Thanks for the kind introduction and invitation, Whitney. So I'm glad that I could be here with the world experts. So I'm going to talk about the efficacy of ECMO for severe respiratory failure. I feel like much of this likely will be a review, but please feel free to interrupt with any disagreements or things that you don't like here. My disclosures won't be speaking about any specific products today. So we're talking about extracorporeal support for acute respiratory distress syndrome. That's really what we're talking about when we say severe respiratory failure. In 2023, our main indications, evidence-based indications at least, for ECMO therapy is really in the severe ARDS patients. We'll talk a little bit about other populations where we're considering things, but that's really what we're thinking about. And of course, fresh and I'm sure everybody's mind in this audience is the deluge of severe ARDS patients that we all just had in the global pandemic from COVID-19 and many, many jurisdictions being really overwhelmed with the number of COVID-19 patients, and some of them very severe and requiring ECMO support. The story of efficacy is actually, again, not a new one because the device is actually not very new. We've had ECMO type technology since the 60s and 70s. It was born with the advent of cardiopulmonary bypass work in Philadelphia from John Gibbon and colleagues. And really this device has been around and really seen as an alternative to mechanical ventilation really, especially in severe cases. The whole idea of the pioneers was that now we have this amazing device where we might be able to save patients who maybe they don't need intubation, maybe they don't need prolonged mechanical ventilation. And certainly we won't need to crank up the amount of ventilatory support in patients where it doesn't seem to be working. And this device might be the magic intervention that we could do, especially in these difficult cases. So fortunately for us, these pioneers, many of them listed here, led by Warren Zaple in Boston in the 70s, decided to launch a clinical trial to really understand if this idea or hypothesis about ECMO replacing mechanical ventilation, especially in very severe cases, would work. And so with the support of the National Institutes of Health here in the United States, they conducted a trial where they randomized 90 adults with severe acute respiratory failure, which at the time would have met the definition for severe ARDS, to getting ECMO with mechanical ventilation or mechanical ventilation alone. And here's what you see is the survival curves, not mortality. Survival was only 10% in both groups, so pretty dismal survival, 90% mortality. So this clearly, unfortunately, dulled enthusiasm for ECMO in adults for a very long time because it told us in the 70s and early 80s that if you're an adult with severe acute respiratory failure, one, you're in big trouble, and two, ECMO wasn't really going to help you. So this really, really stopped enthusiasm, except for in very specialized centers around the world that were doing a lot of the pioneering work on the devices, the technology, understanding the physiology of how it interacted with the lungs, the heart, and this sort of thing. So it's kept going in centers like Ann Arbor, Michigan, in Milan, where Luciano Gattinoni was, at NIH, where Ted Koloboff was, and this work continued. And then fast forward to the, if you will, the previous pandemic, the H1N1 pandemic, which came in 2009, where we, again, saw what we just experienced in COVID-19. A lot of reasonably young, healthy patients, some of them pregnant, presenting to the ICU with very severe hypoxemic respiratory failure and failing sort of best conventional mechanical ventilation and adjunctive therapies. And so we had very prominent case series like this, and this is from the Australia and New Zealand group publishing that among 68, okay, maybe it doesn't work, but 68 patients who received ECMO, this is a case series, they had 71% survival. And at the time of publication, some of these patients were still in the ICU, so at the end of follow-up, actually, there was 75% survival. So 10% survival in 1979, 75% survival in 2009. So again, not all due to ECMO alone, this is a case series, very difficult to make firm causal inferences from this kind of study, but certainly suggested that improvements in general critical care or understanding of ARDS, how to ventilate patients, and again, this is a population of relatively healthy patients who have very few co-morbidities at baseline, that ECMO could potentially be very efficacious in this kind of situation. So again, the idea was we need to prove this in a randomized control trial, and our colleagues in the UK, led by Giles Peak, tried to do just that with the CSER trial. So again, you're likely familiar with this study, where they tried to randomize 180 adult patients, again, with very severe ARDS, to getting considered for ECMO therapy versus getting best mechanical ventilation at the place that they showed up to in the UK. The important thing about the CSER study is to understand at the time that it was conducted, there was only one adult ECMO center in the whole of the UK, which was in Leicester, where Giles was working. And so in this study, if you got randomized to ECMO, what happened was you got transferred from whatever center you showed up at, down to Leicester, to be considered for ECMO. Their primary outcome was six months survival without disability, and what you can see here is the Kaplan-Meier curves, and indeed, there was a significant improvement in six months survival without disability. But of course, the challenge here, just like any sort of record, like Barry Bonds having an asterisk beside his home run record, or these sorts of things, that you don't want to have an asterisk on your randomized control trial, okay, and you could see that little asterisk in the ECMO. And the challenge here, of course, is that of the 90 patients that got referred to Leicester for ECMO, about a quarter of them never went on to ECMO, because they got better by other means. And again, for you in the audience who work at these kinds of referral centers, you probably saw this in COVID as well. You get a call, very severe ARDS patient, very refractory hypoxemia, send the patient to my center for consideration of ECMO, but then they arrive and maybe you tweak the sedation, you tweak the mechanical ventilation, you add inhaled pulmonary vasodilators, maybe you try something, they haven't proned yet, you try proning, and the patient improves and doesn't need ECMO. So this was less a trial of ECMO versus no ECMO, than it was a trial that showed the benefits of referring patients who could be considered for ECMO to an advanced center that could provide it if needed. So the naysayers sort of said, well, this doesn't really prove that ECMO is the thing, but if you send the patient to an expert center where they could get ECMO, that's really the way to go. So finally, I think the idea was then to improve upon the lessons learned from CSER and to finally hopefully lay to rest this idea of whether ECMO, VV ECMO, was useful in patients with severe ARDS was the EOLIA trial. So on the left, led by our colleagues in Paris, by Alain Combe, and we were fortunate to be a study center in Toronto for this trial. And this trial set out to really answer the question of, does VV ECMO benefit patients who have severe ARDS? Everything was protocolized in both groups. Expert and experienced ECMO centers were the study sites for this trial. The plan was originally to enroll 331 patients in this study, but unfortunately the study was stopped early after 249 patients were randomized. Here you see the Kaplan-Meier curves, and again, the primary outcome was 60-day mortality. So there's an 11% absolute risk difference in mortality in favor of the ECMO group, but of course the almighty p-value is 0.07. So no time in the 15 minutes today to talk about the wisdom of dichotomizing trial results by a p-value, but again, you know, most people, not most, I shouldn't say that, people maybe who had prior beliefs that ECMO worked found this trial, confirmed their prior beliefs that it worked because the primary outcome was in favor of ECMO, and in fact, most of the secondary outcomes were also in favor of ECMO. There was even less stroke in the ECMO group in this study. NACE errors, and I was sitting beside my former colleague and boss, Roy Brower, when these results were released at ATS, who said, we need another trial because the p-value is negative and the trial is negative. So what you do when you don't have the result that you want in the primary study is that you look for secondary supporting data. And so on the right, what you see is that since that time, we've conducted a number of secondary analyses that all support in a statistically significant way the benefit of ECMO for adults with respiratory failure. So one was this Bayesian re-analysis of the trial led by my colleague, Ewan Gallagher in Toronto, that showed across a spectrum of prior beliefs from very pessimistic to very optimistic. There was over a 90% probability of having an improvement with ECMO therapy in severe ARDS patients. We did a traditional study-level meta-analysis led by my colleague, Levina Munchy. Again, when you combine it with CSER, as well as some quasi-randomized H1N1 studies, you get a statistically significant result for ECMO. Alain and Giles led an individual patient data meta-analysis, also significant benefit for ECMO. And finally, there's now been three network meta-analyses that all show that ECMO is actually amongst the most beneficial interventions for ARDS, along with prone positioning. So all these other syntheses of the available data tell us that VV-ECMO is useful. And now, finally, this has been codified in the most recent ESICM guidelines that were just published earlier this year, that now switch from we couldn't recommend in 2017 for ECMO because EOLIO was still ongoing. And now there is a strong recommendation that patients with severe ARDS should receive VV-ECMO if they meet the inclusion criteria. I can say that we've actually wrapped up an update to the ATS guidelines that also will provide a positive recommendation on this question, so stay tuned. One other thing to think about is obviously the economic evaluation of VV-ECMO. This is a very resource-intensive and costly therapy not available at all centers here even in North America, and certainly when we think about the world in general. This is a nice study done by one of my former graduate students, Callie Barrett, when she was in the UK, showing that actually it's pretty cost-effective, especially for younger patients. So what we have here is to summarize this thing called the ICER, or the incremental cost-effectiveness ratio, of $36,000 Canadian dollars, which is probably a couple pennies in the United States. But for every quality-adjusted life year you save, it costs about $40,000 Canadian dollars. And when we think about health savings, in the United States at least, in this context, anything that costs $50,000 to $100,000 for quality-adjusted life year would be considered cost-effective. This is actually more cost-effective than renal replacement therapy for chronic renal failure, just as an example, okay? So if you're saving the right people, actually the amount that you spend to save their lives, it's actually quite cost-effective. Finally just to say that there won't be another trial, because the challenge is that it's very hard to do these trials. So this is the best data that we have. So as clinicians at the bedside, we have to use this data to make our decisions about feviacmo. So feasibility is the major challenge. CSER took almost a decade to complete, enrolling patients at this rate. EOLIA slightly better at six years, twice the rate. And over the years, even though they're randomized trials, you can still get secular trends, right? So PROCEIVA was published in the middle of EOLIA. We had to change the standard of care to include prone positioning, because the evidence had changed. So there's this bias, even though these trials are randomized, that they take a long time to do. And if you wanted to confirm the 11% absolute risk difference we saw in EOLIA, you'd need over 600 patients. And at these traditional rates, even with 100 units, it would take nearly a decade to complete. So these are the data that we're going to have. And again, now our guidelines, at least based on this data, are recommending the use of ECMO. We shouldn't rely on randomized control trials alone, especially when they're very difficult to conduct. During the COVID-19 pandemic, there was a call for a moratorium on ECMO use, because early in the pandemic, mortality rates in certain jurisdictions in China where they're putting patients on ECMO was reaching 90%. And people were wondering if COVID was something different, maybe we shouldn't be giving ECMO. We need more data. We couldn't organize in the face of a global pandemic to do an RCT, so we had to do the next best thing. And now, using high-quality observational data that was collected all around the world, we can emulate a randomized control trial when you can't perform the trial that you want to do. So we do this target trial analysis. And indeed, in COVID-19, there were three different groups that did that. One here in the United States, the Stop COVID Consortium, one that was done in Paris, led by Matthew Schmidt and Alain Cohn, and one that we conducted in conjunction with John Fraser using the global COVID-19 Critical Care Consortium. And just to say that when you do these emulations with very sophisticated results, you can get things that are very consistent with randomized control trials. So the nice thing here is that when we looked at ECMO for COVID-19, we saw an absolute risk difference of about 7% in favor of ECMO, so pretty close to the 11% that we saw in EOLI. We have very consistent results across subgroups, very consistent with the prior literature, suggesting that, again, when you can't do an RCT, sometimes an emulated trial is important. And then finally, when we're thinking about efficacy, we shouldn't just think about the short term. This has been a challenge for us in the ICU world over and over. I remember as a trainee, if anybody left the ICU, we were patting ourselves on the back because it's another one that we've saved. They've left the ICU. And now we recognize, of course, that leaving the ICU is just the beginning of the journey for these patients, especially some of them who leave with very significant and persistent morbidity. This is just one example, again, from my colleague Carol Hodgson in Melbourne, showing that importantly amongst patients, if you look at these plots on the right here, sorry we don't have a pointer, the one that's labeled A, these are so-called alluvial plots, what you can see is that many patients who started off with no disability, a good proportion of them, if you focus on the tangerine-colored lines, they go from no to moderate to severe disability. So these, again, we select young, healthy, few comorbidity patients to get onto ECMO. And some of these patients who start off with no disability at six months are leaving with moderate to severe disability. So we need to really think about, one, are we selecting the right patients? And two, is it something that we're doing to them in the ICU when they're on ECMO that's leading to this morbidity? We need to understand that better. I want to finish, if I have time, with a few things about what's next. Because we talked a lot about ECMO for severe respiratory failure. But really the key thing that we've learned is that it's not the life-threatening hypoxemia where the big benefit of ECMO seems to be. If you look at the supplement, nobody ever looks at the supplement, but if you look at the supplement of this study, what you can see is actually the group that had very severe hypoxemia, PF ratio less than 66, there's really no difference between the ECMO groups. Mortality was about 42%, 43% in both groups. But the less hypoxemic subgroup, they really benefited from ECMO therapy. And interestingly, it's the patients who had high driving pressure at baseline that seemed to benefit the most. Again, because one of the main benefits of being on ECMO is that we could dramatically reduce the intensity of mechanical ventilation once ECMO is taken over for the native injured lung. And we can reduce the ventilator-induced lung injury. And we think that translates the clinical outcome that's better for these patients. So of course, then the idea, this is something that one of Dr. Eggerstein's colleagues, Daryl Abrams, put together in York. So the idea is that maybe our intent should be to come up with a device or a system where we can reduce intensity of mechanical ventilation, reduce or improve the concomitant hypercafnea or respiratory acidosis that's going on. And maybe it doesn't have to be flow-blown ECMO in patients who don't have severe ARDS, but that juicy group in the middle of moderate ARDS, which is 50% of ARDS patients. If we can reduce ventilator-induced lung injury in these patients with either something like high-flow CO2 removal or low-flow ECMO, sort of call it what you will, but maybe that's the group to target, because maybe we could have the biggest bang for our buck there. Of course, somebody tried this, the REST trial, again, just thinking about in the last year, led by James McNamee and Danny McCauley, they tried to do this with a low-flow CO2 removal device. And again, if you focus on the Kaplan-Meier curve in the top right side, not only did it not help, it actually led to harm. Less ventilator-free days, about the same mortality, many more complications, statistically significantly more intracranial hemorrhages. So even though we reduced tidal volume from six to four mils per kilo and reduced driving pressure by two centimeters of water, it didn't quite work. Maybe we need to find the right patients again. We now understand heterogeneity of treatment effect. This is work led by one of our graduate students, Jose Diante, showing that maybe it's patients who have high dead space, as shown by higher ventilatory ratio, and the less hypoxemic patients that we need to target. These are the patients where the benefits might outweigh the risks, and we target the right patients, we might see that work. So we're going to look at some of this in a trial that we've just started screening for in Canada called the ULTIMATES trial. And again, here we're using mid-flow ECMO, not a CO2 removal device, to see if we can facilitate a reduction in driving pressure as the primary target to see if that would improve ventilator-induced lung injury. And the idea, perhaps even beyond that, is to shift our focus, just like I said, from beyond short-term mortality to this idea, can we get patients to have less morbidity? If they're leaving with severe, moderate to severe disability, we need to do something different. So maybe we need to use ECMO earlier to facilitate the awake, calm, cooperative, and mobile patient to reduce morbidity in these patients. Just to finish quickly, sorry for going over, so ECMO for severe ARDS, including COVID-19, it's not if anymore, but when and in whom, we want to avoid creep and continue to use our evidence-based criteria for candidacy. We're very interested in this idea of whether extracorporeal support can further reduce ventilator-induced lung injury in certain special populations, maybe enrich on the population most likely to benefit, maybe shift our focus from short-term mortality to longer-term morbidity. To paraphrase Winston Churchill, when the facts change, I can change my mind, so let's wait for data, and the data is coming. So thanks very much. Happy to take questions at the end with everyone. All right, Eddie, thank you for that great talk. So we're going to switch gears a little bit and go to a largely data-free zone here, which is a little unfortunate. I have nothing to disclose except the fact that you're going to listen to a lot of my opinion for the next few minutes. So again, I'll skip this slide here. So we're going to go through three different vignettes and talk about less common indications for which ECMO may provide an improvement in morbidity and potentially mortality. So the first will be a vignette looking at high-risk airway procedures, so foreign body aspiration for children and other indications for adults. And then for the surgeons and anesthetologists in the room, we'll look at different forms of cardiopulmonary support for supporting patients undergoing lung transplantation. And then we'll end with more lung transplantation in terms of how do you support somebody with primary graft dysfunction. All right, so case one. So this is a true case that happened at Vanderbilt, and this was actually published. So this is a 65-year-old with a prior history of tuberculosis, had esophageal stenosis that had been managed with a stent for 20 years prior. So had been doing well for a couple of decades, then had paroxysms of coughing, and then acutely thereafter became short of breath. So she was admitted to the emergency department, and you can see the CT here. I guess the pointer is not working. But the esophageal stent is now firmly cemented in the left main stem bronchus, and you have complete atelectasis of that whole left lung. This is the bronchoscopic view below, where you can see that the stent is nicely lodged in the main carina, complete obstruction on the left, and you're basically aerating on the right. So clearly, this stent needs to be removed, but you can anticipate sort of a hemorrhagic bloody process in doing so. So the question is, is how do you best support this patient during that procedure? So of course, this is an ECMO talk, and so we preemptively put the patient on ECMO, removed the stent, patient did great, and then was ultimately decannulated. So because this is a session on evidence for this, what is the evidence of using ECMO to support patients during high-risk airway procedures? And the answer is not much. So there are a few case series. This is one published by the Vanderbilt Group, so John Stokes is a surgical fellow. And this was a case series of two years of experience when our ECMO program first started. Two patients in this had PAP, pulmonary alveolar prognosis, requiring whole limb lavage. The other six had central airway obstruction, the majority due to malignancy. The primary outcome of this was survival to decannulation, and so if you just draw your attention to the entire cohort there on the left column, you can see that survival to decannulation was 100%, survival to hospital discharge was close to 90%. So showing that we can safely manage these patients and at least keep them alive to hospital discharge and potentially beyond. Now will there ever be a randomized controlled trial on this? No, right? And nor should there be, right? You can select patients who you think you can support and hopefully decannulate by the end to get them through a procedure and give them added time. Same is true for children. So I actually didn't know this. Are there any pediatricians here in the audience? Okay. Well, I didn't know this, that a major cause of morbidity and even mortality in children less than three years is foreign body aspiration. And so supporting them through the process of extricating whatever foreign body is there is really important. And this is evidence showing that we can use ECMO to support them through. So this was a case series of four patients, but in addition, they also looked through the ELSA registry and then also did a literature review. So in this series, there was over 50 patients, and again, for the sake of time, I'll just draw your attention to the last row. And the total to hospital discharge was at least greater than 85% in all three of those sub-cohorts. All right. So let's move on to case two. So for any of the surgeons here in the room, and really any of the medicine folks, too. So this is a common occurrence on our service, especially at Vanderbilt, where our transplant program's booming. So we have a 63-year-old gentleman with IPF and moderate severity group 3PH. He's active on the lung transplant list and now has an offer. So put yourself in the shoes of the anesthesiologists and the surgeons. Question is, how are you going to manage this patient? So how are you going to support them through this? So I'm not sure if most of you know, but when we do bilateral lung transplants, the majority of the time, it's bilateral sequential transplants. So you're explanting the native right lung first, generally. And so that patient, without any support, would require gas exchange just through that native left lung. You can imagine if you have pulmonary hypertension, the RV is not going to be happy. So in most circumstances, you need some form of cardiopulmonary support. Most people nowadays won't recognize this machine. It's the old-fashioned cardiopulmonary bypass machine. As you can see from the cartoon, and I know it's hard to see that, but just appreciate the complexity. And the major thing to the bottom left hand is a venous reservoir, which allows you to really control your cardiac output and your systemic blood flow. But that's an open system. And there's a large interface where air meets blood, which then causes activation of pro-inflammatory mediators, as well as pro-coagulation cascade. And so these patients need to be anti-coagulated. Compared to what's becoming more popular these days, which is performing transplant procedures on VA ECMO, this is a closed system, no reservoir. You still need to be anti-coagulated, but not to the same degree. And so because it's safer, it's come into favor. So you actually have two more options, right? So intra-op, you could support the patient off-pump. And this is still a popular approach in many ORs, right? Because you don't need anti-coagulation. If you can support the patient through, there's less bleeding and less rates of primary graft dysfunction. Cardiopulmonary bypass, I think I spoke mostly about the pros and the cons. And VA ECMO really balances both of those worlds, right? So you have the support that you can get with cardiopulmonary bypass, really without the rates of bleeding and hemorrhage that you have with cardiopulmonary bypass. The Canadians are actually exploring, and others, are exploring the use of VV ECMO intra-op, which seems a little scary to me because you don't have the hemodynamic support. But the convenience of this is that the majority of patients are bridged to transplant with VV ECMO. And so you don't have to reconfigure your cannulation. They have really good surgeons in Toronto. All right. So this patient's a 63-year-old. We've decided to transplant him. We went for the VA ECMO approach. After the case, we decannulated him. He's doing OK. It was a little bit of a messy procedure, so it required some blood products and some crystalloid. Six hours later, this is what your chest x-ray looks like. Your PDF is 150, and you need plateau pressures of over 30 in order to maintain your pH greater than 7.2. Well, what do you do now? Is anybody going to paralyze this patient? Can I shake her head vigorously? No. Right. So we're going to use ECMO, right? So the question, though, is when do we initiate ECMO support? And there's a big question mark surrounding this. Different groups vary. So in our group at Vanderbilt, we'll initiate if the PDF is less than 200, especially if your plateau pressures are greater than 30. But other groups will adhere to the EOLIA criteria and really wait until the PDF is less than 100 or even less than 80. And sometimes programs will do the whole kit and caboodle of deep sedation paralysis, proning. But these are fresh transplant patients, right? We really want them awake and ambulatory, ideally. Putting them on ECMO allows you to adhere to low tidal volume ventilation, right? We know this works via the ARMA trial for ARDS. This is work from the Wash U group just published 20 years after the ARMA group that shows similar things. So the blue line is patients where lung protective ventilation was adhered to. The red line is where it wasn't. The x-axis is looking at different types of matching in terms of donor recipient height, which I won't get into that. But you can see that for every ratio of total lung capacity, adherence to lung protective ventilation led to improved one-year survival. So adhering to that, putting people on ECMO earlier so that you can adhere to lung protective ventilation is really important. All right, so timing. So we talked about mechanics in terms of when to put somebody on and P to F ratios. How about timing? So this is work from the Pittsburgh group. It's a case series over 15 years that basically showed, just for the sake of time, I'll go through this quickly. But the only major factor in terms of improving survival is if you cannulate these patients within 48 hours. And that leads to improved survival, not only for hospital survival, but also one-year and three-year survivals. So we know earlier is better. One other question is, does configuration matter, right? So if you were to put these people on VA ECMO and decrease RV preload, decrease perfusion through the lung, can you decrease the rate of ischemia reperfusion injury? This is an older study, right? So it's now over 15 years old. Half the patients, or two-thirds of the patients were placed on VA ECMO, a third were placed on VV ECMO. But it wasn't really apples to apples. These were the outcomes, though. So for those placed on VA ECMO, 30-day survival, 0%. Those placed on VV ECMO, 100%. I think the other surprising thing is the degree of CNS dysfunction for those placed on VA ECMO. So a lot of bleeding, a lot of hemorrhage. All right. So let's take one step back. So let's assume after the case that we didn't decannulate the patient, that we left them on. Can you keep somebody on VA ECMO for a longer period of time and allow the left ventricle, the left atrium to re-equilibrate? And can you diurese the patient and optimize them and therefore prevent the development of primary graft dysfunction? One other thing to remember is that these patients going into transplant, a lot of them have severe pulmonary hypertension. So what happens during that? Well, you have septal bowing, pancaking of the left ventricle, chronic underfilling. And actually what happens is you have atrophy of your cardiomyocytes and loss of contractile proteins. So all these patients develop problems with relaxation of their left ventricle. Well, after transplant, when the lungs now have improved pulmonary vascular resistance, the left side is going to receive a massive bolus of fluid. And that's a good recipe for pulmonary vascular congestion and result in PGD. So, what the group from Hanover, Germany showed several years ago is if you keep these patients on VA ECMO for at least five days and monitor their pressures, at the end of the five days, you will hopefully have remodeling and an improvement in mean PA pressures, ejection fractions, cardiac output, and parameters of LV diastolic function. And that will lead to an improvement in primary gaft dysfunction. So, to the right are people supported with ECMO for five days, to the left, not. And you can see at 24, 48, and 72 hours, incidence of PGD is decreased. So, hopefully, a little whirlwind tour showing that besides ARDS, you can use ECMO for high risk airways intraoperatively for lung transplantation and then to manage your patients who develop subsequent primary gaft dysfunction. Thank you very much. Hi there. Great. Well, thanks so much, Whitney. And for the next 12 minutes or so, we'll talk about another area of a lot of uncertainty in transfusion and anticoagulation practices in fetal venous ECMO. So, this is really one of the unknowns of ECMO. In a field of a lot of unknowns and areas of evolving research, transfusion and anticoagulation practices certainly are two of the biggest ones. So, let's start with transfusion. And really, the conclusion could actually be the first slide, is that the ideal transfusion threshold for ECMO remains uncertain. But now, let's try to contextualize that a little bit and look at what we know and how we know it. So, we know that blood transfusion from ECMO's inception has been a very commonplace practice. This is just the first successful case of ECMO for ARDS done in the early 1970s, and you can imagine why transfusion was so common, right? Look at all that tubing, like those miles of tubing, that huge oxygenator or gas exchanger there, okay? And it had a lot of complications, a lot of hemolysis, a lot of bleeding problems. So much so that you can see some of these early ECMO studies show multiple units of blood being transfused per day just to require normal hemoglobin, two units a day, three units a day, even up to six units per day of blood transfusion has been reported in these early studies. Yet, despite so many technological advances that has allowed ECMO to succeed as it has in our current era, bleeding does remain the most common complication and independently associated with mortality in both VA and VV ECMO, okay? That being said, modern experience has shown that despite this still common complication and common practice of transfusion, those requirements have decreased. The ANZ study that Eddie mentioned, you know, 1,800 mils total of blood transfused in these patients with H1N1, less than one unit per day, down from two, down from three. A small series that we published several years ago, about one unit of blood transfused total per the entire ECMO run, about 0.1 units per day. And just during the early kind of days of COVID, the Canadian critical care group, oops, excuse me, sorry, EOLIA trial had smaller amounts of bleeding as well. So about 46% of those patients required any transfusion, down from the 100% it would have been years prior. So what are the transfusion guidelines in ECMO in this study? Well EOLIA didn't mandate a certain approach, given this is an area of unknown, but they did recommend a transfusion strategy of hemoglobin of seven to eight grams per deciliter, tolerating up to 10 if somebody was severely hypoxemic with signs of end organ dysfunction. Also currently recommends a transfusion guideline, a restrictive approach to transfusion, though it's not well characterized what restrictive means. They suggest maybe as low as seven, but upwards of 12 may be considered acceptable. And recently a group of Canadian critical care societies recommended hemoglobin threshold of seven to eight, or seven to 7.5 grams per deciliter. So where does this come from in this restrictive approach to transfusion? Well that's extrapolated from what we know for non-ECMO patients in the ICU, right, from TRIC trial, now over 20 years published, where we know that lower restrictive transfusion thresholds are associated with improvement in outcomes in our critically ill populations. And what's the concern with transfusions in this context? Well it's not like we're just replacing our own blood with blood of ours that was stored somewhere else, right? We're transfusing packed red blood cells, right? And there are a lot of problems that occur with that, which is why there is so much debate and consideration. So packed red blood cells undergo biochemical and structural changes as time goes on, right, that have been associated with increased mortality, worsened ARDS, volume overload, trolley, poor wound healing, okay? And it's been these number of transfusions, not baseline hemoglobin, that is most strongly associated with ICU mortality, even extending to very small numbers of transfused blood, that they consider to be discretionary transfusions, or a total of one to two units, okay? Most recently in the ECMO literature, we have a large cohort study of over 600 patients at 41 centers, global centers for ECMO, over a three-year period, so 2018 to 2021, during notably COVID times, and this was really trying to characterize what transfusion practices were in the modern era, with modern technology using overall modern approaches. Now what we see in patients who were transfused versus those who did not require transfusion, those patients who required transfusion had higher REST scores, so possibly sicker in certain ways. They had longer pre-ECMO hospital stay days. They had longer pre-ECMO ICU stays, so maybe they had been phlebotomized already. With that, they had lower starting hemoglobin, not surprisingly, okay? And you can see that there, the hemoglobin starting point was 10.4 in those that required transfusion, 12 in those that did not, okay? Outcomes for this study were that hemoglobin at the time of transfusion, kind of in these global ECMO centers, was about 7.9, so really favoring this more restrictive, modern approach to transfusion, and the overall hemoglobin during ECMO was 9 grams per deciliter, okay? So definitely lower than that traditional recommendation of normal hemoglobin. Overall these patients were transfused about 115 mils of blood per day, okay, on ECMO run, and you can see here in their overall mortality about a 60% survival. Now importantly what they noticed, and what they did know, was that there was no association with mortality until that hemoglobin reached below 7, so in this group. So transfusion didn't seem to benefit these ECMO patients until they got to this extreme degree of anemia, and interestingly, if you read some of the fine print, if the transfusion did drop to 7, not even all centers transfused at that point, right, only about 71% of patients were transfused even on those days where hemoglobin was less than 7. So we look a little bit further about what's going on kind of during these ECMO runs, we see that the greatest decrease in hemoglobin in these patients is precannulation, pericannulation, likely some surgical-related blood loss, effects of the circuit and whatnot. We also see that higher volume ECMO centers, those with more experience, are the ones tolerating lower levels of hemoglobin throughout. Maybe they've seen that they can actually have successful runs at more restrictive approaches to transfusion compared to lower and mid-volume centers. When you look further to characterize these bleeding events, what we see from Protecmo is similar to what's been reported in the ELSO registry data, which is a registry of centers around the world that submit ECMO data to ELSO. We see that cannulation site bleeding is really the most common, there in the sort of red green bar there. Also early on, we see some intrathoracic bleeding, okay, otherwise, oropharyngeal bleeding, et cetera, kind of goes on, but that cannulation site bleeding, especially early in the run, is where the majority of bleeding events come from. Most patients who had hemorrhagic events had one episode, okay. And so the critical considerations here are really that, as Protecmo shows us, we need to treat the patient, not the number. It seems like a restrictive approach to transfusion can be successful, provided that patient is otherwise well-supported. But there are some considerations, and that's patients with signs of hypoperfusion may benefit from a higher hemoglobin target. Let's say that hemoglobin is 7.2 in a patient, yet their lactate's rising. Maybe they're morbidly obese, maybe they have an extremely high cardiac output. There are cases where transfusion may be appropriate at different levels. However, patients with, and there are times where patients may benefit from less or more restrictive approaches to transfusion. Perhaps that pre-transplant patient, some of the patients that Neil might be taking care of, shouldn't be transfused at a hemoglobin of 6.9 or 6.8, right, if they're ambulatory and walking around and have no signs of hyperperfusion. Maybe the risks of transfused blood there that are going to narrow the pool of lungs from which they can draw is actually outweighing the potential benefit in an otherwise well-supported patient. So one approach that we have used was incorporating a restrictive approach to transfusion in the context of a larger blood conservation protocol. So we transfused at a hemoglobin only less than 7, unless there are other signs of hyperperfusion or these things I just mentioned. We also combined that with a low-dose anticoagulation approach, auto-transfusion of circuit blood at the time of decannulation to preserve all that blood that's in the circuit, okay. So hemoglobin transfusion threshold of 7 grams per deciliter is going to be acceptable for patients receiving VB ECMO without increasing mortality, as far as we know, and efforts to maintain physiologic hemoglobin levels during ECMO should not be done routinely. So moving on to anticoagulation, this is going to be another area where there are more questions than answers, and perhaps even more so than in the transfusion area, okay. We are going to really have to be conscious here that the practice is extremely highly variable. There's limited evidence, and the evidence that exists is only based on a small number of RCTs, okay. So I'm just going to start with basically the conclusion first, which is that it is unknown, but we'll go through two very recent society recommendations. So most recently, the International Society on Thrombosis and Hemostasis in looking at approaches to ECMO and coagulation, a very long, very long, very well-written document based on the literature that's available for review, okay, over PubMed searches and whatnot, they come to the conclusion they recommend heparin as the agent, they recommend anti-10A as the target, and they recommend a low-dose approach to anticoagulation, okay. Now remember, though, this is based on literature that's out there, and that our practices, our circuits, all these things have changed so much over time, okay. Also, recently published their guidelines as well, and they basically have said that there's insufficient evidence to guide optimal management, sort of reflecting this ambiguity of which I was mentioning. But they say the appropriate choices might be heparin versus DTI. We recommend a tailored approach, really considering the patient and related factors, whether you're looking at things like PTT, anti-10A, et cetera, but again, also recommend low-dose anticoagulation. And let's not forget, we just have been emerging from the years of COVID pandemic, right, where there have been all these questions raised about the hypercoagulability and effects, particularly in ECMO, of some of these patients who maybe have noted to be at greater increased risk of thrombosis, potentially need more levels of anticoagulation. So why is anticoagulation and whatnot such a, you know, conundrum, an area of mystery within ECMO? Well, we think about kind of the virtuous triad of thrombosis. We have the stasis, endothelial injury, and hypercoagulability, okay, all combined within the ECMO circuit to, again, make these patients at increased risk. You can see, even upon initiation of ECMO, within minutes to early hours, right, you're going to have an increased risk in the thrombotic, sort of hypercoagulable, thrombotic effects of the circuit. Clot deposition is going to occur. These patients often might come to you in a pro-hypercoagulable state, right, due to the pre-ECMO inflammatory state from microangiopathy, DIC, sepsis, and whatnot. And let's not forget the ECMO-related serous phenomenon that we often see, and some of the initiation of clotting cascade in that way. So thinking further and trying to contextualize why we say right agent, right, why is heparin considered the agent of choice still? Well, it's long history of use, okay? It's short half-life, okay? It's kind of inability to turn on and turn off quite quickly, and it's just extreme, extreme broad levels of experience, okay? We're saying that this is really the most common agent used. However, there are limitations. It's got a narrow therapeutic window. It's got a lot of variability based on bioavailability and dose response, and it's going to depend on antithrombin as a cofactor to achieve its effect, which can lead to some variation. DTIs, so bilvalrudin, argatraman, seek to bypass some of these limitations of heparin by acting directly on thrombin to cause effect, okay, inhibiting fibrinogen change to fibrin. And they have more predictable pharmacokinetics than heparin will have. You can see here the two main ones I just mentioned, argatraman and bilvalrudin. Mostly bilval is going to be the most commonly used one for ECMO, okay? So there are no RCTs between the two, okay, except that three retrospective observational studies really suggest that while there is more PTT variation with heparin compared to bilval, the bleeding and red blood cell transfusions and mortality have been basically nonexistent. So anticoagulation in ECMO, what's the right target, okay? I think this is a really fun graph from the supplement of that Protecmo study that shows in the little red dots what centers are using to titrate their anticoagulation. You can see it's all over the map, right? Most people just use APTT still, at least in this study of over 40 global centers, but there's a lot of variation and a lot of people who use kind of a multi-modal approach. So no consensus in monitoring parameters or target ranges and with significant areas of uncertainty and wide variations in practice right here. Just in the interest of time, we'll skip some of this, but I will just mention that ACT, that question has come up and over time, particularly in the more recent era, that's just been shown to not be nearly as accurate as PTT for levels of monitoring, particularly in adult patients due to the variation and the impacts of multiple other inflammatory states in the patient. PTT, while improved, compared to ACT, also can be affected by patient states. So factor deficiencies, clotting times, shortened by acute phase reactants, and again, having some variability to actual heparin dose, whereas anti-10A is going to be the most accurate with the highest correlation of all these agents to heparin level, okay? And a common question that comes up is, what about TAG? What about ROTEM? Well, as sort of interesting and exciting as this is, as sort of a micro-niche area of monitoring, there's just not a lot of evidence in the ECMO literature to support their routine use. There was one small RCT looking at TAG versus PTT in ECMO patients that showed no difference in bleeding or thrombotic complications, but a lot more heparin adjustments needed to be made when using that TAG. So whether or not that actually resulted in any positive impact has even yet to be shown. So what's a reasonable approach to anticoagulation in these patients? Starting with a heparin drip, okay, checking PTT, potentially checking anti-10A, if there's some discrepancy, going by 10A as opposed to PTT, given its higher correlation with the actual heparin dosing and anticoagulation effect, and then reconsidering for there in a tailored approach. Right dose, as I mentioned, that seems to be the most consistently recommended thing, which is low target. EOLA used a PTT target of 40 to 55 seconds, or anti-10A level 0.2 to 0.3, so quite low, and didn't report any significant change in thrombotic events or ischemic CVAs compared to the control group. I will just say that anticoagulation-free approaches are also increasingly used in VBF and ECMO, so stay tuned for more of that as research evolves, and that just circles back to the guidelines that we have, okay. So sort of in conclusion, regardless of modality or measurement, low-level anticoagulation is going to be favored in these patients, and this, amongst other hematologic conundrums are evolving areas of study within the ECMO world. Thank you very much. Thank you, Kara, for that wonderful talk. Okay, I'm going to close out our session talking about liberating patients from venovenous ECMO, another area of evolving study and very limited data. Okay, my name is Winnie Gannon, and I have nothing to disclose, so I'm hoping by the end of this talk you'll kind of understand why decannulation at the earliest, safest possible time might be important, understand the current practices around liberation from ECMO, how we may be able to apply some lessons from mechanical ventilation, and understand the current data around trying to do so, and maybe a little bit about the future and what's ahead. So why does decannulation at the earliest, safest time matter? You know, it seems like it would, but maybe we should unpack that a little bit. So there are a number of complications that can occur for patients on ECMO, really from wrong configuration to problems at cannulation to ECMO-related complications, which is bleeding and thrombosis, anticoagulation and circuit problems, and also kind of long-term sequelae and sort of long-term neurologic deficits and things of that nature. So you know, decannulating patients at the earliest, safest time, you know, might be expected to mitigate some of these problems. Another thing that we, you know, run into is, you know, ECMO is a really kind of resource-intensive therapy, and we know that it's scarce. I think we were kind of hit over the head with this in COVID-19, where there were patients who weren't, some weren't able to access ECMO. And there were data to show that there was a big difference in mortality among patients, between patients who were able to have access to care and those who weren't. And given the resource limitations of this technology, you know, removing ECMO from patients who no longer need it is probably important. It probably will optimize resource allocation and provide ECMO to provide greater access to ECMO. I think we saw this study earlier about cost-effectiveness, and you know, ECMO is a cost, it seems like we have data to show that ECMO is a cost-effective technology. But it is expensive, and it is associated with high hospital costs. So again, decannulating at the earliest, safest time, limiting prolonged exposure to ECMO unnecessarily may reduce costs. I think this is an area where there are not a lot of data, but there have been data to show that prolonged time on ECMO and ECMO duration may be associated with worse outcomes. The study showed at two weeks, outcome worsened, at three weeks and then four weeks, significantly worsened over time. So again, being able to shorten ECMO duration safely and identifying these patients to come off early may improve patient outcomes. So what are our current practice patterns around liberation from ECMO? So we don't have a lot of data to guide it. This isn't really scientifically driven, but we do have the extracorporeal life support guidance from experts worldwide who have been doing this a long time in high-volume centers to help guide us. And these were guidelines from August of 2017. They tell us extracorporeal support to be decreased if native organ function improves. And when support is less than 30% of total native heart or lung function, we really should consider turning the sweep off and seeing how the patient does off ECMO. And for those of you who are new to this space, if you turn off the sweep gas flow, there's no transfer of gas. So there's no oxygenation or CO2 removal. It sort of simulates what it would be like if the patient were off ECMO. But these guidance essentially imply that clinicians need to identify when the patient's ready for this and require iterative kind of gradual reductions in support. And it also says adjust the ventilator to settings you would accept off ECLS, so creating a buffer, some reserve for when you take that patient off, which all makes sense. But again, relies on the clinicians to make those changes sort of on their judgment. These are a little more recent guidelines published in the SCIO a couple of years ago around management of adult patients supported with VV-ECMO. And a part of these recommendations were around weaning and liberation from ECMO. And they gave us a little more insight into when patients might be ready for those off-sweep trials. So they gave us some thresholds of oxygenation, FIO2 consistently less than 60, PEEP less than or equal to 10, and ventilation. So maintaining lung protected ventilation, imaging, when that chest imaging starts to improve. Again, I think these are very helpful thresholds. And I think the right approach, and this is kind of what we're doing, but also not in a very standardized framework, and again, kind of relying on us to identify when patients are ready for these off-sweep trials. So I would say that ECMO is certainly not the first intervention or therapy in critical care where it requires us to wean and liberate. Certainly in the realms of mechanical ventilation and sedation, SBTs, SATs, there have been a lot of work in these areas. And perhaps we can apply some of those things. This was an editorial from 1986, actually, around a paper that was put out about weaning from mechanical ventilation. And it was called, Is Weaning an Art or a Science? And in their concluding paragraph, they state, clearly, further systematic studies are needed before our knowledge in this important clinical area can reach a truly scientific level. At present, weaning is still an art. And I kind of think that this paper could have been written about ECMO at this point. And ECMO is kind of where mechanical ventilation was back 30-plus years ago. And we actually don't have to go through all the mechanical ventilation literature to ourselves to extract all those important lessons, because Riccardo Paradis and Eddie Phan did it for us in an editorial about weaning from VB-ECMO 30-some-odd years later. But here are the things that we've learned from a huge pile of good research and spontaneous breathing trials. We know clinicians underestimate the patient's readiness to wean. We know single predictors to identify readiness or successful liberation are limited. We know daily liberation trials are best. And we know that eligibility determined, like eligibility for an off-sweep or basically spontaneous breathing trial or trial of liberation, is best determined by pre-established pragmatic criteria at the bedside. And that systematic and standardized approach lead to improved outcomes. But we don't know still if any of this really applies to the ECMO space. But it is an intriguing consideration. So are there data that exist around applying these once-daily assessments of liberation from ECMO? There are actually four different papers that came out a couple of years ago, all around within six weeks of each other. One was by our group at Vanderbilt. And this was a feasibility study, a single-arm feasibility study of, can we deliver this protocol safely? Can we do it? Can we enroll? Are we adherent to this protocol? And so this was really borrowed from the spontaneous breathing trial framework of creating a safety screen, so identified failure criteria. So if patients met failure criteria of their safety screen, they did not kind of proceed to the next phases. But if they did not meet any failure criteria, they proceeded in a standardized fashion to a next phase which allowed kind of optimization of FIO2 and mechanical ventilation within certain parameters, lung protection and reasonable FIO2. And then if failure criteria were still not met, the patient received an off-sweep trial and there were failure criteria associated with that. Ricardo Paradis led this study, and Eddie Fan, around the same concept essentially, you know, creating eligibility criteria to receive an off-sweep trial. What I find very compelling about this paper was their criteria were even a little more liberal than ours and found that among patients decannulated, three-quarters of those patients were decannulated when earlier in the day their sweep gas flow was over two, which is unusual, kind of unusual care. But again, this study showed that this was feasible and may have identified patients earlier. This was a before and after study by Elias Pratt and Craig Rackley from Duke, again coming out within the same month essentially. And they generated this protocol very similar. They had a safety screen. They had a phase where you can optimize the ventilator a little bit and then a phase where you actually turn off the sweep gas flow. And they implemented it using their bedside personnel. This was by respiratory therapists at the bedside. And this, you know, is intriguing because, you know, we know from mechanical ventilator literature that involving, you know, non-physician personnel at the bedside to deliver these protocols improved adherence and patient outcomes. And so they compared patients who before this protocol was implemented to after this protocol was implemented. I think there were like 70-something patients prior to and 100 patients or so afterwards and found that in the group that followed the implementation of this protocol ECMO duration was reduced, mechanical ventilation was reduced, and ICU length of stay was reduced, but hospital length of stay and outcomes did not change. And then this was a study that came out again in the same month. This wasn't exactly looking at a delivery of a once-daily assessment of readiness to wean but was really helpful in examining different variables that were associated with an unsafe liberation. And so, you know, you can provide a nice protocol every day, but if it's not well-specified you may not be improving outcomes. You could be causing harm. And so it's really important to understand what those failure criteria should be. And I think, you know, this study is helpful in determining some of those variables. So I will say that even though I think, you know, these data are evolving, you know, single-arm feasibility studies do not equal, you know, the feasibility of a definitive multicenter trial. It doesn't really demonstrate safety or efficacy. And so I think, you know, a part of the future may be kind of how do we study what's better? How do we compare these different strategies? Are there more strategies, other strategies that we should think of? And we do have an ongoing trial looking at this multicenter pilot feasibility study. And so hopefully we'll have results before too long examining, kind of comparing these two groups of usual care to a once-daily assessment of readiness to wean and liberate from ECMO. So thank you very much and thank you to the speakers. I think, you know, we're just at time, but, you know, if there are any burning questions we could probably have like two minutes and feel free to come up and talk to us too. Thank you.

ECMO

extracorporeal membrane oxygenation

severe respiratory failure

acute respiratory distress syndrome

ARDS

transfusion

anticoagulation practices

decannulating

mechanical ventilation