Skip to content
Abrams – ECMO for Respiratory Pandemics episode artwork

Jan 9, 2023 50 min

Abrams – ECMO for Respiratory Pandemics

from Maryland CC Project

Darryl Abrams, MD is an Associate Professor of Medicine and a medical intensivist and ECMO attending at NewYork-Presbyterian/Columbia University Irving Medical Center. He is the Associate Medical Director and Director of Research of the Medical ECMO Program at the Center for Acute Respiratory Failure, as well as Medical Director of Intermediate Care and Pre-Critical Care Services. He presents a lecture during Critical Care Grand Rounds entitled "ECMO for Respiratory Pandemics".

Episode details come from the podcast's RSS feed.

Embed this episode

Ready to play

Abrams – ECMO for Respiratory Pandemics

0:00 50:52
of matches

Transcript Auto-generated, may contain errors

So it's really my pleasure to introduce to everyone today today's speaker. So today we have Dr. Darrell Abrams here talking to us. So Dr. Abrams has done the majority of his training in New York at Columbia and is now an associate professor of medicine at New York Presbyterian Columbia University. He's also the associate medical director and the director of the medical ECMO program in the Center for Acute Respiratory Cellular at Columbia. He recently published a paper in the Blue Journal, which has really sparked my interest in sort of the role of ECMO and respiratory pandemics. And I'm very happy to have Dr. Abrams here today kind of talking to us about this. And he let me know in advance that if anyone has any questions during the presentation, that he would be happy for you to either put them in the chat, which I'm happy to monitor, or for you to unmute yourself and ask them.

So Dr. Abrams, thank you so much for joining us. I'm really happy to have you here. Well, thank you very much for having me. It's really a pleasure to speak to an excellent ECMO center. So hopefully we can all learn something from this talk. And by all means, if you have any questions, please feel free to chime in. Hopefully there'll be some time at the end. As always, I have no disclosures, but I'm open to any offers. So I want to give us a little bit of a historical perspective because we're talking about ECMO for respiratory pandemics. And I'm not sure if you've heard of COVID, but it's something that's recently swept through us for a few years.

But it's not the first respiratory pandemic, obviously, and also not the first one where ECMO may have played or could have played some role. So this is actually a timeline. Interesting enough, the article that she referenced, this was a draft figure we were going to include, and it got scrapped. So this is an Abrams original. But basically, this kind of gives you a bit of a timeline of just where we've been with pandemics and major ECMO trials since the early 2000s.

For those who don't remember SARS-CoV-1, everyone obviously refers to the current one now, that was an outbreak that started in East Asia, ended up in Toronto. There was a little bit of a spread. It was a little bit early at that time in terms of ECMO usage. It was only early at major centers and very rarely used. So there wasn't really much of a role for ECMO there. But 2009 was a bit of a banner year for ECMO in a couple of ways. One, there was the CSER trial. And also it happened to coincide with, although really coincidental, since CSER took several years to complete the H1N1 pandemic. So to just refresh my memories about CSER, and I'm sure many of you know.

It was a trial out of the UK. It was sponsored by the National Health Service. It was a very pragmatic trial. So to some, this answers the question of ECMO for ARDS. And for many others, it was really just a bit of a teaser and had a lot of methodological issues. But it was a randomized trial trial. It was 180 adults with severe, potentially reversible respiratory failure. Not all ARDS necessarily, although certainly maybe by definition could have met criteria. In fact, there was some tuberculosis in there. There was PE in there. A couple other forms of respiratory failure. And the punchline is that six months, the rate of death or severe disability was way lower in the ECMO.

we'll call it ECMO or ECMO referred arm versus usual care. And this was the biggest criticism of the trial is that patients were randomized to referral to an ECMO capable center, which also prescribed optimal standard of care management for whom only three quarters of the patients once they were optimized actually needed ECMO. And for those that were in the usual care arm, which is since it's pragmatic, it was not really forced that they had to adhere to optimal standard of care management. Only 70% would have been considered having received low tidal volume ventilation, which had been around at that point for several years and was certainly the standard of care. It also highlighted a couple of other issues with transport of patients for consideration of ECMO, and that several patients either died waiting or died in transport to get to the ECMO-capable center. So seeing as there certainly was not a uniform, you know, 100% received ECMO in the ECMO referral arm, and everyone in the usual care certainly did not receive optimal care, most people take away from this, this is not an ECMO trial per se.

raises some interesting hypotheses, and we hope that maybe this leads us towards an ECMO positive trauma future, which ultimately we have, or we have an ECMO trial today, but it did not really definitively answer the question, but it happened to coincide with the H1N1 pandemic at the same time. And so there was data coming out saying, hey, ECMO works in these really sick, seemingly works in these really sick ARDS patients, and so they also, you know, piqued a lot of people's interest in using it for H1N1. A lot of young, otherwise previously healthy patients now have severe ARDS from H1N1 influenza A. And there was some conflicting data that comes out of that. So just two of the salient studies to highlight, there was one also coincidentally out of the UK where it appeared that ECMO-referred patients versus those who were not referred for ECMO had a significant survival benefit from receiving referral for ECMO, but not necessarily all receiving ECMO, same kind of problem. And then they tried to tease it out a little bit in France by pairs matching those who received ECMO and those who did not.

And that actually had the opposite direction of point estimate. The odds ratio was against ECMO, but there was a huge caveat that many of these sicker patients who actually seem to have the lowest mortality were not able to be matched against those that did not receive ECMO. So no randomized controlled data in the time of H1N1, only the seizure data that we have and some observational data to try and tease that out. But certainly by center volume, by case volume, through ELSO as one example, started really shooting up. Now that we started to have more data, maybe in support of this relatively novel therapy.

more modern technology than had previously existed. And I'm skipping over the old randomized control data from, you know, the 1979 and 1994, which is a little bit of not quite ECMO data, but ECOR data, which was really very sophisticated technology at the time, but certainly not where we are now. And so a lot of issues with those trials in very early times. A brief mention on that time when it emerged COVID, because it was a maybe... It's not quite an epidemic even. It's just, you know, localized. There were not a ton of cases, but, you know, certainly in the Middle East, it was a coronavirus as an etiology. And there was some data that should this become bigger, maybe ECMO has a role. When they tried to match patients with and without receiving ECMO for MERS-COVID, there seemed to be a benefit with evenly distributed and kind of well-balanced baseline demographics. But, you know, something like survival 60% for those that received ECMO versus, or mortality 60% versus 100% mortality for those that did not receive ECMO. But these are very, you know, selected patients in very small numbers compared to some of these other pandemics.

So now we get to SARS-CoV-2, COVID-associated ARDS. And this is where we really have now the most modern technology. And the question becomes, first, are we talking about the same form of ARDS? Because we're going to talk about a little bit of the olea data in a little bit that probably gives us the strongest data for ECMO for ARDS. And the question quickly arose, is COVID the same? Or does SARS-CoV-2-associated ARDS behave the same as regular ARDS? Some big names in the field, including Luciano Gattinoni, say, no, it's a different form of ARDS. Well, it meets the definition from a syndrome standpoint.

It behaves differently. There's more dead space. There appears to be more thromboembolic phenomena or microvascular disease. They're not behaving the same way on small numbers on his initial experience. And simultaneously, some people that you may recognize, some people I work with, say, we don't have enough data yet. This is very early in the pandemic. For all intents and purposes, it seems to be behaving the same way. More data comes out. Then, by and large, it fits the definition of ARDS, certainly, and seems to more or less behave like non-COVID ARDS. So let's not start going down some road with strategies that are not proven and hypothetical for a maybe new disease.

versus the ones that are tried and true that we've used for many years now that seem to have a benefit in other forms of ARDS. So if we take for granted, and we can discuss this later, but if we take for granted that COVID ARDS is ARDS by any other name, is ECMO appropriate? Should it work? All right, so we only have the CSER data up to a point, but now we have a little bit more. And the rationale behind it, why should we necessarily expect that ECMO is going to actually improve morbidity mortality in ARDS? So before we even launch into COVID,

Do we have enough data to say why mechanistically we think ECMO may help? There's a lot of ways ECMO can help. It can salvage someone from refractory hypoxemia. It can manage refractory hypercapnia and respiratory acidosis. It may have some impact on diaphragm function and myotrauma. But probably the one that's got the most robust preclinical and mechanistic data behind it is this idea that ECMO can facilitate a more lung protective ventilator strategy that allows you to reduce the injurious forces of the ventilator that you otherwise have to be susceptible to by managing conventional ARDS without it.

by reducing volumes and pressures well below the current standard of care, minimizing lung stress, minimizing strain, minimizing all the factors that may contribute to ventilator-induced lung injury. So that's what many of us believe is probably the strongest argument for why ECMO may work, and that most people don't die from refractory hypoxemia, but they probably die from multi-organ failure as a consequence of excessive ventilator-induced lung injury. So... That leads us, before we go into the COVID data, to really just tease out a little bit more, what is the data we have for ECMO and ARDS to therefore use it as a rationale for why we might want to invoke it in a respiratory pandemic such as COVID. So I'm sure essentially here is very familiar with the EOLIA trial, ECMO for severe ARDS, led by Lynn Combs and with my boss Dan Brody, a senior author. And the criteria to enter this trial, this randomized controlled trial for patients with...

meeting one of these three severity criteria with ARDS. So they had to have either a P-to-F ratio less than 50 for more than three hours, very small numbers enrolled in that. You can imagine it's a little hard to sit on someone with a P-to-F of 40 for four hours before enrollment. P-to-F less than 80 for more than six hours. That was far away the most common criteria met for those enrolled in the trial. Or pH of less than 725, basically a severe enough respiratory acidosis, we're not talking just metabolic here, for more than six hours. And this really reflects more of a dead space phenomenon that they have a lot of alveolar dead space.

or that they have very severe respiratory compliance. Because if they achieve those parameters of respiratory acidosis while trying to keep your plateau pressures under 32, you'd meet criteria. Or if you met those criteria and your plateau pressure was under 32, it may just reflect a lot of dead space. So a bit controversial as a criterion when this trial is being designed. We're not trying to enroll COPD. We're not enrolling asthma. But arguably, patients with a severe form of ARDS is not manifested only by hypoxemia. And if we had three hours, we could talk about EOLIA for the rest of the afternoon. But we can certainly leave some questions for that afterwards. But those are the criteria for enrollment. And then here's the punchline.

ECMO arm, of which almost everyone receives ECMO, 35% mortality. Control arm with very high adherence to optimal standard of care, 46%. Did not meet traditional statistical significance based on a power calculation that estimated was going to be a 20% absolute risk reduction. Did not achieve that. You have to imagine performing an ECMO trial where you can randomize the control in ECMO centers because it's going to be capable of performing ECMO if you're assigned to the ECMO arm. It's going to be hard for ECMO centers to swallow unless there's an opportunity for crossover because if they get assigned to control, I don't think those centers are going to be happy letting those patients die if they're dying without at least a chance at ECMO. So there was a crossover option. About 35 patients, 20% actually did crossover. They had the highest mortality, and they had to have criteria of SATs of less than 80% for more than six hours despite every rescue maneuver basically available. Those were assessed as intention to treat, so they were in the control arm.

If you remove them out, the baseline mortality by control was still 41%. It was still higher than the ECMO arms mortality. And you have to figure that at least some of them would have died if not for the use of ECMO. So their mortality could have been even higher had they not had the option for crossover. So on the surface, a clinically meaningful difference, 11%, but not statistically significantly different. But then there were some subsequent analyses, most prominently a post-op Asian analysis led by Ewan Gallagher. And what he looked at was any number of ways you have prior beliefs or prior data, including the CSER data, which is really the strongest data we had prior to this, that whether ECMO works or not, you could be strongly enthusiastic, you could be strongly skeptical.

and you set your anticipated relative risks accordingly, then you input the EOLIA data, and you see what you get out afterward. And basically, based on any given prior, which you can always redesign however you want, the posterior probability of some mortality benefit was present across the board. How much of a benefit? 2%, 4%, 10% is what varied based on your priors, and in particular, your prior beliefs. So even the most skeptical, if you assume there's no difference between ECMO or not, it would move the needle towards ECMO based on the EOLIA data, but maybe not enough to sway you that ECMO is better if the tradeoff is morbidity that you might not find acceptable. But it certainly seems to suggest a benefit.

of at least there being something to a high probability, but the magnitude depends on who you are and what you believe or what data you use going into it. There were other subsequent analyses, including this meta-analysis at Toronto, which basically pulled Yoli and Caesar together, and both are, you know, the point estimates are in favor of ECMO, and lo and behold, the estimates in favor of ECMO when you pull it all together, but with all the caveats of both trials. Likewise, there were network meta-analyses, which is a way of kind of weighting and combining different strategies and seeing if there's a benefit, and what is suggested across 25 RCTs of all different interventions for ARDS, a lot of patients, suggests that both prone positioning and ECMO have pretty favorable relative risks. reducing 20-day mortality in addition to when combined with low tidal volume lung protective ventilation.

So on balance, let's move my window here. So it appears, I'm certainly biased, my fryers are in strongly favor of ECMO, but it appears that ECMO does have a benefit in patients with ARDS. Remember, EOLIA was well before COVID. But how have COVID patients fared with ECMO with that context, right, if it was being used in that context? So now we go to the bottom half of my yet unused timeline and just look at some of the data that started to appear. And it's really, by and large, mortality data of SARS-CoV-2 during the kind of first couple of years of the pandemic. Early on, I mean, the China data, it was so premature. They obviously didn't know what they were dealing with. They tried ECMO probably late, probably in patients who were too sick.

um terrible mortality but you know based on that there were a lot of guys early guidelines that were coming out saying don't use echo that doesn't work mortality is exceedingly high but then everyone starts to take a pause start to get more data as things start to spread to italy and then to the rest of europe so the united states and they start to collect more data and i'm going to go through some of these trials here so you don't absorb all of this but let's kind of work our way through some of this timeline so paris and um the program director was mentioning that you had some of the paris group uh recently visited and talked about um ecmo and they're really you know obviously you know world experts in this um have enormous experience and also very high volume so they early on had a lot of experience with covid and using ecmo for covid 19. And so their early couple-month experience found that 83 patients, they used ECMO in for severe COVID-associated ARDS. At this point, everything's COVID. And what they found was that about a 31% 60-day mortality and a 36% 90-day mortality. So pretty good. And preliminary data suggesting reasonable survival. They looked further. They got a little bit more data. And they're in a network, right? So they've got a few really high-volume centers. And they have a lot of referral centers from around them. And they're very well-networked. So they all coordinate, not so much like the United States with much more siloed hospital systems. So when they looked at all of their data a little bit later of 17 ICUs in the greater Paris area.

basically the same time frame with a lot of mobile ECMO teams. It's what they do, right? I mean, it was a cool design. It helps optimize patients who are being referred and make sure that's the right patients to create standardized criteria for who you're going to consider for ECMO. It's not just which hospital happens to get called, but they centralize the evaluation of these patients. And then they send out their mobile teams and try to bring them to high capacity ECMO centers, or maybe keep them at relatively low volume ECMO centers, depending on capacity and what they can handle. So there were 302 patients on ECMO. You can see the 90 day mortality is not that 31, 36% I just mentioned before, but when you tease it out.

You'll see the bottom left age makes a difference, not surprisingly. How long they've been intubated before ECMO, not surprisingly, is a spread earlier on before maybe too much ventilator-induced lung injury had been acquired. Mortality is better at shorter pre-ECMO ventilator days. But upper left here, the case flow made a big difference, mostly because we were then talking about the paracerbone, the main hub being that high case volume center that's most experienced. That case volume seemed to matter. And it wasn't just about doing ECMO in COVID, but doing ECMO for COVID at places that are really experienced with doing it and have a lot of volume over the years. So that's really kind of like their main center versus everyone else.

not quite simultaneous because it takes time for all this data to be collected, and we'll talk about that for future pandemics, but the Extracorpore Life Support Organization, ELSO, which I'm sure you're all familiar with, started collecting data on COVID-19. And at the point that they analyzed this initially, it was kind of the early first wave of over 1,000 patients, over 200 centers across 36 countries. So really a heterogeneous population, although all ELSO centers, so maybe there's some consistency in how they manage ECMO patients for COVID, seeing as they probably often adhere to the same guidelines. And here we have 90-day mortality, 37%. So again, reasonable survival, much better than that, obviously, initial China data.

Here we have a little bit more data on what kind of demographics portend a better or worse prognosis. So age, you can see here, tracks quite nicely. Again, not a surprise, but we have to be careful in not being ageists about who we select for ECMO, but it's seemingly an important prognostic factor. Immunocompromised states made it riskier that a greater likelihood they were not going to survive. Underline chronic respiratory illnesses, not surprising. All of this kind of goes along with our non-COVID ARDS experiences. Whether they had a cardiac arrest before certainly played into how they were going to do. Kidney injury, initially meeting VA, were all negative prognostic signs for mortality.

Also, around the same time, you can see the same time frame on the first half of 2020, there was a Chilean experience. And this is an interesting one, kind of like Paris, a little different. This is nationwide as opposed to just kind of a network of ICUs and hospitals. But they coordinated across the country to provide more consistent evaluation and management of ECMO patients. So 94 ECMO patients were ultimately managed.

And what they found there was not bad, you know, similar kind of 30-, 60-, 90-day mortalities kind of in line with what we've been seeing, but notably long hospital and ICU length of stay, right, which maybe we're not as used to seeing. This isn't exactly ECMO length of stay per se, but kind of reflective of that. As more data starts to come out, and I'm sure you guys have experienced, patients with COVID-19 don't seem to get through their ECMO runs as quickly as, say, flu patients do or other, you know, bacterial pneumonias, let's say, where it's a known etiology that's easily treatable. A meta-analysis around the same time, 22 studies. It's going to be heavily weighted towards some of the larger experiences, including the ELSO registered data, but almost 2,000 patients, almost all VP ECMO.

And when you look at the pool hospital, Barbaro is the ELSO registry, so obviously that heavily weighs, what they found was a pool in hospital mortality of 37%. And so let's just, you know, pick up sort of the box for where we're at so far with all these early experiences that I've summed up. Paris, 60 and 90-day mortality in the 30s. Chile, Chile, similarly. ELSO registry, similar. And this meta-analysis, not surprisingly, kind of falling in line with the ELSO registry. And just to remind you, EOLIA, 60-day mortality in the ECMO arm with non-COVID disease, 35%. So we're doing okay, right? We think ECMO works for COVID just as it does for non-COVID. Let's keep rolling with it.

Not to mention, obviously, the resource constraints, the capacity issues, everything everyone's dealing with, overflowing with ICU patients, overflowing with potential ECMO patients. So all of that's a caveat that we have to address. Then, as you're probably also familiar, there's some evolutions and outcomes over time as the pandemic evolves. So now let's start making some comparisons as we have some later 2020 data, 2021 data. One of them is out of Spain, in the Nigerian Peninsula, Spain, a little bit of Portugal. What they found is quite a few ECMO patients, over 300.

we can divide waves however you want, right? There's not a clear dividing line in different geographical regions experiencing different timeframes, but how they divided it was 150 or so first wave in the first half of 2020 and looking at the second wave in the second half of 2020. Now, they do survival, but I like to keep things consistent. So look at mortality, first wave mortality, 41%, not too far off from where we were, 60% second wave mortality, okay? So that's the first, you know, some of the first signs that something's maybe up here. There's an ELSO, a Euro-ELSO survey that goes out, surveying a bunch of Euro-ELSO centers that they also looked at kind of

First wave, second wave, again, this time it's a little different divider in September. First wave mortality, 47%. Second wave mortality, 56%. A lot of those patients are still being managed at the time that they're evaluating this, so that's kind of a rolling number. It's a little bit hard to be sure, but then that 56 could have gone up even further after that. Our parasorbone experience, right? So these are the experts, and they had a ton of early experience and gained a lot of experience through their network. They look at 88 in the first wave. That's basically more or less the first experience I mentioned to you. And then later in the second half is the second wave of 71. And here's first wave, 36%, as we mentioned.

Second wave, 48%. And these are the guys who know, right? These are the experts who manage this day in and day out and have early high volume experience. And then we get the ELSO registry data, which is already the largest experience because it's, again, voluntary, you know, it's registry data. Take that for what it's worth in terms of the quality of the reporting, who's reporting it, who has the time to report it, who are the ELSO centers, are these reflective of non-ELSO practices as well? Maybe not. They look at their whole 2020 experience, 4,800 ECMO patients for COVID-19, 350 centers, 41 countries.

Now, they broke it up into very, I think, cleverly, what was the initial experience? Basically, A1 is what we saw before. A2 is if they started doing ECMO for COVID after May 1st, not started doing, if they continued doing ECMO for COVID later in the second half. So A1, A2 is like waves, first wave, second wave of the same centers. The B group is those who only started doing ECMO for COVID-19 after May 1st. So not necessarily new centers, because it could be centers that have been established but just didn't embark on ECMO for COVID, but also may reflect truly new centers that have popped up in the meantime that were not doing ECMO before May 1st.

So we find here is A1 is that mortality that we described before. A2 is same centers, probably same practices. We'll talk about that. Mortality is going up over 50%. And B, the centers that had not been doing ECMO before May and now started doing ECMO for COVID-19 after May, the highest of the mortality. So closer to, you know, A2 and B are closer to each other. And both of them are well above what A1, the initial experience was. In the meantime, there's an update on that meta-analysis because maybe that first meta-analysis is not really reflective of where we're at now later on in the pandemic. So more studies, more patients all over the world. Some are mostly also regions.

You can see the different mortalities, Southwest Asia and Africa, very high mortality. Asia Pacific, quite high. North America, the best, but maybe not statistically so. And what they found is their pool of mortality was 49% versus the earlier 37%. And look at the rise as you look at kind of divvying it up into kind of semi-years, first half of 2020, second half of 2020. By the first half of 2021, we're going to 62% mortality for ECMO and COVID-19. There were several predictors, analyzed predictors of mortality, and among them age, when in the pandemic, as you just saw in terms of the trends.

how many receiving corticosteroids this is something we need to get back to in terms of why we think these trends are occurring shorter ECMO runs which may just reflect you know I don't know just death so I'm not sure what to make of that but certainly some some factors that may be playing into this change over time so that becomes the second question why is it changing I think it's all there it's true but we don't necessarily have the answers as to why first We're under the presumption that COVID ARDS and ECMO for COVID ARDS is the same. It's actually different. Now, the mortality over time is one difference we've noticed. We've seen this long duration of support bear out in the ELSO registry data when they look at A1, A2, and B. I definitely noticed that the A2 and B, the later periods of ECMO for COVID-19, the average ECMO run was going up to 20 days instead of the original period where it was an average of 14 days. So we're seeing a change there. It's taking longer. There's this concern for hypercoagulability, increased thrombosis, and maybe as a consequence, we're using more anticoagulation because experientially we're noticing more clot formation. Maybe that's adding to risk.

that they're observing in patients. They did see in the ELSO registry data that there seemed to be more cumulative incidence of on ECMO events like clotting and bleeding and complications. When they normalized it for the duration of the ECMO run, it actually normalized to similar to non-COVID ARDS being supportive of ECMO. But it still doesn't mean you're not accumulating more problems because you're on longer. So it's a true fact. It's just I'm not sure that COVID is contributing itself to a higher rate, just that you're on for longer and exposed for longer. The immunosuppression and risk of infection is something we certainly need to factor in and get back to because all of that evolved over the course of the pandemic.

And maybe that's playing in. And then there's been described, maybe more RV failure. One experience I didn't highlight in particular, but people may be familiar with, is this rush experience early on of ECMO for COVID. It's kind of like a bundle of interventions that they did. They used the Protech Duo, which is something that basically bypasses the RV. It drains from the RA and pumps right into the pulmonary artery. So it can support RV failure without relying on the RV to pump DV blood forward. Higher antifibrillation and steroid usage.

uh early on they were trying to activate people mobilize people they had a survival a range of like 80 percent and went well out of the range of what everyone else was experiencing um we don't know if it was a trivial to any one of those interventions or kind of a package deal but the rv has been something that people have been throwing around as maybe a factor that's worse than covid uh ards than it is in non-covid um so what we're talking about i'm sorry can i ask you a quick question um and this is probably maybe a very obvious question but were they able to show that their selection didn't get kind of looser as you know i feel like very early on there was so much scarcity of everything, PPE, ECMO, all the things. Did they get less selective as they moved on to the incoming waves? You mean the first wave, second wave kind of changed? Yeah, and did mortality go up because they're more likely to take a healthy 65-year-old than they were in the first wave? So I'm going to list a few hypotheses as why we think some of the mortality rates have gone up. That, I think, could be one of them. It's hard to bear out. I was worried that that was the case. When they looked at the second ELSO registry study, they looked at basically pre-ECMO or baseline demographic risk of death.

you know, what would be, like, you know, kind of, things like, you know, what were things that were going to potentially... who's likely to die more going into it. And actually, the earlier phase or the earlier wave that had the lower mortality would have been predicted to have a higher mortality. So it doesn't, at least in a large registry data, it doesn't describe everything, and it may not have captured all of the baseline demographics that might have increased one's risk. But I'm not sure that people have necessarily loosened their criteria. The rush experience I thought you were first going to be referring to, they adhere to the only criteria. I mean, so I can't blame that on doing it on people less sick who then just did better. A lot of these studies are still small when you're looking at pre and post, I'll call it pre and post, you know, first wave, second wave.

And it doesn't describe all of the potential risks, but it's possible. And I think that's one possibility. So some of the things among the list, as you're mentioning, there's a lot more, as you may have experienced, we've certainly experienced, increased use of non-invasive and high flow and delaying intubation, thinking that if we intubate, first, if you remember early on, like six liters of nasal cannula, now, oh, they're failing, some people, non-rebreather, they're failing, intubate. We're afraid of spreading the disease. We're afraid we don't have PPE. We don't want to expose healthcare workers. We're not going to use high flow. We don't know if non-invasive works. Let's just intubate early. And then we said, huh, this is a pretty decent mortality in the ICU. Maybe we should...

manage them more conventionally like we normally do and see if we could stave off and we're getting some other therapies coming in, treatment options for COVID, let's hold off. But what that may have exposed people to is more patient self-inflicted lung injury. Maybe we're making them work harder for longer. And by the time we do, they do fail and we intubate them. As we know for non-invasive pneumonia, they end up doing worse. So that's one possibility. And there were a couple of studies. The French group noticed definitely a higher rate of non-invasive ventilation use in the second week.

maybe they have more treatment refractory disease. Steroids become involved with a recovery trial. Now, early on, it was very haphazard. Later on, it's much more consistent use of steroids so that when you do get a patient who now fails all upfront therapies, is intubated, and meets criteria for ECMO. Need is a loaded question, loaded word, but meets criteria for ECMO based on at least the only criteria. Now, maybe you're picking for patients who have more treatment refractory disease and just aren't going to do as well versus those that were unselected early on or did not get optimal therapy, optimal or treatment-specific therapy early on in the first place.

I mentioned, I touched upon this in the last slide, but immunosuppression, steroids, and other things like tocilizumab, was that increase in rates of secondary infections. And there are studies coming out that it does look like there were more VAPs, there were more superimposed infections. So maybe that just portended, you know, another hit on patients who were already receiving ECMO for COVID that just led them to do worse later on. One thing that's not clear, one of my fellows is actually working on this with some of the ELSA registry data with some suppositions in there, but trying to get as clean a database as we can on this. We don't have individual patient-by-patient variants, but studying it.

based on where a given variant was the vast majority of the given time when ECMO was being used, so kind of assuming that all the patients had that variant. There may have been changes to the various variants and their disease courses over time as the pandemic evolves, and maybe that portends the worst prognosis. So maybe we'll have a little bit more data on that as this project gets underway. And also, you know, an increased use of ECMO, as I mentioned, I think the B group in the ELSA registry, in these either new centers or low-volume centers, or, oh, ECMO works early on. We say ECMO works for COVID ARDS. Now we can do it too.

let's start doing it because the mortality is favorable we should do it as well but maybe if they're not as experienced and they don't and you saw that with the paris uh group you know network and the spread there between high volume and low volume centers maybe that led to worse outcomes but you still notice that even the quote-unquote experience a1 centers still had worse mortality a2 so it doesn't explain everything but b being worse than a2 maybe that's part of the um part of the equation and as was just brought up maybe that did lead to a little bit more liberal usage oh i can do it in a 70 year old because they see everyone seems to generally do okay um One other study I didn't bring up, but if anyone's seen it, Germany put out, or some researchers in Germany put out a very large experience, an unselected population, because you can pretty much do ECMO on anyone you want in Germany, and they had some of the highest mortalities, 68%, 73% across the whole 2020 into early 2021, with tracking with high ages. But you look at some of the ages they were doing it. I mean, there was not a trivial number of patients greater than 69 years old who were getting ECMO for COVID. So you can imagine what more liberal usage is going to do to your mortality if you remember that age trend with the initial ELSA registry data. So that leads us to the question, you know, the mortality is going up with...

what's the comparator like how do we know if ECMO is actually affected because maybe the mortality for ECMO for COVID is going up but maybe at the same time mortality for COVID without ECMO is going up we're actually still doing okay or maybe we're hurting people or their mortality is no different with ECMO or no ECMO and we shouldn't be doing it because it's not adding any benefits to survival. I mean certainly what everyone would clamor for if we if we could is a randomized controlled trial to know if it's really better than optimal standard of care. I'm telling you right now, it's like impossible to happen. So there's certainly based on EOLIA and those who do ECMO, you're not getting another randomized control trial of ECMO versus ARDS. I think the data is strong enough for those who believe in it to keep doing it. And for those who are skeptical, they're just going to remain skeptical. So, but for, to now invoke that this is a different disease that needs a new randomized control trial, there's certainly a lack of clinical echo points. And forget the fact that it took six years to enroll, six and a half years to enroll for EOLIA. Think of the infrastructure needed to start a randomized control trial in the middle of a pandemic when there's not that necessarily in place and it took a long time to design priors. It can be very difficult to conduct when you're in a pandemic and all of your resources are strained.

And they're diverted to all these other places, right? Even just data collection becomes a very difficult thing to do when the places that want to report the need to and have the data report are also the ones being completely inundated with patients that they're trying to care for clinically. So what are some other options to maybe try and tease out the effectiveness of ECMO versus optimal conventional management, if we can even compare to optimal? So there's matched pairs analyses, those types of analyses, as you mentioned in H1N1. And they did look at this out of the UK. Can we compare those that got ECMO to those that did not? And that's ECMO, it's specialist ECMO centers, and how does that impact in hospital survival?

referral hospitals in the UK who referred for ECMO. There's five UK specialist ECMO centers. In this case, they looked at two of them, in Guy's and St. Thomas's and Royal Brompton. They looked over the course of 2020 to early 2021. And basically, they included 1,300 patients in the analysis, almost 250 of which were transferred with mobile ECMO to an ECMO specialist center, and they compared them to those that stayed behind who did not get ECMO. They did match these, I think, because that alone doesn't tell you anything, right? If they were good enough to get ECMO by criteria, and those that stayed behind, there must have been a reason that they weren't necessarily good enough or they were overwhelmed and couldn't handle it. But you have to keep that out. So they did match pairs analysis.

to see, and you look at odds ratios for in-hospital mortality, they did a few different types of pairs analyses, and basically what they came up with was there was a significant reduction in mortality for those that got ECMO, 25%, 26%, compared to those that did not. Sensitivity analyses, looking at other ways and subgroups, again, the odds ratio was strongly in favor of ECMO, but you've got, and so you can say that referral to an ECMO specialist center, it's kind of like Caesar, right? Referral to an ECMO specialist center for COVID-19 to get ECMO, your survival will be better, but your overall care may be better, right? So there's enormous residual confounding. Is it ECMO? It's a little like these are all over again, which is kind of ironic because it's the UK. So it's the ECMO or is the fact that, you know, the other care that's going on at these couple of centers is also better in improving survival. And also, if you're interested in things beyond hospital mortality, that wasn't going to get at the heart of it. Are there ways to maybe emulate a randomized control trial if we can't actually perform a randomized control trial? And so this was one relatively early in the COVID experience. There was 190 COVID patients who got ECMO across 35 hospitals. And what they tried to do was emulate kind of day after day, looking at the patients and randomizing them, quote unquote, whether they got based on getting ECMO.

having met EOLIA criteria or not. They could have met criteria and not gotten ECMO all within seven days. And they used a PDAC cutoff of 100, even though EOLIA would be used less than 80. And they found a significant survival benefit, very similar to EOLIA outcomes, based on whether they got ECMO or not. My major issue with this study, as nice as it was, is that this is all done at centers that also performed ECMO. So it's like doing one at your own center that does ECMO and comparing them to non-ECMO patients was the reason they didn't get ECMO, even at your own center. So why not? Could they...

Is it simply a capacity issue or are there other factors that weren't captured in emulating this RCT as to why these patients weren't suitable for ECMO? And are those major factors in mortality? And so I think that leaves some things to be desired. But the French group, again, going back to Paris, they did an emulated target trial similarly of mechanical ventilation alone versus ECMO within seven days, a couple of different gas exchange criteria. They looked at 1,200 patients, 164 got ECMO. They did not see overall a difference in 90-day survival between ECMO or not. But when you break it down and you look at what these lines mean here, at high-volume centers versus low-volume, there's a big spread.

In fact, ECMO at low-volume centers seemed to be harmful, whereas ECMO at high-volume centers was the most protected. And it also appeared that there was, in subsequent analyses, that if they got ECMO really early, within four days of mechanical ventilation, had even lower PF ratios, they seemed to benefit most from an ECMO strategy. Now, the one that I was – I mean, I encourage all of you to take a look at this study. This is a slightly different statistical analysis that got published in BMJ and was actually shopped around for a while because I think a lot of other journals didn't really understand what they were doing, and BMJ was promoting this kind of analysis. It's very sophisticated. It's very interesting. I certainly missed it when I first –

kind of read through as far as what they exactly were doing. Basically, they're trying to do a comparative effectiveness and emulating a randomized control trial. But it's this design where you actually, it's all comers. Okay, it's not ECMO versus those who didn't get ECMO. But it's analyzing a strategy where one would get ECMO if they met certain criteria. And the way they censor them, you're basically then analyzing each patient kind of as their own control. That they were going down the road where they would get ECMO if they met certain criteria or they would not get ECMO if they never met those criteria. So it's encompassed, it's the COVID critical care consortium is the source of the data. So there's 30 countries, tons of patients across the world.

And it was analyzing a strategy of ECMO if your PDAF was less than 80 compared with conventional mechanical ventilation without ECMO, not just looking at those who got ECMO with PDAF less than 80. That's where I think you can make a mistake there. And they look at the primary optimum 60-day mortality. And so that's why these mortalities are so low, because it's not about ECMO alone. But basically, if you employ this strategy of triggering ECMO for a PDAF less than 80, it reduced the mortality by 7%, 26 versus 33, in this entire population that was analyzed, and with a relative risk of 0.78. So this might be, honestly, some of the most robust data to suggest that such a strategy if employed.

had a benefit for ECMO based on these kinds of cutoffs if you use them clinically. They looked at subgroups and looked at if there was a further benefit if you isolated certain subgroups. It certainly looked at younger age had a stronger benefit. Some comorbidities, regardless of the comorbidities, seemed that ECMO was better. And then if you look at kind of thresholds also, which, you know, at what point might you cut off the strategy, less than 10 days seemed to be the most effective. You could see this curve kind of like less than seven days is even stronger benefit. They looked at, I mentioned the primary outcome of PDAS less than 80. That's where it was most effective with that risk difference and risk ratio. But there was still a benefit if you used a cutoff of a PDAS less than 120.

but probably harm if you use the cutoff greater than 120 and that ECMO should not be employed with that kind of threshold. And likewise, if you look at driving pressure, which is a hot topic in ARDS and ECMO thresholds and management and strategies once you're on ECMO, it appears that ECMO is most effective when analyzing patients using a cutoff of a driving pressure greater than 15, which does seem to be one of those numbers that bears out in a lot of these observational studies on driving pressure. But that's where you make it the greatest bang for your buck with ECMO and ECMO strategy versus not. So if ECMO is effective, probably, there's still a lot of issues and challenges in implementing ECMO during a pandemic. You're balancing supply. You're balancing demand. And there's a lot of competing interest going on simultaneously for ECMO centers and non-ECMO centers alike. So this is a figure that we put out there kind of earlier in the pandemic. Like, what is ECMO use going to look like in the course of a pandemic? Well, you know, initially, there's a lot of patients. Everyone's got COVID ARDS. Lots of patients meet criteria. ECMO volume should play up, right? Everyone's excited about the prospect of saving people with ECMO. But then you get into the surge range where you're overwhelmed, right? There's more patients than beds.

taking care of one ECMO patient might be the equivalent of being able to have the staffing and resources to manage three other patients. So how do you balance those demands? And so it was too much of a surge. Your ECMO excitement goes down because you really have to balance all the other strains on your system. And then as patient volume goes down, well, maybe that gives you more leeway to do ECMO, or there's just less patients who need ECMO. And so that goes down. And one thing we're observing, and I'd be curious at the end of this, if you guys are observing the same pattern, we're seeing, now that the pandemic's, I don't want to jinx it, but you know, the volume's down, we'll say receding.

We're not getting a lot of ECMO calls for non-COVID stuff. And I don't know if that's a consequence, but maybe over all of this, out of necessity and also appropriateness, centers are learning to prone better and more, and there's just less need for ECMO because proning works. Maybe there's just more small ECMO centers popping up. People are doing ECMO because there was a demand for it during the pandemic, and they're not calling for, you know, referral to major ECMO centers. But there definitely seems to be, and others have observed this too, kind of asking around, that there's a little more of a lull in ECMO volume. And so maybe that's part of this green line. Maybe as people get better at managing ARDS.

and aren't calling for rescue ECMO, that there's just less of a need for ECMO in general. So how do we now, knowing everything we know, and we spend just a little bit of time towards the end of this, how do we prepare for what I hope is not later phases of this current pandemic, but maybe the next pandemic? And to be honest, I think we just have to assume there will be another at some point. And what have we learned so far? So this comes from that review that was mentioned at the beginning of the introduction. So feel free to take a look. We described some of this stuff in greater detail. But certainly, if something first starts to emerge, what Paris does so well, what Chile learned to do pretty quickly.

what I think the U.S. lags in terms of too many just kind of isolated hospital systems and other competing interests in terms of coordination, but coordinating resources across regions, preparing for the possibility that you're going to need to share resources, transfer patients around to balance, and, you know, shunt them to high-volume centers if they can handle it, you know, consolidate where possible, and also prepare from a resource standpoint, supply chain, especially if you're talking about ECMO, which is a relatively high resource, you know, resource-intensive therapy. understand the disease, right? The mechanism matters. If this is something that's going to cause a ton of myocarditis and very little ARDS, well, you're going to have to prepare for a lot more potential VA ECMO and a lot less BV ECMO. This happened to be a predominantly respiratory disease with a little bit of cardiac involvement. How is it transmitted? You know, what organs are affected? Is this going to be something like Ebola where it's a hemorrhagic disease and ECMO is probably the last thing you want to use? Or is it something that's right up the alley of, you know, what we're used to using ECMO for? Defining the scope of the problem and analyzing and coordinating it is a huge barrier. And it's something that argues now more than ever for more automated data collecting and coordinating that data reporting into central...

databases and registries that people can draw from and start to analyze and I think we really need to step up our game and be able to analyze it because I mentioned to collect it because the centers that are going to be the most you have the most data report are also the centers that are dealing with the highest volume of patients and having to balance those those interests and if you don't have dedicated people to do that or an automated system you're going to lag behind and people are not going to have the most updated information um but I think you know is the most, probably the most interesting, and as I highlighted with the Erner study and a couple of the other strategies of analyzing this, understanding the efficacy and the effectiveness, right? Efficacy, you know, kind of mechanistically what works and effectiveness in real world, how it's applied, really prompting early studies and trials. And I think some of these emulation trials, like the strategy of analysis, the methodology of the Erner study, and a couple other platform designs, like we met CAP did. There's a couple of guys, the Toronto group is working on a larger version of a platform design trial for ARDS to be able to plug patients in, you know, with multimodal and different strategies being analyzed all simultaneously. I think that's probably the way for the future.

registry randomized trials, where you can actually kind of trials embedded in registry data. And one that I didn't mention, but it was an idea that was suggested, and maybe in the most extreme of circumstances, is if you have a lottery system, because there's resources that are so strapped, you literally have to just flip up a coin, randomize people, you know, by pure lottery. You can weight it. You can say, this is a vulnerable population. I want to give them a higher probability of getting the treatment. But if you want a lottery system in place for ventilators, or in this case for ECMO, it's a natural randomization.

You can study those that got and those that didn't simultaneously. I think no matter what you study with ECMO, you want to and have to study a simultaneous population who's not receiving the treatment to know what's happening with the natural course of the disease anyway, because if mortality is going up in both, then maybe you're just not capturing the full picture by only looking at ECMO. And so looking at those changes in time is clearly the COVID data has shown us is not static and not, you know, early experience does not necessarily reflect later outcomes. And then how to organize ECMO as a group, both locally and across networks and across countries, developing those models and those relationships now. I think still based on this case volume outcome relationship, trying to concentrate ECMO within experienced centers, it's not to say that you can't start new centers. In fact, there was one nice study where in South Asia, predominantly Middle East, South Asia, they started a bunch of new programs during the pandemic to try and accommodate the demand that could not be met. And they actually had pretty good outcomes. In fact, some of them were better than the more established centers. So some of you can't have a new program, but you have to support it appropriately and make sure the training and the resources are present. And again, it can't be at the expense of so many other patients that it's not, from a utilitarian standpoint, worth it. And simultaneously, based on the data that comes out.

you know, how do you, developing guidance, be it, you know, ethical committees, hospital systems, deciding when to split that switch. When do you go away from ECMO because the mortality is high and it's just not worth it and it's draining our resources towards maximizing, you know, the most patients you can care for and achieving the best overall outcomes in those most constrained times. So with that, I'd like to end it. It does leave us a few minutes, which is great. And I wanted to see, do you guys have any, you know, so if you have any experiences you want to describe or if there's any questions that came up over the course of this talk, I'm happy to answer them.

Beyond Hope Project with Jason Tharp Beyond Hope Project Welcome to 'Beyond Hope Project,' a beacon of inspiration in the podcast world. Join Jason Tharp as we embark on a transformative journey, exploring stories that awaken the will to start anew and find the strength to rise from setbacks. 'Beyond Hope Project' is about lighting the way to your starting line, not dictating your path. We delve into narratives that foster optimism and belief in a future where dreams become achievements. Each episode is a tapestry of positivity, enthusiasm, and comforting support, resonating with the essence of hope. This podcast goes beyond storytelling; it's a movement of positive change rooted in authenticity and transparency. We provide real tools for real journeys, infusing more hope and joy into the world. 'Beyond Hope Project' stands as a testament to the transformative power of hope and the limitless potential of the human spirit. Join us in transcending boundaries and breaking barriers. Here, we prove dai Veteran Chat Project Kyle Turner Our goal is to reach as many Veterans as we can. We noticed a lot of the podcast personalities tend to be unrelatable to the every day joe. Not taking anything away from anyone just hoping to give more people a place to feel connected to something. KZ and Kyle are Army vets who deployed together for OIF 07-09 and OND. Now close to 15 years later we are finding a way to open up about our issues, and find a way to build a community. Paramedic Philanthropy Project Melissa Ade Podcast by Melissa Ade Bigfoot Information Project Podcast Bigfoot Information Project Occasional conversations with leading researchers and investigators regarding the animal commonly referred to as bigfoot.

Frequently asked questions

How long is this episode of Maryland CC Project?

This episode is 50 minutes long.

When was this Maryland CC Project episode published?

This episode was published on January 9, 2023.

Is there a transcript available for this episode?

Yes, a full transcript is available for this episode. You can read the complete transcript on the episode page.

Can I download this Maryland CC Project episode?

Yes. Use the download control on the episode player to save the publisher-provided media file.