Good morning, everyone. Can you guys hear me okay? All right, great. I'm really honored to kick us off today with the first of our few talks on post-arrest care. And my talk is entitled Targeted Temperature Management, What Now? Okay, this is me. I have my email and Twitter account. If you need to, you wanna reach out to me today or after the talk, feel free to contact me. A few intellectual disclosures. I do serve as the site PI for iSCAP trial, which is a hypothermia duration study, as well as a co-I for the SIREN Clinical Coordinating Center. I receive research funding from the AHA, also NIH and Zoll Foundation. And then I've been serving as a member of the L-Corps or International Liaison Committee on Resuscitation Advanced Life Support Task Force, as well as AHA ECC Subcommittee. Few disclosures. So we're gonna focus on adult clinical trial during my talk this morning. And this talk really represents my own interpretation of the available evidence that exists. And finally, I believe that we're still really far from understanding the best practice or approach for targeted temperature management after cardiac arrest. So the objectives of this talk, we're gonna start with some mechanisms in preclinical evidence that might explain the rationale of implementing a therapy like hypothermia in post-arrest patients. We're gonna then follow that by the first half of the talk, an overview of the clinical trials that have been conducted over the past several decades or so. The second half of my talk is gonna focus on interpretations of perhaps some of the variability and differences in the results of the clinical trials comparing to the preclinical evidence. And then I'll wrap up the talk by talking about what now are the future directions. So we're gonna start with some mechanisms and rationale. So the brain actually, what happens after cardiac arrest is during the ischemia or the downtime or no flow state, the ischemic injury from a no flow activates and triggers a whole cascades of inflammatory and also mitochondrial cell death, neural cell death pathways. And then that is actually further exacerbated when your heart is restored. They're doing reperfusion, what we call reperfusion injury. And there, what happens is you can lead to calcium overload, metabolic acidosis, inflammatory responses, and then also micro thrombosis in the microvasculature of the brain. And this is on top of all the mitochondrial dysfunction that already exists. Furthermore, the brain can actually sustain additional secondary insult by a combination of hyper or hyperoxia, hyper or hypercapnia, hypotension or hyperthermia. So the thought that what could the hypothermia do is potentially protect the brain from these insults after cardiac arrest by doing a number of different things. One of which could be from controlling the, or decreasing the overall metabolic needs of the brain after cardiac arrest by delaying energy failure, by decreasing the metabolic acidosis, and then also promotes the pro survival pathway and then perhaps decrease the inflammatory responses. So there's been actually decades of preclinical animal models on this topic, particularly focusing on hypothermia and in the neurologic outcome of cardiac arrest animal models. So this most recent systematic review that was published in 2021 by Ulrich and colleagues basically outlines and summarizes the results of 45 studies over almost a thousand animals of combination of rat, pig, rabbit, and dog studies, which showed there was a strong beneficial effect of TTM, specifically normal hypothermia, computed normal thermia, on the neurologic outcome of cardiac arrest, these cardiac arrest animal models. Through a statistical modeling, they also found that faster cooling rates, lower target temperature within 32 to 36, and interestingly, shorter duration of cooling were all associated independently with an increase of effect size of TTM. So given all of the, what I just presented, the thought is that perhaps if we were to induce hypothermia in these post-cardiac arrest patients that we could in fact improve their neurologic outcome. And before I start with the overview of the clinical trials, I actually want to clarify the definitions of TTM a little bit. So targeted temperature management, TTM, in itself actually could be somewhat vague and not particularly helpful at times because it doesn't really specify how we apply target TTM, targeted temperature management. So in 2021, the LCORE-AOS task force tried to specify that by dividing the term TTM into four different categories. So there's hypothermic temperature control, which involves active temperature control with the target temperature below a normal range. There's normal thermic temperature control, which basically means that there's active temperature control with the target temperature in a normal range. And there's fever prevention temperature control, which involves monitoring their temperature of the patient and actually preventing and treating temperature above a normal range. And then finally, there is no temperature control, which is no protocolized active temperature control strategy. So I'm going to try to tease apart the various clinical trials by using these more specific terms to describe the type of therapy that the patients actually received in the clinical trials. So we're going to start with the landmark studies by Bernard and Hakka. These were published, as many of you guys know, in 2002. And what both studies did was they randomized out-of-hospital cardiac arrest patients with witness V-fib arrests to receive hypothermia or normalthermia. And in the Bernard study, which is the smaller of the two, they randomized patients who received target temperature at 33 for 12 hours versus normalthermia. And then for Hakka trial, which was a little bit bigger, over 100 patients randomized, they received either 32 to 34 for 24 hours versus normalthermia. And what both studies found was that there was, in fact, an increase in survival and survival with neurologic outcome from both normalthermia, hypothermia groups, of almost 20% or above difference. However, the caveat of these two trials was that, in fact, the normalthermia groups for both studies were probably better defined as no temperature control because they were both passively sort of managed and observed and then not really actively sort of managed in a certain temperature range in the normalthermia group. So something to keep in mind as we progress to the other trials. So from there, a decade later, the Nielsen group published a TTM trial in 2013. And this study basically randomized out-of-hospital cardiac arrest patients of presumed cardiac etiology with any initial rhythm except for asystole of unwitness. And so this was a predominantly a 80% shockable rhythm cohort. They were comatose with ROS of at least 20 minutes, and they were randomized to receive either 33 degrees or 36 degrees for 28 hours. The primary outcome for this study was all-cause mortality at the end of the trial, and they randomized almost 1,000 patients. And so the TTM trial found that there was no difference between 33 versus 36 degrees. In terms of the primary outcome, which is death at the end of the trial, you can see almost 50% from both group end up dying at the end of the trial. And also there was no difference between their secondary outcome, which is survival with poor neurologic outcome at 180 days. Fast forward 10 years, the same group then conducted the TTM-2 trial, and this was published in 2021. Again, out-of-hospital cardiac arrest patients with presumed cardiac etiology were unknown cause, and this time slightly, it's still very predominantly a shockable rhythm. 74% of the patients randomized had a initial shockable rhythm, and the patients were randomized to receive hypothermia at 33 degrees or normal thermia, but it was actually a fever prevention of 37.8 or less for 28 hours. The primary outcome was all-cause mortality at six months, and this is a bigger trial compared to TTM. They randomized 1,900 patients this time. And TTM-2 trial basically showed that there was no difference between hypothermia and fever prevention. So if you can look at the primary outcome, about 50% from both group end up dying from all-cause at six months. And then same thing, no difference in the main secondary outcome, which is survival of a poor neurologic outcome, again, a little bit over 50% from both group. Interestingly, in terms of the serious adverse event, the hypothermic group from TTM-2 trial had a higher incidence of arrhythmia resulting in hemodynamic compromise, so 24% from the hypothermia group versus the 16% from the fever prevention group. So there's a few caveats, I would say, from the TTM-2 trials, and you'll see this is kind of emerging as a theme, which is that the majority of the patients from both trials had shockable rhythm and received bystander CPR. So this is a good thing to remember when we're trying to apply the results of certain clinical trials to the patient in front of us who may or may not meet the similar study criteria. And then further, almost 50% of the fever prevention group from the TTM-2 trial actually received some sort of temperature management device with active feedback control. So that could have been a surface cooling device or endovascular cooling device. So the trials I've presented so far have focused on predominantly shockable initial rhythm patients. The question is then, what about the non-shockable cardiac arrest patients? The Hyperion study, which was published by Vasco and colleague in 2019, conducted in France, randomized patients who are resuscitated from both in-hospital and out-of-hospital cardiac arrest who had non-shockable rhythm due to any cause. So it ended up being about three quarters were from out-of-hospital and the one quarter in-hospital cardiac arrest from this study. And they randomized this patient to receive either hypothermia at 33 versus target normothermia at 37 or below for 24 hours. And their primary outcome was survival of a favorable neurologic outcome, which is CBC of one to two at 90 days. And they enrolled about almost 600 patients. Okay, and what the Hyperion study show is in fact that hypothermia group had a higher proportion of patients with survival, with good neurologic outcome, CBC of one to two at 90 days. So 10.2% versus 5.7% of the normothermia group. You can note that this is a much sicker patient with a much lower overall survival rate across the board. This is just one of the figures from the paper demonstrating the breakdown of the patients with different neurologic outcome at 90 days. You can see that the hypothermia group on the bottom there, those with CBC one or two add up to be about 10.2% compared to only 5.7% from the normothermia group. So the conclusion of this study was that among the common test patients who resuscitated from non-shockable cardiac arrest, 33 degrees Celsius for 24 hours had a higher percentage of survivors with favorable neurologic outcome at 90 days than compared to the targeted normothermia group. So how about the in-hospital cardiac arrest patient? Well, this study actually just came out a couple months ago. This is the study that was conducted in Germany. And what they did was they randomized patients resuscitated from in-hospital cardiac arrest patient with any rhythm. So it turned out that about three quarters of them were non-shockable, a quarter of them were shockable, not surprisingly. And these patients were randomized to receive either hypothermia temperature control at 32 to 34 or normothermia at 37 or below for 24 hours. Primary outcome was all-cause mortality at 180 days and 229 patients were randomized. And so this study by Wolfram and colleagues, they found that there was no difference between the two groups in terms of the primary endpoint of death by day 180. And also no difference in in-hospital death and survival with good neurological outcome at day 180. So mortality rate across the board was about 70% between the two groups, as you can see here. So in 2021, the L-Core ALS Task Force conducted a systematic review on TTM. This is results from TTM at 32 to 34 degrees. So this, the review was done including all the studies I've presented so far, except for the Wolfram and colleagues study that just came out a few months ago from Germany. And what was demonstrated in the systematic review is that there was no difference between hypothermia versus TTM at 32 to 34, either with survival at hospital discharge or neurologic outcome at hospital discharge at 30 days. Or the mid to long-term outcome of 90 to 180 days both for survival and survival of favorable neurologic outcome. So from this, the most updated, most recent recommendation from L-Core on temperature management for adult cardiac arrest in 2021 basically includes the follows. That they suggest actively preventing fever by targeting a temperature of 37.5 or less for patients who remain comatose after RSC return on spontaneous circulation from cardiac arrest. And this is a weak recommendation, low certainty of evidence. And that whether subpopulations of cardiac arrest patients may benefit from targeting hypothermia at 32 to 34 remains uncertain. And then comatose patients with mild hypothermia after RSC should not be actively warmed to achieve normal thermia, which is a good practice statement. Furthermore, L-Core suggested that surface or endovascular temperature control techniques when temperature control is used in comatose patients hour after RSC, which is a weak recommendation and low certainty of evidence. And that when a cooling device is used, we suggest using a temperature control device that includes a feedback system based on the continuous temperature monitoring. And finally, they suggest active prevention of fever for at least 72 hours in post-cardiac patients who remain comatose. And this is a good practice statement as well. Okay, so the next half of my talk, we're gonna talk about some of the potential interpretations of why the clinical trials might have panned out that were consistent with preclinical data. So I would say that in general, these are the overall rules of what I call from clinical care trials. And I would say all large trials in general, which is that there's a concept of secular trends. So these large studies with relatively few patients over the course of population tends to take years if not over a decade to do. And over that span of time, when the trials are going on, the secular trends have changed. And so your standard of care might have hopefully gotten better. That might have impacted, in fact, improved your mortality, decreased your mortality and increased your survival rate and survived a good neurologic outcome. So because of that, there's a general overestimation of effect size, which means that we generally need a much larger trial to detect clinically meaningful differences. So hence, we end up one trial after another with neutral effects or lack of differences between the comparison groups. There's several key considerations for therapeutic translation in my mind, which is that what do you need to think about when trying to translate a novel therapy from the preclinical space to the clinical space? So the first one, of course, meaning that we need to understand its mechanisms, how it actually acts to protect, in this case, the brain after cardiac arrest. Second is the dose of the therapy. So in this case, for the temperature, it's actually the duration and target temperature. Third, the therapeutic window. So in this case, it's actually time to target temperature. Fourth, the types of patients, the patient phenotypes, who actually benefit from the therapy. And then do we actually know the pharmacodynamic biomarkers to actually titrate the therapy and to tease apart who actually responded to the therapy versus not? And finally, is it actually feasible to implement a therapy clinically? So what I'm going to do is actually try to highlight all these considerations with some of the trials I've been actually conducted, both in the preclinical and clinical space. So the first one is the concept of duration hypothermia. We already talked about mechanism during the first few slides of my talk. So this is a TTH4 trial, which was conducted by Kirkgaard and colleagues, published in 2017. And what they looked at was actually the concept of dose, so duration of hypothermia, comparing 48 versus 24 hours of hypothermia, 32 to 34, basically randomizing out-of-hospital cardiac arrest patients with any rhythm, but again, a predominantly non-shockable cohort of 90%. Primary outcome was six months neurologic outcome, and they enrolled 355 patients. They found no difference between the two groups, 48 versus 24, so 69% from the 48-hour group survived with good neurologic outcome in six months compared to 64% from the 24-hour group. There was a higher incidence of adverse event from the 48-hour group, 97% versus 91% in 24 hours, and that was predominantly driven by the hypotensive episodes in the 48-hour groups. Not surprisingly, the 48-hour group, in terms of the survivors from those groups, from that group, had a longer time on mechanical ventilation and ICU length of stay. But there's no difference between the two groups in terms of incidence of pneumonia or bleeding episodes. So when you look at the probability of death between the two groups, again, there's no statistically significant difference, although there is a point estimate difference between that 48-hour group seems to have a lower mortality with hazard ratio of 0.8. Another study called CAPITAL-CHIL, it's a cool study name, acronym, study which is done in Canada, single-center study, looked at the concept of moderate versus mild hypothermia, so 31 versus 34. And again, this is a, you know, out-of-hospital cardiac arrest cohort with predominantly shockable rhythm, and they randomized patients who received 31 versus 34 for 24 hours. All cause outcome, primary outcome was all-cause mortality or poor neurologic outcome, 180 days, and they randomized almost 400 patients. No difference between the two groups. The 31 degree Celsius group actually did have about three days longer length, ICU length of stay when we looked at the other outcomes. They tried to stratify the outcomes by shockable versus non-shockable rhythm patients, and again, there was no difference. Although the trend seems to have flipped, so if you look at the left-hand graph where it was a shockable rhythm cohort, the mild hypothermia group seems to have at least a point estimate, better, higher probability survival versus the other way around for non-shockable rhythm, although you can see that the non-shockable rhythm cohort are so few patients, you know, three to six patients at the end of this trial, that it's really hard to make any conclusive statement based on this data. So there's ongoing trial iSCAP, which maybe some of you guys are a site for, which also try to address the concept of dose of hypothermia by looking at from the duration of hypothermia, and this is adaptive study, design study, which basically has been randomizing out-of-hospital cardiac arrest patients of any rhythm to receive, you know, anywhere from six to 72 hours of hypothermia at 33 degrees in attempt to answer the questions of, does longer duration of hypothermia actually improve patient outcome? Can the efficacy of hypothermia be confirmed by evaluation of duration of response? And then finally, does the duration have a differential effect on the safety and quality of the recovery? iSCAP actually tries to also address the issue of therapeutic window, which I'm going to get into in the next few slides, by having an inclusion criteria, that the patients have to be under 34 degrees within four hours of our hospital arise for a 911 call time. So this is an ongoing study that hopefully in a few years we'll have, we'll add to the data that we have. So the concept of therapeutic window actually is probably best described and illustrated by this rodent study that was done by Bob Neumar's lab about 10 years ago. And in this study, rats that had achieved ROSC after X6-scale cardiac arrest were randomized to receive either normalthermia or hypothermia starting at different time points after RSC. So you can see on the y-axis, so normalthermia is the first bar, followed by Th0, or the animal group that basically received hypothermia right upon ROSC versus one hour delay for the Th1, followed by four hour delay and eight hour delay. The light gray bar on the top with a star of each bar actually represents the proportion of rats that had survived a good neurologic outcome. So you can see that we were able to see that neuroprotective effects up to four hours of delay from RSC. And that effect was actually lost somewhere between four and eight hours. And so this suggests that at least in this particular brain injury model that a therapeutic window was somewhere between four to six hours. So the question is then how does that, how do the clinical trials I presented so far, you know, what were their therapeutic window? And you can see that I've just created a table with a summary of all the trials I presented. And they're sort of, by extrapolating either directly from the data that they presented in the paper or on the figures, if you can just focus on the right-hand column in the red there. So all the trials that have been done so far except for Bernard had a therapeutic window times from RSC to target temperature of greater than four hours. So this, you know, and then you can remember the trials actually does show a neuroprotective effect where Bernard Hakka in the Hyperion study. And you can ask, well, the Hyperion study had a very long time from RSC to target temperature of like almost nine hours. How does that, how do we reconcile that? And that's perhaps a combination of the different severity and type of brain injury and cardiac arrest, and also the difference in the secular trend, you know, with the 2002 trials versus the more recent trials. The TTM2 trial group actually tried to look at this by doing a post hoc post sub study. And what they did was they categorize and stratify based on the sites by the temperature at four hours. You can see on the left-hand side, they grouped them into six different groups from the coldest, so 33.8 to the highest 35.3. And they basically showed no difference in the base of the mortality between the two groups, hypothermia versus normothermia. Although there seems to be a point estimate difference in the slowest group, group six, where hypothermia actually seems to have at least lower mortality. But when they looked at the survivor curve, actually a survival probability across the sites, the sites actually had the fastest cooling, actually had the lowest, the patients, I would just say, lowest probability of survival. Again, this is a post hoc analysis. And so one could imagine that the confounder in this case could be that the patients with higher, you know, greater severity of brain injury are gonna be the ones who are gonna be quickest to cool because of lack of ability to shivering and all that. Okay, and then, so this next concept of patient phenotypes is very, I think, very relevant and important. So just kind of briefly summarize in terms of the trials I've presented so far, the proportion of their shockable rhythm. So Bernard and Hakka, really predominantly a very shockable initial rhythm representation, except for the Hyperion trial, which is only non-shockable rhythm, and then the Warfarin trial, which is an in-hospital cardiac arrest patient where they had 24% shockable rhythm. And then if you look at the comparator, so if you compare the overall survivor proportion or favorable neurologic outcome from each of the trials, and just look at the comparator, the control arm, you can see that we've essentially gone from 20 to 30% overall survival rate in the out-of-hospital shockable beef up arrest patients to now more of 50, 60% for the similar cohort. And then the non-shockable patient cohort, Hyperion study was a very sick group of patients with only five to 10% survival rate. And then the Warfarin study had done better because it was actually effectively diluted a little bit by some shockable rhythm patients. So these are all important things to kind of keep in mind when you're trying to, again, apply the knowledge gained from the clinical trials to the patients in front of you. And then, so I think the concept of patient phenotype is actually best described by this retrospective cohort study done by Dr. Cliff Holloway and the post-arrest service group at Pittsburgh. And so what they did, as many of you guys know, they, you know, the Pittsburgh group, they've had a longstanding post-arrest service and they have a huge database of the patients that they follow for over a decade now. What they did was they retrospectively compared the outcomes of these comatose patients in their post-arrest group registry with exposure to hypothermia to either 33 versus 36 or 24 hours. And then their association between the severity of illness, which is based on both coma and organ failure scores and their neurologic injury and their primary outcome of survival to hospital discharge. And over 1,300 patients were included in this analysis. So how they stratify the patient illness is actually by using the Pittsburgh cardiac arrest categories tool, which is the PCAC tool. So this is something that they establish within the UPMC group. And so PCAC one is really the patients who are not very, they're doing very well following commands purposely, have purposeful movements, and they tend to have the really good survival rate. PCAC two are those who are comatose, but with mild cardiopulmonary failure. So on minimal vent setting, minimal pressure support of 0.1 or mics per kilo per minute or less of pressors. PCAC three are those who are comatose, but with severe cardiopulmonary failure. These are those who need moderate to high vent support and dose vasopressor requirements. And then PCAC four are the patients who are deeply comatose with absence of pupil and or corneal responses, regardless of their cardiopulmonary failure status. So what the PICS work group found was that it really does, if the patients have really severe brain injury, it doesn't really matter what temperature they receive. So and how they categorize and define severe brain injury in this study is the evidence of severe cerebral edema on CT imaging, which was defined as great wide ratio, 1.2 or less at the level of basal ganglia, or with highly moving EEG, which they define the absence of cortical activity and or burst with burst suppressions with identical burst. If they then took out these really severely injured patients from their analysis and then looked at the remaining patients, what they actually found interesting was that the PCAC two, so these are the comatose patients with mild cardiopulmonary failure, did better actually with 36 degrees versus the PCAC threes and fours were more comatose, patients actually did better with 33 degrees. Now they've tried to adjust for confounders by doing a bunch of variable adjustment, multivariable adjustments, and also propensity matching, but this is a retrospective study, so there could be residual confounders or something to keep in mind, but this is actually a really nice study to actually support the idea that not all patients are created equal. Okay, so a few more slides on the concept of translation. So pharmacodynamic biomarkers, so we actually have no idea what we're applying a therapy, in this case hypothermia, whether or not it's working. We wait to see if the patient wakes up, but really no data in between. Ideally, though, we want something to allow us to be able to measure and titrate the level of mechanistic target engagement at the level of hopefully the cellular pathway, but if not, at least measure the severity of brain injury perhaps by a combination of blood-based biomarker imaging and or electrophysiology, so that we can actually, first of all, tease apart who is actually responding to the therapy and then also titrate the depth of the therapy accordingly. And finally, in terms of the feasibility of clinical implementation, so this is actually highly related to the therapeutic window of a therapy and then also the systems of care that the patient is actually being brought to. So you can imagine if a therapy, in this, for example, if it's bystander CPR and AED application, that needs to happen within 10 minutes of a cardiac arrest, then that needs to be implemented at the pre-hospital setting, lay responders, 911 dispatchers, and paramedics, first responders. If the window is 10 to 16 minutes, then the paramedics are gonna be probably the folks who are gonna be really highly involved in implementing the therapy. The one to four hours and four hours longer really depends then, again, the systems of care and also resource utilization, ED boarding, so it could be occurring in the ED ICU to IC level. So these are things that it's really important to consider when we do find something novel and hopefully impactful in the pre-clinical space, is it actually feasible to actually implement and what needs to happen in the clinical setting for us to be able to execute this effectively so that our patients actually can see the improvement in the outcomes. So what now? So what I've presented so far and hopefully I've conveyed to you that there's been robust data from animal models that have shown neuroprotective effects from hypothermia after cardiac arrest and that some trials have shown that hypothermia is better than normothermia, so that's HACA, Bernard, and last year, the Hyperion trial, while other larger trials have shown no difference between hypothermia and normothermia, so that's CTM-1, TTM-2, and the Warfront in hospital cardiac arrest patients. No trials have shown that normothermia is better than hypothermia and few studies have shown, such as TTM-2, that there might be increase in adverse event in the hypothermia group and also longer hypothermia, duration hypothermia. So how do we reconcile that? So this is my disclaimer, these are my recommendations. For all post-cardiac arrest patients, at least we should at least perform active fever prevention for 72 hours or greater, preferably by using a temperature control device that includes a feedback system, and that I believe that multidisciplinary engagement to engage ongoing research to better understand the personalized approach to hypothermic temperature control in post-arrest care would be really important to continue. I also believe that standardization and prioritization of post-cardiac arrest care to really help your team lower that cognitive load and actually provide for consistent post-arrest care is actually very effective and important, and that finally data collection to improve quality improvement is actually needed, hugely needed for across health systems. Thank you, I'd love to take questions. Are we gonna do it later or take questions? Let's do one or two questions while we're transitioning and then we have a panel on this topic after my talk also, so lots more discussion on temperature. All right, thank you for your time. Thank you.

post-arrest care

TTM

neurological outcomes

clinical trials

hypothermia

normothermia

fever prevention