Thank you, Dr. Manali, that was an incredible overview of all things cooling. I have the joy of presenting the con position, which is actually the pro-TTM position. And since double negatives are confusing, I'm gonna take the position that induced hypothermia is not dead yet and keep it simple. And before we dive into this topic, I think one of the things I find incredibly confusing about all of this literature is that we use two terms interchangeably. And sometimes to mean completely different things. So I wanna define the terms that I'm gonna use in my talk. Targeted temperature management is any temperature management with a goal range of 32 to 37.5 degrees after cardiac arrest. And specifically, I'm gonna focus in on induced hypothermia, which is cooling to that lower end, 32 to 34 degrees Celsius, versus controlled normothermia, which is maintaining a temperature in a relatively normal range. I think where we've already have discovered we agree is that there's not yet a study in the literature showing us that we can completely ignore temperature management after cardiac arrest. We don't have that trial yet. TTM-3 might actually be the trial to answer this question. But for now, I think we can all agree, we can both agree, that active temperature management after cardiac arrest is still best practice. And I also think we can agree that regardless of what our temperature goal is, the post-arrest bundle, including appropriately timed neurologic prognostication, which is probably the most important thing that we do or facilitate as intensivists and critical care practitioners must be preserved regardless of the temperature. These are my disclosures. I have some research funding from NIH and the other places you see there. I'm also one of the site principal investigators for the ICE-CAP trial, which I'm going to mention. And I am on a couple of scientific advisory boards not related to this talk. So here's my position. My position is that induced hypothermia, which as a reminder, is cooling to the range of 32 to 34 degrees, is not dead. And the reason it's not dead is because the existing literature raises several unanswered questions. First is most of the literature, including the meta-analyses we just reviewed, are really driven by the populations of the TTM, TTM-1 and TTM-2 trials. And those populations may not reflect the out-of-hospital and in-hospital cardiac arrest patients that we care for, especially in the US. I'm going to show you some data around that. It also might be true that patients with more severe hypoxic ischemic brain injury benefit from a different temperature goal. And there weren't many of these patients included in those trials. And finally, I worry that completely de-adopting induced hypothermia and just throwing this in the wastebasket like it's been done on Twitter a couple of times, might disrupt the post-arrest bundle and actually worsen outcomes, unrelated to temperature, but related to some of the other processes of care that are linked to temperature management in the ICU. So the TTM-2 trial, I mean, I don't have much to say negative about this trial. This was an extremely well-done randomized trial by an experienced clinical trials group. I think there are some nitpicky things that you can really dive into. The time to temperature goal was a little bit long in the 33 degree group. There are a couple other concerns I have about how they reported adverse events. This is the only temperature management trial that actually reported increased adverse events with the lower temperature goal. But really that was driven largely by bradycardia, which we know is an expected complication of hypothermia and actually has been associated with better outcome in other cooling trials. But for the most part, we've journal clubbed this one a lot. There's really not a lot of negative things to say about this trial or the group that conducts it because it was very well done. But as always, when you're interpreting a trial and trying to integrate it into your practice, the devil's in the details. When you dive into table one and look at who was actually included in this trial, you'll find that these patients may not look like the patients you care for. They certainly don't look like the patients that I care for in the ICU. So what you see here is an extract from table one of the TTM2 trial, the hypothermia arm, the normothermia arm, and then a study that our group published, which was a group of out of hospital cardiac arrest patients in Seattle where I practice, who were cooled after cardiac arrest. And you'll see a number of striking differences. Our patients are a little bit younger than the ones in the TTM2 trial, all very male. This is a problem in cardiac arrest research that we need to deal with. But if you go down to the bystander witness column, 90 plus percent of the patients included in the TTM2 trial in both arms had a witness cardiac arrest. The bystander CPR rates are off the charts. We pride ourselves in doing pretty well in this column in Seattle, but if only 56% of your arrests are witnessed, it's hard to have a bystander CPR rate that's much higher than that. So fewer than half of our patients got bystander CPR. And similarly, majority of patients, over 70% in the TTM2 trial had a shockable rhythm, only 38% in the cohort that we examined. Some of the other physiologic characteristics are actually pretty similar, but a high proportion of these patients going on for cardiac interventions in the TTM2 trial, reflecting a very cardiac nature of this population, many fewer in ours. And this is probably true in many of the places that you all practice as well. We know in the US that the incidence of cardiac etiology and ventricular fibrillation leading to out-of-hospital cardiac arrest has been declining since the early 2000s, and we've seen skyrocketing etiologies of other types of cardiac arrest in my region, namely driven by opioids. And the populations in this trial really don't look like the populations that I'm caring for in my ICU. Now we do have a randomized trial where the patients do look a little bit more like the patients that I care for, and that's Hyperion. As Dr. Manali mentioned, this was a multicenter randomized trial that included 584 patients, and it had both out-of-hospital cardiac arrest patients and in-hospital cardiac arrest patients, all with a non-shockable initial rhythm. These patients got randomized to induced hypothermia in the range of 33 degrees versus controlled normothermia, and this is one of the large randomized trials in the cooling after cardiac arrest space that actually did show a benefit for cooling to 33 degrees, about a 4% absolute difference in their primary outcome of neurologically intact survival. There are a lot of criticisms of this trial too, the biggest one being this concept of fragility index, where if you take a few patients from the 33 group and you move them into the 37 group, you get a different result. But as somebody who does clinical trials, I actually don't think the fragility index is a very good concept, because if you power your trial appropriately and you do the right power calculations, have the right stopping criteria, you should not enroll many more patients in the trial than you need to to show an effect size. So I actually think this trial was extremely well done, very different population, but one that looks a little bit more like mine and actually showed benefit for cooling. Looking at all the other big temperature trials going all the way back to the Bernard and Hacke studies in 2002, you'll see that the flavor of these trials was very heavily primarily shockable initial rhythm. Even in the TTM first trial, 80% of patients had VT and VF, and then like I said before, in the recent TTM2 trial, 73%. The only recent trial, including a large portion of non-shockable patients, was this Hyperion study that had a 4% absolute difference that did reach statistical significance. And this isn't the only trial or study to get to the idea that patients with more profound hypoxic ischemic brain injury might do better with lower temperatures. This is an observational study, so not a randomized trial, using data from the really robust University of Pittsburgh post-cardiac arrest service data set. They included about 1,300 patients who had out-of-hospital cardiac arrest and who clinicians chose to cool to either 33 or 36. That's important, this is not randomized data. They then looked back retrospectively at patients who were cooled to these goals and stratified them according to an initial illness severity score called the Pittsburgh Cardiac Arrest Category. The PCAC score, the Pittsburgh Cardiac Arrest Category score, of one is basically someone's awake, so they weren't included in this study because they wouldn't have gotten cooled. A score of two are comatose patients without significant cardiopulmonary dysfunction. Three is comatose, but they have cardiopulmonary dysfunction. And four is comatose with loss of brainstem reflexes, so a higher score is worse. And what they found in the overall cohorts is that there really were no differences between the 33 and 36 degrees patients. Similarly, no differences in neurologic outcome. In the patients with lower scores or more mild injury, the 36 degree goal actually did better than the 33 degree goal. But as you go down to patients with more severe brain injury, the relationship actually flipped, and patients with higher severity scores did better at 33 degrees than at 36 degrees. Similar story for neurologic outcomes. Patients with more brain injury did better at 33 degrees than 36 degrees. This is, again, not the only study to show this idea. This is a study from Japan where they stratified post-cardiac arrest patients based on their initial lactate level as a marker for sort of arrest severity. They found that patients with the mild lactate level, which was a lactate less than eight, actually there's no difference between being cooled to the range of 32 to 34 degrees or 35 to 36 degrees. In the moderate group, which was a lactate of eight to 12, similarly, no difference between the two groups. But as you get to the more severe range, which was lactate greater than 12, the 32 to 34 degree range actually looked better than the 35 to 36 degree range. Another study implying that patients with more severe arrest characteristics or less favorable arrest characteristics might do better at a colder temperature. And finally, another study from Japan that used the RCAST score, which is another illness severity stratification tool that looks at baseline demographic and arrest characteristics as well as initial lactate level. These patients were stratified based on cooling to the range of 33 to 34 degrees or 35 to 36 degrees. Lower RCAST score is better, so no difference in the low group. Interestingly, in this study, no difference in the high group either. It was the patients in the moderate category that did better at the lower temperature. And the authors posited that perhaps the patients in the high category were too severely brain injured to actually benefit from any temperature management. And then even looking at the data from the TTM and TTM2 trials themselves, this is the meta-analysis, a patient-level meta-analysis that included the primary data from each trial. 2,800 patients total, 1,400 in the induced hypothermia arm at 33 degrees, and then they combined the control arms of the other two trials. No difference overall in the primary outcome, but if you look at the force plot for all the subgroups, the group that got no bystander CPR, that was a subgroup that actually favored hypothermia. Now, this could just be an artifact of multiple comparisons, which happens when you do a lot of statistical comparisons, but it's kind of in line with some of these other studies showing that patients with less favorable characteristics might do better with a lower temperature goal. We've mentioned the ICECAP trial. This is an ongoing trial in the U.S. where patients are all cooled to 33 degrees and then randomized to a different duration of cooling. Now, how could a different duration of cooling trial actually help us get at this idea of whether this goal or this concept of cooling patients after cardiac arrest helps? On the surface, this is a trial of finding the right duration, but really it's designed to answer a couple of other questions. The most important one, perhaps, is a dose response. If there's no duration that's beneficial, if the dose response curve is completely flat in all patients, it's unlikely that this therapy is beneficial to any patient. So this trial is also looking at efficacy of TTM by looking at the shape of the dose response curve. This trial is also designed to look at heterogeneity of treatment effect. Are there subpopulations of patients that benefit from different durations? On the most kind of, the biggest buckets that are being used in the trial are shockable and non-shockable, but there's some really nice advanced phenotyping work that's being done as sub-studies of the trial, seeing if we can really nail down the subsets of patients most likely to benefit from different doses of cooling. The patients being included in this trial are comatose after out-of-hospital cardiac arrest. Importantly, they have to be cooled really rapidly, within four hours of the 911 call in order to qualify for the trial, which is different from a lot of the previous randomized trials. And then you can see the other inclusion criteria there. And the primary outcome is neurologic outcome via the modified Rankin Scale score at 90 days. So moving on to some of the process measures that we saw change after the first TTM trial. And my concern here is that if we throw out the entire bundle of managing temperature after cardiac arrest, what else could potentially go wrong? So this was a large study from Australia and New Zealand, from the Resuscitation Outcomes Consortium group there, that looked at over 9,500 out-of-hospital cardiac arrest patients before and after publication of the first TTM trial. The patients were pretty evenly split before and after, and they found a couple of interesting things. First, they found that when the TTM trial was published, the average temperature, or sorry, the lowest temperature achieved for patients after cardiac arrest was going down over time. TTM was published, and the slope of that line changed. Now, these are their axes, not mine. Their axes are a little wonky, right? This is 33 and 36, so the absolute change is relatively small. But the slope of the line changed, implying that people did change their practice after the trial was published. There was also an increased incidence of fever after publication of the first TTM trial. Perhaps most concerningly, there had been a secular trend of improving mortality after cardiac arrest in Australia and New Zealand up until the time the TTM trial was published. And that slope abruptly changed after publication of this trial. Now, this is large registry data. It's impossible to say if this is causation or correlation or completely unrelated, but certainly the timing is concerning. It's also hard to say if this was related to temperature itself or if seeing a large null trial in cardiac arrest led to increased nihilism in other areas and people were doing things like withdrawing life-sustaining therapy earlier because they didn't think they could do anything to help these patients. We saw a similar phenomenon happen in Seattle. This is a study that we did with our experience implementing a 36 degree goal. So from 2002 until about 2014, April 7th to be exact, we were cooling patients to 33 degrees. After that TTM1 trial was published, we changed our goal to 36 degrees. The distribution of temperatures changed appropriately. And we actually found that compared to this later period, patients cooled during the 33 degree period had a higher odds of favorable neurologic outcome adjusted for a number of pre-arrest and post-arrest factors. So why might this be? Well, one possibility is that this is truly a temperature related effect. And I think it's possible that in patients with more severe brain injury, as we've discussed, the lower temperature is actually the right thing to do. It's also possible that since the bundle of post-cardiac arrest care is kind of formed around temperature management, that if we throw out temperature management, we throw out other components of the bundle inadvertently. So this is our post-cardiac arrest care guideline. This is a beautiful infographic made by one of our neurology residents, a shout out to Khasra. And so our current protocol is that we're cooling patients to 33 degrees over the first 24 hours, unless they're in that ice cap trial, they make it a different duration. And then we rewarm them over a subsequent 24 hours. That forces us to not do much else, except for some imaging and some biomarkers over that time period until the patient is rewarmed. The neuroprognostication process doesn't even really begin until 48 hours to 72 hours after cardiac arrest, where we start to get additional diagnostic tests. And then eventually we'll sit down with our neurology consultants and come together and triangulate around what we think is a likely prognosis for these patients and have a discussion with the family. By taking the temperature management and the cooling components out, I worry that many centers would shift this entire curve leftward and start having these prognostication discussions and decisions earlier. We don't actually know that these tests work at early time periods. In fact, we know the false positive rates are higher and we could risk making premature decisions around withdrawal of life-sustaining therapy without the structure that TTM creates. So my take-home points, I think induced hypothermia is not dead. The TTM1 and 2 trials were really well done. I think they're valid in the populations they were studied, but the populations of the trial don't really look like mine, and I suspect they don't look like yours. There are some additional trials underway like ice cap that will hopefully help us answer some of these questions. I worry and I think that patients with more severe hypoxic ischemic brain injury might benefit from a lower temperature goal based on the data that we discussed. And then I also worry that if we rapidly de-adopt induced hypothermia, we may disrupt that entire post-arrest bundle, especially the process of neurologic prognostication and worsen outcomes inadvertently. And I will stop there.

pro-targeted temperature management

induced hypothermia

cardiac arrest patients

severe hypoxic ischemic brain injury

post-arrest bundle