Good morning. My name is Martha Curley. I'm the COCA Chair in Pediatric Nursing Science at the University of Pennsylvania with a joint appointment at Children's Hospital Philadelphia. And I'd like to welcome you to this really exciting session. It's really an important topic within our field. It's something that we continuously struggle with each and every patient that comes into our intensive care units. We have a stellar cast. The title of the session is Pediatric ARDS or PARDS? Do we have a consensus? I've asked each of the speakers to hold to their 20-minute presentations, preferably ending about 18 minutes into their talks so that we could take a couple of questions from each speaker. And then at the very end, we'll have a 20-minute panel to integrate some of the conversations that we'll be having today. The first person up to lecture on the topic, kids are not little adults. The challenges of defining PARDS is Robbie Kameny. He's from Children's Hospital of LA. He is an associate professor at USC. And he's a PI of two, I think, of the major studies that are happening in the field, PARDI and RedVent. So welcome. Thanks, Martha. Good morning, everybody. Thank you all for being here. It's a privilege to be here. Thanks for the invitation to talk. So I'm going to talk today about challenges defining pediatric ARDS. I'll begin by saying here are my disclosures, none of which are really relevant for this talk. So I think it's important to remember that there is an S in ARDS, that this is a clinical syndrome that's characterized by diffuse inflammation in the lung. It results in hypoxemia from injury that occurs either at the alveolar epithelial surface or at the endothelial surface. But unfortunately, there's no single diagnostic test. There's not a biomarker that's going to be positive in all cases. So in both adult and in pediatrics, we are left to try to define this clinical syndrome based upon the signs and symptoms, which we hope are reflective of the pathophysiology of the disease. So these include things like hypoxemia, which we think reflects the degree of intrapulmonary shunt, poor respiratory system compliance, and loss of end-expiratory lung volume, that the process is happening diffusely, albeit not homogeneously, throughout the lungs, that there is endothelial injury as well as alveolar epithelial injury, interstitial and alveolar edema, as well as increases in alveolar dead space. And for the last 30 years, we've had clinical definitions of ARDS in the adult literature. Our most current definition is the Berlin definition, which I'm sure everybody in the room is familiar with, within one week of a known clinical insult, bilateral opacifications on chest imaging, and then respiratory failure that's fully explained by cardiac failure or fluid overload, and then mild, moderate, or severe ARDS, which is based upon a PaO2 to FiO2 ratio with a minimal PIPA 5. And of note, those that are on non-invasive ventilation, by the Berlin definition, are categorized as having mild ARDS. But unfortunately, there really is limited validation for this in children. Now, around the time that the Berlin definition was published, the PALICC group was being convened, convened by Neil Thomas, Philippe Jouvet, and Doug Wilson, trying to address really nine major areas in pediatric ARDS diagnosis and management, and to create, for the first time, a pediatric-specific definition for ARDS. So why do we need a pediatric-specific definition? Because there may be differences between pediatric ARDS and adult ARDS, particularly with regards to our practice patterns that may affect the diagnosis. So there were seven areas that we considered for the pediatric definition, and I'm going to talk in detail about three of them. The best criteria to stratify risk, how do we deal with patients on non-invasive ventilation, or those without arterial lines, and then the radiographic criteria. So hopefully, you all have seen the PALICC definition, which was published now in 2015. So we're excluding patients with perinatal-related lung disease to differentiate those with neonatal respiratory distress syndrome from those with PARDS. Timing and origin of edema is identical to the Berlin definition. Our first major departure came with the chest imaging. So this is a new infiltrate consistent with pulmonary parenchymal disease. And then like the Berlin definition, for those on invasive ventilation, we've stuck with mild, moderate, and severe, but have changed to using the oxygenation index instead of the PF ratio and allowing for the oxygen saturation index if a PAO2 is not available. And then for those on non-invasive ventilation, we're simply labeling them as having PARDS. There's no severity stratification. And this is based upon either a PF ratio or an SF ratio. So let's talk first about the risk stratification. So here, for measures of oxygenation, we recommend using the oxygenation index, which I hope everybody's familiar with. But it's the FIO2 times the mean airway pressure times 100 divided by the PAO2 instead of the PF ratio for those on invasive ventilation. And we've kept the PF ratio for those on non-invasive ventilation. And why did we do this? Why the OI over the PF ratio? Well, here's just an example. Here we have a patient that's intubated. He's on a PEEP of 5. PAO2 is 60, and FIO2 is 60%. So a PF ratio of 100. By Berlin classification, this patient would have severe ARDS. Somebody comes in, changes the PEEP on the ventilator. 15 minutes later, this is now what the new blood gas is. PF ratio is 240. Has that patient's lung disease severity changed from severe to mild in 10 minutes? No. It's affected by what we do on the ventilator. So we need to capture that to some degree. And here we're using the mean airway pressure to capture that. And you'll note that the oxygenation index for this patient is, in fact, the same at both of these points. So why couldn't we use PEEP? What they did in the Berlin definition was consider a minimal PEEP threshold. And one of the problems is that PEEP is not the only way we manipulate oxygenation on the ventilator in pediatrics, for good or for bad. And John, I think, will talk about some of these concepts perhaps soon. But when we look at observational data in pediatrics, and this is 1,000 pediatric ARDS patients from six different ICUs, what we see is there's basically a flat relationship between PEEP and FiO2. So here's PEEP. On this axis, here's FiO2. This is the median PEEP that's chosen at a given FiO2 in the interquartile range. So while we escalate PEEP marginally, we don't escalate it a lot as the FiO2 is increased. And this is in contrast to what would be recommended, for example, by the PEEP-FiO2 titration grid in the ARDSNet protocol, which is given here in green. So that's why we felt PEEP alone was inadequate and why we went to the oxygenation index. But we didn't know what the right cut points would be for the oxygenation index. So here we look to the data. This is data from CHLA that served as the derivation set. And we saw that there were these sort of natural break points at OIs of 4, 8, and 16, which had stepwise increases in mortality. We checked this in a derivation set of about 600 pediatric ARDS patients from six different studies. And what we saw is about a third of the patients in each of the groups of mild, moderate, and severe down here, and mortality that went from 12% to 22% to 30%. Now, one of the things we did not address in the PALIC definition is timing. And these are two studies that have come out since the publication of the PALIC definition, which highlights that at the time of ARDS diagnosis, the oxygenation index, or even the OI-OSI criteria, may not have as good of a risk stratification for outcome as what you would get, let's say, 24 hours later. So you see here, there's almost no difference in mortality with mild and moderate. But by the time you get to 24 hours, there's more of a stepwise increase in terms of the risk of mortality. And I'll come back to this in a second. So recently, we engaged in the PARTY study, which Martha alluded to. This was an international cross-sectional study to try to understand the implications of the new PALIC definition. We enrolled 145 international PEDS ICUs from North America, Europe, and South America, as well as some from Asia and Australia. This was designed as a main study plus some ancillaries. We did five continuous days of screening over 10 distinct study weeks. So we had 50 different study days. We identified patients that were newly diagnosed with ARDS, looked at their clinical variables in the first three days of ARDS diagnosis, and then followed their outcomes. We screened about 23,000 ICU patients, half of which were on mechanical ventilation, identified 744 as having ARDS, and 708 of them had complete data. And what is interesting is about a quarter of the patients fall into each of the groups, NIV, mild, moderate, and severe, based upon their initial oxygen impairment at the time of PARDS diagnosis. But interestingly, mortality is very similar between NIV, mild, and moderate. And it's not until the severe group that you see this large increase in mortality at 30%. This is another way to look at that. And this is restricted to those that were on invasive ventilation at the time of their PARDS diagnosis. The number of cases in blue, the diamonds here represent the mortality rate with the confidence intervals, oxygenation index here on the x-axis. And what you see is that the mortality is relatively low and linear, or flat, I would say, here at the OI of less than 15. And then there's this jump up that happens at an OI of 15. And that happens to correspond to the severe group, which was an OI of 16. Let me orient you to this slide, because it's a bit busy. Here we have the four different time points, PARDS severity, 6, 12, and 24 hours. And what we looked at here, a number of cases on this axis, and then mortality here. This is the PALIC-OI-OSI groupings of mild, moderate, and severe. This is the Berlin groupings, if we were to use a PF ratio of 100, 200, and 300, or substitute an SF ratio. Do not require bilateral pacifications here. But one thing that you'll notice is that really, all time points, more patients would be classified as severe using the PF-SF cutoff as compared to the PALIC groupings. And in PALIC, more are in that mild group. And the mortality for the PALIC severe group is almost always higher than the mortality that you see with the PF-SF cut points. And as you go past the first point to 6, 12, and 24 hours, there also becomes this stepwise increase in mortality from mild to moderate severe, especially within the PALIC-OI-OSI groupings. And if we look at the mortality discrimination, comparing the PALIC cut points to the Berlin cut points that sit from 6 to 24 hours, you see that the PALIC breakdown appears to better stratify risk as compared to Berlin at really all of these time points. So now the second thing we'll talk about are how do we handle patients that don't have a PAO2. And here we recommend using the oxygen saturation index, which substitutes the SpO2 for the PAO2 in that equation, or the SF ratio instead of the PF ratio for those that are on non-invasive ventilation. There have been multiple studies now that have shown that these parameters are linearly correlated and can be highly predictive of one another, the OI to the OSI, or the SF to the PF, or perhaps with a transformation. As long as the saturations are less than or equal to 97%, because you want to stay off that flat portion of the oxyhemoglobin dissociation curve. And it's important to include SpO2-based criteria because if we don't, we're gonna select for a different patient population because our practice patterns have changed. We don't put arterial lines in every patient on a ventilator, right? And here is data that supports that. So this is 1,800 mechanically ventilated patients. We calculated an SF ratio on them within the first 24 hours of ventilation. Now, 600 of them, we could not because they had a SAT above 97% the entire time. But what we see is that about 40% of patients have an arterial blood gas within that first 24 hours. And those with an arterial blood gas had significantly higher mortality, roughly fourfold higher mortality. Where's that mortality coming from? Well, interestingly, if you look at those with mild or moderate degrees of oxygen impairment, only about a third of them actually have an arterial blood gas. It's only till you get to the severe group that 70% of them have an arterial blood gas. And of course, 50% of patients with an arterial blood gas are on a vasoactive agent as compared to only 10% that don't. So if we restrict ourselves to requiring a PF ratio, we're gonna select for a different population of patients, a sicker population with combination of hypoxemia and probably more importantly, cardiovascular dysfunction. And we saw that in the party study. So here, what we looked at is the 708 patients that were diagnosed based on PALICC criteria. And at the time of diagnosis, only about a third of them met all Berlin criteria with bilateral opacifications and a PF ratio less than 300. There was a large chunk that still had bilateral opacifications, but did not have an arterial line. So they would meet criteria if we use the SF ratio. And then about a quarter of the patients that had radiographically unilateral disease. But this changes over the first three days of PARDS diagnosis, where now there's about 60% overlap between the Berlin definition and the PALICC definition. And you'll note that those patients still have significantly higher mortality. There are still about a third of the patients that never get an arterial line and a small percentage, about 12 to 13% that never develop bilateral opacifications. So that leads us to the last point, which is about chest imaging. So chest imaging findings of a new infiltrate or infiltrates are necessary to diagnose pediatric ARDS. So we got rid of the requirement for bilateral opacifications, and there were several reasons for that. The first is there's high inner observer variability. I'm gonna come back to that point in a second. The chest X-ray is poorly sensitive to detect disease and often lags behind the degree of hypoxemia. And we saw that in that previous slide that you see many patients may develop those bilateral opacifications over the first three days, but they don't have them at the time of the initial diagnosis. It may also be that the pathophysiology of ARDS is occurring in patients that have what appear to be radiographically unilateral disease, and our techniques are just not good enough to be able to detect that. And finally, we wanted to figure out if there was an important mortality risk stratification based upon bilateral opacifications. And the problem with this is that since most definitions of ARDS required bilateral opacifications, there are very few studies that have looked at patients without bilateral opacifications. But when we looked at one of them as an example from our CHLA group, what we see is that about half of the patients had bilateral opacifications, but 60% of those with a PF ratio less than 100 had bilateral opacifications. So what's important here is that the hypoxemia and the bilateral opacifications are confounding each other, right? They go together, which I think makes sense. So once you stratify by the degree of hypoxemia, there's in fact no difference in mortality between those with and without bilateral opacifications, although maybe there's a hint of it in the more mild patients, although there are very few patients here. So we looked at this again within the party cohort and found something similar, although interesting differences. So within the mild and moderate group, there's no difference in mortality between those with bilateral versus unilateral disease. But interestingly, those with NIV PARDS, if they have bilateral opacifications, their mortality is almost four times higher than those with unilateral opacifications. And in the severe group, there's also this signal of significantly higher mortality, although not statistically significant. So we may ask, maybe it's wrong to do bilateral opacifications, but could we look at quadrants of consolidation instead? And what's interesting about quadrants of consolidation is certainly four quadrants of consolidation appear to impart a higher risk. But interestingly, there's no difference between two and three on those on noninvasive ventilation. And you look again at these mild and moderate patients, there's basically no difference at all based upon quadrants of consolidation. And in the severe group, although four quadrants is associated with very high mortality, there's also high mortality for those that only had one quadrant. So there's not a real consistent relationship here once you control for the degree of hypoxemia. And we still have the problem of poor inter-rater reliability on the interpretation of the chest X-ray. And this is 702 simultaneous interpretations as part of the party study between an intensivist and a radiologist. And although almost 80% of patients had bilateral infiltrates within this cohort, the kappa value, so once you correct for chance-related agreement, is only 0.3. So this is sort of barely better than what would be expected by chance alone. And ultimately, no radiographic variable had an independent relationship with mortality after controlling for other factors. So I'll just conclude by saying PALICC identifies about 40% more children as having ARDS compared to Berlin, and identifies this 12 to 30 hours sooner. The OI-OSI groups appear to better stratify risk than the PF-SF ratio groups. Regardless, patients that have severe PARDS really at any time point have very high mortality risk, over 30%. But perhaps a delayed marker of PARDS severity, let's say six to 12 hours later, may be a better way to stratify risk. And that might be important, for example, in clinical trials. And then finally, the chest imaging criteria don't have an independent relationship with mortality and have high variability. Thank you. So this paper's up for questions, anyone? I'll start by asking a question. Yeah. So as a clinical trialist, how do you think these operational definitions have really impacted the field on both a positive and a negative slant? Yeah, that's a loaded question, Martha, as a clinical trialist right here. So I think the definite positives are it identifies a common set of operational definitions that we can all use in clinical trials to say, okay, we're gonna target only severe patients, and so we're gonna use this accepted criteria, for example, to look at that severe patient population. I think there are limitations still, one with the timing issue, which I alluded to, and I think you have probably experienced this with prospect is that the patient who's severe initially may not be severe six or 12 hours later. So that has to, I think, be considered, especially for clinical trials. And then I think the other important piece, which I'm looking at Mary and Heidi here, is just because somebody is severe doesn't mean their phenotype is the same and their response to a therapy is gonna be the same, right? So it may be that this severity classification is capturing some component of the mortality risk, but it might not be capturing the treatment benefit for a specific therapy, and that might come from some other factor like the recruitability of their lungs or if they have an inflamed phenotype, as an example. Yeah, so it's very frustrating when the subjects, potential subjects, have not read the PLIC guidelines. And we're trying to get them into a trial, and what's happening is that they're hitting the initial bench with an OSI, but then they miraculously get better, which is good because if they're getting better, they shouldn't be enrolled into a trial enrolling moderate to severe PARDS. But what about the heterogeneity of PARDS? Do you think the PARDS diagnosis in and by itself has helped or hurt our advancement in the field? Yeah, that's a really good question. I think the PARDS, personally, I think it's helped because it, again, gives us a common language, but it does potentially put a blinder on that component of the heterogeneity where treatment effect may be different based upon something that is completely not present within the definition. So I think that's where we need to keep going in terms of our advancement. And whether those phenotypes, so to speak, need to be part of a definition, I'm not sure, right? But they certainly need to be considered when thinking about therapy decisions. I mean, there's this highly inflamed PARDS phenotype like the sepsis-PARDS overlap patient, which may be very different than the aspiration PARDS patient, as an example. And the therapies may apply, may be different. That being said, there are some therapies that are likely going to universally be applicable, like, for example, low tidal volume ventilation and a conservative fluid management strategy, as an example. But maybe there's going to be a differential response to PEEP, as an example, within those patients, yeah. Other questions, anyone? Go ahead. Hey, Alex. I have a great talk. I'm Alex Rota from Duke. Obviously, the definition has gone a long way to creating uniformity for research and so on. When you look at etiologies, such as sepsis versus RSV, is it time to have a separate PARDS for the very low mortality subsets? Yeah, good question. And this is something we spent a fair amount of time talking about as part of the first pallet group. And I think ultimately, to me, it still goes back to that same issue of heterogeneity and what you're going to do with the definition for treatment. And so you would differentiate an RSV patient where you may not consider a very high risk therapy, for example, in that patient, because their outcome is going to be good, from the septic, multiple inotrope severe PARDS patient. To me, they both still have PARDS, but the therapies we may apply to them are going to be potentially different. Some are shared, like the tidal volume ventilation. Some are potentially very different, especially as we get into higher risk therapies. And I think you can't capture all of that with a definition. Thank you, Matt. Yeah. Okay. Thank you. So at this time, I'd like to welcome John Arnold to the podium. His talk is PEEP and Recruitment Maneuvers, When and How. And as he's coming, I'd like to say that he's a professor at Harvard. He directs both the ECMO and the respiratory therapy program at Children's in Boston. And when I was a baby clinical nurse specialist, it was really fun to work with John and John Thompson, who was the director of respiratory therapy at the time, in a unit that was pretty much running by the seat of our pants. But anyway, welcome, John. Jeepers, Martha. That's the nicest thing you've ever said about me. That's it. So thank you to the organizing committee and I guess to Martin for the invitation to speak. Although I don't know how I got this topic. And I apologize, I've got a frog in my throat. I'm like about three quarters of you recovering from the plague. So I'm gonna cover a lot of data because Martha wants to leave enough time for a food fight at the end of this session. So I'm gonna run fast. How many of you think that PEEP is important to provide lung protective ventilation? Okay, how many of you think recruitment maneuvers are important to provide lung protective ventilation? Yeah, I'm the same, okay. Okay, good, we're on the same team here. I think all of us who use a mechanical ventilator have this cartoon in our minds. It captures the phenomenon of hysteresis. It emphasizes the fact that the lung is very sensitive to its recruitment history. And then if you open the lung, you can allow the lung to settle on the deflation limb and provide optimal gas exchange. And you've seen many, many, many versions of this slide simply demonstrating that the lower inflection point, the lung looks like this. Once opened on the maximal equatorial curvature point, the lung looks like this. And I think, at least in a cartoon, illustrates the importance of lung opening. Now, this is some interesting data that some of you probably have seen. This is Gary Neiman when he was at Syracuse in the Department of Surgery. He used a lung injury model and then he used a little camera, a telescope with a camera attached just inside the pleural surface. And he actually looked at alveoli. He was able to outline the alveoli both in the normal lung and then the tween injured lung. So it's kind of cool. Let me just quickly show you some videos. This is the normal lung. Pretty boring. But his model is that the I to E, we can kill that volume if you want. The I to E, in-sport x-ray change in area is an important correlate of lung injury. So that's the normal lung. So here's lung injury followed by a recruitment maneuver. And I want you to focus on how long it takes for these alveoli to open. So we're now, jeepers, 10 seconds in the recruitment maneuver. It was a 60-second total. 15, I guess I do have a watch. 20. And only now do alveoli begin to open. So a single recruitment maneuver takes a long time to open alveoli. Now, here's the same animal now injured. This is five sonometers of PEEP. And it shows the inspiratory to expiratory change in alveolar area during conventional ventilation. That's five of PEEP. And here's 10 of PEEP. So the I to E change in area is much smaller in the higher PEEP group. So Gary went on to conclude that recruitment increases alveolar area, decreases alveolar instability, which I think is what we're all trying to do, whatever type of ventilation we're applying. And he showed, I think convincingly, that inadequate PEEP leads to derecruitment in alveolar instability. Now, Gerhard Wolf, when he was with us, did some very interesting work using electrical impedance tomography to basically guide PEEP titration. It was a six-hour study, saline lavage-induced lung injury. The EIT-guided group used continuous EIT imaging to increase PEEP and reverse atelectasis, and then to decrementally reduce PEEP to limit overdistension. And importantly, this was a dynamic process where Q hourly, we repeated the maneuver to continuously identify the optimal PEEP in these animals. The EIT-guided animals had significantly higher PEEPs applied during the entire duration of the protocol. And Gerhard nicely illustrated the significant improvement in histopathology that EIT-guided PEEP titration produces. So this is the EIT group. This is the control group. And then on careful histopathologic analysis, Gerhard showed that it's primarily hyaline membrane formation that was reduced when you use EIT to guide PEEP titration. So that's the physiology. That's sort of convincing, but now we need to look at the data. So you all probably remember there are three large trials of high-low PEEP in adults. This was the alveoli trial. This was the areas network study. This was a big study. They showed no significant difference in overall survival or survival to hospital discharge using higher PEEPs. This was the LOVE trial, the lung open ventilation trial. This is Maureen Mead's clinical trial group. These are legit adult clinical trialists, as you all know. This was in a thousand patient study. And also they show no significant difference in the high PEEP versus low PEEP group. The high PEEP group also included a recruitment maneuver at initiation, as you probably remember. And this was the EXPRESS trial. I think the best way to summarize these data is using a forest plot. So here are the three trials, and here's the cumulative individual patient meta-analysis. Importantly, the 95% confidence interval includes the null outcome. So even combining three large, well-conducted adult clinical trials of high versus low PEEP, we cannot convincingly demonstrate that high PEEP is better than low PEEP, at least in terms of mortality. What about the pediatric population? And Robbie and his crew provided some intriguing data from their, gee, 1,100 patient, eight center study, showing that, in fact, if you use PEEP lower than what was suggested by the high PEEP ARDS network protocol, there was a significantly higher mortality. And importantly, that driving pressure was also associated with a higher mortality. I can't not mention that, in fact, in this data set, if you were treated at CHOP, you had a better outcome than if you were treated elsewhere. So maybe we should send all of our cases to Natalie and Martha. Okay, so two recent trials that you've probably looked at quickly, but I want to point out some features that are important. So this is the alveolar recruitment trial in ARDS. This was, I think I can say this, this was a bad-ass trial. This was centered in Sao Paulo. It was multi-centered. They recruited 1,000 patients. And look, in Brazil, you can't get away. Zero lost a follow-up. Three lost a follow-up. So this was really a thoroughly performed study. Now, the ART intervention arm is worth talking about. So there were stairwise recruitment maneuvers used, using a driving pressure of 15 on tops of PEEPs of 25, 35, and 45. Then a decremental PEEP trial was done all the way down to 11. Static lung compliance was measured at each PEEP decrement. So static lung compliance on the way down the PEEP titration. Optimal PEEP was then defined as the best, PEEP at best static lung compliance plus 2. And this is their intervention. So this is the stairwise recruitment maneuver protocol. This is the decremental PEEP protocol. And then you got newly recruited to 35 sonometers plus 15 driving pressure. And then you settled in on ventilation. Now, I should notice, you'll notice the number here is 35. Initially, I said 45. They had three unexplained cardiac arrests in the intervention arm. So at an interim analysis, they adjusted their stairwise recruitment maneuver to maximum PEEP of 35. But importantly, everyone got the same recruitment maneuver. And importantly, this was an early intervention, less than 72 hours in a moderate to severe ARDS population. To everyone's amazement, they showed that this badass lung recruitment and higher PEEP titration using static lung compliance was actually associated with a higher mortality. That stopped a lot of people in our tracks. And when you carefully look at subgroups, this effect was everywhere. Every subgroup analysis showed that lower PEEP was better than higher PEEP with a recruitment maneuver. So the primary outcome, death less than 28 days within six months, number of ventilator-free days, all significantly different in the intervention arm and worse in the intervention arm. And importantly, there was more pneumothorax and chest X-ray evident barotrauma in the intervention arm. So there was a cost in the intervention arm. So how did this happen, this careful study with lots of preclinical data suggesting that this is the way to open the lung? Well, as evidenced by their need to adjust the intervention, they had a lot of overdistension, which produced hemodynamic impairment, cardiac arrest. And as we've said, I think on these podium for several years, that low PEEP, low tidal volume strategy works. We're reducing mortality in adults in a pediatric patient simply by adhering to protocolized conventional ventilation. And I think Robbie alluded to this. There was no test of PEEP responsiveness. We know there are patients who we take care of whose lungs are not going to open. If we knew that they weren't going to open, we would not subject them to an aggressive lung recruitment maneuver. So although this is a dirty trick when you're testing the effectiveness of PEEP, I think it will exclude a lot of harm from patients who wouldn't benefit from this intervention. So this is the FAR-LAP study, sort of published in the wake of the ART trial. This was a small trial, 115 randomized. Again, they tried to identify moderate to severe ARDS early on in its course. The FAR-LAP group had daily recruitment maneuvers, again, stairwise recruitment maneuvers with incremental PEEPs and then a decremental PEEP titration. Their optimal PEEP was based on oxygen saturation. And their control group, again, was a low PEEP ARDS group protocol. So stairwise recruitment using 15 sonories of water, 20, 30, 40, you've seen this in other trials. Then PEEP titration decreasing until there was derecruitment evidenced by a change in SAT. And then when derecruitment was reached, a brief re-recruitment maneuver was repeated for two minutes. Now, importantly, whether you responded or not, and they didn't actually define response, you got that recruitment maneuver every day for randomization for the first five days. And here's their data, their outcome data. No significant improvement in outcomes using the FAR-LAP intervention. Less INO use, I must say, less prone positioning use, less ECMO use in the intervention group. So less rescue therapies. But no significant improvements in outcome. And their primary outcome was ventilator-free days. So what happened in the FAR-LAP trial? Well, first of all, the ARD, the ART trial data were leaked, were not leaked, where they were presented at a national meeting. And I guess they were leaked, weren't they? So the DSMB said, well, you've got to stop because we're going to kill people. So early terminated, you know, offenders, they were trying to go for much larger numbers. FAR-LAP did reduce driving pressure on day one. That may be important, particularly in pediatrics. It did improve oxygenation and reduce new adjuvant therapies. But it also increased the frequency of new arrhythmias. And so, again, we're seeing, at least in adults, adverse cardiopulmonary interactions when we use aggressive stairwise recruitment maneuvers. And the investigators acknowledged that their recruitment maneuver sequence was complex, it was time-consuming, and it was potentially unsafe. And they also suggested in their discussion that there should be some attempt to identify patients who were, in fact, recruitable before using recruitment maneuvers. So in pediatrics, what's a safe recruitment maneuver? Which maneuver most efficiently opens the lung? And how do we adjust PEEP after an effective recruitment maneuver? So I will say right now to all three questions, I don't know. There are some data, and this is a paper that John Keir published when he was our fellow. And it's a small study, but it shows what happens in a group of pediatric patients, what happens when you use stairwise recruitment. So we started with a pressure-volume curve here. We then did baseline ventilation, volume cycled, 6 mL per kilo. We then performed a sustained inflation, 40 times 40, back to P-flex plus 2 in volume control. And then we used incremental recruitment maneuvers. Fifteen, consistent driving pressure of 15, but 15, 20, 25, 30, 35. Importantly, we stopped increases in PEEP if the lung opened, at least by gas exchange criteria. And this was the Marcello Amato insight provided from his very similar trial in adults. Once the lung opens, at least by gas exchange, you stop. And you then, in theory, do not subject patients to additional barotrauma, additional adverse cardiovascular interaction. So this shows PO2s on the same FiO2 during sustained inflation. Sustained inflations, as you all know, don't work. But during stairwise recruitment, the PO2s steadily rose. And interestingly, the PO2s stayed significantly higher than at baseline during PEEP titration in a decremental fashion. So lung opening had been achieved, and we were able to keep the lung open in this small group of patients as we decrementally reduced PEEP. This is illustrating that beware if you're using stairwise recruitment maneuvers, because you're going to greatly over-extend the lung, at least during the maneuver. VDVT ratio is going to go up. And if you don't keep an eye on alveolar ventilation, you're going to produce tremendous hypercapnia, which we discovered early on. So we would have to adjust during stairwise recruitment ventilator rate in order to compensate for the increased VD to VT ratio. Now, this is just one patient. I want to show you what actually went on. These are EIT images, which some of you will recognize. So here we are, stairwise recruitment maneuver. Plateau pressures of 30, 35, 40, and then 45. And whoops. This is the PO2 when we started, 109. Now we need to open this lung. So pressure's up to 35, 102. This was our first patient, by the way, enrolled in this protocol. 40 sonometers of water, 130. What kind of lung opening should we stop? That's pretty high. No, it doesn't meet our definition for stopping. Let's go to a plateau pressure of 45. And wow, the lung opened. I remember when Keir, using his little eye stat at the bedside, said, jeepers, the PO2's 450. I said, no, it's not. Send a blood gas. And sure enough, the lung had opened. So stairwise recruitment maneuver, at least in this patient and in selected patients, does work. So the recommendation, the reveal. Robbie alluded to this. We need to be better at using less oxygen. The OSI, and of course the OI, are very powerful predictors of outcome. So if you keep a saturation between 90 and 93, certainly, but always less than 97%, you can track the OSI beat to beat. And you can have immediate information about the patient's predicted risk of mortality. I think we should be setting PEEP initially based on lung mechanics. I think we've been chasing oxygenation for a long, long time. Remember the oxygenation paradox in the ARDS-Net trial? That small tidal volume group had significantly worse oxygenation. And I don't know if you've talked ever. Some of them are still alive, the ARDS-Net investigators. There was talk about shutting the trial down, because they're blue, and they look like crap. They're sweating. There's a big adult breathing in assist control mode. Heidi remembers this. And bravo to the ARDS-Network investigators. They continued, show that the only intervention we've ever proven to be effective, low tidal volumes, improves outcome. And we can't chase oxygenation. So I think that's an important message to keep reminding ourselves. I do not recommend that anyone routinely use recruitment maneuvers. I really don't. I think you should only use sterilized recruitment maneuvers and titrated PEEP when you're on toxic ventilator settings. So I'm going to make these up partly from data from Robbie and his group. A plateau pressure greater than 28, a driving pressure greater than 13, and tidal volume of greater than 8. If you cannot provide acceptable gas exchange, and guess what? You get to define what acceptable gas exchange is. If you cannot provide acceptable gas exchange with settings less than these, then it's time to do something else. I think that's a careful sterilized recruitment maneuver. Now, let me point out, this is the mortality plot from Marcelo Amato's publication in the New England Journal. This is delta pressure, driving pressure versus outcome. The line of identity, the relative risk of one is crossed up here about 16 solenoids. I'm not sure whether that's the right driving pressure limit. I said 13, because there's now currently a trial ongoing in Europe looking at an intervention where you limit driving pressure to 13 in the intervention arm. So there will be data. I think coming up soon in the adult population to tell us what the proper driving pressure is. So let me finish by, again, cautioning that we really need to get away from oxygenation. It feels good. The oximeter is right there. A SAT of 100 is your friend. Not anymore. We've got to get away from that. We should be measuring lung compliance. We can do it. Ventilators will measure any kind of lung compliance you want, and we should be tracking this information. I think in terms of clinical trialists, Martha, we need to not only have sustained moderate to severe ARDS, that is 12 hours of meeting criteria, but we should test PEEP responsiveness. And we can at least post-talk and analyze patients who are PEEP responsive versus not PEEP responsive. And then finally, I'm holding on to the dream. Where's Gerhard? EIT imaging is a bedside technique that allows us to visualize the lung opening. And what Gerhard showed in several papers is that before you reverse atelectasis, you over-distend the lung. You undoubtedly over-distend the lung. So that if your lung opening intervention, steroids recruitment maneuver in this case, is not limited, you are going to damage the lung. And if you repeatedly apply steroid recruitment maneuvers to patients, you're going to damage the lung by virtue of overdistension. All right, so enough table-thumping. Thank you for your attention, and I hope for some interesting questions. up for questions. Anyone? I wish we all had an IT machine in our unit so that we can, you know, pop it on to see exactly what's going on within the black box, but we don't have that. Can you talk a little bit about the maneuvers and the time it takes to do these maneuvers? Yeah, they're very complicated and I think we, after John Keir's paper, we rolled out sort of a guideline as to how we use recruitment maneuvers. You have to have stopping rules and they have to, you have to have hemodynamic stopping rules and then you have to have lung opening stopping rules. So I think as long as you develop a team around these maneuvers, they should be done during the light of day and they should be done with a specialized team that you can safely attempt to open the lung. Now not all lungs are going to open. Interesting, that patient that I showed you had failed two trials of high-frequency oscillation. I came on service and said, nah, that person doesn't know how to oscillate. Let me try. The lung didn't open. So we were all that patient thinking that we were probably going to cannulate and, as you saw, at a plateau pressure of 45, that little girl's lung opened. So I think it's something that you need to approach cautiously, but you ought to probably have a protocol available in your unit to use recruitment maneuvers. I definitely agree with the non-responder. Yeah. Some kids do and some kids don't and sometimes you just keep pushing them through to do when they don't want to do that. So one question and then... And I think that's the flaw from FARLEP and ART is that they did recruit maneuvers in everybody. Just one question, sorry. Now hold it for the end. Okay. I can't get it. Oh, is it on? Okay. Yeah. Hi, I'm Jill Bybee. I'm a PITS intensivist from Holland DeVos in Grand Rapids. That was a great presentation. Thank you. You said acceptable gas exchange. You get to set your own targets. I was wondering what your personal targets were. Really? You want me to say that in public? Uh-huh. A SAT of 88. And, you know, we've all taken care of patients with SATs in the low 80s and we say, patience, patience, patience. So I think a persistent SAT less than 88, you're really on the steep end and any plugging or something else is going to produce a whole kerfuffle. Plateau pressures are 28, I think. I don't know the driving pressure number, but I do know that we should be talking about it because there are going to be data soon. And we use an FIO2 of 0.6. I know people are fond of 0.55, but 0.6 in our unit is where we stop. Thank you, John. All right, so next up will not be Martin Niebuhr. He was unable to leave the Netherlands because of a family matter. So the operational definition of a friend is 24 hours before going on the podium to take the call and say, remember those slides that you did? Can I borrow those slides? Not only can I borrow them, but I would be happy to do the talk. And I said awesome. So this is Heidi Flory. She's an associate professor at Michigan. She's at Mott, and she's a phenomenal friend. Thank you. Well, thanks for letting me talk today. And it's quite the illustrious podium. I know there's many people in the audience and definitely on my group of speakers that would probably be better at this than I will. But I've done ARDS for a while now, so here we go. Bear with me, and I will hope to defer to many of our colleagues in the audience for questions. So the topic is conventional versus high-frequency oscillation, and does it matter? And my slides aren't moving. There we go. Here are my disclosures. The active ones reflect NIH funding, and non-advisory role reflect any conflict here. So this is going to be a talk about why, what, and how. So we're going to talk about why we need definitive testing, and my speakers before me alluded to much of this as well. What can we do better? And the speaker before me also alluded to some of that as well. We'll go into that for a little bit. And then we're going to talk about how. How can we potentially do this? And we're going to talk about that rigor that we are doing as it applies to PROSPECT and the PROSPECT methodology in specific. So until PROSPECT, we have not really had a large multi-center treatment trial comparing mechanical ventilation versus conventional ventilation versus high-frequency ventilation, or comparing multiple mechanical ventilation strategies in the low tidal volume era. And of course, most of our strategies have been extrapolated from adult studies, including low tidal volume ventilation. So we don't have any data for us to report, but I think I'm preaching to the choir in the audience. So this is one of my favorite photos throughout history, is on the right, on your right. But the question comes in, why? So why for adults versus kids versus infants? Should we be treating the child with BPD, and the healthy infant flipping us off, like this adult ARDS patient? Even within our pediatric population, if you're looking underneath, and that's something that Mary and I do often, should we be treating the five-year-old with community-acquired pneumonia, and no past medical history, and no restrictive chest wall, and probably heterogeneous disease, the same way we treat somebody with a complex, diffuse, homogeneous disease? Let's say a patient with severe aspiration pneumonitis of an older age group, and with complex comorbidities. And then I knew that Robbie, I knew one thing, Robbie was going to talk about bilateral disease. So the question is about this bilateral disease, and we know PALICC now allows us to have severe unilateral disease with enough of an oxygenation defect qualifies you. So what do we do in these instances? And this is data that's now impressed by Dr. Joseph Kony at our institution, who analyzed patients with specimens after a restore. And what we do see is, even after controlling for severity of oxygenation defect, you can actually see that these patients with bilateral disease have a significant early inflammatory response. And how does that impact how we conventionally or otherwise ventilate patients and treat them? We don't know that information yet. And this is older data, but I think it's very worth presenting here, that we're also not sure about mild PARDS versus severe PARDS, and if they respond differently to this low tidal line ventilation strategy that we've adopted. And so Simon Erickson and his group reported some time ago that maximum tidal volume correlated inversely with mortality, even after adjusting for oxygenation defect. This is Robbie and colleagues' data. Again, particularly at that day one time period, higher tidal volumes may be associated with shorter durations of ventilation, but also particularly in those patients who have good compliance, that milder cohort. And this is data that comes out of China that also hearkens to all of this same idea that maybe more ventilator-free days in the patients with the higher tidal volume. So that is some of the why. Now some of the can we do better. What if we think we have best practices? Are we good at doing that? And are we good at doing that around the clock? I mean, all of us are slaving in-house overnight as faculty, so are we using up our our mojo there in the middle of the night? Although Dr. Arnold will say we're not going to be doing these art recruitment maneuvers in the middle of the night, so don't try this at home. But this is a slide that you'll see a ton in all of the talks, perhaps, except maybe sedation. I don't know, Natalie. But we do know this, and we've adopted this from a long time ago. And one of my pearls, maybe Dr. Arnold will appreciate this, is go back to the physiology that was learned in the 70s, 80s, and 90s. There were some interesting pearls in there that might have gotten washed over. And so we have this lower inflection point. We have this upper inflection point. If you're over distended all the time, you're in big, big trouble. But if you can find that optimal map, an optimal peep, then you can potentially stay in this safe zone and not slip down again into this closing point. And then the sicker you get, the more this hysteresis narrows, and this flattens on this side, and this point is even harder to maintain. So it's very hard to get there, and it's very hard to stay there all the time. And so one of the things, as was alluded to just before, is that we can end up with severe hypercapnia and increased PVR if we're way over distended all the time, severe acidosis and those heart attacks in adults. And we can also end up having this kind of physiology at the bottom, and then we're basically pulling a lung open that's kind of like Velcro, or kind of like a sponge with Krazy Glue in it, as opposed to a moist sponge. And so we really want to try to stay up here with the smallest tidal line as best we can, and this optimal peep on conventional ventilation, or an optimal map on the oscillator. So we talked about this already, and I think the only point I really want to make is, when the speakers are talking about low peep in adults now, they're talking about this. So this is even what we're not so good at attaining. So don't think about low peep being where we're living here. But one of the things that wasn't said about this multicenter trial is that in all comers, the patients had a much higher survival if you were able to stay on target all the time. But particularly in this very severe group, look at this mortality difference is enormous here, and quite reflective. So I think this is a very important take-home message. This is work by Sean Ward, presented a couple years ago. And again, we're not really good at saying what we think we're doing, hearkening Dr. Morris from ARDSnet. So even when we look at these, our four treatment trials for acute hypoxemic respiratory failure, acute lung injury, for a total of 315 patients. But even at the 24-hour mark, we're barely getting 60 percent of these patients with ARDS on a tidal line strategies that's at least 8 mLs per kilo using ideal body weight. So we're not so good at doing what we think we were good at doing. And then there was this 2013 banner year on the high-frequency oscillation in adults, and that pretty much cracked the code for many things. But how do we interpret this as well? And so we know Oscar, definitely line over line, there was definitely no difference, but there was definitely no harm by the oscillator, but it didn't help. Most of these patients had direct ARDS. But in oscillate, we were very concerned because this is a mortality difference, and this is one of the only studies that embarked in a mortality difference in this group. But pH in this group was met on the oscillator, was managed by adjusting frequency, and who in this audience uses a hertz of five and a half? We don't. And after that recruitment, the mean airway pressure was pegged at 30, and this does not allow for individualization either. But the control group, now they were held to a constant. They managed this all the time, and they did very, very well. In this meta-analysis by Eddie Phan's group, if you are maintaining this low tidal volume strategy overall, there's a favorable impact on mortality for conventional ventilation. But if you're all over the place, you're kind of stay all over the place, and it's a little bit unclear. So outside of Dr. Arnold's trial that enrolled 70 patients comparing conventional ventilation versus high frequency over 25 years ago, now we don't have that RCT to tell us what to do. We have matched data from the VPS cohort that did not suggest good things either, but didn't suggest necessarily severe harm. This is a post hoc analysis from Restore that Scott Bateman and his colleagues did that also showed that these patients on the conventional ventilator or with late oscillation actually stayed on the ventilator for a shorter amount of time than those who were randomized earlier. But if you look at further at this data, you get to see here the vent-free days, or the duration of ventilation I should say, was longer in the early oscillator group compared to the conventional group, particularly if you took a sicker quintile, and then that was quite statistically significant. But the mortality difference didn't quite get there, and as my good friend and colleague Dr. Iris Sheffitz would say, maybe prolonged duration of ventilation on the oscillator represents us of doing the opposite of tweaking the knobs too much, but maybe we just sit on our hands too much, and maybe we're not weaning assertively enough. I didn't say aggressively because you don't want to get too overzealous, but how often do we say let's do something once or twice a day, and there's where your days go. So we had pretty good agreement for allowing to keep the oscillator on as an alternative mode of ventilation for PALICC. And then this past year, well 2019, Martin and his group finally published their results of this observational cohort of 115 children, where in the Netherlands they are enrolling these patients or treating these patients with the oscillator very early on, using a very high initial frequency and high delta P, again with staircase recruitment maneuvers. And in this group of patients, in this slide, the orange bars are all the severe PARDS groups. And so once placed on the oscillator, yes, OI was pretty high, but it tapered off into a really nice range here, even in the severe PARDS group, PF ratios rose reasonably. CO2 clearance, despite the high frequency and the high delta P, was actually quite good, even in the severe PARDS group, as also evidenced by the pH. And then, most importantly, we're thinking about safety, because we don't want kids to be having untoward hemodynamic effects. And so we see here that it was safe, that heart rates actually came down over the days. And this, in here, orange is infants, and blue is toddlers, and greater than, and gray is going to be actually five years or older. So heart rates were reasonable, mean arterial pressures were fine, CVP was good, and lactic acidosis didn't tail off either. So now we're going to talk a little bit about how. How should we do this? How could we do this? Well, the best way to evidence this is PROSPECT, and we're looking for more sites, so see Dr. Curley at the end to start your site up as well. So in PALIC, we recommended to titrate for 88 to 97 percent, 97 is the new 100. In PROSPECT, since it's a moderate to severe cohort, we're looking at 88 to 92 percent here. pH values that we pulled into PROSPECT in the acute phase of PARDS are exactly the same. So this is what we said, but now we're going to do it, put our money where our mouth is. And predicted body weight we suggested, and now we're going to do that as well. So there are four limbs to PROSPECT, conventional supine, conventional prone, high-frequency supine, high-frequency prone, and now we're enrolling high-moderate to severe PARDS, so an OI of 12 times 2, or an OSI equivalent. I don't know how to go back, but anyway, right-click, there we go, thank you. Or 1 for a fast responder, one OI of 16 or greater. And we're stratifying by age, so this is going to be as personalized of a randomization as possible, but then ultra personalization with open lung strategy that's very dynamic and active, pushing towards permissive hypercapnia and pushing towards permissive hypoxemia and a 24-7 treatment algorithm. So you've seen this before, and the take-home message is you want your lung to look like a chicken, and not like a piece of liver, right? And so the idea in all limbs for PROSPECT is to maintain the lung open, open, open, open, in the acute phase defined as the OI greater than or equal to 8. Patients are going to have arterial access, and they will have neuromuscular blockade for the first 24 hours. So how we're doing this in PROSPECT, at the start of conventional, you're on pressure control or PRVC. We're not using APRV in this trial. We're goaling a 5 to 6 per kilo tidal line, and this is a moderate to severe and severe group, and we're including that for pressure-supported breaths. We're going to try to goal a peak pressure of 28. This is not a plateau, but 32 if you have a high-resistance chest wall. And we're going to adjust rate to keep ourselves in this good pH, and we're not going to adjust based on PCO2. We'd like not to inverse the IDE ratio. Again, we're not doing APRV, and this is going to be stepwise or staircase PEEP titration. Then we're going to keep doing it. Every 6 hours at least, we're going to reassess these patients. We're going to titrate that tidal volume. If we're having problems with peak pressure, we're going to march our way down towards 3 per kilo tidal volume, again with the respiratory rate. And here's the PEEP-FO2 grid. That's the low PEEP grid from ARDSNet. And then we will continue to adjust that. But we need to be safe, and we need to know what to do when the patients don't respond as interest as we hope. We can increase to 100% if we need to, but if we're not doing anything severe for less than 10 minutes, let it go. But if we are doing something that we're not happy with, with a D-set of 85% or lower for more than 10 minutes, we want to make sure we examine that patient. We increase the O's judiciously, and then we do another titration. If we get significantly acidemic and we've adjusted our rate up, but now we're worried about air trapping, then we are going to potentially march our way up in tidal volume. In the oscillator limb, again open lung strategy and oscillating on that deflation limb, again by using stepwise or staircase recruitment maneuvers, using the highest possible frequency. And that's to deliver the lowest possible tidal volume to these patients. And also to initiate at 9 to 12 as opposed to our other typical lower Hertz that we may have used in some institutions. We're going to use a high fixed power setting initially with a delta P of 70 to 90. But again, if we've optimized the frequency, if we are able to stay on that descending limb of the PV curve and we're not over distended, then this shouldn't be delivering injurious ventilation to the airways. And remember, most of your impedance is at the level of the endotracheal tube as well. At the start, we're using the 3100 A and B. We are suggesting transcutaneous CO2 monitoring fluid bolus so that as we go up, because we're starting the mean airway pressure 5 to 6 above the current conventional mechanical ventilation map, then we need to make sure we can counter that and make sure we're perfusing the lung. And again, then we do a recruitment maneuver. In the ongoing phase, we fine-tune again every six hours and we titrate O's till we get to 40% here. We titrate frequency and power to maintain the highest frequency that attains that goal ventilation on that descending limb, not de-recruiting and not over-inflating. And if our pH is suggesting we have room, we're going to try to go up a little bit and nudge a little bit further to see if we can maintain that highest optimal frequency and that lowest tidal volume for those patients. If we've gotten to the limit, then we will decrease our power. Every 12 hours, these patients will all get a decremental titration maneuver. We're not going to sit on our hands and say, oh, we're going to sit on this map overnight and try again tomorrow. And then once your map is less than 15 to 20, then we're going to switch over to the conventional ventilator and use that PEEP-FIO2 grid. What if we've deteriorated in the oscillating limb? Well, if we have difficulty weaning, if we're very acidemic, we do stepwise prescription. Deflate the cuff if you're not already deflated, increase the power, decrease the frequency. If that's not working and your frequency is 8, decrease the bias flow, increase the eye time. So a stepwise prescription and an algorithm to follow in each of those instances. If we're not oxygenating well, then we're going to make sure we assess the patient. We judiciously increase those a bit until we're in a good range. Then we look at the gas. If our PCO2 is rising, it suggests we're hyperinflated. If we're hyperinflated, what do we need to do? We need to decrease the mean airway pressure. And that might just pop everything open and put us in that optimal map where we need to be. If we're not overdistended, then we do the other and we do an incremental recruit maneuver to find that optimal map again. So there is a strategy for recruitment maneuvers to be done in each of the limbs. Titration generally on the conventional only at the start or with a deterioration. And the oscillator, a recruitment maneuver to find that optimal map at the start, but then a titration to try to see if we can nudge our way and make sure we're not wandering out here where we don't need to be and wander into the optimal map each time. But if we're overdistended, making sure we come down in a titration. If we're under distended, making sure we pop ourselves back up to where we need to be. If we fail after maximal attempts in either the conventional or oscillator limb with significant desaturation or significant acidemia for four hours, then we failed in that limb. And then the investigators are allowed to choose another limb of their choice, although we're exited from the trial. And if there's ever a safety concern, then we call Dr. Barbaro in our institution and we consider cannulating for ECMO or doing whatever you think you need to do if there's a failure. So what else can we learn from prospect? And I know I probably only have two minutes left, but I can't not talk about this when I'm asked to give a talk. So anyway, one of the things we're also doing, and Dr. Dahmer's in the audience and my co-PI on this, is that we've recently got our notice, a grant award for Ensnare to study endotypes in children with severe PARDS. And so we know that heterogeneity does exist, as Robbie was pointing out to earlier, that a study might show no benefit, but there might be a subgroup that may benefit. But a study might show benefit for a whole cohort, but there might be a group that is harmed. And so we need, we have understanding that we need prognostic and predictive enrichment for these things. And I think everybody in the audience has seen this figure somewhere in their past. Carolyn Calfee and her colleagues, good friends of ours, have recapitulated this image in four ARDSNet treatment trials. I wanted to point out, though, what you may not be aware is all this data, the biochemical data and the clinical data, represent data before randomization. So this is what the patient comes with to the trial, not after the trial has been implemented. And there are, consistently in adults, two hyperinflamed and a hypoinflamed group. Hyperinflamed have high inflammatory markers and significant acidosis. And in the alveoli trial, which was higher PEEP versus normal ARDSNet, the higher PEEP limb actually offered a 10% mortality benefit to that group of patients. But that trial was stopped early for futility. So our hypothesis is exactly that. We do think that there's pathophysiologic differences and response to treatment. And our aims are basically that. We're going to try to do exactly the LC analysis. In fact, Dr. Calfee and Dr. Delucchi are great mentors for us in this. We're going to collaborate with Drs. Nader, Yahya, and Hector Wong on a transcriptomic analysis of RNA expression. And we're going to try to merge those two things together to find the ultimate refined endotypes. And so it takes a village, and I appreciate everybody. And thank you for your time. Did you all get that? Okay. We could take one question. If not, we could hold the rest of the questions to the end of the session. All right. One question, yeah. Everybody, that's a very good question. The question was, who can touch the vent in the study protocol? And you actually do have to be tested on understanding the protocol, and all clinicians who are going to be taking care of the patient where it's RT, RN, or physician, or nurse practitioner, or whoever, including resident if you wanted to, would have to be tested first, in which case then you can be part of the team. We have a nice model going on where, at Riley, where the nurse practitioners have been trained overnight to do the titration, so it's worked well, but it's a case-by-case basis. So thank you so much, my friend Heidi. All right, next up is balancing sedation and mobility with Natalie Napitano, who's a research clinical specialist at Children's of Philadelphia. She's also a PhD student at Drexel. She's trying to live through all her stats courses. So please, lots of applause, thank you. Good morning, everyone. I'm going to flip to the last page of the book real quick for this one, because we have no consensus on balancing sedation and mobility in pediatrics, let alone in pediatric ARDS. So instead of trying to provide you with that evidence going forward for the next 15 or so minutes, I'm going to try to convince you that we need to do research on this, and we need to have some sort of consensus, and that it's important for us to really be taking a look at. So as a respiratory therapist researcher, I do have conflicts. However, none of them have to do with sedation or mobility. So hopefully you'll find that at the end of this as well. So we'll do kind of like a quick background as to why this is important and why do we really care in general, and specifically in ARDS, take a look at some of the consensus statements that we have, and if they make any sort of a statement about how to minimize your sedation or optimize your sedation, as it might be, to assist with mobility and mobility in general. And the ability to use sedation to facilitate early mobility, what are the scores that we have, what are the agents that might be helpful? And then what does early progressive mobility really look like, and what are the options for pediatrics? So starting with our background, really the most of the background that we have is really from our adult colleagues. So the Canadian Clinical Trials Group did a prospective study looking at adult ARDS patients, and they looked at their survival after ICU and hospital discharge. And this was really that first look of what happens to our patients after they leave us, and maybe they survived, but survived to what end? And what is really their quality of life? What does the morbidity of their disease process look like now? So they followed these patients for 3, 6, and 12 months, and then after that first year, they decided they wanted to look farther and extended the study out to look at them for five years. So what they really saw here was it took about 12 months in order for them to regain all of the weight loss they had from being in the intensive care unit. They then also saw that they're looking at their PFTs, and they did full PFTs at 3, 6, and 12, and then as they followed through, while their lung capacity was all still there, the residual volume was all still there, they didn't lose lung, their functional capacity was significantly limited and actually never got back at that 12-month time frame to normal ranges for their age. And then when taking a look at their exercise capacity, their 6-minute walk tests, they never got back to baseline at 12 months. And then looking at their quality of life scores, they also never got back to baseline, their physical scores, at 12 months. And only 49% returned to work in one year. So we really have to take a look at not just the clinical medical management of ARDS during this time frame, but what are the things that we can be doing that wouldn't inhibit that process that could help them function in the long term, get back to what they perceive to be the best quality of life for themselves. Out to five years, we then saw, I apologize, this is a little bit hard to see on the screen there, but we saw that their, again, distance walk tests, when you went out farther than one year to five years, still wasn't up to baseline. We saw that their survival actually went down. So even though they survived to discharge, as the years went on, more and more of them, their disease processes kind of took over. And then looking at both their mental and physical quality of life perceptions, their mental kind of got back up to normal, and it was at least in that normal range, but their physical quality of life really never got back up to normal, even after five years. So there's definitely something that we can be doing, hopefully, to make this better. So mobility in general, or immobility, is associated with muscle loss. And this begins within 48 hours that they've seen in adult patients, and is greatest at about that two or three week during critical illness. And up to 40% muscle loss occurs within that first week. So there was a trial looking at children that looked at muscle weakness, and it was identified about day four of ICU admission. Their proximal and distal muscle wasting occurred in 57% of children with intensive care unit acute weakness. And then at three months, follow up, 57% still had persistent proximal muscle weakness, and 14% of them still couldn't walk independently. So reportedly, from a surveillance standpoint, in a lot of the studies, 25% to 100% of adults get this ICU acquired weakness, and then 30% of children. And that 30% surveillance for children really is a cautious number, because there's really only been two studies thus far that tells us post-discharge what's really occurring. So we definitely need a lot more data on that. The severity of diaphragm atrophy is something that's also being looked at to directly correlate to the severity of limb weakness in adults. And this also has shown that the length of mechanical ventilation also has a direct correlation with the atrophy of the diaphragm and the atrophy of the muscles in the limbs. We did see one pediatric study looking at this that looked at pediatric patients, and they did see an increased severity in this diaphragm atrophy in patients that saw neuromuscular blockade during their illness, as well as that had low levels of spontaneous breathing. So we really looked at those who were doing a lot more of their spontaneous breathing versus those who did not. So really, this ICU acquired weakness, as it's been termed, is an independent predictor for length of mechanical ventilation and ICU duration in adults. We don't really have any good pediatric trials yet looking specifically at this, but that's something to look towards in the future for all the clinical trialists in the audience. So looking towards our definitions and our consensus statements, we've talked about PLEEC a lot lately, particularly in this session. And really, PLEEC says we should have minimal use of sedatives and pain. We should be using our scales, using our goals, depending on whatever scale it is that you use. And those are all strong agreement, but there is absolutely zero mention of progressive mobility in the PLEEC definition. So there was really no data for them to go off of at the time, so that was probably partial to it. So SCCM rolled out our ABCDE and then the addition of the F bundle to address a lot of different things to help with this, and particularly for this talk, the delirium monitoring as our D, and management, and then early exercise and mobility for the E arm. And this has been applied in adults, and how well this works within the team, and how well using this bundle in general really helps the functional outcome of patients afterwards. And they saw that in adult ICUs, there was a reduction in the duration of mechanical ventilation, length of stay in the ICU, and cost per case. But we really don't have any outcome-based evaluations of using this bundle in pediatrics yet. So there has been some feasibility, how does this work within our daily rounds, evaluations that are more observational, but nothing looking at the functional outcome of that bundle as of yet. So the use really for our sedation options that can help facilitate mobility, it's really hard to make that determination when you look at pediatrics as a whole. And looking at your infant to your 18 years of age, it really is going to be person-dependent, but trying to keep that in the minds of when you're using your whichever scores you use for that phase of illness, your pain, your sedation, your withdrawal, your delirium, and making sure that we're using the least amount of agents possible that is clinically appropriate for their phase of disease, and that can help facilitate some sort of mobilization. One medication that has been the magic pill for a lot of us in trying to do a lot of things is the dexamethamidine, which has really gotten vast use overall in PICUs. We don't have any really good data and safety on that specifically in pediatrics as of yet, but that's something that definitely can help with giving that bit of sedation and not inhibiting your respiratory drive so that you could actually help facilitate some understanding and mobility in children. But again, no evidence on use of this. Where's the best scale number per scale that could be used to help facilitate mobility that really needs to be patient-dependent and depend on the disease? So looking at the application of physical activity within the first two to five days of life is really, or two to five days of critical illness, excuse me, is really what we're talking about when we're saying early mobility. So they're in that acute phase process within the unit. They're the sickest of the kid, and now we want to get them moving. And that's what's really what we think is going to help. So options for progressive mobility, really, progressive is a building process. So we can start with a lot of the passive mobilization pieces, getting an OT or PT consult when they hit the door, and let them help guide that process and what they can handle within the goals of care that you've already provided. It can include a passive range of motion, just moving them around, massage. These are things that we can actually kind of help get parents engaged in at the bedside. They want to do something, helping to massage their legs. And this has been shown to really help to decrease pain in adults, stiffness, swelling. And it also has been demonstrated to have improved functional exercise capacity and perceived functional status later on when you actually just keep movement instead of just lying there. And then work up to more active mobilization pieces, active range of motion if they're able to participate, resistance exercises when they're able to participate, and then just sitting up in the bed. Sitting up without something behind you really helps to create that core muscle strength that's going to help with your diaphragm. It's going to help with all of your ancillary respiratory muscles to help wean better from the ventilator. And then dangling, sitting in the chair. There's a lot of things that we get to before we're trying to actually walk these patients if they're someone that can walk. If they're a three-month-old, we're not going to try to walk them, obviously. But they should be able to do maybe what normal three-month-old babies do. They should be grasping. They should be looking at things and trying to do additional stuff within the confines of a therapy session. So we do have two pediatric trials to take a look at. One was a prospective cohort study that looked at active mobilization using more of a virtual reality gaming consoles. And they looked at just eight patients. They looked at feasibility. They used them at least once per day. And they had no adverse events. But they did show that there was a greater upper limb activities, obviously, as they were doing things. But a lot of the Wii and the Wii Sport and the dance moves and the different ones that they can do is what they were kind of working on. And there was a second study that came out, which is another proof of concept trial, 25 patients, ages 3 to 18 years of age, that also used active and passive. They used some of the cycle ergonometry that you can get. A lot of the adult hospitals will have these little cycle movers. They're used mostly for your knee replacement surgeries in adult hospitals. But they move both legs. And it's passive. So you just hook their legs up to it, and it keeps them kind of moving. And they don't have to really participate or be awake to do that. But we're still moving their muscles around. And then they also threw in some of the virtual reality interactive gaming consoles when they were awake enough to be able to do those things. Both studies show that it was feasible to do these things, particularly child life has all these games already that we're using. We can just use the different games within the consoles to kind of help with some of that. And then no adverse events for any of these. None of them were really active. They weren't trying to get patients up and walking them as of yet. But at least safe for what they were doing. So in conclusion, there's really no literature to go over for you that is very specific to increasing mobility and pediatric ARDS. There's a little bit of data in pediatrics in general that early mobilization can be beneficial. So we're really looking from our adult colleagues to really show what it is that can be done and what the benefits could be within children to give us an idea to start looking at those within the pediatric population. It obviously would need to be something that is geared towards that age patient. And that is one thing that OT and PT folks are really good at doing. The use of sedation and delirium scales with treatment goals to facilitate mobility when they're clinically stable to do so are helpful in keeping that in mind on rounds, keeping that within the bundle of our ABCDE bundle that it's always talked about as to when we can keep moving forward. And I think doing actual passive movements with them helps keep it on everybody's mind that this is something that we should be doing. So the strongest recommendation that I could potentially make is really to get OT and PT involved early on. There are a lot of things that they can do that we don't even think about. And they're also there to help kind of push the team to say, hey, can we do this? Can we do this step next? How do you feel about this? And oftentimes, we found, at least within our hospital and our severe chronic lung disease group, that we have very strong OT and PT involvement within those patients. And when we have them on non-invasive ventilation, using them as a guide to help us know when we can wean. Because we wean, we think they look good when they're just sitting there doing absolutely nothing. But as soon as you try to sit them up and move or just do grasping and actual normal care for an infant, they get tachypneic. They start getting diaphoretic. So you can't really, if they can't do anything but lie there and breathe, then we're not giving them adequate respiratory support. So it can really help guide that process so that you know you're weaning appropriately and really help you move back up and step up their respiratory support before they really decompensate on you. And that's it. So open for questions. Thank you. I was wondering if Sapna Kutichar was here. Did she leave, take her picture and leave? She did. I think she did. She came back in and then she left again. OK. So right now, I know Sapna Kutichar from Hopkins is really leading the charge behind Pick You Up. And I know that she recently submitted a step wedge designed to study the effect of early ambulation on clinical outcomes. So we'll see how she fares on scientific review. Anyone here currently doing the ABCDF bundle in your PICUs? A few. And what kind of experience have you had with it? Yeah, I think that's one of the biggest pieces, and I'm excited to see the results of her trial and actually as things start to progress forward, that starting with the passive pieces just keeps on everyone's mind. And really keeping that specialist involved in the care, OT and PT, both one, whatever you're allowed to have, can really help kind of move that forward. But I know it's challenging to get them involved as much as we want them to be in the ICU. So starting and asking for it and trying to say that this is important to us really helps them get more people involved in the ICU as well. And along those same lines, with the same ask, in our own institution, our nurse practitioner team, Danielle Van Dam and colleagues started early mobility algorithms so every patient is assessed every day for their phase of care, so that the whole team can identify at least within their certain acuity of illness what we could be doing passively, and then each day as they move on to try to be able to mobilize them more. And so that's been a really great help, and now trying to sustain it and do it all the time is sort of the focus that we're at. But doing something repeatedly and making a part of your standard work, I think, is potentially going to be an important way to go. Any other? Okay. Thank you. All right. Shirley, not the last person to come to the podium, but really Ryan Barbaro from the University of Michigan, who also leads the ECMO program at Michigan and also the ECMO registry. So he manages that. So assistant professor, but I think really a rising star within the critical care environment. And I have to say, and I want you all to applaud that he submitted this wonderful proposal to NHLBI, and I think it's going to hit the pay line. And I hope that he's going to be sharing that today. Thanks a lot for that very, very generous introduction. And so I am going to talk about something that's near and dear to my heart, ECMO and pediatric ARDS. And yes, while I was vaguely waving my hand at some of the research that we need to do, what is implied in my slides I will more explicitly speak to, as Martha has already kind of spoken to as well. So thank you. By way of disclosures, I do want to acknowledge that I am the ELSA registry chair. So I do think about this stuff a lot from an ECMO perspective and have an investment in it. I also get the support of a K-12 called Tactical, which is an implementation science grant. And so just first, I mean, everybody in the room will know, but I think it's worth stating, right, when we decide that we're going to put a patient on veno-venous ECMO, we're committing to systemic anticoagulation and cannulation. And so just walking through this, obviously, we're going to be draining venous blood, powered generally by a centrifugal pump, pushed to a membrane lung, which removes CO2 and adds on oxygen. And then that oxygenated and decarboxylated blood is generally returned to the right atrium, where the native heart pushes it out to the patient, gets to the lungs, where minimal gas exchange occurs, and then out to the patient to be able to utilize it. And then much of the theme, as we've talked about here, when we're talking about what's the evidence of ECMO support in severe pediatric ARDS, we're really borrowing from two bookends. We're borrowing from our neonatal patients and also from our adult patients. And specifically, I'm going to focus primarily through these slides on two randomized control trials. So this is a remarked study in neonates, the UK neonatal trial, which is now old enough to be a medical student, published in 1996. And also the EOLIA trial, which just came out in 2018 by our friends in Europe that looked at adults with severe ARDS. And so as I said, we're talking about neonates here in the UK trial, less than 28 days, but also with perinatal lung disease, so something explicitly excluded from the PALICC definition. It was meconium aspiration, congenital diaphragmatic hernias, as well as pulmonary hypertension. These patients were intubated for a short, or needed to be intubated for a relatively long time compared to the adult trial, but less than 10 days. Also they needed to have significant poor derangement of gas exchange, so an oxygenation index of greater than 40, or a PaCO2 greater than 90 for at least three hours. EOLIA, much more recently, again looked in severe ARDS, as I mentioned, also limited mechanical ventilation to less than seven days. They had three entry criteria from the perspective of gas exchange, so you could have a PF ratio of less than 50 on 80% oxygen for three hours, or more, a PF ratio of less than 80 on 80% oxygen for six hours, or a pH of less than 7.25 with a PCO2 greater than 60 for six hours. And so what were the findings of these studies? So looking just at the UK trial first, ECMO had a roughly mortality rate of 30%, whereas conventional mechanical ventilation, one of 60%. That's a relative risk of .55 with a statistically significant P value. In EOLIA and our adult colleagues, the mortality rate in ECMO was 35% compared to 46% in those treated on conventional mechanical ventilation for a relative risk of .76 that was not quite statistically significant. And so EOLIA was actually stopped early because it had met a pre-specified futility specification at the kind of fourth interim analysis where they had enrolled 240 of a planned 330 patients. One of the things that the study didn't incorporate as much is what the crossover rate would be. And so there ended up being around a 28% or 35 patients crossed over from the conventional to the ECMO, and that probably weakened a little bit of their power analysis. And so subsequently Gallagher, who's been cited here before for others of his work, did a post hoc Bayesian analysis at EOLIA trial. So when we talk about Bayesian analysis, most of our statistics from the randomized control trials that we do are frequentists. What a Bayesian analysis is doing really is it's recognizing there's prior information. Oftentimes the Bayesian analyses that you read actually take a really just unspecified prior. It doesn't consider too much what is known about the world. But here they actually added information from three previous studies. So the CSER trial in 2009, as well as two propensity score analyses, one done by NOAA published in JAMA and one done by FAM published in the Blue Journal. And they incorporated that into an intelligent prior. And then they looked at, hey, with that information and adding in increasing amounts of EOLIA, what's our belief about what ECMO actually did? And so you can see a relative risk reduction of just anything positive, significant confidence. They used a mortality rate of 46% here, which is what they found in EOLIA and that controlled them. So this would be an absolute risk reduction of around 4%. There's a lot of confidence there. But then as you increase this, so this would be like what they found in that initial ARTSNET low versus high title volume. This is what they found in the CSER trial. These are large risk reductions. There was less confidence about that. And so what does, so all right. In the neonatal population, it looks like, yes, we did have a clearly statistically significant benefit of ECMO. In the adult population, I think that various different people will perceive this differently. But it seems most likely that there was a positive effect of ECMO there. So what does it mean to talk about early ECMO? Well, I think first we just need to talk a little bit about where does early respiratory ECMO fall. And so we're not talking about the things that most of us have been talking about the rest of this talk. So this is OIs of 4 to 8 is what we said was mild, moderate is 8 to 16. OI of 16 is severe ARDS. That's where we've been stopping. I'm all the way over here at 40, almost like three times what we've been stopping at. And that's early ECMO, right? That's where they enrolled patients. So I mean, early is a little bit of a funny word to use. But I think we can talk about early in two ways, right? We can talk about early in terms of exposure to hypoxemia or exposure to mechanical ventilation. Here I'm just going to talk about hypoxemia. And so just to orient you on the top here, same SAT, same PAO2. It just helps with the numbers coming through. But this is a patient on an FIO2 of 60%, PEEP of 10, peak inspiratory pressure of 28, amine airway pressure, say, about 17. That person has a PF ratio of 97, and they would have an OI of about 18, so just barely severe ARDS. And I think that most of us would be like, yeah, that's not a patient I'd put on ECMO. But when we talk about a patient who has those same oxygenation parameterizations with an FIO2 of 100%, a PEEP of 16, a peak inspiratory pressure of 30, and a MAP of 24, with a PF ratio now of 58, and an OI of 41, some of us would put that on. Some people might say, hey, I'm going to wait a little bit longer. This is what I'm talking about when I'm talking about early. And so when we talk about early as an exposure to mechanical ventilation, it's so conflated with what's going on with oxygenation that it's really hard to think about in terms of a randomized control trial. But what we do have is observational data from the ELSO registry. This was published by Zabrocki and colleagues in 2011, coming out of the ELSO registry, like I said. And you can see that, really, there was a substantial step off at 14 days. It was statistically significant, but wasn't really observed earlier. And then just to remember, so that UK neonatal trial said, hey, you had to have been mechanically ventilated for less than 10 days. EOLIA said less than seven days. And those were probably good marks, because it looks like at 14 there's something different going on. All right, so now early. Is there evidence? So in the perinatal lung disease and in the ARDS, there does actually look like ECMO might be better when it's done earlier. And so this is just the OI. So when your OI is between 40 and 60, you can see here that actually your risk of mortality was farther to the left good, versus when it was greater than 60, it was starting to get over towards actually maybe not having as much of a benefit. And then when you look at the EOLIA trial, so this is just, they just took the median and then looked at it plus minus. And so here, three days was the median number of mechanical ventilation days versus greater than. And that didn't seem to have much effect. I think, again, thinking back to that Zabrocki slide, that's not surprising. But when you look at the PF ratio here, so 66 versus greater than 66, those patients who were less ill did better. So when their PF ratio was higher, 66, they had a relative risk of 0.56 and a value of about 0.05. And this is for an interaction between the mortality outcome and the PF ratio. All right. So does ECMO research in perinatal lung disease and ARDS apply to pediatric ARDS? And so, I mean, I guess specifically we would say that by the PALIC definition, no, we excluded it. So this is a pretty convincing argument that in terms that kids aren't little adults. But also throughout our talks, I think what we've come to is also still borrowing from that evidence. But here, I want to just turn to, well, do we do ECMO the same way in these populations? So this is data just directly from the ELSA registry over the last five years. These are neonates. And when we, in the ELSA registry, speak to neonates, we're speaking as should be zero to 28 days. But when I'm saying the word pediatric, I'm actually carving out those neonates 29 days to 17 years. So this is all children together. And this is adults. And what I'm showing you here is in respiratory ECMO only, what are those patients who had their carotid artery cannulated? As I start to talk to my adult colleagues more and more, I realize that when we say the word VA ECMO, they assume we mean somewhere other than the carotid artery. But that's not the case. We mean the carotid artery most of the time. And so 80% of our neonates have their carotid artery cannulated. That's about a third outside of the neonatal period and less than 1% in our adult colleagues, which is why they don't think about it. So what about the complications that we see across these age groups? So our neonatal pediatric and adult using that same kind of age frame here, so zero to 28 days, 29 days to 17 years, and then 18 and older. So you can see that in terms of the hemorrhage that we're having in the brain, that that is much more common in the younger age groups than it is in the adults. Obviously you're often seeing things like intraventricular hemorrhage, which is maybe different, but still much more common. Infarcts maybe about the same. Seizures maybe about the same, maybe more commonly in neonates. Here within the lung, for whatever reason, it looks like we are observing or at least reporting more pneumothoraces. And then dialysis is just overwhelmingly more commonly used than it's used in the adult population. When we talk about mechanical complications, these aren't necessarily surprising just based on the way that we're doing it. So these are usually open cannulations, and our cannulocyte bleeding is much more common here. In terms of hemolysis, again, you have smaller cannulas, you have higher RPMs, you have more hemolysis. Our circuit failure, you have lower flows, so you're getting more clot and things like that. None of these things are surprising, they're just different. And our oxygen and pump failures look like they're about the same. So when we're talking about transitioning from the mechanical ventilator to ECMO, we're talking about really weighing risks and benefits. So we can certainly apply what we've learned in our adult ARDS and in our perinatal lung disease. But in the absence of evidence, we can also think about, in our patient population, what's going on. And so as severe pediatric ARDS is progressing beyond an OI of 40, so we're talking about really sick lungs now, you are going to be at a place where you're going to be starting to think about, are we having more ventilator-induced lung injury? Are we having clinically significant hypoxemia? Is this causing organ injury and potentially even leading to death? And so when I think about this for ECMO, really, I think around this on a continuum. And so if you're talking about, even before we're talking about early ECMO, and you're starting to think about exposing somebody maybe with just an OI of 16 to ECMO, you may be worried more about that you're going to actually expose that patient to increased brain bleeds without maybe the benefit that we need. When you're talking about the way we observe patients being treated on ECMO, you can also say that these patients do seem to have longer critical illness or longer ICU stays. They also seem to get exposed to more sedation, both comatose sedation as well as neuromuscular blockade. And so maybe those things are going to have some downstream effect in terms of their functional status overall. When we're talking about increased lung injury and hypoxemia, as I alluded to, maybe that's going to lead to some increased organ failures. So we're making trade-offs. And so when I talk about ideal early veno-venous ECMO for pediatric ARDS, really what I mean here is that, ideally, we would have evidence in our pediatric ARDS population to make informed decisions at the bedside that also take into account what happens long-term to these patients. We just don't have it. So then it comes to the question of, are you going to base that decision on perinatal lung disease and severe ARDS in adults, or are you going to base it based on the complications and the various different strategies in which we deploy ECMO to our children that have pediatric ARDS? And I think if I could look into a crystal ball, I would know those patients that would go on to have progressive lung disease at the OIF-40, and those patients I would put on ECMO because, based on what I can see in the perinatal lung disease population, what I can see in the severe ARDS population, it seems like that would do better. And since they both kind of came to a threshold of initiating ECMO, though articulated differently that seems to capture a similar patient population, I'd say that probably is right for children. But if you're talking about before that OIF-40, and you're talking about patients that right around that threshold may actually stabilize or plateau in their disease, and you might be able to avoid that carotid artery cannulation for many of them, and that systemic anticoagulation, I'd say it's nice to be able to avoid that. Maybe you do shorten some of the exposure to their ICU stay and to the sedation and things like that that may have long-term effects on their outcome. And so that's all I have. But I will just, before I conclude, just speak to what Martha had said. And so this is something that I care deeply about. I think we do need to inform this within our patient population. And so I bothered to actually put in a grant to ask this. And so we were piggybacking off of the PROSPECT study, where they have very specifically characterized how patients will be delivered mechanical ventilation, whether that's on the oscillator or the conventional ventilator, and also specified where thresholds of failure meet. Those thresholds of failure do happen to be beyond the oxygenation index of 40 usually. And so there's an opportunity using the ELSO registry for us to look at patients who are in your natural practice receiving mechanical ventilation and seeing if those patients actually, when they go on, because they do just typically end up going on closer to an OI of 40, if they fare differently, both in terms of their short-term mortality as well as their long-term outcomes. And by that, I mean because PROSPECT is looking at 1, 3, 6, and 12 months FSS, POPC, PCPC, and health-related quality of life, we will do the same. And the group of investigators that will help us do that are the same group of investigators that executed the ROSE trial, that randomized control trial, that looked at this in adults after neuromuscular blockade. They're really skilled in doing that. And so hopefully we'll have an opportunity to build on some great science with some more science in a place where I think we have some shortcomings, but hopefully we can fill those gaps. Thank you. I definitely think our adult colleagues will push it to the left, lower. And so the reason for that is two things. One, I don't think that adult ECMO is the same as pediatric ECMO, beyond the cannulation. Also when you talk about the flow rates, right, when you're flowing at five liters a minute and you have large cannulas with no step-offs, you just have less risk for clotting, you have less risk for hemolysis. They can run, and I think they will be running in the next 10 years, circuits with much less or no heparinization at all. And we can't get away with that. And when you look at the complications, you see that. And so I think that hopefully, whereas I think that the pediatric population gifted ECMO to our adult colleagues in the way that we sustained it through the UK neonatal trials and built upon it, I'm hoping that our adult colleagues will help us to understand how we can better anticoagulate as I do think they'll shift this to the left now more and more. Other questions? It's interesting when you were struggling with the design of your trial. Can you just talk a little bit about that? Yeah. I think there were so many things I struggled with. And so I guess, you know, one of the things obviously- Let me just interrupt. So there's, I mean, optimally, you can't do an RCT here, you know, so, and to convince everyone without an RCT that you should cannulate within 24 hours of coming into the unit is a little bit differently than, you know, what Perlow Work used to say is that the patient's going to die. And that's why we need to put the patient, you know, on cannulation. So trying to get the reviewers away from an RCT, because we can't do that, you know, Jim Fackler tried to do it 20 years ago and nobody would randomize to where you are now. Yeah. I think, you know, I guess, so I think what Martha has spoken to here is a reality, right? Like when we ask us at the bedside to make a decision about whether we're going to put a person on ECMO or not, we feel like just that has to be a clinical decision rooted in what we think is best for the patient because we feel the stakes are so high. And in fact, in your study long ago when you asked parents that, that's exactly what they said back to us is that is how they feel. They're accurately reflecting what we feel as clinicians. And I think that in that case, right, no, we're going to have a hard time randomizing until we have better clarity on what's going on here. And so, yes, I think that taking advantage of a place where we do have an opportunity where somebody has clarified prospect exactly where that threshold is, and then looking at the random noise that we see in our patients as we all clinically make that decision at the bedside and often more early, that hopefully we do have an opportunity to see one way or another whether these things lead to a better effect. I do need to acknowledge, right, that because prospect is completely characterizing the mechanical ventilation strategy that you're receiving and that we all provide mechanical ventilation differently. It is not just the threshold that we're looking at, but also the way that you're delivering mechanical ventilation. And so it's, yes, taking advantage of an opportunity that exists to help fill a gap and hopefully build us forward to answer more actually interventional trials down the road. Thank you. Should we grab a couple of chairs or something? Yeah. I think we're fine. Nader, you had a question, and I blew you off, so you can come back now, please. It's all right. John can't hide forever. I really enjoyed that peep talk. I noticed a couple of things. I was hoping you could clarify something for me. I heard the PF ratio got better when you took the plateau to 45 in that one example. And so that's oxygenation. I saw an EIT image, and you mentioned imaging over distension and derecruitment. You could do ultrasound, EIT. All of these things are potential. CT is probably not going to be feasible. You mentioned lung mechanics, which I'm grossly interpreting as either driving pressure or compliance, and probably globally, okay, as opposed to using ventilator measurements with or without an esophageal catheter. I saw a VDVT or dead space, and there's like a fifth one, but I think those are the big ones. You're getting my point. There's four or five different things that one could optimize to get the ideal peep. They will likely, I suspect, give different peeps. And while there was a comment toward the end in your summary to maybe deemphasize oxygenation as much as maybe prioritize some of these other things, like imaging over distension or focusing on lung mechanics, I can't help but notice that in the one specific patient example you gave, okay, it was actually the PAO2 jumping from 130 to 430 that actually made you guys feel that the lung had opened. It's suggesting to me that a lot of us, you know, despite the best instructions, will pay attention to oxygenation. So I was just wondering if you could clarify that. Like what, like if these different metrics will give different answers for optimal peep, which one would you suggest we focus on or prioritize? So that's what happens when you share your dirty laundry with smart people. No, I'm, this is a mea culpa. I think we've all focused for too long on oxygenation. It just makes us feel good. So that small trial of steroids recruitment that John Keir did, that was a long time ago. And the PAO2 is a stopping rule. We just wanted to mimic, actually Motto did the same thing. I do think we should be focused on compliance or driving pressure as a surrogate. I think that's where the future's going to be. And I don't know if that driving pressure trial is enrolling yet, but it's been approved and sites are accumulating. I think that's, we need to wait for that data. But for the moment, we've got to use a lower FIO2. We need to keep SATs around 90. You pick the number. And we need to follow OSI. You know, there's compelling data that the OI, and now the OSI in patients who don't have an A-line, is a very powerful predictor of mortality. So if the OSI or the OI are rising, something's not going right. Can I add also, I think that, as John pointed out in the talk, there are so many things available to us on the machine, right? You can get a compliance. You can get a dead space. And when we look, for example, at our other favorite topic, sepsis, we don't just look at one feature in deciding what to do with that patient. We use all of those features. And so that, we never say things like volumetric capnography, oh my goodness, right? But that's another nice piece we could add as well in thinking about what to decide. And I do want to put just one other piece in terms of driving pressure is that, ideally, the patient's not participating when you're measuring driving pressure. And we're doing a lot of a wake up and breathe. And so we have to make sure that we're careful when we just willy-nilly decided. So use a lot of metrics and a lot of data, but use it appropriately. Make sure you know what your gold standard is and not just an interpretation. Thank you, Heidi. The concurrent session that I am missing, because I love you, is, in fact, sepsis. And should we measure, there's one dude talking about SCV02, somebody talking about lactate, somebody talking about peripheral perfusion, and about two or three other metrics. So this exact conversation is parallel to sepsis. And while we do use all those informations for setting up a trial or for operationalizing some of this stuff, there's a balance between feasibility of follow this one metric. This is your surrogate marker for response or non-response. Because everybody up here talked about identifying responders as the one that you should be trialing and you're not going to see a signal in a heterogeneous population. So that's usually easier to operationalize with one consistent, clear response metric. And so that's kind of the point. You're going to get different answers if you pick the wrong one. I mean, I think, too, with the PRONE study that we ran, the first PRONE study that we ran a long time ago, we were able to show the numbers got better. But the outcomes did not. So that's why we're back here about 20 years later combining therapies in a way that is essentially reflective of practice. But we have only like eight more minutes. And I wanted to give everyone one last opportunity to give the audience one take-home message. If you were to give one message about, you know, the topic, PARDS, essentially severe PARDS, what would that be? You each get to say something. I'm not going to pick which one goes first. Sure. I think, for me, the point was sort of brought up by the questions that came up in the definition talk is that the definition is only so good at identifying a cohort of patients for us to look more closely at. But really, as we're starting to think about what therapies we're going to apply that needs to be individualized for that particular patient. And it may be very different upon either their phenotype or their responsiveness to test maneuvers to see if they're going to benefit from certain strategies. We can't capture all of that with the definition. So I want to support what Robbie, a lot of us have said. This is a very heterogeneous disease. And I don't know what the right parameters are to follow or optimize. But we're doing very well with mild and moderate PARDS. And I think that we need to focus on the patients who are in toxic ventilator settings and or who have severe PARDS. And I think that's where a protocol-driven and intervention-driven trial will make a difference. It may be ECMO. And, you know, there's North American early intervention, which is 40. There's European early intervention, where that's the best lung protective intervention we have. Take them off the ventilator. Use NovaLung. So I'm going to be very interested if someone actually publishes data, not just the NovaLung registry, about early intervention of ECMO. There may be adult data coming from Europe. Trying not to repeat Robbie, because that was going to be my point. I know, right? I got to go first. I know. So I definitely think we need to be individualizing our approaches based on the age of the patient and what type of ARDS they have, as it is a syndrome. And they all kind of respond differently. And there should be some sort of an individualized algorithm to determine if they are a responder or not, and depending on whether or not they are and what their baseline lung disease is. Because a lot of these studies are looking at these very nice, clean-cut ARDS with no comorbidities. Like, our exclusion criteria for prospect is quite long. And unfortunately, we still have those patients that come in with ARDS, and we need to try to figure out, how do we ventilate them appropriately, and how do we treat them appropriately? But we have no studies for them. So we kind of need to look at how all that works. And I think to Nader's point, we do need to be using EIT more, which we've talked a lot about here, but is still yet FDA-approved in the US. Although we have a machine, and we could be using it more. It's not FDA-approved yet, so we need to get that consent on everyone. And esophageal pressure monitoring and a lot of these other pieces that we can bring back to help us make those determinations for that PARDS with all the additional comorbidities on top of it. I think it's very rare for us to consider the ventilator the treatment, right? We think the ventilator is the support. But the ventilator is actually the treatment, and I think we need to be able to recognize it as such, and recognize that we don't want to be too aggressive and pull things back when support is still needed. But we need to be able to recognize what we're deciding on in terms of whether it's hurting the patient or benefiting the patient, because it makes my numbers look good, and I like that better with what I'm looking at. And that's hard to do. It's hard to toe that line, especially around the clock. And I think that treatment is being given around the clock, and I think we need to pay attention to that. And we need sites to participate and prospect, because we are good at mild and moderate PARDS, and we are good at getting that crappy-looking patient in, and then doing some nice work on them, and then they're no longer in that severe PARDS category. So I think that's one thing. I guess for me, my take-home would be, you know, in pediatrics, I think we're continuing to start to do this, but in pediatric ARDS, I think because we don't have a good convenience sample, we really need to improve in looking at what the long-term effects are, and that's especially true for those patients who receive ECMO support. And so hopefully, with prospect, and in all of our experiments and studies, that we can continue to look beyond our ICUs about how those patients are doing once they've left our hospitals. Thanks for adding that in, because I guess that's probably what I should have had me probe. I think it's important to all of us. So it's interesting, if you think about historically where we've been with ARDS. We finally have PARDS. We finally have pediatric ARDS, because we finally have enough data to be able to at least operationally define it. We've been struggling for a long time, like on BEST-PEEP, on recruitment, and how best to do that. You know, where does the oscillator fit in, especially when we get all these negative trials coming from the adult world? And now we know that maybe we can make a significant impact in how we can lead the field, and how to better do this. I think there is a place for ECMO. When kids come in and they're dying, you know, a set of cannulas is probably not as bad as, you know, some of the illustrations that you had in your grant application, so great. But I think we are all challenged to pull this off together as a community. And we've been talking about PROSPECT. It is, you know, right now we've got 41 sites actively screening. We've got 35 patients who have been enrolled. I think we had two during, as we're sitting there reading, you know, we've got two who are coming in. We need to get to 100 by June 1st, if not within the NIH structure, by funding. So yes, you know, we thought we were going to run a 48-site clinical trial. We have yet, the UK is coming in, they're coming in now, but we need to get the patients enrolled. And one thing as a trial is you learn how to eliminate a disease is to start studying it. So yes, we are studying it. Hopefully we can eliminate disease. But hopefully with a PROSPECT clinical trial, we'll at least have some data to parse out the heterogeneity of a fact that plagues this. I personally think the definition of PARDS now needs to go into a 102, a 103 version, because once you define it, the treatments vary about how you got it in the first place and what that's all about. And looking at numbers really doesn't make things better. It's really about being at the bedside 24-7, being at the bedside, because a half an hour, two hours, four hours of toxic settings is not going to turn the patient around. So we need to do things differently, and I think we're challenged, but I think we've made some progress. Maybe not enough, but we're going to keep going. Anyway, thank you all for participating.

pediatric ARDS

challenges

defining

treating

consensus

PALICC definition

individualizing treatment

ECMO

severe pediatric ARDS

early intervention

worsening oxygenation

lung injury

optimal timing

criteria

long-term outcomes