Therapeutic Hypothermia in Post–Cardiac Arrest

The 2014 Therapeutic Hypothermia and Temperature Management meeting in Miami brought together experts in the field of cardiac arrest. Dr. Kees Polderman, Department of Critical Care Medicine at the University of Pittsburgh, moderated a series of state-of-the-art lectures on current targeted temperature management (TTM) strategies for post–cardiac arrest patients. Dr. Niklas Nielsen, Department of Anesthesiology Intensive Care, Helsingborg, Sweden, presented an overview of the TTM trial comparing 33°C versus 36°C after cardiac arrest that was conducted in intensive care units in Europe and Australia. This trial has generated a lot of discussion in the scientific field regarding what is the optimal temperature to target in these severely injured patients. After a brief introduction and framing of the discussion by Dr. Polderman, Dr. Nielsen reviewed the rationale behind various aspects of the patient recruitment protocol for the TTM trial as well as outcome measures. Dr. Carmelo Graffagnino, director of Duke Neuroscience Critical Care Unit, professor of Neurology and Medicine at Duke University Medical Center, next lectured on temperature management of post–cardiac arrest patients in the post-TTM era. Based on the findings that there were no significant differences between the 33° and 36° temperature groups in the recent multicenter trial study, many clinical groups are questioning how best to treat these patients. Finally, Dr. Marvin Wayne, clinical associate professor, University of Washington, EMS medical director Whatcom Medic One, and ED PeaceHealth St. Joseph Medical Center, Bellingham, Washington, discussed community hospital experience in utilizing hypothermia in cardiac arrest patients. Dr. Wayne presented findings of what we know and how best to make a difference in these patients and ongoing challenges of prehospital and in hospital care. He emphasized that many knowledge gaps remain in this area of clinical investigation. It was clear from the various presentations and comments from the attendees that this is a very important subject that requires further study in terms of additional clinical trials as well as basic science research.

Question: I'm a bedside ICU nurse. What were your shivering and sedation levels in the controlled 36°C and 33°C study?

Dr. Niklas Nielsen: Shivering is reported in the supplementary appendix of the published article. Unfortunately, we had to move it there due to space constraints. It is similar in both groups. It is 30% in the 33° group and 34% in the 36° group. No statistical difference. Regarding sedation and muscle relaxation paralysis, we do not have the exact data yet, but will have that in the future as we currently collect it.

Question: So it wasn't protocolized?

Dr. Niklas Nielsen: Sedation and management of shivering were not protocolized; the study protocol is published for anyone who would like to read it. It was a pragmatic trial and so we did not ask the institutions to follow a specific protocol for sedation. They could use the drugs that they were used to. But we emphasized treating the groups similarly.

Question: I have a question for Drs. Graffagnino and Wayne. In Los Angeles, I can testify that the reaction to the trial has been to assume that temperature management is no longer necessary. So it has been an almost ubiquitous implementation of no further temperature management after cardiac arrest. I'm curious if you have seen that experience in the southeast and in the northwest?

Dr. Carmelo Graffagnino: At our institution, they've heard me harp for decades on fever, and so I think they appreciate both from our personal experience and the TTM trial that we are not going to let a patient become febrile. What it has shifted is a number of cardiologists even in our own institution thinking that tight 36° control would be more palatable if the differences aren't going to be significant between 33° and 36°. They are looking at ease of delivery, and it is much easier to cool at 36; at least, that is what they think. My experience in the Neuro ICU, in fact, is that it's easier to cool somebody to 33° than keeping them at 36° and awake. So that prompted us to persuade our regional colleagues that we can do the next trial because now we have equipoise.

Dr. Marvin Wayne: I do travel around, and my goal is to motivate people in the community hospitals, and today even in some of the bigger university-type hospitals. The study isn't saying, don't cool. It's saying don't warm, and that cooling has some variability. We may not know the perfect answer, but figure out a number that your team can work with and live with. Part of the problem is that in the prehospital setting we're all talking about cooling. Well guess what, if you don't save anybody, you don't have to worry about it. If your resuscitation rate is 2% in your community, that's 2% that gets to the hospital. What does that mean as opposed to if your resuscitation rate is 40% that gets to the hospital? So it's got to start from the bottom and work its way up. Part of that is to emphasize to people what the study does not say. Niklas is the first one to say that in this study, this does not say don't cool. It says don't warm or don't let the patient get warm.

Question: I think that what is happening is most of my cardiology colleagues are saying, let's wait for the fever and then if that happens, then we'll treat the fever. This leads me to my second question for Dr. Nielsen, and Carmello you alluded to this. The differences that you've shown between the groups are mean temperatures over time. But there is a way to look at the area under the curve. I'm going to present some methodology this afternoon in the stroke session where we've published this method. Have you looked at area under the curve or total minute degrees to see exactly how different the groups might really be or not be?

Dr. Niklas Nielsen: We haven't done the analysis yet. There is a group in Italy that will look at all temperature-related questions and break down the different groups and assess the area under curve and cooling rates. Of course, only hypothesis-generating results from that, but it could be really interesting to look at.

Comment: Hi, I'm from Panama, and Latin America is only to commend you. Panama is probably the place in Latin America where we do more therapeutic hypothermia for cardiac arrest than most countries in Central and South America, and for neurological reasons working in the network of critical care physicians for Central America and South America. When you published your article, I received during the first week more than 150 e-mails from colleagues over Latin America saying to me that we probably should stop cooling. This is only to let you know. In Panama, we decided to continue using therapeutic hypothermia.

Dr. Niklas Nielsen: That is always an unfortunate effect of research that people read whatever they want to read. I can only say that the trial didn't show a difference between 33 and 36. It actually didn't show that treating fever is beneficial as many have interpreted. We cannot conclude that from the TTM trial either. We can only conclude that 33° and 36° was equal in terms of survival and neurological function based on the CPC and MRS scales. That is basically it. But, of course, a trial like this will generate review articles, position statements, and guideline updates. There will always be a reaction and then it will settle down, hopefully in the right direction.

Question: Thanks, these have been really outstanding presentations. Combined with yesterday morning I think they have given us a good insight to start putting some of these data together. I do think, though, that even if many of us have heard this presented several times, there now seems to be a dueling view as opposed to a clear indication of what are the next steps. So my question is really a definitive and declarative statement, but it is informed with this observation.

The way that we as healthcare communities make decisions are not based on the way we report trials. The Fuquentous view, they are really based on Bayesian analyses that say the following. The community priors, it's a set of behaviors. There's a new observation that provides the likelihood ratio of change. Then based on the strength of that observation, there is a set of probabilities that we should change our behavior. So let's ask the simple question, 11 years ago when the HACA and Bernard studies were published, the community was in the position really of equipoise; we didn't know if this really works. So these two powerful observations moved the community probabilities of behaviors. Even though cooling wasn't broadly implemented, there was a general sense that this probably works, and then we have an implementation problem. So the question is if we are applying the same standard at looking at this new evidence in a way that makes any sense at all. So if we use the same behaviors that we did with the first two very important studies, we'd say, the data that is now presented us clearly should move the community posterior probability behaviors back to some median position. Let's take 36° as a target but not stop cooling, and yet I think if we took a poll of this room, we'd find there are actually very divergent views. I'm going to go home and do nothing different, or I'm going to go home and guide everybody to do 33°. So could I ask particularly, perhaps the whole panel, what is your explicit thoughts about how to apply this evidence to clinical practice?

Dr. Kees Polderman: Who will start?

Dr. Niklas Nielsen: I can just say that in the sites participating in the TTM trial, we are doing 36 right now. What I think also is that many sites that have done 33 and are very confident that this is the thing to do, they will also change to do milder hypothermia when they have problems and they will feel safe. But I also think that the main take-home from the TTM trial is that we should not be so sure about dogma, what is the right thing to do. I think the TTM trial has shown the possibility to test the hypothesis of temperature management more thoroughly and has opened the community for this question. Perhaps we should not settle with the evidence we have with the TTM trial, but rather use it as a stepping stone for further research in this field. But I really like the thought of the Bayesian approach to it. Looking at the point estimates, they are actually in the direction for 36°.

Dr. Marvin Wayne: Let me comment from a community hospital perspective, or maybe a semi-real world. I want to go back to this team approach of critical care for critical patients. The concern I have is that what the study says or does not say may be misinterpreted in what it says. I'm already hearing that we don't have to be so “critical” caring for these patients. It doesn't say that at all. So we've elected to use 32–33° and allow a little bit of flow to 33–34. They work hard on these patients, and I certainly don't think we're harming them, and I do think we're helping them. But again, the misinterpretation. It does drive me crazy that people are concluding we just don't need to cool. But what are you going to do? So the key is the application of critical care in critically ill patients. Cooling is a component, but it's only a component. I think these gentlemen would agree with that. But my fear is that we've got so fixated on the temperature that we are going to start allowing their critical care to fall apart. Just an observation.

Dr. Carmelo Graffagnino: I'm curious to ask the audience a question. Last year, when we were meeting at the Neurocritical Care Society, we were discussing whether we would join TTM while it was still in progress. Clearly, three-quarters of the members of the Research Committee felt it was unethical to randomize to 36° versus 32° or 33° because it was so clear that 33° was effective. We didn't think 36 was more than a placebo. Now that we have TTM, I would ask the audience if a trial was being planned where you are being asked to randomize your patients again in this country to 33° versus 36° with a different strategy to guide duration and outcomes. How many in the audience would be willing to randomize patients to a 36? 33? (Many hands.)

Dr. Niklas Nielsen: And to turn that question upside down, who would not do it? That's fewer.

Dr. Carmelo Graffagnino: So the other guys are ambivalent? No, I mean, I think this has completely changed the tables. Because, as I said, 2 years ago, there was a strong feeling that we could not do this kind of study. I think it has reintroduced that possibility in our country now. I think we need to add supportive data and expand. At our institution, we are currently doing 33. There is an extremely strong pull from the Chief of CCU to go to 36. We've all agreed that we are going to have entire CPO order sets. Our hypothermia team has not been dismantled. We still get a 100% neurology consult on every cardiac arrest. So the entire critical care team continues, and whether we target 33, 36, or 34, we've even open labeled as we are exploring this. We've cooled people for 2 and 3 days when they look physiologically unfavorable, and seen anecdotally very good outcomes in those. So, we will probably go to 36, but individualize until we get our trial going as an individual basis. Because we have that range now. We can go 33 to 36. If somebody is having an enormous amount of difficulty maintaining blood pressure, under pressers, maybe we can go to 36° and be just fine and that individual may need that until we have more data on how to pick the right dose and the right durations.

Dr. Marvin Wayne: One of the issues that we've been working hard on addressing is the Kim study addressing yes or no on prehospital cooling. We have much longer transport times than Seattle does. We cover a larger area with longer transports. So we are actually starting our cold saline during the resuscitation. We don't know if that makes a difference either. But it seems to clue the team in that we are bringing a patient that is resuscitated, with cold saline running. So the ED, in a community hospital, is geared up, and the cooling equipment is getting prepared. Everybody's mind is moving in a team-oriented cooling direction. So I'm not sure the temperature alone is critical. Do others feel that this is more than again just a number?

Dr. Carmelo Graffagnino: I would think one of the important things to keep in mind, because I think this was lost on the Kim study. The Kim study tested the hypothesis that prehospital cooling using fluid is advantageous to using an emergency room. We don't know what the effects are of two liters of fluid on coronary perfusion. We saw a trend in that study toward repeat arrest. There was a pulmonary edema trend that may not have caused mortality but certainly affected morbidity. There are others that are looking at other methods of inducing cooling without volume. That study was testing two things, a high volume load and temperature. I think the lack of a benefit may or may not be due to lack of the cooling part. So I think that has to be very clearly understood too, rather than just say, there's no benefit to prehospital cooling.

Dr. Marvin Wayne: If you want to take a contrary position, we think that fluids are beneficial in people who have RV infarct. So maybe we are actually helping these patients. Again, what you are hearing are a lot of unanswered questions that we keep asking each other. I hope we'll come up with some answers to give us more direction.

Dr. Kees Polderman: There is some animal work showing that there is a risk of decreased coronary perfusion with high volume loads. It may vary per patient and per type of infarction.

Question: Just a question on the methodology of the statistics from the TTM trial to Dr. Nielsen. Can you comment on the percentage of PEA and asystoly in the trial and how you came up with the power analysis for that?

Dr. Niklas Nielsen: The power analysis was on the whole subset. We used the Hypothermia Network Registry and the INTCAR registry to simulate the inclusion and exclusion criteria of what kind of population we would end up with. We were within percentages of the estimated population in the final randomized population, and so I think we were quite successful with that estimate. We could also simulate the mortality, and we had a mortality as predicted in the sample size calculation. In the end, it was only 20% with a nonshockable rhythm. So again, this is a trial of mostly shockable rhythms more similar to the earlier trials. We don't have more than you can see on the breakup in the forest blot analyzed for the nonshockable and shockable rhythm groups separately. But as you can see, the intervention effect does not seem to be different in the nonshockable group. They are exactly aligned.

Question: A couple of questions. One, in the TTM trial, I'm curious if you have or will be presenting data on the type of cooling, intravenous versus surface cooling? If there may be any difference in the two. Also, what the timeline was for initiation and achievement of goal with those two treatments. If I may, a second piece of that is the timing of the angiography and STEMI. With or without STEMI, I guess, because 40% had STEMI, 60% went to the cath lab. For those who had STEMI, were they cooled afterward? Those who went to the cath lab who do not have STEMI, what was the timing there?

Dr. Niklas Nielsen: There is a group from London, St. George, that worked with angiography data together with us in Copenhagen and Sweden. So we will look at that. We will also look at the intravascular versus surface cooling. We have not done that yet. But we have all that data. We have all the different types of machinery used in the trial as well.

Question: I'm very interested in also looking at cooling without fluids. Are there ongoing trials of whether it is opening of windows in the Northern Hemisphere or putting ice bags on. I'm also curious talking to people in LA if anybody has looked at the ambient temperature of those arrests that occurred in 90° urban areas versus five below zero rural areas in northern Michigan. I was just curious.

Dr. Niklas Nielsen: I think there is registry data from Austria where they have looked at admission temperatures and seasonal variation, but they did not find any impact on survival.

Question: I've got three questions. First, don't you think that the outcome of the TTM trial is more a product for the prehospital project than the difference between 33 or 36 degrees?

Dr. Niklas Nielsen: I'm sorry, can you repeat the question?

Question: The outcome of the trial is 50–50. Don't you think it's more, what happens prehospital before the patient comes in than the differences between the two degrees? Because it's so much different from all the other trials that you have prehospital. There are results from America that show a 60% increase in survival only by changing the prehospital effects.

Dr. Niklas Nielsen: I think that the prehospital phase is probably where we should invest our energy and then provide the best intensive care that you can give.

Question: So you are looking at everything about the temperature, the blankets, and things. Are you also looking into what happened prehospital, if that could make a difference?

Dr. Niklas Nielsen: Yes, the groups were balanced. We have done adjusted analysis as well. The survival actually skews a little bit more toward benefit for 36 when you do the adjusted analysis for the small baseline differences there are. But the groups were balanced.

Question: Last question I have, do you believe in neuroprotective effects of hypothermia?

Dr. Niklas Nielsen: Yes, I think it's been shown in animal models, and I think the evidence from newborns is very compelling. Even if I talk to Dr. Marianne Thoresen about that, and she said that yeah, we have not controlled to 36 in these babies. So that is a trial that may have to be done after the TTM trial. But I think also that you might have a relationship between benefits and harms of cooling and that in the end it balances for the clinical population.

Comment: But you said you changed your temperature protocol to 36° in your hospital, but you're a believer of hypothermia. I don't get it.

Dr. Niklas Nielsen: I mean that we have shown that it is equally beneficial, and it's easy, and it doesn't induce any problems with urgency of rapid cooling. You can just cool and stay at 36°.

Dr. Marvin Wayne: May I take off on a quick comment, and I'll be quick with this. How many of you are involved with prehospital care? A small number. To be honest, I think the point that needs to be made is that if you don't save people, it doesn't matter what you do. So if you are cooling 20 patients a year, because that's all you save, versus if you are cooling 80 patients a year because you are saving that many more. So maybe all of you need to think about the prehospital care systems in your communities.

Comment: This is more of a comment rather than a question. In the UK, in our center and in parts of London, the thing that the TTM trial has brought in and has focused our minds on is actually on the prognostication and withdrawal of therapy and protocolized withdrawal. I think that's really important because certainly we've spent many hundreds of thousands of pounds in bed days and things on patients who simply aren't going to do well. I think having an evidence-based trial to allow us to legitimately do that is hugely important in our healthcare system. I think that's the huge benefit that we are going to see.

Comment: Dr. Nielsen, I just wanted to follow up on your comment about 36° making it a little easier. As a bedside nurse, I find that 36° is harder to deal with in regard to shivering, even at 33°.

Dr. Niklas Nielsen: Is that a sedated patient or intubated and sedated?

Comment: Yes, our ROSC patients. I wanted to get feedback from you, and I'd love to see the report on the sedation levels of those as well.

Dr. Dalton Dietrich: I would just like to say in closing in preclinical models of traumatic brain injury, stroke, cardiac arrest, spinal cord injury, one relatively brief period of mild hyperthermia changes the whole rest of that animal's life, and so if we are talking about preventing hyperthermia as the strategy that might be best in the majority of our subjects, we are going to have to be very proactive in making sure that temperature does not rise, because once it rises those neurons don't like it at all. All of a sudden you have something going on that you already had an effect on in terms of the postinjury setting and even a brief period of hyperthermia when the nurse says, “Uh oh, the patient is running a fever,” it may be already too late, and we need to be proactive, and that gets back to the whole concept of temperature management being so critical in our patient population.