Clinical Studies Targeting Stroke

The potential benefits of therapeutic hypothermia have been tested in many experimental models of cerebral ischemia as well as in patients with stroke. Although much has been learned regarding the potential beneficial effects of cooling and targeted temperature management, ongoing clinical investigation studies continue to test whether therapeutic hypothermia is beneficial in stroke patients. A series of state-of-the-art lectures presented at the 2014 Therapeutic Hypothermia and Temperature Management meeting in Miami brought together experts in the field of therapeutic hypothermia targeting cerebral ischemia and stroke. Dr. Fred Rincon, Thomas Jefferson University Hospital, spoke on the safety and feasibility of targeted temperature management after intracranial hemorrhage. Dr. Rincon emphasized that mortality remains high in this patient population and many clinical trials have been negative. More research is required to improve outcomes after intracerebral hemorrhage, and current results utilizing targeted temperature management will provide information on the applicability of temperature management strategies in this patient population. Dr. Dimitre Staykov, Department of Neurology at the University of Erlangen-Nuremberg, provided the attendees an update of the European multicenter, randomized, phase III clinical trial of therapeutic hypothermia plus best medical treatment in patients with acute ischemic stroke. The proposed trial will include blinded outcome assessments of 1500 patients, with individuals being randomized with hypothermia. This trial continues to progress satisfactorily, and the research community awaits further information. Dr. Patrick Lyden, Department of Neurology at Cedars-Sinai Medical Center, updated the group on the phase II/III study on intravenous thrombolysis and hypothermia for acute treatment of ischemic stroke. The rationale for conducting these clinical studies and current antishivering protocols were also reviewed. To date, 95 subjects have been enrolled, with 16 sites actively recruiting subjects. These discussions resulted in a very informative question-and-answer session that touched upon many aspects important to individuals involved with treating these patients.

Question: I have a question for both of the stroke studies. You've showed and we've seen elegantly how the pathophysiology of injury is not a yielding model. Something that I think is missing from that picture is that it is also not static in time. So you have certain injuries that occur during acute ischemia, and then you have a separate concurrent set of injuries that occur with reperfusion. Yet, the studies that have been designed don't seem to be testing those two hypotheses. Because you are including people who do reperfuse, which is only about 25% of the tissue plasminogen activator (TPA) patients, and you are including patients that don't reperfuse. My concern is that you are going to dilute the effect of mitigating reperfusion injury versus neuroprotection. To my best understanding, if it is not done preischemia or extremely early postischemia, hypothermia is probably not neuroprotective so far as I can understand that. And that it's more of a reperfusion protection, which is what we are seeing in cardiac arrest, in TBI, and now with focal stroke. So that's my first question.

Dr. Patrick Lyden: Let's do the first one first. In the next session, we are about to get an earful on exactly that point. That cooling after reperfusion may make a lot less sense than cooling before reperfusion. So I have two answers to this question. The first one is to tell the anecdote of our TPA for stroke study. So when we were designing the National Institute of Neurological Disorders and Stroke (NINDS) trial of TPA for stroke, we were told by the experts to exclude lacunes from the trial design. Everybody knows that lacunes will not respond to a thrombolytic because the mechanism of lacunes is known. Miller Fisher told us the mechanism. There is no data, but he told us the mechanism. The mechanism is lipohylanosis and cystic medial necrosis. It's not thrombosis. Therefore, you should exclude lacunes from the NINDS TPA for stroke trial. Now, as you know, that trial was positive. It's still the only positive trial in stroke, and the subgroup of lacunes was the most responsive. They responded the best. So we now understand that in fact small vessel strokes are thrombotic, and you can treat them with thrombolytic therapy. So I'm very hesitant to design a trial too much based on what we think we know. And I'm inclined to design the trial a little bit more broadly to allow for the fact that we may not know everything. There may be things we don't know. We have to have enough patients in the study to explore and learn from them. Secondly, in the animal models, which are mostly done in young, healthy males, we don't really know that this applies to our stroke patients; there are data that there is a benefit if the artery is not reopened. And there is benefit even if the cooling is started after recanalization. It's just not as good as if you start it before recanalization. So the ideal study would be, you get vascular imaging, you demonstrate recanalization, and you start cooling prior to that recanalization. It's very time-consuming and very technically complex, and our studies are testing the hypothesis that hypothermia in a large group of all comers with post hoc vascular imaging will be effective. And we may very well miss it based on the argument that you make.

Dr. Dimitre Staykov: Well, I think there is also the point of practicability. You know, you have little time in such an acute disease, and you have to decide what you want to do first. And I think it's a good approach to start hypothermia before you have all the diagnostics.

Question: But you could design an event-driven trial where you keep enrolling until you have enough patients who reperfuse to explore your primary end point. And then you have the nonreperfusers to explore your secondary, rather than picking and assuming that you know what the benefit is. You keep making it event driven.

Dr. Patrick Lyden: That's a very valid alternative design.

Question: The second question comes regarding what we've heard earlier about the potential risks of high-volume cold saline in the cardiac patients. We know that, in the subarachnoid hemorrhage patients, there is very elegant work done by Derringer and also by Laroux in focal ischemia for vasospasm, which I know is not the same thing, but it is focal ischemia. That high-volume fluid infusions, which is why we've dropped the H and the triple H of hypervolemia, actually resolve in decreased PPTO₂ higher lactate pyruvate ratios and more ischemic injury to the brain because you've hemodiluted the oxygen-carrying capacity. So, a concern I have with preloading your patients with 2 liters of saline who are already ischemic is, could you potentially be causing a greater amount of ischemia—and these are human studies of cerebral physiology when you give high-volume fluids in focal ischemia—that may actually hurt some of your patients.

Dr. Patrick Lyden: So we debated this extensively. And in ICTUS L, there was no saline load at all. But we were not happy about our time to target, which was in excess of 6 hours; the median time to target was in excess of 6 hours from stroke onset. So we put in the iced saline load hoping to get to target sooner. As I mentioned yesterday, and I can't mention the numbers, I can only say the result and then we'll publish this shortly. Looking at the first 50 cooled patients in ICTUS 2 compared to the 40 cooled patients in ICTUS L, with and without a bolus, the two populations are comparable. Same catheter, same protocol. Mostly, the same study sites. The only difference was the 2-liter saline load. We had no increased risk of harm. There were no complications that we could identify. Our DSMB tells us this. I haven't looked at this part of the data. But I do know that we are not getting to target 1 minute sooner. So by doing a 2-liter saline bolus, we are not getting us to target cooler. We have an extremely powerful catheter. Most of our patients are at target within 20 minutes of getting their catheter. So it may be that the catheter is so powerful it doesn't make any difference, but we would like to eliminate the 2-liter bolus for some of the reasons that you are suggesting, but in a small amount of patients so far, we haven't found any harm. What do the Europeans feel?

Dr. Dimitre Staykov: Well, we have actually no other data except the data Dr. Rainer Kolmar provides with his 10 patients in ischemic stroke. But comparing this to the triple-H treatment, which is a multiple dose of this, what we are giving, and considering the fact that those are patients who come from home, they are not so fluid-overloaded as ICU patients who have been there already a couple of days. It is maybe a minor effect as compared to a triple-H treatment. So I think that it is possibly safe as the preliminary data show.

Question: So, gentlemen, I appreciate the science I have been hearing; it is kind of mind-exploding. But I go home tonight, go work in the ED over the weekend, and I get the 48-year-old male, 58-year-old female who come in with an acute ischemic stroke. What can I take home without waiting for God knows how long until you guys get these studies done? Is there anything I can take home from all of this to use now, not in 2 or 3 years?

Dr. Patrick Lyden: Fred, what do you think about that?

Dr. Fred Rincon: TPA.

Question: OK, we've got TPA. Keep going.

Dr. Fred Rincon: So beyond TPA, admission to a stroke neurointensive care unit requires good assessment of vital signs, hemodynamic profile, sugar control, and statin therapy. Consider antiplatelet therapy for stroke prevention. And the normothermia part, it depends if you are in an ICU that is aggressive at treating fever; I would suggest to do it as long as the patient has clinical deterioration in the setting of fever. Beyond that, I think you are doing what the guidelines tell you to do.

Question: So I agree with the guideline concept that you throw out. But you guys are talking about all this neuroscience that says maybe we need to be doing something, especially for the penumbral area, that we can potentially salvage. I think that we need, just as somebody here said, fever control. I think that is the one message that is not getting out. Because we are sort of letting patients then wander after we do their thrombolysis or whatever. And even if they recanalize, we're letting them wander into kind of an abyss of where you are going to go is where you are going to go.

Dr. Patrick Lyden: So I think this question raises a couple of points, because some of us are going to go home and have clinical trials in which we want to enroll. And some of us are going to encounter neurointerventionalists; they know how to open the artery. They know that recanalization is going to work. And so my colleagues and I published this hierarchy to help all of you make decisions when you get home from this meeting. And we believe that the most ethical thing you can do for your patient is to offer standard therapy. That would be therapy that has been proven in a randomized controlled clinical trial to work. The next most ethical thing you can do is, if available, offer that patient to join a randomized clinical trial. If there isn't a randomized trial, the next most ethical thing is to offer a registry. So these go by different names. Basically the idea being that you do what you want to the patient, but at least you register that experience so that there is some knowledge gained of what you did. And then the least most ethical thing to do or the last thing on the list is to personalize the treatment for that patient—best empiric therapy. Use your judgment for that patient, which the statisticians have renamed “make a guess.” So it's an educated guess, but it is still a guess. Now the reality is that the topmost item is the least frequent that you are able to do, and the bottom-most item is the most frequent that you have to do, because most of the time there are no data to drive the decision.

Question: I presented some data yesterday from our pilot, where we tried the approach that was discussed before—patients who undergo an intervention with clot removal and try to get to hypothermia before the opening or during the opening of the vessels. And the other advantage that that has is that you are pretty sure there will be reperfusion with reperfusion injury. So you could show an effect in a smaller number of patients. This has its own problems, because of course we have the negative trials from clot removal. That's a different discussion. So in your trial, you will likely have maybe a reperfusion rate of 40%, and it will be better if TPA is given earlier. I'm wondering if you are going to look at this also. If you are going to do subgroup analysis to see if you get more of an effect of hypothermia in patients who have earlier TPA.

Dr. Patrick Lyden: Yes, certainly those subgroup analyses are actually planned. Absolutely. Your question and Carmello's question really raise or beg the longstanding tension that we all have whether we're doing stroke, trauma, neonates, MI, or cardiac arrest, between trying to show a balance today in favor of a therapy versus waiting for all possible technical advances in all possible developments and enhancements of the study. At what point are you developing a new treatment? And you just have to stop the animal work and the R&D and go for it and try a trial versus at what point are you jumping in too early. It could be argued that the Merci and Penumbra devices were tested too early. That we jumped into an IMS 3 trial before the field had really matured. And the same can be said for carotid arterectomy back before distal protection devices. But on the other hand, TPA worked. TPA is not the ideal thrombolytic, tenectoplase is, but we did it. Carotid surgery worked despite the fact that technically today there are much better approaches. So, how long do you wait, and how complicated do you make the trial? There is no right answer. It's a tension between doing everything and waiting for the ultimate treatment versus trying to get a trial going today.

Comment: Absolutely. I think we need both, and I think the model we have or what we're going through is very similar to what happened in cardiology, when we had the same type of discussion about TPA, and when we had the same type of discussion about coronary intervention, clot removal, and then which was better. The cardiologists have been better at getting funding and doing large trials. Maybe we can learn from that as well.

Question: I have a question about what we've been talking about the last couple of days concerning a 33°, 36° trial targeting cardiac arrest. I'm just wondering specifically in terms of intracranial hemorrhage, where you stress the common occurrence of hyperthermia, and yet your trial, if I understood it correctly, is comparing 33° to standard of care. Could you clarify? Is that standard of care at 36.5? Is that what I saw?

Dr. Fred Rincon: It's not really standard of care. The standard of care for ICH patients is not temperature modulation. So basically, the way that we practice in the ICU is we look at these patients and we react to the fever. So if you're in an aggressive intensive care unit like my unit, we will see the patients, and our attendings are more aggressive than myself, for example. That would just preemptively treat the patient for one fever in the first 24 hours if the device is on. But the normothermia arm in my clinical trial by no means is a standard of care. In the clinical trial, we start normothermia before the onset of fever; that's what makes it not the standard of care. Why did we design the study with a control arm? In my mind, there is no ethical reason to compare the hypothermia arm against a fever arm. So, I followed like the same similar pattern for the cardiac arrest study, where I needed to show that I was trying to find an effect against an arm where there was no detrimental effect. I can pick up controls to show that fever is bad and then compare them against my two arms, but for the clinical trial I didn't think that was ethical in my mind.

Question: I'm just concerned at the end of the day, you'll look at your control group and there will be a large variation in temperature. And then it will not be statistically different from the cooled because it is just so highly variable. I would just like to see the non-temperature-controlled—where a high number of your subjects will be 38, and then you have a 36.5 or whatever, and then you have the 33—to really emphasize exactly what we're doing with that group.

Dr. Fred Rincon: We could easily pick a standard-of-care arm, either prospectively or retrospectively, from my own institution, and that would be very easy to show. Now, I'm not saying that normothermia or hypothermia appears to be helpful, I just don't know that. In my experience, what I've seen, and I think that Dr. Badjatia has a recent article on ICH and normothermia, shows that normothermia as a therapeutic intervention has also some tradeoff in terms of the side effects of it. You need a lot of medication to keep these patients down. I have a similar experience in what Dr. Lyden was showing. Once you reach 33, these patients are very easy to manage in the ICU. They require less medication; you can examine their reactivity by serial neurological assessments. And they are quite different from those in the normothermia arm.

Dr. Patrick Lyden: Just to follow up on that, these two trials were designed over several years and went through many rounds of regulatory review. Since that time, statistical methods have advanced, and the so-called Bayesian or adaptive designs are now more acceptable to the FDA, and I don't think to European authorities quite yet. But if there were a corporate partner who was particularly interested in supporting a novel trial design, we could look at all of these questions for the same sample size. We could answer the question 36 versus 33 versus 24 hours versus 12 hours and get to an answer much more quickly.

Question: Have you taken one of these alternative trial designs to panel and experienced the reaction of the KOL physician community to these alternative trial designs?

Dr. Patrick Lyden: It's a very good point. These adaptive designs have been proposed and have been shot down by individual panels. However, there is at least stroke study that has now been approved for funding and regulatory approval. So I think whereas up until maybe as late as three years ago there was significant antipathy toward this, things have shifted. And I think it might be worth taking another shot at it.

Comment: I think, as we talk about it, we had different therapeutic windows for targeting ischemic stroke. I mean, obviously during the ischemic period, if you could manipulate the pathophysiology, that is probably the most critical window. Then we just talked about reperfusion injury as another critical window. But even after 24 hours, you have these inflammatory cascades, these M1/M2 different types of macrophages being formed that we know we can manipulate in the late phase to protect against ischemic injury. Although it's best to treat early, I don't think necessarily you have to have the hypothermia on board. I mean, it may maximize your effect, but as Fred has shown, if you have a delay in terms of inducing hypothermia, maybe 6 hours after ischemia in a rat, if you cool for 24 hours you get pretty good protection. So, duration of cooling versus when you start, I think that's what's attractive. You have these different windows of opportunity. And depending on the importance of that pathophysiological phase—be it the ischemic insult, the severity; be it the amount of reperfusion injury; be it the inflammatory cascade—you know on a particular patient or a particular rat, that is what is going to determine how effective your therapy is at the end of day. It's a little complex, but it's also good that we have different windows of opportunity.

Dr. Patrick Lyden: I agree completely.

Question: I just wanted to ask Dr. Rincon a question. I didn't get the details. How large are the hemorrhages that you include?

Dr. Fred Rincon: More than 15 cc.

Question: 15, okay. I just wondered how you decided to cool them for 3 days? Was it an empiric decision or…?

Dr. Fred Rincon: Very good question. I thought 8 days was too long, and the newer technology or the newer versions of the technologies there enable us to do a slower rewarming. So even though I'm stopping the cooling for hypothermia at 72 hours, I'm still guaranteeing that the patient is getting at least an additional 72 hours with the rate of rewarming. And the second thing is that I think the battlefield as I mention in the talk is within the first 3 days. So for neuroprotection, I think that you know 72 hours and a slow rewarming may be enough, and 8 days I think may be a little bit too much. There are some animal models, for example, comparing short, mid, and then long-term, seeing that the long-term animal models have not had that much of benefit and functional outcome. But those are animal models.

Dr. Dimitre Staykov: I like that very much with a slow rewarming and the stepwise rewarming, not in one term back to normal, but in several steps, and that's the way we do it. Because in ICH there is at least from a pathophysiological point of view, a critical point that's exactly at 72 hours when it's hypothesized that the erythrocytes that are extravasated release their hemoglobin, iron, and other toxic products, and open a new cascade of pathophysiological injury. So I think it's a good design that you rewarm them slowly in order to go over this period.

Dr. Patrick Lyden: Let me close by thanking Dalton and Helen for organizing this session, and let me thank my copanelists for being part of this, and thank all of you.