Intraoperative Temperature Management

Intraoperative temperature management is currently a critical procedure for treating patients with a number of CNS injuries including traumatic brain injury (TBI), intracranial hemorrhage (ICH), and the management of elevated levels of intracranial pressure. A series of state-of-the-art lectures presented at the 2015 Therapeutic Hypothermia and Temperature Management Meeting in Miami brought together experts in the field of intraoperative temperature management and therapeutic hypothermia. Dr. Ross Bullock, Department of Neurological Surgery, University of Miami Miller School of Medicine, presented an update on the Hypothermia for Patients requiring Evacuation of Subdural Hematoma (HOPES) trial targeting severe TBI. Dr. Bullock summarized the history of therapeutic hypothermia in preclinical and clinical studies of severe TBI and emphasized the limitations of previous clinical multicenter investigations. Exciting new data were presented to emphasize the potential importance of early cooling in severe TBI patients undergoing craniotomy for hematoma removal. Based on previous clinical studies and new preclinical data, it appears that this specific subset of severe TBI patients may best benefit from early cooling strategies. This multicenter clinical trial is ongoing in the United States as well as Japan. Dr. Javier Provencio, Neurological Intensive Care Unit, Cleveland Clinic, Cleveland, Ohio, presented temperature and contemporary management strategies targeting ICH. Dr. Provencio emphasized the importance of controlling fever in ICH patients and targeting inflammatory cascades for biomarker assessment strategies. A current protocol including the testing of systemic normothermia in ICH was also summarized in this presentation. Finally, Dr. Fred Rincon, Thomas Jefferson University Hospital, Philadelphia, Pennsylvania, summarized current strategies in controlling intracranial pressure (ICP) after TBI and in other neurological problems. Dr. Rincon summarized previous studies using pharmacological agents to control ICP and predictors of refractory ICP in response to pharmacological and targeted temperature management protocols. Active clinical trial studies are testing the effects of therapeutic hypothermia in specific patient populations at multiple stages of therapeutic management. These studies should provide useful information regarding how best to treat our severely injured patients using experimental as well as standard-of-care protocols.

Question: Ross, what do you think about using cytokine biomarkers, like the Licor machine, for the HOPES trial?

Dr. M. Ross Bullock: That is really exciting data that you showed and intuitively it seems a really logical way to go forward. How quickly can you get the information out of that machine; can you get it in a useful clinical time frame to be able to influence how you rewarm, for example?

Dr. Javier Provencio: It is really not a long assay, but it specifically says it is not to be used for diagnostics. We still don't know which brain injury–related cytokines really help us. Hopefully in the future it will be a relatively rapid and clinically useful assay. At present it's not a quantitative assay; it's a semi-qualitative assay. You can assess cytokine trends relative to other clinical varaiables, but you can't obtain absolute numbers from this assay.

Question: Javier, have you thought, since you are challenged with so many patients who have had surgery for their intracerebral hemorrhage, to just create a new study with patients who have had surgery first? It's the majority of your patients.

Dr. Javier Provencio: We have thought about that and one of the problems we have had is that surgery is not commonly used in ICH except for apparently in our center. The idea of doing a study that is not particularly generalizable to the rest of the world is somewhat problematic. My assumption in designing the study was that a minority of the patients that present with ICH get operated on but maybe that is not true. If the experience is that most patients with ICH get operated on, then maybe it is worth doing. I've shied away from it because it becomes an esoteric study that doesn't really help people.

Question: It seems like sometimes we go back and forth between narrowing down our study patient population or our study design to a fairly small study group, but as we discussed, the initial trials in hypothermia were only done on out-of-hospital patients with ventricular tachycardia/ventricular fibrillation (VT/VFIB), and then we had difficulty generalizing to all forms of cardiac arrest. We also did some studies in TBI that maybe were too general so that we missed detecting a treatment effect that could be beneficial in some subsets. How do we find that sort of “sweet spot” of people that aren't so sick that we can't help them or aren't so mildly injured that they are going to be fine without the test therapy? How do we find the patient population that we can best focus on for trials, and find some answers that can still help generalize our therapies?

Dr. Javier Provencio: The head trauma trials were clearly based upon a heterogeneous group of patients, and if you have too much heterogeneity, then the statistics just don't work unless you have thousands and thousands of patients. For us to do a relatively small study in a single center, we wanted to get the most homogenous patient population we could get. That is why all the effort is made about exclusion/inclusion criteria. For this trial, we wanted patients who weren't previously ill, we wanted patients whose hemorrhage wasn't due to tumors, arteriovenous malformation, or some other problem. We were really trying to get at what the pathophysiology of ICH is, and then maybe the next study will be more generalizable and more practical. Whether we get to those other things, I don't know. The head injury trials for me were a great lesson that if you take all comers, then it's a statistical mess and it's really hard to find differences.

Dr. M. Ross Bullock: Just to follow up on that point, I think mechanisms has been the holy grail of translational research over the last 30 years. It has been to identify a mechanism in laboratory model studies, and then try to find a therapy to address that mechanism and to apply that therapy in a specific subgroup of patients. We have broken that rule in TBI over and over. The biggest amount of funding that has ever been issued for TBI research in history is awarded to Geoff Manly, Andrew Maas, and others, to try to address this problem, and hats off to them for doing that. Their hypothesis now is to try to use “big data” in big numbers of patients to identify some of these subgroups. I still think we have to go to the lab, get a mechanism, and then try to apply that to subgroups. This HOPES study is just one example of trying to do exactly that.

Comment: I will follow a little bit up on what Sam said, and I would suggest to you that perhaps part of the problem is that hypothermia works. However, but we have forgotten what we learned with myocardial infarction 20 years ago, when we knew that fibrinolytics worked but we had a lot of unresolved questions about timing dose, duration, and, importantly, concurrent medications. Some of the drugs that we commonly use in these patients have effects that may be counteracting the intended benefit of hypothermia.

Dr. M. Ross Bullock: A huge truth is that modern cardiac surgery is predicated on hypothermia, and a lot of that didn't happen with randomized control trials. It just evolved as a practice evolution in the 60s and 70s. It changed the world of cardiac surgery, maybe randomized control trials are not everything, and perhaps practice evolution is a way to go forward.

Comment: I think we can still do randomized trials to test some of the concurrent interventions. There was a small but elegant study where they showed that continuous paralytics were beneficial as compared to bolus paralytics in patients with cardiac arrest, reducing shivering, reducing length of stay, and reducing morbidity. I think we need to think a little bit about doing some of those single-center studies to figure out what the different components are before we go back to doing a large pragmatic trial to test this intervention, which continues to be promising, but we haven't shown work in either cardiac or neurological/neurosurgical populations.

Dr. Javier Provencio: I think one of the things that is unstated is that the ICU and the OR are the worst situations to do studies in because there is very little evidence to support what we do. There are huge practice differences that occur on wards for the most part. If you look at the practice variation across the neurosurgical world or the neurointensive care world, you never know where those variables are going to come from. You wonder if any of those variables can unravel a study, and this is why multicenter trials have so much trouble in getting patients that are homogenous. The STICH trials are a perfect example of this in that every hospital was doing things so differently that it made it difficult to randomize. Maybe what we have to do is just demand that everybody does the management the same way. It's hard because it is like “herding cats,” and the neurointensivist is blaming surgeons for this, but we are the worst. We all do it differently.

Comment: I would like to add a comment about randomized control studies or other types of studies. You are all aware of the CENTER-TBI, which is the multi-European collaborative study based on neurotrauma patients. They will use comparative effectiveness research based on previous data from impact analysis. I was wondering whether this will be a good way to go also for other types of pathology, like SAH and ICH? Maybe to gather a large number of patients and see the different types of pathology like TBI and hematoma, whether by reducing temperature there is a positive effect. It is also another way to go to identify these new tools in statistics and then add an intervention like temperature control therapies, and then maybe we can gather additional information that we can then translate into clinical practice. Changing the research paradigm away from randomized control studies—and you don't need to do any extra research intervention to patients—could theoretically actually be harmful as well.

Dr. Javier Provencio: The problem with comparative studies, which I think in theory were great, is that the difference between the two comparators has to be bigger than the noise. So if the variability in the system is greater than the difference between those two comparators, they will never see a difference and I think that is one of the problems in the ICU and the ORs. The other way of doing this research, all the way at the other end of the spectrum, is to say instead of testing two interventions hypothermia versus nonhypothermia, that we will put people into “camps” or cohorts, and we say we are going to test a management strategy based on this hypothesis, but the rest of the management is left open to the individual site investigator. This way you don't actually care about the noise in the system because you are actually testing a single hypothesis, in a different “ecology”. That is sort of the other end of the spectrum. I am not sure which is right, but this addresses whether you have good control on how you do things in a trial, and if the “ noise” is not very high then you can see differences, due to the intervention.

Dr. M. Ross Bullock: I agree absolutely. As practice has changed in TBI, there have been big reductions in mortality and bad outcome. Many of those changes in practice have come from comparative effectiveness and nonrandomized trials. Maybe these two huge comparative effectiveness research studies, TRACK-TBI and IN-TBI, which are happening on both sides of the Atlantic now, will confirm this. The aim there is just to collect all the data, on huge groups of patients, with different interventions in different places, and see at the end of the day which are the subgroups that do well.

Dr. Fred Rincon: I was going to say the same thing. I think we need to compare at least these three trials, which are going on right now, and see if we can derive meaningful information.

Question: I have two questions; one is for Dr. Rincon and one is for Dr. Bullock. First question is for Dr. Rincon: In your Center's ICP control paradigm, hypothermia is the final intervention right?

Dr. Fred Rincon: In my mind, yes, but it really depends on the disease. To this end, for TBI it should be. I think you have to first use some other things: sedation, analgesia, and head elevation. I don't think hypothermia should be the first ICP control intervention because, as I said, there are patients that in 2–3 clinical trials have been shown not to respond to prophylactic hypothermia.

Comment: Would you use barbiturates before hypothermia? You would not use barbiturates plus hypothermia, right?

Dr. Fred Rincon: Sometimes we have done that and used both, yes.

Comment: If you use barbiturates with hypothermia, a very severe hypokalemia may occur and lead to severe arrhythmias. Does this make sense?

Dr. Fred Rincon: Yes, I agree that is a risk. We know that, and this is why patients are in the intensive care unit, where we can actually monitor all of these side effects. One thing that I forgot to mention during my talk is that most of these clinical trials have been done in centers that are excellent in terms of hypothermia management. These are not centers that do not have experience in terms of these side effects. Another critical variable that needs to be considered when you are trying to study the effects of hypothermia in a clinical trial is that it has to be designed or applied in a center that has a lot of experience.

Question: Dr. Bullock, my question is about the protocol where you had mentioned a concern of occurrence of thrombosis if you place a femoral vein catheter for a long period of time. In Japan, as you know the rate of rewarming is 1°C per day, and maybe this is related to your concern of thrombosis on the catheter, right?

Dr. M. Ross Bullock: These catheters, as far as I understand, are approved by our FDA to remain inside of your patients for 4 days. The longer that you leave the catheter, the more potential risk of a deep vein thrombosis (DVT); we do a lot of surveillance because of that concern, and so far we have not found it to be a problem except for this issue that you get the “fake DVT ultrasound.” It looks like a DVT to the ultrasonographer just because there are cooled saline-filled balloons in the femoral vein. We have to solve that problem. Getting back to your point of slow rewarming, of course, I think everybody agrees slow rewarming is essential, it's necessary, and you cannot do rapid rewarming or you undo all the benefits of the hypothermia treatment.

Question: Javier, why did you keep this ICH hypothermia trial as a single-center trial? Are you planning to expand that trial to other centers at some point?

Dr. Javier Provencio: The immunological and other studies we did here are actually pretty involved and require quite a bit of processing, using MRI to measure relative edema turns out to be quite dependent on the MRI system and image interpretation, for example. This has been kept as a single-center trial so that we can have much more control over the information before we actually spread it out to other places. Maybe it was a bit naïve to think that we should do that.

Question: For the panel, The Brain Trauma Foundation guidelines for adults and children are very different in terms of ICP management, especially when it comes to hypothermia and barbiturates. It seems the pediatric guidelines are a step behind and are very specific regarding barbiturates, but not hypothermia. I think Fred told us he is having trouble convincing the neurosurgeons at his center not to use barbiturates and use hypothermia instead. However, if you look at The Brain Trauma foundation guideline for ICP management for the adult population, it says to use barbiturates. So I ask the group, how can we change that?

Dr. Fred Rincon: I will make a comment on your first question/comment. I agree with Javier that the study is designed very well, but I don't think it answers all of the questions. If you look at the inclusion criteria for MISTIE, it pretty much included patients that are not that sick. In ICH there is always different flavors and phenotypes of patients to study blood pressure control. Is blood pressure control good for all of them? Possibly, I don't know. But what about the other patients that have hematomas of 40 or 60cc that are sort of lingering there in the ICU. Can we target them? Is MISTIE going to answer that question? I don't think so. Now, what about those patients that MISTIE targets once the hematoma is out? Is that the end of the game? I think we need to look at all of these things that we are discussing here, such as inflammation. In terms of the second question, I only see adult patients in my intensive care unit, and the data that I showed today in terms of observational studies supports that there is more information in using hypothermia than on using barbiturates after TBI. They have a similar profile but there is no data in terms of, for example, infection with the use of barbiturates. There is more data on hypothermia that actually tells me it may not be that significant. I also think that in order to change the guidelines we need to wait for the studies that were designed on the basis of that premise. Eurotherm is one of them and is not using or allowing barbiturates in that particular study population. If that study shows that hypothermia in fact is a good tool for TBI patients where barbiturates are not used, that is going to change the whole thing.

Comment: Children are not small adults, and the mechanisms of brain injury in children are actually quite different from what they are in adults. The pediatric brain trauma research community is actually very well involved and does really good studies. If you are going to put together a guideline you are basically faced with whatever data you have available to make your recommendations. It turns out that the data they have in children leads them to those recommendations, and in an adult there may be different recommendations, and we are stuck with that. Whether it is the same physiology or not, we make recommendations based on what clinical trial data you have.

Question: For Dr. Provencio, and perhaps comments by the others, your study is going to be really illustrative as far as the mechanisms leading to secondary injury that occur. Where do you foresee this going? It seems like past attempts for neuro protection in search for the silver bullet have had poor translation from bench to bedside. We are here today because hypothermia is not specific; it is a wet blanket over the problem. Do you think that what you learn will be more directive as far as patient selection, or are you also holding out hope that there is still some type of specific pathway that “can be hit by the silver bullet” that will have better applicability in selected subjects?

Dr. Javier Provencio: In ICH particularly there is not as robust an animal literature as there is in other diseases like stroke and TBI. The problem in translating research into actual clinical care is that we pick one little piece of data without trying to understand the whole mechanism. At least in the natural history part of this study—that is, in the control population—we can find out which pathways are really affected. The hypothermia group is really interesting because you could see what hypothermia actually does to those pathways and whether it's actually important. What I would like to do with this information in addition to doing more human trials is going back to an animal model that we've developed and looking more carefully at what those mechanisms we see in humans really do in rats and mice.

The problem with modern medicine at least in the United States is that to test a drug or a therapy is ungodly expensive and most people only get to do it once; you get one shot at it. If you don't understand the physiology really well, then you fail. As an example, look at factor VII—great drug, clearly works but tried in a patient population with hemorrhagic stroke/IVH, where maybe it was given a little bit too late and maybe not quite the right dose. We are never going to get that drug tested again for IVH. I think that understanding this fully makes a lot of sense. I also don't believe that just doing things in rats and assuming that humans are the same makes sense. I always like to start out with human trials, figure out what happens, sort of observational stuff in humans, and then go back and model it in animals. We are pretty far away from understanding this well. Hypothermia is nice, because it seems to work and it's pretty broad. I think it is a good window to what's going to happen.

Dr. Fred Rincon: Particular to the area of ICH and temperature, for example, how can we move forward in terms of checking inflammation? I think we have concentrated too much in hematoma. We just don't understand the interactions between all these markers, for example, inflammation, hemoglobin, oxygen radicals, etc. What about cerebral edema? Is that also something that we need to target? Once we are done with hematoma, for example, the hematoma is out because the surgeons prove that hematoma evacuation is good, then what about medical therapy for the stuff I mentioned before? Will these inflammatory markers help us concentrate on other things like lung injury and renal dysfunction? I think that is uncharted territory and things that we could study in a better fashion.