Ladies and gentlemen, please welcome Dr. José Díaz-Gómez. Good afternoon. It is my pleasure to introduce this year's Peter Safer Memorial Lecture being presented by Dr. Daniel Talmor. Dr. Talmor is the chair of the Department of Anesthesia, Critical Care, and Pain Medicine at Beth Israel Deaconess Medical Center, and the Edward Lauenstein Professor of Anesthesia at Harvard Medical School in Boston, Massachusetts. He's a longtime contributor to the Society of Critical Care Medicine. Dr. Talmor's research interests are centered on the early identification and treatment of critically ill patients, with a particular focus on the optimal delivery of mechanical ventilation. He has published over 200 peer-reviewed papers, as well as multiple book chapters, editorials, and abstracts, and has lectured extensively around the world. Dr. Talmor has served as principal investigator on a number of national lung and blood institute-funded and is currently an investigator in the NHLBI Network for the Prevention and Treatment of Acute Lung Injury. Today, Dr. Talmor will present the lecture, ARDS, Will Precision Approaches Move the Needle? And will be available for live questions and answer session immediately following this presentation. Ladies and gentlemen, please welcome Dr. Daniel Talmor. Good afternoon, everyone, and thank you for joining me. I'd also like to thank the organizers of the conference for inviting me to give the Peter Safar Memorial Lecture. My name's Danny Talmor from Beth Israel Deaconess in Boston. Before I begin, I'll say that I have no really relevant conflicts of interest to this talk, so then I have a lot of opinions, and I'll be stating them during the talk. Now, I think it's fitting to acknowledge the times we live in right now. And I think back to this book, The Tale of Two Cities, by Charles Dickens, and what is maybe the greatest first paragraph in any book in the English language. It's the best of times and the worst of times, the age of wisdom and the age of foolishness. And we as critical care physicians are living all of that. We're living the worst of times. We're seeing hundreds of patients in our ICUs extremely sick. But on the other hand, it's the best of times. This is the time when everything we've learned and trained for in our entire careers is coming to bear, when we are as valued as could possibly be by society and are facing an intellectual challenge like none other. And at the center of that challenge is really our old friend, ARDS. COVID-19 manifests itself in acute respiratory distress syndrome. This is a disease that's been around for more than 50 years, one that we've battled with often and one in which we're making great progress. And I'd like to describe some of that progress to you today. And I think it's always important to remember where we came from. And in the beginning, in ARDS, we focused on restoration of normal oxygen and CO2 levels. So this was our initial treatment modality for mechanically ventilated patients. And it was based on studies, published studies, the excellent studies for their time. For example, this one in the New England Journal in 1963, which demonstrated that patients under anesthesia faced a large drop in their oxygenation. But that using continuous large tidal volumes or periodic deep breaths, hyperinflations, we can prevent the increase in the variable shunt that is atelectasis, that is hypoxemia. And really, this was our approach to mechanical ventilation for over 30 years. We delivered high tidal volumes in an attempt to maintain normal oxygenation. That was reinforced after ARDS was first described in this paper in CHESS by Tom Petty and David Ashbaugh, the people who first described ARDS. They had learned that high tidal volume ventilation with both inspiratory plateau and positive end-expiratory pressure improves oxygen transport across the lung. So this is the state of knowledge really until the mid-1980s. At that time, we became more aware of the importance of preventing ventilator-induced lung injury. Several smart critical care physicians began to describe how the ventilator itself can injure the lung and actually exacerbate the ARDS. And how was that treated? So this is the famous graph from the ARDSNet low tidal volume study showing an improved outcome when patients are ventilated with lower tidal volumes rather than higher tidal volumes. So here we begin to shift our paradigm from obtaining normal numbers to preventing further injury. And this paper was actually controversial at the time, if you remember, in that it did not – the use of one-size ventilation for everyone seemed to be against the way physicians actually practiced. And in this very interesting editorial from Critical Care Medicine in 2005, investigators looked at the mortality among patients in the ARDSNet trial who were excluded for various reasons. And they found that the patients who were excluded from the trial actually had the same mortality as those randomized to the 6 mLs per kilo tidal volume. But the mortality in the 12 mLs per kilogram tidal volume group was much higher. So this led them to conclude that the trial was flawed in that it allowed the clinicians in the trial to apply only one of two tidal volumes, and that this was very different from what clinicians do in practice. They found that when patients with – who had a lower pulmonary compliance were randomized to a higher tidal volume, randomized to the 12 mLs per kilogram group, they had an increased mortality rate. On the other hand, those with better compliance, randomized to the lower group, had an increased mortality rate. So by eliminating any variability, we actually increased mortality in both groups. This lesson was reinforced to us a couple of years ago in the LungSAFE trial, a large survey of mechanical ventilation around the world. It found that clinicians weren't necessarily using 6 mLs per kilo tidal volume in patients with ARDS. In fact, many patients had more than 8 mLs per kilo of tidal volume. So the question was, and the question raised at the time of the paper, are these clinicians just ill-informed? Are they making a mistake? Are they inappropriately treating their patients with low tidal volumes, with higher tidal volumes? In fact, maybe this represents clinician wisdom and rejects the one-size-fits-all solution to tidal volume. So while focusing on the reduction of ventilator-induced lung injury, perhaps we were bringing on other problems. And the same could be said, perhaps, for PEEP trials. So here's a summary of the major PEEP trials in the literature, all of which were negative except for the ART trial, which seemed to show harm from higher PEEP. But all of these trials were also essentially flawed because they did not take into account the individual variability of the patients. If you look at what happens when we increase a patient from a lower PEEP to a higher PEEP, we increase the plateau pressure, and we may or may not be benefiting the patient. And when we randomly increase the PEEP in patients, we don't know if that patient is recruitable or not. We don't know if they can possibly benefit from a higher level of PEEP. So among those patients in whom PEEP is increased, there will be some who are non-recruitable. And in those patients, increasing PEEP will increase the plateau pressure, increase the driving pressure, and likely the injury from PEEP will outweigh the benefit. On the other hand, there are recruitable patients in that study, patients in whom increasing PEEP recruits lung, plateau pressure doesn't necessarily increase, and driving pressure will go down. In those patients, the benefit outweighs the injury. And by recruiting patients, by applying empiric PEEP to patients, we may be applying it to those who could possibly benefit and those who could possibly be injured. One way to individualize the response to PEEP is to use the concept of transpulmonary pressure, the actual distending pressure of the lung. To remind people, the transpulmonary pressure is the airway minus the pleural pressure, and that can be elucidated using an esophageal balloon. The transpulmonary pressure, as we showed in this paper, may be very different from the airway pressure. And without that measurement, we cannot individualize the level of PEEP. I'll give you an example. So this is sample data from an anesthetized obese subject, and I'm going to focus in on the area of that black line and expand out the time scale. So you can see in this patient, we've disconnected the patient from the ventilator over here. We then reconnect him and begin to inflate the lung. So we're delivering pressure all along here, but you can see no volume is going into the lung. In that first area, even though the pressure is going up, no volume is delivered. It's only when the pressure applied to the lung is greater than the esophageal pressure do we begin to see the volume of the lung increase. So only when we overcome the weight of the chest wall are we seeing a breath delivered. The same goes on the downside. When we deflate the lung, as soon as the pressure in the lung or the PEEP decreases below the esophageal pressure, the lung will begin to collapse. So while we focused on prevention of ventilator-induced lung injury, we didn't necessarily do it in the smartest way. We used one-size-fits-all approaches that were neither what clinicians did in practice or what made sense from a physiologic standpoint. Over the last decade, more attention has been paid to patient self-induced lung injury, the risks of spontaneous ventilation, and the risk to the diaphragm. Now spontaneous ventilation is obviously necessary. This doesn't work moving directly from controlled mechanical ventilation to extubated. We need a middle step of spontaneous ventilation. So we have to learn how to use spontaneous ventilation in the smartest way. We need to learn how to bring the patient into the best possible synchrony with the ventilator. And dyssynchrony can be deadly. I'd like to show you again on these tracings what the effect of dyssynchrony can be on tidal volume and transpulmonary pressure. If you look in this area here, on top there is the airway pressure, and you can see the patient is getting a delivered breath there. But down here, the patient is also taking a spontaneous breath. So on the esophageal pressure tracing, you can see an inflection, a decrease in esophageal pressure, which indicates the patient is taking a spontaneous breath. What happens then? The airway pressure increases as that spontaneous breath stacks on top of the ventilator delivered breath. That leads to an increase in volume and an increase in transpulmonary pressure, both of which can be injurious to the lung. So dyssynchrony, spontaneous ventilation, not in sync with the patient's ventilator pattern can be extremely dangerous. I'll show you a second example here. So this is a patient under APRV ventilation. You can see on top the airway tracing. You can see on the second line the esophageal pressure tracing. The patient is taking large breaths on top of their APRV. And that's leading to very large transpulmonary pressure. So the transpulmonary pressure measured on this patient was greater than 35 centimeters of water. That's a highly injurious level of transpulmonary pressure as this patient breathes spontaneously on the high plateau levels of APRV. When we look at what that does to the volume delivered, those spontaneous breaths can lead to tidal volumes of much greater than 10 mLs per kilo. So spontaneous ventilation over delivered mechanical ventilation called dyssynchrony can be extremely dangerous. And that's been demonstrated in a number of studies which have shown the relationship between dyssynchrony and mortality. So here, this study by Louise Blanche showed that when the asynchrony index was greater than 10%, so more than 10 dyssynchronous breaths, you can see that the mortality was increased. ICU mortality greater than 50% in these patients. On the flip side, lack of spontaneous breathing can lead to atrophy of the diaphragmatic muscles. And this is an area where we've paid a lot more attention in recent years. So this very interesting study by Ewan Gallagher showed that patients under mechanical ventilation had a decrease in their diaphragmatic thickness over the first eight days of mechanical ventilation. Now some patients actually had an increase in diaphragmatic thickness, and those were the patients who continued to have spontaneous ventilation and make inspiratory efforts during the ventilation. So this would indicate that some patients actually benefit from reduced support, greater degrees of spontaneous ventilation, and their diaphragm is preserved. Now how do we square this? This is an extremely interesting study by Takeshi Yoshida, where he took patients with two levels of injury, mild and severe, and animals with two levels of injury, mild and severe, and then allowed them either to have spontaneous breathing or no spontaneous breathing. And you can see from his results that those who were mildly injured, who had no spontaneous breathing, actually did worse than those who were allowed to breathe spontaneous. On the other hand, those with a more severe injury did better if they were not allowed to breathe spontaneously. So maybe that's the answer, is we individualize our level of ventilation, we individualize our control of spontaneous ventilation to those who are more severely injured. So having focused on prevention of ventilator-induced lung injury, reduction of patient self-induced lung injury, the risk to the diaphragm, we arrive to our current understanding of ventilation for ARDS, that there are phenotypes of disease based upon mechanics, clinical factors, and imaging that we can use to personalize our mechanical ventilation. And first, we have to acknowledge that ARDS is a heterogeneous disease, we've known that all along, and really that even accentuates the need to personalize ventilation based on the patient's factors and the disease factors contributing to that patient's mechanical ventilation. I'd like to show you an example of individualization in practice. This is a patient from our ICU with moderate ARDS from trolley and aspiration. We ventilated this patient at three levels of PEEP. So first on a level of PEEP of six, we measured the patient's pulmonary mechanics, we looked at their esophageal pressure and their airway pressure, and we calculated from those the trans-pulmonary pressures. So this patient on a PEEP of six had a low compliance of 14 centimeters of water, a high driving pressure of 18 centimeters of water, and that was explained by a negative trans-pulmonary pressure at expiration. So this patient was collapsing after every breath, and by allowing them to collapse, we decreased the compliance and the driving pressure went up. When we examined this same patient on a PEEP of 18, empiric PEEP of 18, the patient's driving pressure was even higher, the compliance was the same, and the trans-pulmonary pressure was now very positive, positive six. So we can over-distend the lung by increasing PEEP. But only by using the patient's individualized mechanics can we choose the optimal level of PEEP, and in this case it was 12. When we placed the patient on a PEEP which gave them an end-expiratory trans-pulmonary pressure of approximately zero, we achieved the optimal driving pressure, 15 in this case, and the better compliance, 18. So this is individualization in practice. And we've shown, and this is a study under review so you won't have seen it before, we've shown using data from our EPVENT2 trial that patients who were maintained with a trans-pulmonary pressure of plus or minus two centimeters of water did better, had a better survival than those in whom the trans-pulmonary pressure was either higher or lower. So maintaining a trans-pulmonary pressure in that optimal range of zero led to an improved survival in patients with ARDS. That shows us how to deal with PEEP and oxygenation. Dyssynchrony is a more difficult problem. We know that clinicians are not good at recognizing dyssynchrony. It can be extremely difficult based on observation of respiratory tracings. So the individualization of treatment for dyssynchrony probably lies in automation. And this is data from my colleague Elias Beydorf-Cassis showing automated detection of various patterns of dyssynchrony. And the way this will eventually integrate into a ventilator is using neural networks and pattern recognition to find and identify different phenotypes of dyssynchrony and then suggest modifications to the ventilator settings in order to eliminate those dyssynchronies. Beyond the ventilator itself, we'll see enhanced use of imaging. Now we're all very familiar with studies of use of CT scans in ARDS. That led to a significant improvement in our understanding of ARDS. The entire concept of baby lungs and ventilating baby lungs comes from these CAT scan studies. Unfortunately, CT scanning is not practical at the bedside and it's extremely difficult, at least in the United States, to move a patient to the CAT scanner for a diagnostic study. CT scanning was... We refocused on CT scanning during the COVID epidemic. So this is data from Dr. Gattanoni demonstrating the low and high type phenotypes of COVID ARDS. So the low was highly compliant lungs. The H type was very noncompliant lungs. He differentiated these as needing different types of mechanical ventilation. But we do have a bedside modality and that's EIT. Now EIT, I've been to, of course, many critical care meetings. It always seemed to be just around the corner, but it is actually here. Bedside EIT machines are available. We can use them and we can use them to assess ventilation and perfusion in patients with ARDS. And in fact, we know that EIT has excellent correlation with CAT scan studies. We can view the ventilated and perfused areas of the lung. We can identify areas of collapse and over distention. And here you can see that using EIT, they've been able to identify the optimal PEEP which minimizes over distention and minimizes collapse. So if we have a PEEP lower than 10, we see an increase in collapse. If we have a PEEP higher than 10, an increase in over distention. And at a PEEP of 10, we see the maximum compliance of the lung. This is something you can do at the bedside using EIT imaging. And as I say, it is here. Finally, we should recognize that different phenotypes of ARDS can be identified using biomarkers. And biomarkers have been measured, of course, in ARDS for years, but we haven't exactly understood how to use them. And great progress has been made over the last 10 years in identifying biomarkers associated with hypo- and hyper-inflammatory ARDS. And this is a paper by Sinha in intensive care medicine a few years ago showing the biomarkers associated with hyper-inflammatory ARDS here and those associated with hypo-inflammatory ARDS here. So using latent class analysis of these biomarkers, it's possible to identify patients who are at higher risk of mortality with the hyper-inflammatory subphenotype or those with hypo-inflammatory ARDS with a lower risk of mortality. So biomarkers, again, measured for years, but something that we're just understanding how to use. Using these phenotypes of the hyper-inflammatory and hypo-inflammatory subgroups of ARDS, we can then go back and look at the results of clinical trials that had previously been done in heterogeneous populations. So this is a reanalysis of the data from the HARP study, which tested the use of simvastatin in patients with ARDS. So you can see here that this study was overall negative. And if you look at the hypo-inflammatory patients, there was absolutely no difference in their outcome. However, if you take the hyper-inflammatory patients, you can see that those who received simvastatin had a better outcome than those who did not. And that outcome, that improvement in survival was consistent out to 90 days. So again, this shows us that taking large trials in heterogeneous populations does not make sense in the era in which we can individualize our treatment to specific population sets, positive trial. So use of biomarkers can lead to a strategy of enrichment of clinical trials. And enriching our clinical trials using pulmonary mechanics, imaging, and selective use of biomarkers can lead to smarter trials, trials that are more likely to succeed, and trials that are more likely to target the appropriate patients. And I'm sure like all of you, I'm sick of negative trials in ARDS, and the solution is not to power through and just recruit larger numbers. The solution is to be smarter about who we recruit into our clinical trials. So beyond the clinical trial, taking it to the bedside, using imaging, in particular ultrasound and EIT, using pulmonary mechanics measured at the bedside, esophageal manometry, compliance of the lung, driving pressure of the lung, using smart ventilators which can assess the synchrony and help us deal with that. And finally, biomarkers will allow us to develop an integrated approach for prospective decision making in our patients with ARDS and improve outcome in those patients. And finally, taking that personalized mechanical ventilation, personalized phenotyping of our patients, we should be able to develop fully automated mechanical ventilation. And again, this is something that we've been hearing for years, it's just around the corner, but I really believe it is, and the ventilator companies are beginning to bring these modalities to market to develop the first prototypes of closed-loop ventilation. They have been tested in Europe. We have not yet tested them here in the United States, but it is here. So to sum up, personalization of mechanical ventilation using patient phenotypes is the wave of the future. I believe we will see it fully developed over the next five years, and this is an exciting time for discovery in ARDS, it's an exciting time to be a clinician in the ICU, and I look forward to seeing how all this develops. I'd be remiss if I didn't thank the people I work with. On the top left-hand side is Steve Loring, who was my mentor in pulmonary physiology. The others are Jeremy Beidler and Elias Beidorf-Cassis, my friends and colleagues here at BIDMC, whose ideas have contributed much to my career and to my thinking on this topic. So I'd like to thank you all for listening. And I believe that after this, we have a open question and answer session. Thank you very much. And now please join us to continue the conversation with our speaker. Good afternoon, everyone. My name is Jose Diaz Gomez. And once again, I have one of the highest honor in my career, being able to moderate this session with Dr. Daniel Talmor. For those of you that do not know, I was associated with Dr. Talmor about 12 years ago. So actually he has recommended that I should tell him, Danny, like the day we met, the day that actually he invited me for lunch with cardiothoracic surgeons and really impacted my career on echocardiography. So it is an absolute privilege to moderate the session. And Dr. Talmor, Danny, I want to basically read probably one of the most influential individuals in our country for research in this field. Dr. Gaji from Mayo Clinic. His question is actually, he starts saying, congratulations, Danny. There is false dichotomy between the proven benefits of standardization and obvious patient context heterogeneity and I'm wary of the message that usual care delivered by expert clinician is better. It is not. Wouldn't evidence-based medicine, could book medicine be more acceptable and more effective if the recipes are a slightly more sophisticated than unfortunate, non-evident based. PIP, FIO2 table, digital world provides opportunity to do that. So thank you, Jose, for that kind introduction and thank you, Auggie, for the question. I obviously agree 100% with you. There isn't a dichotomy between using standardization and patient-centered care. The issue is what is the standardization around? So to use six or 12 mLs per kilogram on all patients is oversimplification and standardization to a point where you lose all meaning to what you're doing. A much smarter approach as I tried to lay out in the lecture is that we should identify patients who are appropriate for an intervention and then use the standardized intervention. So for example, here, we use, as an example, esophageal balloons to see who is appropriate for a higher versus a lower level of PIP. You could use an intervention of measuring a standardized panel of biomarkers to see who is an inflammatory versus a non-inflammatory phenotype of ARDS and then target your therapy specifically to the phenotype identified and then use standardized therapies. So you're right, it's a false dichotomy. We need to be smart about applying standardized therapies to the right patients. And really, the key is to go back to the way we design clinical trials. So clinical trials are often designed to power through questions that make no sense to begin with. So it really makes no sense to apply the same tidal volume to every patient, but if we recruit enough patients into the trial and choose the right control group, we will eventually show a difference. And instead, we should be designing our clinical trial to identify the right patients for the right intervention and then test the standardized interventions. So I think you make an excellent point. We need to be smarter about the way we apply standardized interventions, but we also need to be smarter about how we select these using more appropriate clinical trials. Thank you, Danny. Excellent, excellent response. Danny, when we met back in 2009, that was one year after your landmark trial on the esophageal pressure monitoring. And I remember it was part of the pride of your prestigious Department of Anesthesia. However, there is a question here from the audience. It said, why the esophageal balloon has not become widely used more than 30 years after have been introducing into clinical practice? Yeah, that's an excellent question. One I kind of asked myself also. So there are technical challenges. And when we started with this, esophageal balloons were not widely available. Ventilators did not have the software to measure esophageal pressure and display the results in a matter that made sense. So there are definitely technical challenges, although ventilator companies are now building this into the ventilator. There's also the learning challenge. So most people were not brought up learning how to use this and weren't as fortunate as I was to meet the right mentor at the right time. But there is a natural resistance to the new or the new old. And it takes time for this to be adopted. There's also conflicting evidence. I think that we muddied the waters more than we helped with our second EP vent trial. And we're hoping to correct that with some of the new data that I showed that's now in review. But there is no doubt that it makes sense physiologically to apply PEEP at a level which overcomes the pressure collapsing the lung. And I think that the more people who try it, the more we'll use it. I think that it will eventually be integrated with the other modalities, the imaging and the biomarkers to provide a more complete picture of the patient's lung injury. So I still use it. Our ICU still uses it. And we used it extensively in our patients with COVID-19. All of us, maybe not all of us because some of the younger people might not have, but at one point we all learned how to place a SWAN. And placing a PA catheter is much more difficult and prone to nuance of interpretation than placing an esophageal balloon. So I suggest that everybody at least investigate how they can bring this to their ICU. It's a tremendous tool for patient care. It's also frankly a tremendous tool for resident and fellow education. It allows you to truly understand a patient's lung injury. So I don't have a great answer why it hasn't been more widely adopted, but I hope it still will be. Excellent. Danny, we are getting now over 430 people and there are many questions. So this is great. And a follow-up question is the following. You were actually including this recent article as that's the EP event to study published in JAMA in 2019. So how do you want to elaborate about your last statements regarding esophageal pressure and the results of this study? Yeah. So on first and simple read, the EP event to study is negative. It did not show a benefit to use of esophageal guided pressure ventilation to over routine use of high peak. And in retrospect, there are a number of flaws with the study design and that's the way clinical studies are. You make decisions and seven, eight years later, you're living with the consequences. So our comparison group of the high FIO2 table is not actually what's used in clinical practice. We know that clinicians tend to favor lower levels of peak. So the control arm does not reflect current practice in ARDS. It seemed to reflect the trend at the time, but not what is currently going on. And because we chose that control group, we didn't achieve adequate separation between the groups. So the levels of peak and the levels of transpulmonary pressure between the two groups were quite similar. And given that it was not likely that we would find a difference. However, the secondary analysis, which I shared in the talk, showing that the closer you were to a transpulmonary pressure of zero at end expiration, the better your survival, it seems to indicate that there is benefit to using esophageal balloons and achieving adequate peak. If you were under peaked or over peaked based on the transpulmonary pressure, your survival was worse in the study, but that's a secondary analysis. It's under review currently, and I expect that it will be published soon. Thank you, Danny. It is actually a very interesting question and more into practicality. And our colleague is asking, would you please speak more about the use of driving pressure compared to PEEP to personalize vent settings? Yeah, so driving pressure is an extremely interesting concept and a good way to pull together a number of different factors influencing the volume. Unfortunately, there isn't a dial on our ventilator for driving pressure. We can manipulate it really one of two ways. We can either manipulate the PEEP or we can manipulate the tidal volume. We can probably make small adjustments to it by using the rate by flow, but really it comes down to tidal volume, pressure and PEEP. So driving pressure right now is more of an outcome or a surrogate outcome. We would adjust the ventilator to reduce the driving pressure and hope that as it was shown in the New England Journal paper, that would influence outcome as well. We still have not seen that in any interventional trial. So we've seen trials. We have not seen a trial using driving pressure as outcome. And we haven't seen a trial that I recall that has shown that manipulating driving pressure affects outcome rather than driving pressure just being a marker of poor lung health. So I think there's more to learn on the topic of driving pressure. It's certainly an interesting way of looking at things. I think at the risk of sounding like a one trick pony, driving pressure of the respiratory system does not make a lot of sense. Really what we're interested in is driving pressure of the lung for which we need to place an esophageal balloon and separate the lung and the respiratory system mechanics. So driving pressure is a concept that needs to be explored. I think it has legs, but we just don't know how to use it yet. Thank you, Danny. As you might know, the anesthesiology section has been following you very closely. And our leader actually has a very, very good question. This is Dr. Nurok and he's adding a personal statement saying terrific presentation, Danny. Thank you. Can you comment on the potential for non-invasive mechanical ventilation to cause lung injury in patients generating very negative intraocular pressures? How we may measure potential for lung injury in non-intubated patients and how these questions frame your approach to deciding to intubate a patient in respiratory failure? Yeah, Michael, excellent question. And really we could reframe this as the question over the last year in the treatment of COVID-19 patients. In the first wave here in Boston, like I think many places around the world, we were very fearful of patients' self-induced lung injury through excessive spontaneous ventilation. And we had other considerations, spread of the virus and so on, but it was clear that patients with COVID-19 respiratory failure were extremely dyspneic. They were taking very large tidal volumes and they had the potential to further injure their lungs. In this last wave, at least here, the pendulum has swung the other way and we're much more liberal in leaving patients on either non-invasive ventilation or high flow nasal cannula for much longer. But the question is, are we leaving them too long? And we are seeing at least anecdotally here in Boston that we leave patients on high flow nasal cannula for extended periods. When they wind up intubated eventually, they are extremely sick and often unsalvageable. So where is the middle ground? How do we decide in which patient it's appropriate to leave them breathing spontaneously and in which patient should we intubate, take control of their ventilation and eliminate spontaneous ventilation? Unfortunately, after that long wind up, I do not have a one size fits all answer for you beyond be aware of the risk of spontaneous ventilation induced lung injury, patient's self-induced lung injury and examine the patient. When you see the patient taking large tidal volumes for an extended period of time, reconsider your strategy that patient may need to be intubated and ventilated. Well, Danny, this is a very important answer you just made. And if you don't mind, I want to probably take advantage of the question with the following. There have been some probably new learning about finding the right moment to actually implement VV ECMO support on COVID-19 patients as well. And we have tried to follow this natural history of the disease more closely and kind of making that call. Can you comment on that, like the natural history itself of ARDS secondary to COVID-19 and that precise moment you are saying you need to intubate versus you need to start VV ECMO. And the reason why I'm saying that is because there are individuals who believe that you should actually do lung rest or there wants just to do rehabilitation or actually extubate the patient and just leave them on ECMO. I'm sorry, it's probably on a question on ECMO related, but it's coming, I'm not doing a good job if I'm ignoring the natural history of the disease that you are actually an authority on that. Yeah, thank you, Jose. So first of all, let me admit, I am not an authority on ECMO. And I believe that ECMO is a useful tool. I do not believe that it is a routine tool for the treatment of lung injury from COVID or not from COVID. It's a rescue therapy. And it's a therapy that I think should only be used after an adequate trial of mechanical ventilation and not jump from spontaneous ventilation to tube them and put them right on ECMO. We are a referral center for a large network and we received numerous referrals for ECMO and actually very few of them landed up on ECMO. Most of them were managed with appropriate ventilation, including the things I described in the talk. So ECMO is a useful tool. It's not a panacea. And the evidence, the best evidence, the EOLIA trial, which to my mind is a positive trial. I mean, the P value was 0.07, I believe. That's positive. And particularly in that patient population. But it was positive in patients who were super sick who had received an adequate trial of mechanical ventilation. I'll go out on a limb here and say that I'm not a huge believer in the concept of total lung rest either. So from a, if we look at the history of critical care, in no organ has responded well to total rest. So we rest the gut by not feeding it. That didn't work out so well. We rest the kidney by immediately putting the patient on CBVH, but we found out that that's not necessarily needed either. We rest the brain by sedating the hell out of the patient. Well, that doesn't work as well. So why should collapsing the lungs down and letting them completely rest, why would that work? I think we have to be careful about our ventilation, but to go to complete organ rest hasn't worked anywhere else and I don't really see the rationale or any proof to be honest, that that works in patients only. Great answer. Thank you, Donnie. And our members are still excited about the esophageal manometry. And let me try to put some of those together. We have many questions, but this is actually by one of my colleagues, Baylor Amy Butcher. And she's asking, in the absence of esophageal manometry or balloons, are driving pressure trials a good surrogate to determine optimal PEEP in ARDS patients? So, you know, my short answer is, I don't know. My long answer is, I think so. I mean, let's design a trial design that will optimize driving pressure and see what happens. It's a much smarter target than to just increase the PEEP randomly or decrease the PEEP randomly or give a higher or a lower tidal volume. We're optimizing towards a physiologic parameter that we know is related to outcome. So Amy, I would say, let's design the right trial to test that concept. As I said in a previous answer, we can't adjust driving pressure on the ventilator. We'll have to come up with an algorithm of how we do that as part of a standardized procedure in a trial. Thank you, Danny. And actually, another anesthesiologist section member is asking Dr. Giavelli, what recommendation do you have for implementing esophageal manometry at centers with enthusiasm, but without prior experience or local expertise? Starting from scratch, what would you do? So I would say one thing we did show in our event two trial is it's possible to generalize this across multiple centers. We had more than 10 centers involved at various points in the trial, and each of them were able to, with minimal training, place balloons and to interpret the results. If you have a ventilator capable of displaying esophageal pressure measurements, it essentially will give you the interpretation. There's some, placing the balloon can be finicky. There's some nuance to it, but you've learned how to place a swan. I can assure you that you can learn how to place an esophageal balloon. And my email is easy to get, and I'm very happy to send out training materials to anybody who's interested to do webinars for anybody who's interested to show them how to do this. I think it's a terrific technique, not only for clinical care, but for teaching and education at the bedside. I love that answer, that analogy, Danny, and thank you again for engaging our SCCM members, not only that, showing once again how humble you are. And I can tell you, even he's a chairman of one of the most prestigious anesthesiology departments in our country. He has replied to me within 24 hours most of the time. Thank you, Danny. There is this question, it says, between esophageal manometry, biomarkers, lung ultrasonography, diaphragmatic ultrasonography, and EIT, there is an ever increasing number of ways to monitor and potentially customize care for ARDS patients. Do you think there is a place for all of these modalities? How would you envision this customized care? And it is possible only at few select locations. If not, do we need to pick the easiest or the most generalizable and try to go forward with these and spread it widely? Yeah, so excellent question. You have this Chinese menu of options. Which one do you choose? Is it three, seven, and nine, or is it all of the above? And I think really it would depend a lot on which patient you're dealing with. So for example, in our practice here, an obese patient gets a balloon. There, it's very difficult to predict what the pressure being applied to the lung is because of the weight of the chest wall. On the other hand, a young patient with a new bilateral infiltrate, it may be that the priority would be to getting an inflammatory markers panel on that patient. I think part of this will be answered by the ventilator companies that will eventually integrate automation into ventilation. And we've seen the beginnings of that from a number of companies. And we'll integrate interpretation of esophageal pressures, identification of the synchrony, for example, into the ventilator algorithm. And it may be that eventually they ask you a couple of questions about the patient. So is this patient have inflammatory ARDS or non-inflammatory ARDS? And that may be a different algorithm. There's a lot of possibilities. And unlike many talks I've heard over the years, I really believe that we are on the cusp of a major change in our approach to ARDS, more individualization based on all of the things we've discussed and that it eventually will be supported by smarter types of ventilation to manage the patients. Thank you, Danny. And once again, one of the things you need to know about Danny is that his resourcefulness is unparalleled. And the reason I said that, and the follow-up question in that regard, Danny, do you feel comfortable with just continuing with the same definition for ARDS? And I'm talking about the work that you did in Rwanda in the last five years in that paper, where were you actually able to increase the sensitivity of recognizing ARDS? What would you propose for new classification or we need to wait for more of this personalization in ARDS? Well, that's a really good question, Jose. And ARDS and acute hypoxic respiratory failure are essentially the same thing. ARDS is a convenient term to encompass acute respiratory failure, but the requirement for bilateral infiltrates, it's not essentially relevant. I do think that we should come up with a better definition, a more patient-centered definition, a better way of classifying patients. And that's really the first step to, as Augie said at the beginning, smarter testing of interventions in order to match the intervention to the patient. So I don't have a better alternative right now, but I do think that ARDS is, it's 51 years old or 52 years old. It's getting close to retirement age. We just need to find the appropriate, we have to do the appropriate succession planning first. Thank you, Danny. My Congress co-chair, Dr. Emlett from Pittsburgh, want to ask, do you routinely assess recruitable lung combined with lung ultrasound during the course of ARDS on daily rounds? Also, do you feel that biomarkers will have a clinical trial to determine its use? So, unfortunately, as chair of a department, I don't round daily in the ICU anymore. I try to get in on weekends. The use of lung ultrasound to identify recruitable lung, I think is interesting. We have people in our department who believe in that. We don't, none of us do it on a daily basis. What we have found, again, getting back to my hobby horse, the esophageal balloon, is once we recruit patients using transpulmonary pressure, they stay recruited. You don't have to make frequent adjustments. So I could see using lung ultrasound at the initial ventilator settings for the patient, deciding which PIT to use at those initial, placement on mechanical ventilation. I doubt that it's required on a daily basis, but I have no evidence at all to back that up. So, ultrasound, I'm a big believer. I think it's useful, probably useful in this context on day one. Biomarkers and clinical trials. So yes, I showed you the reanalysis of the statin trial, where if they had isolated the use of statins to patients with inflammatory ARDS, they would have actually seen a positive benefit to use of statins. And it makes you wonder where many interventions that made a lot of sense abandoned because we just recruited two heterogeneous group of patients into our clinical trials. So, Auggie and I were part of a trial for the use of aspirin. Maybe it would have been smarter if we had done a biomarker panel on those patients and selected out the ones with inflammatory, with a positive inflammatory panel who had ARDS mostly related to inflammation. Maybe they would have benefited from the use of aspirin. So, I think we need to be designing our trials with that in mind, who are the right patients who may respond to the intervention? Excellent answer, Danny. We have one of our leaders from Cleveland Clinic, Dr. Minerva Escobar-De Vila. He's actually the director for the Pulmonary and Critical Care Fellowship there and a good friend of mine. He's asking the following question. While we get to the point of doing personalization based on data, for instance, neural networks, et cetera, acknowledging this is gradual and a developing field, how do you suggest we implement personalization into practice? Yeah, so great question. And I think this is gonna be in a stepwise manner. Right now, we can take a ventilator, input the patient's height, weight, gender, age, and then allow the ventilator to train itself on that patient's respiratory mechanics and automate the mechanical ventilation settings. So I think that's gonna be the first step in this and the ventilator companies are already very close to this and they've rolled out preliminary modes that take advantage of that training on the patient's personal mechanics. So I think that will be stage one. Integrating biomarkers into that, as I said, it may be as simple as inputting into the ventilator, is this patient hyperinflammatory or not? That may change the settings on the ventilator. It may be linking an EIT machine to the ventilator. I could see that happening in the not too distant future. We already integrate esophageal balloons into EIT machines to get a more complete picture of the patient's respiratory mechanics and the patient's lung imaging and perfusion. All of that could be connected to a ventilator to deliver smarter and more intelligent ventilation. But I think the first step is going to be the ventilators training themselves based on the patient's personal data on the appropriate mechanical ventilation settings. Thank you, Danny. And actually there was a question by one of my colleagues that about, how would you do esophageal pressure monitoring if you don't have a Hamilton ventilator? They're asking that. So there are other ventilators that have this technology integrated into them. There is at the last resort, you can connect an esophageal balloon to a blood pressure transducer and just transduce it on your bedside monitor. There are ways around this, and I'm happy offline to help anybody that would like to learn about it. Thank you. Dr. Giavelli is asking, we have seen in ARDS and other disease states that improve physiologic endpoints, for instance, compliance, oxygenation, do not always translate to improve patient-centered outcomes, mortality. How did you view esophageal manometry against that context? Yes, so great point. And just to remind people who may not be aware, in the ARDSNet low tidal volume study, the patients with lower tidal volume had worse oxygenation, but they had better outcomes. So looking at that the other way, the higher tidal volume improved your oxygenation, but you were more likely to die. So we definitely have to be careful about these intermediate endpoints. And that's why our EPVENT2 trial was designed not for an intermediate endpoint, rather a composite endpoint of time on the ventilator and mortality. EPVENT1, which was stopped early, was designed for an outcome of oxygenation, and that's why it definitely was not a definitive trial. It was a exploration of the concept and a proof of the concept. So I think Dr. Jaboli also raises a good point when we talk about trials targeted to driving pressure. So driving pressure seems to be related to patient outcomes, but it's not. I would say that that manipulation of driving pressure hasn't been related to change in patient outcomes. And we need to really look at that carefully before we use driving pressure as an outcome in a clinical trial. So it's a very good point, and it goes to the difficulty of clinical trials in ARDS, because in order to get enough patients to show a mortality outcome, we need a lot of patients. But if we're selecting patients based on specific criteria, getting to a lot of patients may be impossible. So we're gonna have to figure out how to do these trials in a smarter manner that allows us to test more specific groups against the interventions. Thank you, Danny. Another quality that you have is to be a great global collaborator, especially in Europe. And it is my pleasure to present this question by one of our leaders in the UK, Dr. Zakmani. And his question is biomarker-driven phenotypes could help us drive non-ventilatory management. Do you think they will be helpful when looking at different ventilatory approaches, APRB versus conventional? Yeah, so excellent point. I think that biomarkers will influence our choice of drugs, the inflammatory versus hyperinflammatory versus hypoinflammatory phenotypes. I do think it may have a role in use of the ventilator. So if we know, or we suspect that hyperinflammatory ARDS will be associated with stiffer lungs, so-called pulmonary ARDS, and hypoinflammatory may be, this is getting back to ancient concepts, right? Non-pulmonary ARDS, then it may have a role in identifying those more clearly. And we know that, for example, pulmonary ARDS is less responsive to PEEP. And using biomarkers may be a smarter way of identifying a lung that is less responsive to PEEP. So I think you're right, that the use of biomarkers should really drive the pharmacological interventions, but it may also have a role in helping set the ventilator. Thank you, Dani. And similarly, Dr. Abhijnath Patel is asking, in reference to the hypoinflammatory versus hyperinflammatory ARDS phenotypes, many of the biomarkers used to differentiate, for instance, IL-6, TNF-alpha, are not at present easily assignable for real-time intervention. Is there any role in using the non-specific inflammatory markers, such as CRP, ferritin, or D-dimer, that are available in real-time? So, excellent point. First, I would challenge all of us as a critical care community, if we find the use for the more difficult-to-measure markers, we may find the technology to measure them more quickly. I know that, for example, in COVID, a lot of centers were measuring IL-6 because they thought that the use of TOSI would help the patients, and we should chose those patients that would respond. Regarding the other biomarkers, so the question I think is not, and I'm not a biomarker expert, is not whether one or the other will help identify, it's what is the panel of the biomarkers that will identify patients with hyperinflammatory ARDS? And I do think there is room for, you know, the so-called bedside biomarkers to play a role in that. Great. So, so far I have been able to demonstrate all these wonderful things about you, Danny, but there is one that really I want to emphasize at this point. You are a leader, and there is this question about leadership. It's coming from one of our future co-chairs for Congress, Heather Meisen from Emory. And the question is about changing culture. And she said, I see many fears about losing the ability to control mechanics and allowing the ventilator to make the decisions. Closed-loop ventilation artificial intelligence. How do we move past that fear of losing the control? Thank you, Heather. I think that's the million dollar question in technology in the ICU is all of us went into critical care medicine because we're hands-on people. We like to touch the patients. We like to touch the machines. We like to understand every nuance of what's going on. We're all detail-oriented and to a large extent control freaks. So how do we get around this? First of all, the technology needs to be proven and we're not there yet with really any of this artificial intelligence or closed-loop ventilation. It's on its way. It is being tested, but we're definitely not there. Second, it's still incumbent upon all of us to understand the tools we're using. So I would never say, okay, let's get this ventilator in, hook it up to the patient. And we don't really know what, we don't need to know what's going on. We absolutely have to understand the technology. And I'll give the, it's an overused analogy, but we'll go back to aviation. The pilot probably does not understand every, all of the programming in the autopilot of the airplane or the technology behind every doodad that gets the plane into the air. But the pilot understands how the whole thing works together and knows when it's working and when it's not working. So I think, and has a deep knowledge of the technology, of course, too. It's not that they're ignorant. So we need to see the trials. We need to understand the technology. We need to test it out in our own centers and our own hands. And we always need to be the pilot. We need to know when the nose is going down, how to pull it out and to take care of the patients. So it's, I think use of technology and use of closed loop systems is not application of responsibility. It actually increases our responsibility because we need to be able to supervise all of these new systems and know exactly what they're doing and exactly how to deal with any issues that may come up. Thank you, Dani. There is an anesthesiologist intensively, it's Dr. Lytle, and she's saying the following. Is there a thought that using a sulfagyl manometry could assist outcomes in this patient while they are under operative conditions? As an anesthesiologist, I've been often wondered what is best for my operative patients with lung injury, but not sure how to optimize it in any given patient other than following the ICU bed settings. Yeah, great, great question. So we have not really explored use of the sulfagyl manometry in the operating room. Although there are clearly patients in whom adequate recruitment is a crucial issue. And it can be the patient with lung injury. It could be the patient who's undergoing a laparoscopic procedure and has a reduced lung reserve. It could be a patient who's in deep Trendelenburg. It could be a patient who's undergoing robotic surgery, which is essentially all of the above. So I think there is room to explore this. I'll refer you back to a super interesting paper by Amato and company in anesthesiology a couple of years ago, looking at EIT in patients undergoing laparoscopic and open surgery and showing that there is a definite opportunity to optimize ventilation in the patients undergoing laparoscopic surgery using EIT. So I expect that sulfagyl balloons would be equally useful in many of these patients. As an aside, one of our enemies in the operating room is routine. And something that we have to fight with our trainees all the time to understand that while we may fall into routine, our job as anesthesiologists is really to recognize immediately when this patient is not routine. And teaching our trainees to look at a patient and to individualize the ventilator, that's really a area where anesthesiology can improve massively. So yes, I do believe that sulfagyl balloons may have a role in the operating room, as may other modes of imaging of the lung and improvement in the ventilator settings. Thank you, Danny. As you can imagine, of course, we reach out even our pediatric critical care colleagues and Dr. Goldhart is asking a question about the PIDS patient population. I know you don't work in a pediatric ICU, but I'm curious about your answer to this question. Do we see similar variability in phenotypes in pediatrics? Are you familiar with any studies looking at utilization of some of these strategies in PID ARDS? So Emily, I'm afraid I'll have to disappoint you. I'm not a expert in PIDs by any means. And I know that there is a study undergoing in Los Angeles looking at more individualization of mechanical ventilation, but I can't really give you details. And I can't tell you much on this topic. Sorry, I don't wanna ramble on too much. Once again, you're one of the most honest person I have met. Thank you, Danny. You're gonna make me blush now. Is it? Dr. Emlett is asking a question regarding training of our fellows. And as we teach bend management, there are more lawn simulators of varying interaction with the machine. How does that affect the development of adopting ventilatory approach? Sorry, no. How does affect the development of adopting intelligent ventilation system with closed loop protocols? So basically lawn simulations, varying interaction with the machine and how does affect the development of adopting intelligent ventilation system with closed loop protocols? Any? Sure. So I'll tell you a story. On my first night of calling my internship, I was in internal medicine and there was a room with four ventilated patients. And they were signing out the department to me and they said, okay, here's four ventilated patients. Good luck. And I had never seen a ventilator before. So I asked, can you show me what to do? And they said, just push this button if there's any problems and it was the alarm mute. So that was the state of the art in ventilator education in 1992 or 93 or whenever it was. We've come a long way since then. And I definitely do see a role for simulation in training and use of ventilators and use of smart simulation. Any simulation needs to incorporate also hemodynamic simulation because we know that what we do with the ventilator can affect patient hemodynamics. So we would not allow somebody to put a, well, we may allow, but ideally we would have somebody intubate a mannequin before they intubated a human. Ideally, we would also have somebody practice ventilation on a simulator before they go and place a ventilator on a patient. So I think there definitely is room for increased use of simulation in critical care training around ventilation. It may increase our confidence in the use of closed loop ventilation. If we can place a closed loop ventilator on a simulator and see what the settings come out as and how that influences the patient's hemodynamics and oxygenation. So I do think, just thinking through your question, that use of simulation, trying out this closed loop ventilation, seeing what it actually does on a simulator before it's applied to a patient may be very important in the adoption of these methods. So that was me thinking through your question on that one. Thank you. We have a question actually by one of our attendees from Europe. His name is Lazer Axelman. And it's a very interesting question. He's saying, I work in the pre-hospital environment as a critical care paramedic in rural Eastern Europe. What it could easily be much of a day to get a patient to any sort of definitive care. While I have relatively decent transfer ventilator, it unfortunately can measure most of the parameters you dissected and pointed toward as reference for determination of phenotyping or treatment optimization. I also have zero access to labs of any kind. Ben optimization in this world is entirely based on pulse oximetry, entitled CO2, and a very keen ongoing assessment. Is there any sort of down and dirty way to determine any degree of these ARDS subtypes to attend early optimization before providing four to 12 hours of optimal ventilatory support? Yeah, so Lazer, thank you for the question and thank you for the work you do in such a austere environment. You basically pointed out the most important component of how I would approach this problem. And that's very keen ongoing assessment. So there is no substitute for a keen and intelligent and dedicated physician at the bedside. That is the number one monitor for any patient. And you already have the most important thing there then. But how would I look at it? So I would try to think about this patient and what is going on with them. Now you're talking essentially about an interim period of ventilation. In an interim period of ventilation, I would try to reduce the tidal volume. I would look at the patient and assess what I thought their PEEP need would be. So if the patient is obese, I would increase the PEEP level. If the patient is a large woman, I would increase the PEEP level. If the patient is very fluid overloaded for whatever reason that will increase the weight of the chest wall, collapse the lung, and that patient likely needs a higher level of PEEP. So I would reduce the tidal volume. I would incrementally increase the PEEP based on my observation of the patient. If I thought the patient was really struggling on the ventilator, so they had that sort of fish on a hook look about them, I would strongly consider paralyzing the patient if that's available to you and getting control of their ventilation for this period of transportation. I would remember that a pulse ox saturation of 92 is just fine for this patient. You don't need to fight to get the saturation up to 100. That end tidal CO2, we do not need to totally normalize the patient's CO2. And in fact, doing so by increasing the rate or increasing the tidal volume may be injurious. So I would allow the patient's end tidal CO2 to drift up to maybe 45 if that was happening. So it's that very keen ongoing assessment that is most important thing. It's the trying to figure out in your own mind what is the patient's phenotype and remembering that we don't need to correct all of the numbers in order to keep this patient alive until we can get them to better care. Thank you, Danny. I cannot conclude this session without taking advantage of your vision as a researcher. I'm being impressed by the last developments on the research, basically evolution. And I want your thoughts regarding the role of adaptive clinical trials like the REMAP-CUP. Would you apply that to ARDS and how you can encourage our younger, junior or mid-career individuals to kind of go in that direction? So I think adaptive trials are definitely a positive development in the field. There is no reason for us to embark on 800 patient trials when it's clear after a hundred patients that this is not gonna work. And by utilizing the mechanism of adaptive trials, we'll be able to cut our losses before trials go on too long. I think the idea of a platform trial which cycles out interventions in and out based on patient physiology is a great one. We need to work out how that will exactly look. And there are problems, we're maybe not having the perfect control arm, but these innovations in clinical trials are really promising for testing individualized therapies because they'll allow us the large platform in which to test in specific populations the right therapy for the right patient. And to your last point, I would highly encourage everyone. We are, as I said at the beginning, this has been the best of times and worst of times for critical care docs. It's also been the best of times and the worst of times for critical care research and COVID research. There is a tremendous need for young investigators to come up in the profession to get involved. And the best way is the way I did, which is by a clinical question at the bedside and a great mentor. So seek out the mentor, seek out the questions. And I think we are a growth profession now and there's lots of opportunities for all of us to get involved in great research. That's an excellent way to kind of trying, going to the end of our presentation here, but there is one last component that I think I saw that on your career, which is how to achieve that effective global collaboration. Right now, you're probably one of the anesthesiologist intensivists who has more robust relationship with other specialties around a disease that primarily never have been our friends in our specialty as an anesthesiology. And I want you to share with us how had you been able to grow as a leader in that regard? Well, so I'll be the first to admit that I've been very lucky in my career. I've met great people, I've interacted with them and I've learned from them and I've grown with them. I think start with the questions, start with the hunger to do research, talk to everybody you can, develop your network and it will come. I think that critical care researchers that I've known are some of the most generous people I've ever met. I'll tell another story, I see we only have two minutes left. I walked up to Art Slutsky in 2000 in the ATS in Toronto and started chatting with him. And he took me under his wing and provided amazing mentorship to me through my career, even though we never worked together directly, he was just always a friend and somebody I could bounce ideas off of. Most critical care docs are, he's exceptional, but most critical care docs are generous people and they want to advance the science. So just get out there and talk to people, come to meetings like this. Hopefully they'll be in person and people will be able to interact more, but reach out and talk to people, develop your ideas and if I could do it, anybody could. Well, having said that, I just will encourage all members who connected on this session to really follow his instructions, reaching him out. And once again, the same story he was sharing about Dr. Slutsky and him, happened to me with him 12 years ago. So I do practice critical care echocardiography because Dr. Talmor gave me the opportunity to be an elective fellow visiting Beth Israel Deaconess Medical Center in 2009. Even though that was basically, I think it was his hobby, his real research was in ARDS. So once again, it has been a privilege moderating this session. Please stay in touch with Dr. Talmor and thank you again for participating so well with all these wonderful questions. I learned a lot and thank you for sharing with us your knowledge. Thank you, thank you Jose and thank you to everybody who tuned in. Thank you very much. Thank you. ♪♪♪

COVID-19

intubation

patient management

ventilation-induced lung injury

tidal volumes

high flow nasal cannula

VV ECMO

mechanical ventilation

ARDS

PEEP

adaptive clinical trials

critical care research