us with the most appropriate data that could be generalized in other circumstances.
Another component of the exception from informed consent requirements has to do with public notification. Now this happens two times. Once is prior to the initiation of the study and public notification is a one-way communication in which we let the community in which the study is about to proceed know that the study is going to happen.
Now after the study, it is our responsibility to engage in a second type of public notification and that includes telling the community in which the study has gone forward, what the results of that study was. We also notify our scientific and academic communities with our publications that come out after the study is completed. But when it comes to public notification to the community that has been somewhat of a difficult thing to grasp and to figure out the best method by which it should be done. So some options that were discussed in our panel it is distinct to remember from community consultation, remember community consultation occurs before the study and is a two-way communication that allows communities to voice their opinion about a study. We call that community consultation. It's not community consent. It's not a community veto, but it provides IRBs with information about what the community is feeling about that particular study. So public notification is distinct. It's one way.
Methods that could be used to provide public notification regarding the rolling out of a study would be to talk about it to local governments. We can provide reaching out services to community individuals that are of importance in a particular community. We can talk to city governments. We can publicize in their methodologies at their public meetings. We can also use the big four, TV, radio, Internet, and newspaper to try to reach out to the community again recognizing that it is unlikely that we will reach and penetrate the entire community either with public notification or with community consultation. Already active proceeding in this fashion provides the community with a feeling that we are respecting their opinion and we're respecting them as potential subjects for our research project. We also need to understand that any attempt at community consultation and public notification is expensive in terms of not only time but also money and that needs to be factored in when you budget for your particular study.
There are other ways to reach out to the community to let them -- inform them of studies. One suggestion was an informational letter to specifically to study participants after the study has completed in which you relate to them how the study proceeded, and what the results were. Other ways are available now because we have an expanding social network of communications, including Twitter. There are a number of study specific websites that provide information for interested individuals in the community and as well as the website, clinicaltrials.gov. It's important to make sure our community becomes aware of these options so that if they seek to follow up and decide that they want to find out what has happened they know that such outlets exist for them to obtain information.
So again then we specifically talked a little bit about community consultation and we acknowledge and the literature has acknowledged that community consultation has been very difficult to achieve, but perhaps even more so it's difficult to measure its success, because we have no specific metrics to know whether or not things are proceeding in the fashion. The process of community consultation is not always described in detail when the results of exception from informed consent projects are reported. So we encourage sub-studies within these larger clinical trials that look at community consultation, the process itself. Whether or not IRBs should request changes to protocols after community consultation, whether or not IRBs restrict the enrolment of patients from one reason or another or whether or not they actually decline to allow such a study to go forward. So the large clinical trial should consider including sub-studies about community consultation.
Community consultation is a two-way communication and just as a side bar, I think most investigators think that this is at the hardest part of the ethic process. After sitting through this conference for the last 2 days and hearing about not having outcome measures and not having the model and being an emergency physician and never having seen a patient beyond 28 minutes let alone 28 days, I think that this ethic stuff is a snap compared to some of the issues that you're dealing with.
What I would recommend though is that because this can add a layer of complexity to a trial that you intend to do some thought leaders recommend that in addition to a principal investigator on a project sponsors should consider adding a human subjects' investigator to that project as well. Principal investigators have an awful lot to do with maintaining the scientific integrity of the trial and it's difficult to add to that role, that requirement to know and to appropriately execute and track and survey all of the ethical requirements for the exception from informed consent as well.
So to communicate consultation two-way communication between the communities that may -- in which the study will be done as well as communities that represent potential patient target population. Some methods have been discussed. We talked about random digit dialing as a method. The use of existing meetings and focus groups, methods of community consultation should be considered and dictated by local IRB requirements. In most of these ethic studies the investigators would do well to establish an ethic plan describing how they're going to meet the regulations of this, the 50.24 and then discuss this ahead of time with their IRB to make sure that the plan that they intend to execute is appropriate.
Finally, we talked a little bit about balancing rigorous scientific design and the ability to obtain informed consent and trial design options that might be possible. This has to be study specific and has to be examined on a case by case basis to support the proof of concept as well as the prospect of direct benefit when it comes to deciding whether or not an ethic trial is necessary to solve a particular clinical problem.
When we are about to propose an exception from informed consent trial to the FDA, we need to provide some data that looks at previous studies either any preclinical trials or applicable clinical trials in other consenting populations that might indicate the safety of the agent we're intending to investigate as well as the potential efficacy in the proposed study population. We recognize that animal studies don't replicate results in clinical environment. Elective surgery doesn't really replicate conditions of trauma and that every single trauma patient has perhaps a unique pathophysiology. So it is a difficult transition from animal elective surgery to phase II -- I'm sorry -- phase III trials under this regulation.
Finally, we talked somewhat about data retention issues. This was brought up by a question from the audience. The scenario that was described was a patient treated in a single dose or short-term exception treatment study later chooses to withdraw, how do you use the data. The FDA regulations require that data collected in a clinical trial be retained in this study database because not using this data could introduce bias. It might affect the safety and efficacy results of the study. If it appears that there were ethical issues in obtaining the data, then the FDA would consider the patients or LAR's request to withdraw that data from the database.
When a family or subject decides to withdraw from the research, we would encourage investigators to indicate to the family or to the survivor specifically why it would be important to continue to collect certain safety data regarding the investigational agent that this patient has received. If the patient or their legally authorized representative consents to allow continued collection of data but does no longer wants to receive the intervention then it is reasonable to go ahead and continue to collect that safety data. But at any point if a subject withdraws and they do not want additional data collected then it is our obligation to honor that request. And those are basically the talk points that came up in our discussion. Thank you.
MR. MANN: Thank you.
MR. SHIRE: Martin Schreiber from Portland. I've a question specifically about 50.24. We saw in two slides yesterday, one of the requirements for being able to do this type of study is that the intervention, specifically the slide says the intervention has to provide the potential for benefit to the subjects. We have seen in multiple trauma studies with the Glue Gran, Recombinant Factor VIIa, hypertonic salines.
In these trauma studies the subjects are committed to evidence-based guidelines on how they're to be managed. Nutrition, head of bed elevated, VAT guidelines, et cetera, et cetera. And what we've seen in these studies repeatedly is that by participating in the study, the subjects have a lower-than-predicted mortality. My question is, is that evidence of potential for benefit, if by participating in the study whether you're in the control group or in the intervention group, you're going to be committed to these standards of care which improve your survival. Does that meet the requirement for that specific --
DR. BIROS: My impression is no. But I'm going to let Sara expand upon this.
DR. GOLDKIND: So you're correct in stating that under 50.24 when we look at prospect of direct benefit it is specifically in relation to the intervention that is being studied. The study intervention. And also on that same slide that I -- as included in your packet it talks about the prospect of direct benefit and Dr. Golding alluded to this in his earlier remarks that we are not in our consideration of prospect of direct benefit not only are we using the evidence base that we can glean from a variety of different sources but also we're thinking about the risk benefit analysis of that disease being untreated and whatever is available recognizing that available treatments are unproven or unsatisfactory for this condition when we do that prospect of direct benefit analysis. And that is to protect people from being enrolled in trials in which they may not be getting the best, you know, they may have -- there would be some harm to them from being enrolled in the trial.
What we would expect to see as a general, overall matter both for ethical and scientific reasons is that they get the best care they can possibly get if these guidelines are appropriate and they should be available to guide subject treatment in both the control and the experimental arm and that is part of the design considerations for the study. But when thinking about prospect of direct benefit under 50.24 it is focused on the intervention that's being studied.
MR. SHIRE: Thank you.
DR. GOLDKIND: So I hope that helps.
MR. MANN: I guess if there are no other questions then this part of the session is closed.
DR. BIROS: Thank you.
CONCLUSION AND ADJOURNMENT
MR. MANN: And before I turn the chair over to Dr. Vostal, I would like to first congratulate the federal government for its co-participation in this endeavor. Clearly, trauma is a major concern and one that transcends many different organizations within the government and so this co-participation by FDA, the Army, and NHLBI, et cetera, are absolutely essential to pursuing this pathway. And in many respects in this complex area this is a model for how intergovernmental cooperation and research I think should occur. If anybody remains -- remembers the fable of Stone Soup from their grammar school days, this is sort of an illustration of everybody bringing their resources to the pot to effectively maybe make a contribution that is in the national interest. And so I would like to congratulate Dr. Epstein, Dr. Hoots, Dr. Hack, and Dr. Golding for bringing their resources of their agencies together in this enterprise. And I think this meeting together with the meeting that occurred in April basically are tremendous illustrations of intergovernmental cooperation. I would also like to thank once again Dr. Vostal, Dr. Lindsey and their staff for their diligence in organizing this. I think a turn to heat up would be good, but --
MR. MANN: -- apart from that, I think everything has run very smoothly and thank you.
DR. VOSTAL: Thank you, Dr. Mann, and thank you all the speakers for their wonderful presentations. So that brings us to closing. But you may think you're done, but since this is a government-run operation there is always paperwork to be filled out. And if you'll notice in your packages, there is an evaluation form. It would be greatly appreciated if you could fill that out. In case you're wondering whether we thought that the workshop was a success, our objective was to bring out a lively discussion and a lot of opinions on specific topics and we're able to do that, especially with the panel format.
We also then wanted to be able to distil some of that information into a manageable format, and we really appreciate the panel chairs were able to do that in a very short period of time. I would also like to thank the steering committee -- if I could bring this slide up -- the steering committee worked diligently for a long time. It took about a year to plan this workshop and their input is greatly appreciated. As the workshop got closer and closer, the administrative staff and logistical support took over and did the heavy lifting and we really appreciate that. And finally, again, we'd like to thank for the financial support from multiple sources which speaks to the collaborative effort in this area. So and that concludes this workshop and thank you very much and have a safe trip home.
(Whereupon, the PROCEEDINGS were adjourned.)