Ramesh: Hello everyone, welcome back to our next episode of Crit No Podcast. here with me is my co-host, Dr. Bharat Kumar, who's a senior clinician and researcher at Apollo Hospitals. So this week we thought we will talk about the new trial, Bharat, the logical trial, that looks at two different oxygen saturation targets in the post-cardiac arrest situation. So just a brief brief introduction to the logical trial is there has been a lot of data in the past, and we will review some of the data saying that hyperoxia may be worse. During cardiac arrest, there is hyperperfusion to the brain, and immediate post-resuscitation after you get spontaneous circulation, there is always a risk of reperfusion injury to the brain. And having hyperoxia during that phase, specifically a PaO2 of 300 millimeters mercury or higher. Been associated both in animal studies and in some human studies to have adverse neurological outcomes because of oxidative injury that happens from this high level of oxygen. So, over a period of time, people have realized that giving too much oxygen may not be a good thing to the immediately resuscitated brain. And so I think people have looked at whether we need to be using a conservative strategy. And then this study fits in that because it is specifically going To look at all patients with cardiac arrest, both out of hospital and in hospital, who are unresponsive after the return of spontaneous circulation. And this trial specifically looked at two different oxygen saturation targets and evaluate the outcome. Before we go on into the trial, Bharat, I think it may be worthwhile for us to just quickly review the existing literature so that people understand the context of the study.
Bharath Kumar: Sure, sure. Yeah. Thank you. Thanks, Ramesh. And good to be back another episode of CritNO. So specific to this question, there have been a few studies, but we will focus only on the major ones. So there are a couple of postdoc analysis of previous trials. So we'll start there. The first one is a postdoc analysis of the ICU ROCKS trial. The ICU ROCKS trial asks the same question, know, conservative versus liberal oxygen targets, but in all mechanically ventilated patients. In that a subset of patients were also post cardiac arrest with a HIE. So there were about 165 patients who were post cardiac arrest. So the investigators analyzed this in a post-doc manner and published the data separately. And they looked at the primary outcome of unfavorable neurological outcome. So what they found is that with the conservative arm, was a better or less proportion of patients had an unfavorable neurological outcome, 55 % as compared to 68 % with the liberal strategy. So there was some signal that the conservative strategy might be beneficial in that particular cohort of post-cardiac arrest patients. Then there is another post-hoc analysis of the hot ICU trial. Now this is a trial that evaluated this question, among acute hypoxemic respiratory failure patients, know, rather than all
Ramesh: Mm-hmm.
Bharath Kumar: mechanically ventilated patients. Again, there is a cohort of post-cardiac arrest patients in that trial, about 330 odd patients. And they analyze this separately where the primary outcome was 90 day mortality and they don't show a difference in the 90 day mortality. There were, however, adverse events or serious adverse events in the conservative oxygen group. So these are the two post-doc trials that suggested that there may be some benefit from a conservative oxygen strategy. We then have a trial called the EXACT trial, which is published in 2022 in JAMA. It's important to know that this trial stopped early because of the COVID pandemic. This trial asked the question of whether conservative or liberal oxygen targets are which is optimal in out of hospital cardiac arrest patients. But interestingly, the trial was conducted in the out of hospital setting as well. So paramedics and the EMS staff were randomized to maintaining these targets. So in the conservative arm, patients would be kept in between 90 to 94%. And in the liberal arm, between 90 to 100%. until they arrived in the ICU. So the intervention was delivered completely outside the hospital. They included about 425 patients, again, this trial is from Australia. Their primary outcome was survival to hospital discharge.
Ramesh: Mm.
Bharath Kumar: And this was worse in the conservative arm, 38 % versus 48%. So the primary outcome was actually worse in the conservative arm, at least in this trial that enrolled patients outside the hospital. there were more hypoxic episodes in the conservative arm as well. And of course, the reason that they postulated was that there was probably a difficulty in monitoring the oxygen targets closely out of hospital. So this is the exact trial. We then have one final major trial, which is the BOX trial, which was published in 2022 in NEJM. I think many of our listeners and viewers will remember this trial. This is a two-into-two factorial trial. that looked at oxygen saturation targets but also at BP targets post cardiac arrest. So they included 790 patients from Denmark across two centers in the oxygen targets question. The primary outcome was death or hospital discharge with severe disability or coma. They used the cerebral performance categories scores and they did not show a difference. So this is the sort of landscape of trials or major trials that have looked at this question. So it does, mean, the numbers enrolled so far, until logical aren't perhaps large enough to arrive at a definitive conclusion. And they've, of course, also been done in different settings. So that's really the lead up, I guess, to logical, Ramesh.
Ramesh: Yeah, so I mean I think the points are all very fair, right? But the other studies did not have large enough numbers or specifically look at two oxygen targets as the primary analysis. Most of them were subgroup analysis with some signals. So I think I think this was one of the leading questions. And I th I think a recent resuscitation council guideline statement basically said that this is one area of research which is a high priority in the immediate post-cardiac arrest status. So I think all of that fits in nicely into why the study needed to be. Done. Now going on to the study Barat, I just thought I'll quickly summarize what they actually did. The study was done in 53 ICUs across Australia, New Zealand, and Ireland. And I think if I remember right, one of the earlier data on the hyperxia causing harm came from the Australian registry. And I think probably they have just taken it further with this study based on that. Originally, they decided to do a 1400 patient study. Yeah. But then subsequently, to control for dropouts and follow-up, they increase it to 1840. They assumed the 32% control event rate of good neurological outcome and survival at 180 days, but they also wanted an 8.5% increase in chances of good neurological recovery. The patient studied were all post-cardiac arrest, any kind of rhythm, both in hospital and out-of-hospital cardiac arrest. They should be unresponsive post-ROSE. By unresponsive, they said anything other than not following commands was considered as unresponsive to these patients. And the other criteria is that the clinician should deem them appropriate for this trial. If they think the prognosis is too bad or or they think they're going to do no withdrawal of support, then those patients could be excluded. The clinician should feel that hypoxic ischemic encephalopathy is the predominant reason for their unresponsiveness. And randomization should happen within. 12 hours of being in the ICU. So those were all the criteria how the patients were picked, and they were all reasonably straightforward. In the intervention group is the conservative oxygen strategy arm. They basically said that the lowest oxygen saturation you can have in both groups is 90%. In the conservative group, the upper limit of oxygen saturation is 95%. So you can have a saturation of 90 to 95 percent. And the FIO2 was tapered down to as low as 21% in patients who have mechanically ventilated. And if they were extubated, they were completely weaned off oxygen in the conservative arm. in the liberal arm, the bottom target was still the same, 90%, but there was no upper limit. The clinicians decided whatever saturation the patients had had to be maintained, and the FIO2 was not allowed to be tapered below 30% in the in those particular in that particular subgroup. And generally the clinicians were advised not to tweak the PEEP and just work on the FIO2 in both the arms. They would even though there was no protocol, but the recommendation was work on the FIO2. Rather than the PEEP, if the saturations were good. I think that was just a safety clause that they kept in there. The only deviations that were allowed is if the patient was transported out of the ICU or if the patient had a procedure, bronchoscopy, endoscopy, whatever, then the saturations, even in the conciliator arm, can be cranked up to avoid any desaturation or anything else happening to those patients. So I think it it sounded reasonably fair. And They did randomize patients, but they used a newer technique, adaptive randomization. We will talk about it a little bit once I just finished the methodology here. The primary outcome that was evaluated was a extended Glasgow outcome scale at 180 days. So you have to be no the Glasgow outcome scale goes from zero to eight, zero being dead and eight being very good neurological recovery. So that was the primary outcome. Study looked at several of secondary outcomes, including no the mortality at 180 days, the lowest outcome scale and the highest outcome scale, ICU length of stay, hospital length of stay, percentage of patients discharged to home from the hospital directly, and so on. So I think the study specifically answered the question we talked about, Bharat. But I I just thought I should pick your brain on what do you think about this adaptive randomization. How does it work? I know in one of our earlier podcasts we talked about adaptive power analysis and this is an adaptive randomization. So could you expand on what it actually means and why would you do that and what are the pros and cons of that part?
Bharath Kumar: Yeah, sure. So as you said, yeah, thank you. We did talk about sort of adaptive sample re-estimation, sample size re-estimation in one of our earlier episodes. And I also wrote a blog on adaptive trials, sort of a primer to adaptive trials. So this sort of is a nice continuity of that discussion. So here in this trial, they have used a technique called as
Ramesh: For the sake of our audience, but at just just for the sake of our audience I think worthwhile mentioning your blog so the where they can go and look
Bharath Kumar: Yeah, yeah.
Ramesh: it up. Yeah.
Bharath Kumar: It's ccevidence.blog. I'll also add a link onto our channel. So in this trial, they used something called response adaptive randomization. So essentially what we do here is that the randomization ratio is altered during the course of the trial, but it's done in a way that does not compromise the integrity of the trial. So with all adaptive trials, is the rule. Whatever you adapt should be done in such a way that the integrity of the trial is not compromised. So why would we do this? The main reason or the most important reason really is that if an arm is doing better, it's important that more patients are randomized to that arm. So that's the whole concept of response adaptive randomization. So this needs to be, of course, done in a blinded way, as in the investigators, the trial steering committee, et cetera, should not know about it. It should be done in coordination with the DSMB and by an independent team. But the main reason is an ethical reason to say that if an arm is doing better as the trial moves along, more patients should preferentially be randomized to that arm. Now, in the current trial, have kept that as at a very sort of, it's not a very, the increase is not substantial. It's only the ratio they use is 1.05 is to 1. So what they do is they start the trial with 1.05 is to 1 in favor of the conservative arm based on the data from their ICU ROCKS trial which showed that the conservative arm is doing a bit better. And then at each interim analysis based on the performance of patients and the outcomes the ratios change to either 1 is to 1 or 1 is to 1.05 depending on which arm is doing better. The maximum is 1 is to 1.05 is to 1. in favor of either arm. So it's not a huge increase, it's actually a very marginal increase, but slightly skews it in favor of the arm that is doing better. So as I said, the rationale is primarily ethical. You want to have more patients in the arm that is doing better. This, however, does come with several downsides. The 1 is to 1 ratio is the most efficient ratio in terms of randomization, statistically, in terms of power, et cetera. So any ratio other than that decreases the efficiency. And another thing is that to change the ratio, your outcomes have to be ascertained fairly quickly. So in this case, they probably were able to because they chose in-hospital mortality to change the ratio. If they had to choose, for example, 180-day outcome, you can't change the ratio effectively at all. Every patient, you to wait for 180 days. It's going to be difficult. So outcomes have to be ascertained fairly early. This may not be a problem in this trial because the ratio is only 1.05 to 1, but if the ratio is much higher, let's say 2 is to 1 or 3 is to 1, then it brings some risk in imbalance in numbers between the two groups. It brings some challenges with respect to interpretation. And also if the trial is enrolling over several years, then a number of things will change over that period, right? Outcomes are improving also because of care processes improving and so on. So it can sometimes create some tricky situations because we don't know which arm is doing better and necessarily because of what? Is it the intervention alone? But it's a very, you know, it's sort of a cool innovation from an ethics kind of a perspective. But if you don't do it right, then it can bring some problems. In this trial, I don't think it created major problems because, you know, the ratio itself is not very, very different, not substantially different.
Ramesh: Okay. But as a as a person who is not a full-time researcher. just a few questions based on what you told me. Does it even make sense to have a ratio of 1.05 to 1? I mean to me doing this adaptation really doesn't make any clinical sense. I mean I understand if it's 1.5 is to one or two ish to one, then I I think everything you said made sense. But here 1.05 to 1 to me sounds like doing an innovation for the sake of it rather than actually getting any meaningful methodological advantages or ethical advantages to it. I I mean what you think I mean is there a reason you can
Bharath Kumar: Yeah, yeah, yeah.
Ramesh: you can you can
Bharath Kumar: Yeah, don't know. That's actually a very valid point. I think the reason they've chosen it is, so this ratio was chosen in the context of the larger Megarox trial, which is a 40,000 patient trial. And logical is a small sort of a nested trial within it. So probably for the larger trial, which has a general cohort, a sepsis cohort, a non- ischemic encephalopathy brain injury cohort, and so on, maybe the tweak is still important, right? In a very large trial. Here, you are right that within the context of logical alone, there is a risk of it looking like a token attempt at increasing the ratio. Yeah, so
Ramesh: Yeah. Symbolic one. Yeah. Yeah.
Bharath Kumar: I do agree with you on that.
Ramesh: The other thing I I thought about randomization is because they did adaptive, I don't think they did a randomized block randomization or something like that. So the limitation of no this kind of ratio that we keep talking about, right? I mean we moved on from simple randomization to block and then a ra and and a random block size, right? So those are disadvantages still exist with this kind of a model, right? But
Bharath Kumar: Yeah, yeah, I don't think they're stratified by site or did block randomization. Yeah, yeah.
Ramesh: There is no block randomization here from what I could tell. I think they just did this altered ratio.
Bharath Kumar: Correct. Yeah. That's correct. Absolutely.
Ramesh: Yeah. So I and I I think that nicely takes us to what they actually found. So we'll just start off with first of all talking about the adaptive randomization. So the first interim analysis was done at 557 patients, and they basically said it was 1.05 is to one in favor of conservative oxygen therapy, and and the ratio didn't change. Overall, thousand eight hundred and forty patients were randomized. But despite this 1.05 is to 1, I think more patients ended up in the liberal oxygen arm than conservative arm purely by chance rather than anything else. But having said that, the baseline characteristics were very similar across the group. So a quick recap of the baseline characteristics here. Patients were on an average about 60 years old, 70% were men, 75% roughly, 74% was out of hospital cardiac arrest, and 25% in hospital, which we would expect. Patients were very sick, and Apache 2 score was 28. So, really sick patients post cardiac arrest, I think they improved. 51 to 52% of the patients had a shockable rhythm. So equal amount of shockable and non-shockable, making it nicely extrapolatable. Most of the patients had a cardiac arrest because of a medical cost. 88% had a at a medical cost as a baseline characteristics. Mean FIO2 between the two groups is 0.5. Time to enrollment in the trial from the on from the onset of return of spontaneous circulation was around 7-7.3 hours. Time to randomize. From ICU admission was about three hours. The median starting PAO2 was 98 across the world. So I think a reasonable cohort population that you and me would see in the ICU. I think getting these people randomized within three hours is very commendable. It's not an easy thing for people, you know, people who have done clinical trials would know it's not that easy to randomize patients within that short period of time. So that was good. When you come to the outcomes, Barat, I think the primary outcome, as we talked about, is 180-day extended Glasgow outcome scale. And the outcome was available only in 1709 patients. About 112 patients, or roughly about 4%, missed the primary outcome for various reasons. There was no outcome differences between the two groups. The primary outcome happened. Favorable outcome. Again, I want to know specifically mention this and probably come to you to get your views on this. The Glasgow outcome scale scale is an ordinal scale. It goes from zero to eight. Zero is death. Eight is very survival with very good neurological recovery. So it is an ordinal scale: 0, 1, 2, 3, 4, 5, 6, 7, 8. They dichotomized this scale into two. 0 to 4 was considered poor outcome, and 5 to 8 was considered good outcome. And that's how they have decided on the primary outcome. So if you look at the good outcomes, which is a Glasgow outcome scale of 5 to 8, control a conservative arm had 38.2 percent having that, and the liberal harm had 39.2%. Obviously, the difference was not statistically different. And as I mentioned, they had looked at a whole bunch of secondary outcomes. Alive at 280 days, no difference. 48% versus 49%. The worst Glasgow outcome scale, which is death, 55% in the conservative arm and 53.7% in the liberal arm. The other extreme, upper and lower good recovery happened in 19.7 in conservative arm and 22.6 in the liberal arm. ICU length of stay and hospital length of stay was pretty much similar. Four days ICU length of stay and hospital length of stay seven days. The proportion of patients survived or and discharged to home was about one-third, 33%, 34% in both arms. So whatever secondary outcomes you looked, I think there was no difference. In terms of the intervention itself being very different, they looked at the percentage of patients in ICU who had a hours of saturation more. Than 97%. In the conservative arm, remember that we kept the saturation at 95. Only 21% of the time their saturation was above 97%, versus in the liberal arm, it was it was present 53% of the time. And if you think about PAO2 more than 100, again as a marker of separation between the two groups, it happened only 57.7% of the time in the conservative arm and 78.4% in the liberal arm. On the other hand, if you look at the PAO2 of less than 60, conservative ARM had 43.4% of patients had at least one episode of PaO2 of less than 60, while the liberal HAM had only about 27.5% of the patients had at least one episode of PaO2 of 60. So looks like there is a decent separation in terms of the oxygen targets that happen between the two groups here. So that's what the the study and the outcome showed, Barat. Now going back to you, what do you think about dichotomizing an ordinal outcome as they have done here? Why would you do that? I mean, I understand it's easier to analyze the data if you dichotomize, but apart from that, why would you do that and what are some of the pros and cons the reader should think about when they are interpreting a paper like this, Parad?
Bharath Kumar: Yeah. No, thanks. Thank you. Thanks, Ramesh, for summarizing the results. The only thing I want to point out is that I think they did, if I'm not wrong, they imputed the outcomes for those that had missing outcomes as well. So the analytical cohort still has all of it, but they did not have the outcome for about 112 patients or so. Coming to the question on the ordinal outcomes, generally, any continuous or ordinal outcome, it's not a good idea to make it into binary categories. You lose lot of statistical power, you lose efficiency. So in general, it's not a good idea. The reason they may have done it is, as I said, one, probably ease of analysis, two, maybe the ease of interpreting for most readers. But it certainly takes
Ramesh: Yeah.
Bharath Kumar: away the efficiency and the power because obviously collapsing so much information into two categories loses information. But also clinically you lose information because you know these categories are not, although it's not an outcome, they're not necessarily, the difference between the categories is not necessarily the same. So actually having a better idea of the difference across each of the categories is useful I think from a reader's perspective. The way that this could have been done is to report something like a proportional odds ratio. which is basically estimating the common odds of being in a more favorable outcome when you compare the two arms. I think they did a sensitivity analysis that did something similar to it, but maybe not exactly a proportional odds. And they reported that those results are not different, but it would have been nice to have that as the primary analysis. In general, if you have a continuous outcome or any ordinal outcomes or any multiple categorical outcomes, not a good idea to make it or collapse it into
Ramesh: Yeah.
Bharath Kumar: a binary outcome.
Ramesh: No, other thing I was just looking at the Glasgow outcome scale here, Bharat. The four, if you get a score of four, it it is called as Upper severe disability where you need partial assistance in your activities of daily living. So somebody can help you bathe or just help you know, take walk you to the restroom, basically. So that's all. And five, which is considered a good outcome, is basically independent but unable to resume work or school or all previous social activities. Basically, you're independent with ADL. I would argue that four and five are not that different from a clinical context point of view. But here one would fall as a good outcome and one would fall as a bad outcome. And I
Bharath Kumar: Correct. Correct.
Ramesh: mean, I guess you have to draw the line somewhere, right? I mean, if I am a survivor and you tell me that somebody will walk you to the restroom or help you take a shower, I think families will still be happy taking that outcome compared to somebody who's dead or a persistent vegetative state, right? So I think it dilutes. the clinical significance
Bharath Kumar: Correct.
Ramesh: o of that particular outcome I felt just look l looking at that. I thought I'll just point that out. What do you overall think
Bharath Kumar: Yeah, yeah, yeah, absolutely.
Ramesh: about the study, Bharat, in terms of the no the strengths and limitations and what what would you know what would you take from the study?
Bharath Kumar: Yeah. No, thank you. think, I mean, the strengths are pretty clear. This is a stellar research group in terms of both the area of work, but also the high quality trials that they keep churning out. It's a nested trial. So as I was saying, it's nested within the MegaRocks program of research, which is efficient because you're trying to also answer multiple questions within a large trial. And so that's good. I think it's generally extremely well executed and you pointed it out in your talking about the time to from admission to randomization and so on. Also from a generalizability standpoint, as you said, it included both out of hospital and in hospital cardiac arrest as well as shockable and non-shockable rhythms. So I think those to me were shrinks. There were a few concerns that I had. is, despite many of the oxygen metrics they report, if you look at the overall FIO2s and saturations, the separation is still modest which perhaps you know it reflects a general practice of not going too high right like most of us even otherwise probably don't go high to 98 99 and so on so maybe it reflects that and and the only thing you could do in the liberal arm is restrict your FIO to not lower than 0.3 which of course they did effectively but still the saturation separation or the PAO2 separation is perhaps not very high. The sample size, thought postulating an 8.5 difference is perhaps too optimistic. Of course, they do increase the sample size for other reasons after that. We discussed the problem of dichotomization of the primary outcome. Now, while the generalizability improved by the shockable, non-shockable out of hospital in hospital, I was wondering though whether the in-hospital cohort is a different beast, right? The way they behave, their pathophysiology, the reason for arrest, their immediate post-arrest management, right? Because in out-of-hospital, obviously, especially in our context, it's very different, right? These are very, very different because in this trial, 78 % got bystander CPR. Not going to be the case for
Ramesh: Yeah, yeah.
Bharath Kumar: us, So I thought those, yeah.
Ramesh: Absolutely. And I was going to say that this this co this co-outcomes cannot be extrapolated to outcomes here. Even for them, good neurological recovery happened in thirty five to thirty eight per 38 and 39 percent of patients in both groups. That's quite good. I don't think we can expect that in our country with the rates of out of hospital CPR that happens in cardiac arrest that happen outside the hospital.
Bharath Kumar: Did you think of any other concerns that you had Ramesh?
Ramesh: Yeah, yeah, a couple of things. One, in terms of the generalizability, I I found that 25% of patients who were in the ANZICS registry who were put in after a cardiac arrest were included in the trial. So that tells you that it's a very good generalizable population from across Australia and New Zealand, number one. Number two, couple of things that I would like to share here is there have been protocol deviations in about 4 to 5 percent of patients, and protocol otherans in terms of titrating oxygen was also suboptimal. For example, in the conservative ARM, about 26.3% of the patients did not get their oxygen reduced as they should have, and in the liberal ARM, 8.4% did not adhere to the oxygen titration protocols that they should have. So there is a fair amount of protocols. otherance violation. So if you take conservative arm not bringing down the oxygen and liberal arm probably also doing that, they kind of can dilute the effects that happen across the both groups, and that could be one of the reasons why why this could have happened. second, my my thoughts are like we talked about the missing outcome and how it was imputed, and 6% of missing outcome is although acceptable, is quite a lot considering all this. I think 6% imputation can kind of tilt the balance anywhere. My concerns were see, the the clinicians and the bedside caregivers obviously cannot be random, cannot be blinded to the targets. So there is some amount of performance bias, if you will, right? For example, my end-of-life decisions or my even though neuroprognostication they say was similar, there was no standardization of how neuroprognostication was done. So there could have been biases in terms of the clinical care withdrawal of support between the two groups because the clinicians were completely unblinded. And if I withdraw support earlier, then obviously one group would have a different outcome compared to the other. They have negated it by having neutral assessors who were blinded to the randomization. The assessors were all trained to get the Glasgow outcome scale and they were completely blinded. So there is some amount of objectivity. But considering end-of-life decisions and other care decisions could have been made by the clinicians who knew the randomization, it still adds a little bit of skepticism to the overall results in terms of whichever way it goes. It's a negative study, but still could have confounded some of the results. So that's something that I would like to point out. And you mentioned this in the earlier trial. In the in one of the earlier trials, in the pre-hospital study, where they looked at oxygen saturation targets, conservative liberal, the conservative arm did poorly. And they hypothesized that patients who are on conservative arm would have had multiple hypoxemic episodes en route to the hospital, that would have caused some of the adverse outcomes. If you look at this study here, Bharat, the PAO2 of less than 60 happened at least once. 43% of patients in conservative arm, which is almost half of the patient population in that arm, versus only 20 odd percent in the liberal arm. So, one end I'm trying to avoid hyperoxia, but being conservative and getting them to 21% oxygen potentially may have added to some of the hypoxemic incels and negated any benefit that I would have gotten from preventing hyperoxia. So that could have been one of the explanations for why we probably didn't see a good result. And and the third point I want to make is that. that from the time of achieving Rosk The randomization patients took about seven hours to be randomized into each arm. As we talked about, three hours is a great time from ICU admission. I don't think you can do any better than that. But still, from ROSC to randomization, seven hours is a long enough period of time that whether it is hyperoxia or hypoxia, the damage is already done to the brain in the immediate post-resuscitation phase, and maybe we get the patients very, very late. And that could have again been a factor that made the study potentially negative. So these are some of the thoughts I had in terms of why we probably had a negative result and why we need to have a little bit of skepticism before we just blindly say we are going to do this particular target or not. So over to you, Bharat.
Bharath Kumar: Yeah.
Ramesh: What would your practice change be? First of all, would there be a practice change? Number two, would you what would it be if you did change your practice after this trial?
Bharath Kumar: Yeah, no, thanks, Tameesh. Before I get to that, I want to actually bring
Ramesh: A
Bharath Kumar: up something which I thought was interesting. I was going through the supplementary appendix. The neuroprognostication
Ramesh: Mm-hmm.
Bharath Kumar: that they have done is essentially primarily clinical, because the proportion of patients
Ramesh: Mm-hmm.
Bharath Kumar: that get a CT or an MRI or a neuron-specific analysis are all pretty low, 15%, 10%, which I think was an interesting thing for me to note, right? Because at least where we practice pretty much everybody gets at least a neuroimaging, if not a neuron specific enolase and other things, as part of our neuroprognostication. There the neuroprognostication, at least looking at the appendix, seems to be predominantly a clinical neuroprognostication.
Ramesh: Clinical
Bharath Kumar: I just wanted to bring that up as something that was interesting. In terms of my take on the trial or how do we go forward, I mean I... My view is that probably my practice would not change in the sense that I'm currently targeting anywhere between 92 to 95%, which is where I probably would stay. I would definitely not risk going below 90, which again, I think we currently don't do. And if it's a PAO2 value, I'm thinking between 65 to 90 or 70 to 90. And there is the MegaRox results that are... know, awaited and they're going to have obviously add that there's probably about 8,000 or 9,000 patients if I'm not wrong, that will be in the the HIE group. So we might get some more information either end of this year or early next year. But until then, this is sort of where I think my practice will be. What about you, Ramesh?
Ramesh: No, I completely agree with you, Bharat. One, we need to understand that there are several factors that affect the brain injury, secondary brain injury that happens after resuscitation. And reperfusion injury and oxygen targets are only one of them, right? I mean, you can have autoregulation problems, mitochondrial problems, inflammation, fever, seizures. There are so many other things that could have confounded the results of this particular study, which we don't have the data. I completely agree with you. We are talking about Australia, where the nursing-patient ratios Far better than what we see here, and in their hands, almost 43% of the patients had PAO2 less than 60, which I probably would say would definitely hurt the brain, and therefore to me 90% seems too low. And I would definitely keep a buffer in a setup like ours. So I agree with you, having a saturation target of 92 to 95 seems reasonable, and I exactly echo your points of keeping a PAO2 of 70 to 90, but being conscious that both Both extremes would probably hurt the brain and and then go from there. So, my take home again is the same as you target a saturation of 92 to 95, and then no do the rest of the stuff that we would do for cardiac arrows, postcardiac management, and hopefully
Bharath Kumar: Cheers. Yeah.
Ramesh: minimize further secondary brain injury and go from there. So I I think overall it's a fabulous study, like you said. No, it's a tough question.
Bharath Kumar: Absolutely. I get. Yeah.
Ramesh: Relevant question. We need to definitely have answered to this question. I think they have addressed the answer very clearly and given us a direction. Let's look out for the MegaRox trial. I I think with that we can close this particular podcast, Barat. And for for new viewers who have who are not used to our podcast, Crick knows available on YouTube, Apple, and Spotify. So please do check check us out. And Dr. Barat's blog is there, CCM Evidence. So again there'll be a link with write-up about the adaptive random adaptive randomization that we talked about in this trial and look for look forward to your comments and inputs and for us to get keep getting better and keep doing this further. So thank you all. Have a good one.
Bharath Kumar: Thank you, thank you so much.