So, we're going to start with talking about the use of non-invasive ventilation following mechanical ventilation, and I'm going to talk to you about, I'm going to take the pro stance about using NIV after patients have been extubated. Maybe. So, my disclosures are here for you, I have several research grants, none of them directly apply to the topic at hand today. And my educational objectives for you are to learn about clinical outcomes and how they might be improved by the use of NIV following extubation, to become familiar with the population that may benefit most from these interventions, and to understand the contraindications and risks to using these non-invasive ventilation modes after extubation. Jonathan Casey will be taking the opposite stance in a few minutes. So what do ICU practitioners think about use of NIV post-extubation? This was just pre-COVID, this article was from 2020, and it was a survey. I went to medical school and grew up in Pittsburgh, and I was dismayed a little bit because this came out of western Pennsylvania, which I thought was a good place, and maybe is, but these 487 ICU physicians and nurses and respiratory therapists of that group, only 20% of them thought that respiratory failure post-extubation was a common problem. So in my ICU it's a big problem, and I'm going to show you why I think it's a really important problem for us to talk about. 84% of them were really confident that they could recognize who these patients were who failed, and that's great, I'm a lot less confident than that. Sometimes I struggle knowing which patients are going to succeed and which patients are going to fail. Finally, 54% of them thought that NIV reduces the risk of extubation, failure in high-risk patients. And if any of you from western Pennsylvania are in the room, I hope to share my information and change your mind a little bit about that. So this is a slide which demonstrates an article that was published in 2011, and it informs us about why it's so important to try to, why failed extubation is so important and avoiding failed extubation is important. You'll see that they included over 1,000 patients in their study, and of those patients when they extubated them, and these patients met traditional criteria to be extubated. So when they were extubated, if they did not require re-intubation, their mortality rate was only 6%. If they were treated with non-invasive ventilation and they didn't require re-intubation, their mortality was 8%. Those who did require re-intubation, whether they got NIV or not, were looked at in a second group, and of those people who were extubated a second time, their mortality if they weren't re-intubated was 52%, and if they were re-intubated again and then failed that extubation, their mortality was really high. So the message is that if you fail extubation, the mortality rate for patients is really high, and it's something that's really worth avoiding if we can do so. I don't believe that the reason that the mortality rate is high is because of something that we're doing. Usually we select these patients in an appropriate manner, and we know if we use our proper criteria that there's going to be a certain percentage, 15 to 20%, who do require re-intubation, and that group really is defined mostly by their severity of illness and by other factors and not by things that we do like the process of re-intubation. But nevertheless, if they fail, their mortality is very high. So the question is, does NIV improve extubation failure rates? That's really what we're going to talk about today, and I think the answer is yes. This is from a meta-analysis that we did as part of the CHESS guideline, and this is an example of the effect of NIV on extubation success. We looked at five randomized controlled studies when we looked at this data, and there was clear evidence in these studies that NIV was favored in terms of preventing re-intubation. Does non-invasive ventilation improve ICU length of stay? It does. This, again, is another element of our meta-analysis from that guideline showing, looking at four randomized controlled studies and demonstrating that length of stay was improved. And does it improve short-term mortality? The answer to that seems to be yes, too. Again, four randomized controlled studies. All of the studies were small, which is why the quality of evidence is low, and they were largely—a couple of them were single-center, so—but again, NIV favored short-term mortality. When one looks at long-term mortality, it is even possible that NIV improves that. So when we did our meta-analysis, there were only two RCTs that informed this. There seemed to be some evidence, though, that especially in patients who are very high-risk, that long-term mortality might be improved. This is a study—one of the studies that we looked at in the meta-analysis from Ferrer in Spain. It was published in 2009. There's a low N on this study. There's only 100 patients. In this particular study, they selected people who had chronic respiratory disorders. So these patients had a respiratory acidosis. When these patients were randomized to NIV versus standard therapy, there was an effect on long-term mortality in the patients that were at—these patients being at very high risk, suggesting that in the highest-risk patients, that NIV is beneficial. So NIV seems to improve a variety of outcomes post-extubation, including re-intubation rates, ICU length of stay, and short-term mortality, perhaps long-term mortality as well. Are there other modes of non-invasive ventilatory support that we can use that are effective in preventing re-intubation? The answer to that is yes. So this is a study from Hernandez in 2016, looking at high-flow nasal cannula compared to NIV, and whether you look at time from excavation to re-intubation or to death, looking at either of those two outcomes. In this particular study, there was similar outcomes, suggesting that high-flow nasal cannula may work. If you can't give your non-invasive ventilation to your patients continuously, which is certainly true of me in my ICU, the patients don't always wear the mask for the entire period of time, they have to eat, what can you do in the interludes? Can you give them high-flow nasal cannula in the interludes? And the answer is yes. This is a great study by Thiel that was published in 2019 in JAMA. In this study, which was a multicenter study, they looked at 30 ICUs in France. All of the patients who were enrolled had passed a spontaneous breathing trial, and they were at high risk of extubation failure. They excluded a couple of conditions which you can understand below, chronic neuromuscular disease patients who are on long-term non-invasive support. And in the control group, they used heated high-flow alone at 50 liters. In the intervention group, what they did specifically was they used NIV immediately after extubation for at least four hours with a minimum duration of 12 hours, all night long. And you can see what the targets are that they used in terms of the NIV settings. Between the NIV sessions, they used high-flow. And what they found was that the group that received NIV plus high-flow actually had improvements in outcomes compared to the standard groups. So I would suggest to you, and in fact, this is the protocol that I try to employ for many of my patients, is to use a combination of NIV and high-flow when it's possible to do so. So in high-risk patients, the use of NIV plus high-flow reduces reintubation rates compared to high-flow alone. The limitations in this particular study is that the control group was high-flow instead of standard oxygen. It wasn't blinded. And the specific protocols that they will use may have influenced the outcomes. Fernando last year published a meta-analysis of post-extubation strategies. They did a network meta-analysis, which is a meta-analysis technique that allows one to compare multiple different interventions instead of just a couple. And they looked at 36 randomized controlled studies. The interventions they looked at were conventional oxygen, NIV, high-flow, and alternating NIV and high-flow. And the specific outcomes that they looked at were excavation failure and short-term mortality. This is just a map or a scheme of the different comparisons that they used in the meta-analysis. And what they found when they looked at their specific data and compared different interventions is that NIV compared to conventional oxygen seemed to be a winner. High-flow compared to conventional oxygen also seemed to be an excellent choice. When they looked in the network meta-analysis comparing high-flow to NIV, there seemed to be little difference. And then you can see the other specific things there. The combination of the two, which is the fourth block down compared to conventional oxygen also appears to be a good choice. So NIV, high-flow, combination of NIV and high-flow all seem to be reasonably good choices compared to standard therapy alone in terms of improving outcomes from mechanical ventilation. That first chart looked at re-intubation rates, and this is a similar chart from the same meta-analysis looking at short-term mortality, finding similar findings comparing the groups. So NIV reduces re-intubation rates compared to conventional oxygen. High-flow also does. The effects are most prominent in patients at high risk. Existing data, I don't think, clearly defines between these three strategies. If you went to the year in review, there was a suggestion by one of the speakers that maybe high-flow is better. Right now, I don't think the data is there. I think what we have today suggests that all of these strategies are effective. And the important point for all of this, and I'm going to show you a little more data in a second, is that these strategies have to be employed immediately after extubation to be beneficial. If you wait until the patient is developing respiratory failure, then you're likely to have more problems and not less. Should you use non-invasive ventilation in all patients following extubation? Of course not. So there are some patients you can't use NIV for, right? So there are patients who have clear-cut contraindications to using NIV. And some of those things are listed here for you. The other thing is that most of the studies that have been done have very heterogeneous populations. And it's pretty clear that one needs to pick patients at high risk. The high-risk elements that often are chosen in these studies are respiratory acidosis, COPD, and heart failure. So those are the kinds of patients that seem to benefit the most. But it's clearly the high-risk patients that are most at risk. I'm not going to talk about Dr. Casey's trial, because he's going to talk about that with you. I think it's a great trial. It's a single-center study. I congratulate him for having conducted such a robust study at one center. You can see that in this study, they used protocolized support compared to usual care. I'll let him talk about that. There was no difference between the two. The protocolized support used either NIV or high flow. I'm not going to go into the details. But there clearly are some studies that show in broader populations that you don't have to use non-invasive ventilation to prevent re-intubation. Dr. Casey likely will tell you that many of their patients did have risk factors. So that's for you to sort out when you hear his data. But we'll talk about that some more. So I think the use of non-invasive modes for prevention of re-intubation and not rescue is important. This is an early study from Esteban in 2004, I think it was. And what it mentions is here. And in that study, what they did was they waited until the patients developed respiratory failure to apply NIV. And in that study, the NIV patients had higher mortality. So if you wait until they develop respiratory failure, that's a problem. This is a network analysis that just appeared earlier this year looking at NIV, high flow, and different outcomes. For one thing, they concluded that in low-risk patients that neither of them were necessary. They also found in their network analysis that they called prophylaxis the intervention being deployed immediately after extubation. And treatment meant that they deployed it when the patient developed respiratory failure. And clearly, the evidence that they found in this network meta-analysis was that you need to use it as prophylaxis or immediately after extubation. And that when you waited, the benefits were not as robust. So yes, use non-invasive ventilatory modes in appropriate patients to prevent re-intubation. My references. Well, thanks, Dr. Lett, for inviting me today. Thanks to everyone for being here on the last day of the conference. I know that I can't beat his fashion sense, but I hope I can beat his arguments. So I'm going to talk today about the use of non-invasive ventilation following extubation to improve outcomes and take the con position. And I think you'll hear a lot of the same data, but maybe framed a little differently. So I am an assistant professor of pulmonary and critical care medicine at Vanderbilt. And the director of the Pragmatic Critical Care Research Group. The only disclosure that's relevant to this talk is a travel grant from Fisher & Paykal to talk about a lot of the concepts we're talking about today. The learning objectives for my talk are to discuss the data behind the use of non-invasive respiratory support. And I want to be clear that what we're talking about, and I think what Dr. Lett's lecture focused on as well, is the patient and the general medical surgical population. There's another rich area of evidence for post-extubation support in post-operative patients that we're not touching on today. So Dr. Lett, I think, did a nice job of addressing, what's the problem that we're talking about? Why do we care about this? The majority of patients in the ICU experience mechanical ventilation. And post-extubatory respiratory failure is common. So about a quarter of patients you extubate will have some level of respiratory distress. And 10% to 15% of those patients will require reintubation. And patients who experience reintubation have very high rates of pneumonia and increased risk of death. And there are only two therapies that have been proposed to prevent reintubation. Noninvasive ventilation and high-flow nasal cannula. Again, I think just to drive home what phase of respiratory support we're talking about, you can use high-flow nasal cannula or noninvasive ventilation at various phases. You can take a patient who fails a spontaneous breathing trial and extubate them anyway. That's contrived in COPD patients, so-called facilitative noninvasive ventilation. That's not what we're talking about today. You can take a patient who you extubate and you wait until they get in trouble and they fail. That's called rescue or therapeutic noninvasive ventilation or high-flow nasal cannula. Again, that's not what we're talking about. What we're talking about is this phase here in the middle where you extubate a patient and right at the point of extubation, while they still look good, you say, should I be applying something preventatively to prevent them from getting in trouble? And that's really the focus of the lecture today. You already saw some of this data. There have been a number of really seminal studies on the use of noninvasive ventilation among high-risk patients. Two of them were from Ferrer. The first published in 2006 in the Blue Journal showed a benefit among the broad population of high-risk patients with regard to reintubation. But when they drilled down, it seemed that that benefit was particularly among the hypercaptic patients. So that follow-up trial that he did and that was published in Lancet in 2009 was really limited to those hypercaptic patients, which I'll highlight as a minority of our ICUs. Less than 10% of our patients are hypercaptic at the time of extubation. But for those patients, there's a big benefit. And it's not just with regard to reintubation, but a mortality benefit for those patients. I also want to highlight that this recommendation was really strong in the 2017 guidelines that Dr. Goulet helped to write. I'll also highlight for you all that when you're invited to a pro-con, hopefully you aren't paired against the guy who literally wrote the guidelines. It's not a good position to be in. They are six years out of date, I'll say. So maybe some new things we can learn since then. In the 2017 guidelines, they recommend non-invasive ventilation for all high-risk patients. And note that the high-risk criteria have been defined variously, but they suggest hypercapnia, COPD, CHF, and other serious comorbidities. They also referenced in these guidelines that there was this new data from hyponosocannual that had recently come out and wasn't considered for the guidelines. And Dr. Goulet referenced these as well. These were two trials done in Spain by Hernandez. And they looked at two populations. They said, for high-risk patients, we think they have to have something. And is hyponosocannual as good as non-invasive ventilation for non-hypercapnic high-risk patients? And you can see some separation occurs, but it wasn't significant. So they said, yeah, hyponosocannual is probably as good. And they were the first ones to look at, well, what about everybody else? What about the low-risk patients? Should they receive anything? So they looked at low-risk patients and randomized them to hyponosocannual or conventional oxygen, just a standard nasal cannula or not a breather, and showed that even for those low-risk patients, they were able to have the rate of reintubation. That has been incorporated into subsequent guidelines, like the ESICM guidelines that suggested hyponosocannual as opposed to conventional oxygen for any patients with high-risk features. And then the very recently released ERS respiratory journal practice guidelines, in which the ERS task force suggested the use of hyponosocannual over conventional oxygen in non-surgical patients after extubation. And then subsequently, they also recommended non-invasive ventilation over hyponosocannual, which we'll talk about. So I think a summary, an interim summary, might be that for hypercapnic patients, non-invasive ventilation decreases mortality, the strongest evidence we have for the use of preventative post-extubation support. For non-hypercapnic high-risk patients, non-invasive ventilation decreases reintubation, and hyponosocannual might be non-inferior to non-invasive ventilation. And then for non-hypercapnic low-risk patients, hyponosocannual also decreases reintubation. So that's everybody, right? So those three groups together is everyone we take care of. And that raises the obvious question, should every patient receive something? I would argue that maybe, but maybe not. I'm not sure that those trials really represent all the patients we care for. I think as a clinical trialist who does trials like this, I worry that these trials cherry-pick patients who are the most likely to benefit from these therapies. And a few pieces of evidence is that these trials took years to enroll a pretty small number of patients, even though half the patients in their ICU were undergoing extubation or were receiving chemical ventilation and undergoing extubation. And then they had pretty high rates of reintubation even in their control arms. So in study after study, the reintubation rate in large registries is 10% to 15%. But you can see the reintubation rate in the control arms in these studies was really high. Even the low-risk study had a higher reintubation rate at 12% than many overall registry studies have shown, which suggests to me that maybe they were picking patients most likely to benefit. Further, I think there is some evidence regarding how we practice that the use of these therapies is pretty low at most centers. So we are certainly applying these therapies, but we're only applying them to a small minority of people. Dr. Roulette mentioned the surveys from practitioners that show only half of people believe noninvasive ventilation is effective. And even after all the trials we've talked about, if you go to UpToDate today, they give a very strong recommendation in the opposite direction of what these trials might suggest. We do not routinely extubate patients at risk of post-extubation respiratory failure to noninvasive ventilation or high-flow nasal cannula. However, select patients at the highest risk may benefit. So I think this is an area of ongoing controversy. You've got a number of trials that suggest potential benefit. People aren't doing it. Maybe this is an opportunity to improve outcomes by really testing a strategy of, let's apply this in everybody and see if it works. And that's the context in which we conducted the proper trial, the Protocolized Post-Extubation Respiratory Support Trial. It was a prospective, unblinded, pragmatic cluster crossover trial conducted in the Vanderbilt Medical ICU between October 1, 2017, and March 31, 2019. The inclusion criteria was everybody, everybody undergoing extubation in our ICU. The exclusion criteria were those who were on the vent for less than 12 hours, so really, really low-risk patients, patients who had a D&I order, who couldn't experience reintubation, and patients who had already been in the trial, had already been extubated out of hospitalization. Our ICU was divided into two geographic clusters, front hallway and back hallway, and then we had three-month clusters where patients in one half received protocolized support, and the other half received usual care, and the groups alternated every three months. In the protocolized support group, everyone received something. So patients were all extubated to either noninvasive ventilation or high-flow nasal cannula. Patients with suspected hypercapnia received noninvasive ventilation. All other patients received high-flow nasal cannula. The therapies were left in place until 5 a.m. the next day when patients were assessed for their readiness to be weaned from those therapies, and all their aspects of clinical care were determined by the clinical team. The other group received usual care, which was every decision at the discretion of their treating clinicians. The algorithm that we used, I won't go through this in detail, but I do want to highlight that even for patients who had contraindications to noninvasive ventilation, they were supposed to receive something. So if you were hypercapnic but had a facial fracture, you would receive high-flow nasal cannula. Primary outcome was reintubation within 96 hours. Secondary outcomes included number of ICU free days. Exploratory outcomes were all cause and hospital mortality, ventilator free days and time during intubation. During the study period, 1,517 patients in our ICU received mechanical ventilation. Of whom, about a third died prior to extubation. Of those who were extubated, about 200 were excluded for receiving ventilation for less than 12 hours or for having a D&I order in place. And ultimately, 359 were assigned to protocolized support, and 392 to usual care, and all patients were included in our intention to treat analysis. You can see the baseline characteristics of the patients in both groups here. I hope these represent the patients that you take care of in your ICU, median age in the mid-50s. Majority of patients were obese. About half of patients had sepsis or septic shock. About a quarter of patients had failed an SBT prior to their first extubation. And about 70 to 80% of patients had at least one criteria that deemed them to be high risk based on the criteria used in prior studies. 92.2% of patients in the protocolized support group received protocolized support compared to 16.8% in the usual care group. I think it's worth highlighting that the major difference in the way the respiratory support was used was a difference in hyphal nasal cannulae use, that both groups received noninvasive ventilation in 15 to 17% of cases. Here's the separation between group slides showing how much support was given in the protocolized support group compared to usual care. And then what do we find? So doing this 18-month study where we excluded every patient to a device, we found that it didn't affect re-intubation within 96 hours, which occurred in 15.9% of patients in the protocolized support group compared to 13.3% in the usual care group. And secondary outcomes were similar between groups. This is the same data shown in a competing risk analysis, looking at the competing risk of death, which also showed no difference. Many trials in the past have used durations of support of 24 or 48 hours, so is it possible that we just didn't use it long enough? It didn't look like that was the case, that there was no effect modification by duration of support. And similarly, when you looked at the results by number of risk factors for re-intubation, it didn't appear that protocolized support was more effective in either low-risk or high-risk patients. There were no difference in other exploratory outcomes, including indication of re-intubation, re-intubation of 28 days, or in-hospital mortality. So our summary is that protocolized support failed to prevent re-intubation when a broad population of critically ill patients wasn't affected by duration of support or by the number of risk factors for re-intubation. There are limitations of the study. It's a single-center study. We don't routinely check our chilled blood gases on a spontaneous breathing trial, so it makes it harder for us to characterize who the patients are with hypercapnia in this study. And that the trial intervention was compared to usual care. There was the use of these therapies in the control group, particularly noninvasive ventilation, with the main difference between the two groups being driven by a difference in hyponatial cannula. Since we've conducted the proper trial, there are a couple other studies that I think add to the richness of our data in this area. One that was just published is the RINO trial. This was by Maggiore, an Italian researcher who did some of the earliest studies in the use of hyponatial cannula following excavation. This 494-patient study randomized patients to hyponatial cannula or Venturi mask and found no difference in re-intubation, which occurred in 13% of patients in the hyponatial cannula group compared to 11% in the Venturi mask group. And then this study by Arnaud Thiel and Jean-Pierre Frat was mentioned by Dr. Ouellette, and I'm going to interpret it differently, which is that they said every patient skinning out to flow does adding noninvasive help. And the answer was yes, that the group that received noninvasive ventilation in hyponatial cannula had a lower re-intubation rate than people who received hyponatial cannula. It's also important to note that this group didn't exclude hypercapnic patients like many of the other prior trials have, so it might be a little surprising that they would enroll and randomize hypercapnic patients to hyponatial cannula alone. And when they looked at their own data, the effect, again, was driven by the patients with hypercapnia. There is maybe some signal for benefit among the patients without hypercapnia, but it's not significant. So I think the summary of all this data put together is that there's no data to support excavating every patient to something, as we try to do in the proper trial, that patients with hypercapnia clearly should receive noninvasive ventilation. That's some of the best evidence we have with regard to mechanical ventilation in the ICU. So we should be identifying those patients and excavating them directly to noninvasive ventilation. Most patients without hypercapnia can probably be managed with conventional oxygen therapy, but I think it is clear that when you add up these trials, there are some patients without hypercapnia who benefit from noninvasive ventilation, but who are they? In the proper trial, 80% of our patients had at least one risk factor. And as Dr. Ullett said, I think it's very hard to identify who really is high risk. So I think that's where the next work should be. Who is high risk? How do I identify them? And then testing that group more specifically for noninvasive ventilation, I think, is the next step. I think it's unclear if any patient should be excavated to high frequency cannula. So thanks again for the invitation. It's been fun talking. All right. Thanks, Dr. Ullett, for that introduction and for inviting me to participate in this session. I'm Melanie Garbarino. I am an assistant professor at the University of Cincinnati, and I don't have anything to disclose yet because I'm just not that cool yet. I'm going to be presenting the pro side of this debate, and I hope that I'll be building a slew of points in favor of using the Rapid Shell Breathing Index as a tool to help predict successful ventilator liberation. Here are learning objectives. We're going to talk about some historical ventilator liberation strategies. We're going to outline how the RSBI was derived in the early 1990s. We're going to talk about some of the physiologic interests in the RSBI as a tool. We're going to define some advantages. We're going to outline benefits. We're going to discern future directions. And at the end of all of this, I will have you convinced, I promise. Mechanical ventilation has been a life-saving tool, widely employed in critical care settings since the polio epidemic of the 1960s. No good thing comes free of cost, however. And over the years, we have learned that both prolonged mechanical ventilation and premature ventilator liberation carry deleterious effects, including increased morbidity and mortality. This raises the question, when is the right time to liberate the patient from mechanical ventilation? Prior to the 1990s, there were no consensus strategies applied for ventilator weaning and subsequent liberation. Practices relied mostly on clinical gestalt with high variability in patient outcomes. Some other historical strategies included vital capacity, minute ventilation, maximal inspiratory pressure, and static and dynamic compliance, shown here at the bottom of the slide. I've compiled the performance of various methods from as many studies, historical studies, as I could access. And we can appreciate the wide variability with all strategies. Note the wide ranges of positive and negative predictive values for minute ventilation and maximal inspiratory pressure. Several methods also required special equipment and patient cooperation for special maneuvers, which makes it difficult for reproducibility across studies. Others, like vital capacity, were commonly used prior to the 1990s, but really had no data to support them and have no data in the literature. They just happened to be custom. And so what we can ultimately see from this table is that historic criteria for ventilator liberation were really not much better than a simple coin flip. In response to this historical heterogeneity, the Rapid Shell Breathing Index has become one of the most ubiquitous components of contemporary ventilator liberation strategies. It was first described in a two-pronged prospective study by Drs. Yang and Tobin that was published in 1991. In this initial trial, 100 medical ICU patients clinically stable for ventilator weaning were placed on a T-piece, and data was collected after they breathed room air through this T-piece for a minute. During the T-piece trial, the tidal volume was measured by a handheld spirometer attached to the patient's endotracheal tube. And as we all likely know, the RSBI is the ratio of the respiratory rate or respiratory frequency to the tidal volume expressed in liters. The RSBI was compared to several of the strategies shown on the previous slide, as well as another index created by Dr. Yang and Tobin that I won't spend much time on, other than to tell you it was called the CROP Index. Dr. Yang and Dr. Tobin reported that an RSBI less than 105 predicted successful weaning with a sensitivity of 97%, a specificity seen here of 64%, as well as a positive predictive value and negative predictive value of 78% and 98%, respectively, which we can see here in this blue box. They also performed receiver operator curve tables. And you can see here that the RSBI less than 105 predicted weaning success. Oops, I'm sorry. The RSBI performed much better than any of the other strategies used. And you can see here that they had an R-value, ROC value of 0.89, which is almost unheard of in the literature to have an R-value, ROC value that's so close to 1. So they concluded that the RSBI less than 105 was the single most accurate index in predicting success or failure of ventilator liberation. Dr. Tobin and many other investigators have noted that beyond the promising statistics from this initial study, obtaining the RSBI was simple and intuitive. It was easy to remember, which lended favor over more complicated indices that required special equipment or patient cooperation to obtain. The description of the RSBI marked a turning point in ventilator liberation. The RSBI as a ventilator liberation technique was repeated and validated by many other investigators. Compiling these subsequent studies, it certainly appears that the RSBI is better at predicting successful ventilator liberation than the random coin flip associated with the historical strategies I showed you earlier. After its definition and validation, several groups of investigators derived a physiologic explanation for the predictive power of the RSBI. This was based on the assumption that the primary reason patients fail liberation from mechanical ventilation is due to an imbalance between the work imposed on the respiratory system and the ability of that specific patient's respiratory muscles to meet that demand. Several factors, including high resistance, low compliance, and malnutrition contribute to the maladaptation of the respiratory muscles. And this maladaptation is the likely cause of the rapid shallow breathing we see in patients that are highly likely to fail a trial of extubation. Indeed, in 1997, a group of researchers within the trauma and critical care division at the University of Cincinnati studied the relationship between the RSBI and the work of breathing in 15 surgical ICU patients. They compared the RSBI at different levels of pressure support ventilation to the patient's work of breathing. They found, shown in the two tables here, that the RSBI directly and proportionately correlated to the patient's work of breathing at all levels of pressure support. As pressure augmentation was increased, the RSBI decreased in a linear relationship. The authors also noted that as direct measurement of a patient's work of breathing can be expensive and is invasive, it's difficult to be widely applied and thus can be easily misinterpreted by unfamiliar clinicians. These investigators concluded that the RSBI is and will be familiar to most clinicians. It's easily and reliably obtained. And it varies directly with work of breathing, making it a more meaningful tool for physicians when caring for mechanically ventilated patients who are otherwise ready for ventilator liberation. Since the 1990s, the RSBI has become the cornerstone of ventilator weaning and liberation protocols in ICUs across the country and the world. As early as 1996, Dr. Ely and his colleagues designed a randomized trial protocolizing a daily spontaneous breathing trial in patients that had received or had been determined to have recovered enough to undergo ventilator weaning. This trial, or this protocol, was driven by nurses and respiratory therapists. It included direct report of the result of the patient's SBT and the result of the RSBI to the physician team. Dr. Ely and his colleagues found that institution of such a protocol required minimal additional time, incurred no extra costs, and was easily incorporated into pre-existing nursing and respiratory therapist routines. They also found that the protocolized discussion once daily between the nurses and the respiratory therapist greatly improved communication between the staff and physicians. The AWAKEN breathing trial of 2008 also demonstrated similar findings. Protocols that include the RSBI as a key part of an SBT are easy to implement because no special monitoring or respiratory equipment are needed. There are no additional expenditures in cost or time, and no laboratory studies are necessary. Given these benefits, incorporating the RSBI reporting into SBT protocols can be widely applied in all settings where mechanical ventilation is utilized, rural or urban, community or university. SBT protocols incorporating the RSBI have become so ubiquitously applied that ventilator manufacturers have included algorithms that calculate and report the RSBI on ventilator screens. You can see here highlighted in the red box that this is a patient performing an SBT in my ICU. And based on the results of 47, at the end of 30 minutes, they're doing extremely well. The take home here is that the RSBI has become ingrained in modern ventilator protocols and is such a familiar tool that it's used in nearly all ICUs. It's incorporated into most ventilator algorithms, and it's taught to all trainees who are taking at least USMLE Step 3 or internal medicine boards. Nearly all studies looking at the RSBI over the last three decades have shown that measuring the RSBI is safe, and there have been no reported complications or incidents of harm during the measurement of the RSBI. So the RSBI is very low risk to obtain. While the RSBI has many things to recommend it as a tool in ventilator liberation, it has by no means reached its maturity and should be further refined. I propose three primary questions that we should be asking in our clinical trials that may solidify and improve this widely applicable, easy to use metric. First, as T-PIECE is being used less commonly across ICUs, is there a right way to measure the RSBI? Secondly, if we aren't using the strategy that was used to define the RSBI at its inception, should we determine a new threshold that's applicable to modern spontaneous breathing trial methods and protocols? And finally, if we improve standardization of the RSBI in ICUs with 2020 metrics, will the predictive accuracy of the RSBI improve? So several studies have looked at the best strategies for vent weaning, and it's been suggested in multiple studies that pressure support is a superior weaning strategy to T-PIECE, to T-PIECE, CYMPIE, and even CPAP, which is used, I know, in many ICUs today. This figure on the right is from a prospective trial by Dr. Sabira, which I think nicely demonstrates early separation in successful extubation and shows favorability if you utilize a 30-minute pressure support trial over a two-hour T-PIECE trial. Other studies further suggest that utilizing pressure support ventilation as a weaning method is associated with fewer vent days, fewer adverse events, and fewer extubation failures. There are also various studies looking to measure the RSBI when the SBT is performed with different strategies. In a nice review by Dr. Karthika and his colleagues from 2016, they outlined that an RSBI obtained after a 30-minute SBT on low pressure support ventilation is less stressful for patients and is superior to other methods. Finally, the most recent CHESS, ATS, Ventilator Liberation Guidelines also recommend using pressure augmentation when performing SBTs, but they don't make any recommendations on how the RSBI should be applied or incorporated into our ventilator liberation strategies. It has also been suggested that perhaps the historical threshold for the RSBI of less than 105 is too high for modern SBTs, which much more often utilize CPAP or pressure support instead of the T-PIECE strategy. There have been several prospective trials showing successful ventilator liberation. It RSBI is much lower than 105. And in some, they even show more extubation failures when this traditional threshold is applied. In 2009, a group of investigators out of the University of Vermont looked at various strategies of obtaining and calculating the RSBI in 67 mechanically ventilated patients. They demonstrated that when calculated on a CPAP of 5 centimeters of water, the RSBI was 71 compared to 90 obtained on a T-PIECE. And this difference in this obviously small study did reach statistical significance. In another study that was performed in China in 2014, the investigators compared RSBI values obtained on pressure support ventilation and T-PIECE at 3 and 30 minutes in 208 mechanically ventilated patients. They showed that after a 30 minute SBT on a delta with a pressure augmentation of 5 over a CPAP of 5, the average RSBI was about 68 compared to approximately 100 on T-PIECE. They then evaluated the predictive power of the RSBI at varying levels on pressure support ventilation. And based on their results, an RSBI of 75 had the best combination of sensitivity, positive predictive value, and accuracy. This group also challenged the traditional T-PIECE threshold of 105, utilizing a 30 minute spontaneous breathing trial when 105 was initially defined on a two hour T-PIECE trial. And they found that this traditional value was associated with more extubation failure. So as there are various ways to obtain the RSBI and various thresholds reported throughout the literature, most meta-analyses are very hard to interpret or are impossible to perform due to the high heterogeneity between studies. Many of the larger multicenter trials lead the spontaneous breathing trial strategy up to the individual institution and not all report the SVT method used in their multivariate analyses. Given the ease of obtaining, intuitive nature of, incorporation into pre-existing routines and protocols, and safety profile of the RSBI, I imagine that it will not change and remains really an immensely useful tool in ventilator liberation. These aspects of the RSBI are fixed in the RSBI and should not, in my opinion, the RSBI shouldn't be abandoned. But we should focus our further efforts and attention on honing this tool so that any improvement can be easily and immediately implemented into ventilator liberation protocols and help improve our accuracy of predicting who is ready to have the tube removed and who is not. In summary, since its inception in the early 1990s, the RSBI has become a vital component of ventilator liberation. Among the reasons the RSBI is a simple, safe, non-invasive, easily-protocolized, and cost-effective surrogate for identifying when the demands of the respiratory system no longer outweigh the capability of the patient to need them. While more modern strategies for ventilator reading have called into question the threshold value for the RSBI, this does not mean that the intensive care community should abandon our old friend, but that given how integral the RSBI is into contemporary practice, future efforts should seek to redefine the optimal threshold for this value. Thanks very much for your attention. Good morning, everybody. As I get my presentation uploaded, I just wanted to thank you all for coming and thank you to Dr. Ouellette for inviting me to speak today about the controversies in mechanical ventilator liberation. I will be presenting the counter-argument to what you just heard about the Rapid Cell of Reading Index to predict successful ventilator liberation. Again, I'm Katie Gardner. I'm a pulmonary and critical care attending at the University of Tennessee Medical Center in Knoxville, Tennessee, and I have nothing to disclose. Our learning objectives today will review the predictive success of ventilator liberation using RSBI as a weaning parameter in the clinical decision making. I'm going to look at the validation of the RSBI calculations and also talk to you more about the predictive value, as well as some areas for further direction and how to use this tool. Deciding to extubate is a significant portion of our work in the ICU. About 40% to 50% of the duration of the mechanical ventilation of our patients is spent in this weaning parameter, and that has significant outcomes on length of stay, mortality, morbidity, increased risk of infections, increased risk of tracheostomies and complications for our patients. So assessing readiness as soon as the patient's reason for acute respiratory failure has resolved is imperative in our management of our patients. Extubation failure can occur in up to 10% to 20% of patients, and so using strategies to help identify these patients at high risk for extubation failure are essential to improve our management and outcomes. As has been identified in this talk so far today, identifying those high-risk patients is very challenging and limited, and I'm not an expert in that either. So any tool we have to help predict this, I think, is valuable. Failure of planned extubation is associated with prolonged mechanical ventilation and higher mortalities, about 35% to 50%. Some of these risk factors have been identified as listed on the screen, but that can be very, very variable among the heterogeneity of our populations in our ICUs. As has been mentioned, the origin of RSVI had begun in 1991. There was 100 patients that were receiving invasive mechanical ventilation, where it was found that after unsuccessful liberation, patients were developing a rapid shallow breathing pattern. And when studied, it was found that the rapid shallow breathing index of less than or equal to 105 could be predictive of that extubation failure. The threshold less than that was where extubation predicted to be more successful. But the sensitivity was elevated, although the specificity was a little lacking. The positive predictive value was only about 78%, but the negative predictive value is where the utility of this tool came from. So highly predictive of weaning failure, not necessarily as successful liberation with those nuances in mind. So using the RSVI to predict ventilator liberation, there's been several clinical observations and variable studies with various outcomes. They all validated that the RSVI is a tool itself, but the heterogeneity amongst the patients was variable for that success in liberation. Looking at some of the studies from 2006, there were several factors associated with re-intubation for patients despite passing a successful spontaneous breathing trial. In 37 hospitals with patients that intubated for over 48 hours and deemed ready for extubation, they identified RSVI as one of the best predictors associated with extubation failure. However, there was a litany of other parameters that also impacted their care, including positive fluid balance for the 24 hours preceding extubation, pneumonia, respiratory muscle fatigue, diaphragm atrophy among others. And all those parameters pooled were all very poor predictors of successful extubation. It just happened to be that RSVI was at the top of that list as being the best. I do think that as technology has changed throughout the years, is there anyone in the room that still pulls out a right spirometer to measure their RSVI in their clinical practice today? Probably not. So most of us look at that number that's calculated on the mechanical ventilator screen. And so there were some studies that looked at validating this. So in 2021, there was a report of measuring the RSVI using the traditional spirometer compared to the value that's displayed on our ventilators and calculated by the programming. They enrolled 47 patients. And they used the RSVI calculated by the ventilator and compared it to that standard on every single patient. They actually found that that ventilator displayed reference was over 17 breaths higher than the spirometer used, which is from the original 1991 trials. So that significant overestimation was actually statistically significant. They actually found that averaging the values of that displayed parameter at 0, 15, 30, 45, and 60 seconds was found to better correlate with what we consider the standard from Tobin and Yang. Now, taking that difference in the values, does it matter clinically? So in 2006, they looked at including that parameter in the clinical decision-making about extubation. This looked at a little bit bigger parameters of 304 patients. And they looked at using that measured RSVI on the ventilator compared to not using it in your decision to extubate. They found that in the patients passing their 2-R SBT that were otherwise eligible for extubation based on their clinical care, they did find a shorter duration for weaning time of one day in the population where they did not use the RSVI in their clinical decision-making compared to the group that did use it. So although they had a shorter duration in their mechanical ventilation, the other clinical outcomes as far as in-hospital mortality, tracheostomies, unexplained extubations were not changed. So one day shorter, but not other significant outcomes. Another group in 2022 produced a very good summary of the different variables that affect this RSVI in our clinical practice. Several of them are ventilator settings. And as you heard summarized in the previous presentation, looking at modalities including TPs, CPAP, pressure support ventilation as those parameters do influence that RSVI calculation. They found that RSVI was typically smaller or a lower number in patients on CPAP or pressure support ventilation compared to TPs. So that's further moving us away from that original trial with Tobin. Another parameter was the timing of measurement of the RSVI if it was calculated at the beginning, middle, or end of a trial, whether it's 0, 30, or 2-hour time periods. And that could influence your decision-making in that value. There is some suggestion that, for example, COPD patients may fail later into their trial at 2 hours. And so using that RSVI versus the initial one may impact your successful liberation. There's also several patient factors that influence the RSVI, their cardiac status, their pulmonary status, their age, their gender, the reason for mechanical ventilation, if they even have abnormal respiratory physiology, or if they're just on for what's colloquially called airway protection. So all of these factors do influence it. And there's been some directions to help improve the accuracy of our RSVI. One of those is the serial rate of change of the RSVI. That has been found to be a little bit more predictive in patients that have been on mechanical ventilation more than eight days or in LTACH facilities for over 21 days. But otherwise, in their acute care ICUs, there has not been a lot of significant changes. In 2022, there was a very large meta-analysis published by Trivedi. And it included 48 studies that summarized well over 10,000 patients. This did show that the pooled RSVI amongst all studies showed a moderate sensitivity of about 83%, but a poor specificity of 58% for predicting extubation success. They did do significant subgroup analyses. But due to the heterogeneity, there was no creditable effects on the RSVI thresholds, measurement techniques, or patient characteristics. So as discussed, would changing the RSVI parameters influence our outcomes? If we change that number to less than 80, looking at 80 to 105, or any numbers less than 105, the sensitivity does stay roughly the same. There's a little bit of an increased specificity when we look at changing the RSVI value down to less than 80. But which patient population does that apply best to is still unclear. And those outcomes did not have any difference on mortalities. Looking at the sensitivity and specificity on TPs, pressure support, or CPAP, was grossly the same across subgroups. They also did not see a difference in subgroups of chronic obstructive lung disease patients or patients in multidisciplinary ICUs, so medical, surgical, neurologic ICUs. So again, highlighting the heterogeneity of these pooled summaries is difficult to account for those differences, RSVI and the parameter on its own in isolation. There are significant measurement technique differences. Study level patient characteristics was very variable. And then even the outcomes to define weaning failure and that time period, whether it was 12 hours, 24, or 48 hours, there was a lot of variability. So in summary, it did find that there is some positive predictive value for RSVI measured in patients with more prolonged invasive mechanical ventilation. But prior to that date, it's a little bit variable. In summary, RSVI may be at the lower utility in patients with a very high or very low pretest possibility for extubation success. So that middle group is where it may be more applicable. It's a simple measure, but it doesn't account for enough variability in our patients, in my opinion, for this presentation. One way I do suspect that we can improve upon this number is using diaphragmatic rapacellar breathing indices. There's actually several interesting articles this past couple of years looking at use of diaphragm analyses to evaluate our patients' respiratory mechanics. There's been other discussions at this conference that have talked about pulmonary function testing and COPD patients and using training of diaphragms to help improve those values. And I think there's a way to apply it also in our mechanically ventilated patients. This study looked at using the diagnostic accuracy of what I'm going to label as D-RSVI as a predictive weaning outcome to compare to the standard RSVI calculated on the ventilator. It was a prospective observational study, and they calculated this using a single operator using a curvilinear ultrasound, looking at the right side of the chest, looking at the diaphragm between the mid-clavicular line and the anterior axillary line, referencing the posterior aspect of the diaphragm and scanning upwards as the diaphragm contracted to be able to calculate the diaphragm displacement. And so it was a single operator using that same tool for interoperator reliability. Diaphragm displacement was calculated using the M-mode on ultrasound. And overall, a couple other studies reported that dysfunction of less than 10 millimeters was a reliable predictor for respiratory weaning failure. The annual 50 patients who included 45 of them had passed a successful SBT and met all other parameters for successful extubation. 41 patients of those were successfully extubated and did not require rescue or reintubation, which accounts for a weaning failure of about 18%, which is consistent with the literature. The patients were mechanically ventilated for more than 48 hours and were ready for their first SBT at the time of analysis. As you can see, they did measure some of those variables that we used to help with ventilator liberation, including maximal inspiratory pressure, RSBI, diaphragm displacement, as well as a couple other transdiaphragm pressures. So there was a significant difference in the maximal inspiratory pressure, as well as the diaphragm displacement, in the patients that were successfully liberated and those that failed extubation. They found that the diaphragmatic RSBI had a positive correlation and greater diagnostic accuracy than the traditional RSBI alone. The diaphragmatic RSBI actually had a sensitivity of 90%, specificity of 100%, positive predictive value of 100%, and negative predictive value of 69%. So that's a significant increase from about 40% to 90% and 26% to 69%, respectively, compared to the RSBI. So in conclusion, I do feel that the RSBI is an important predictor of weaning outcomes. I do not feel that the RSBI should be used alone in selecting our patients for successful liberation, but can be used to help stratify those at risk for extubation failure. And I think using it in the context of the heterogeneity of your patient, your ICU, their reasons for intubation in the first place can help you be more successful. And I think using more dynamic tools, like an ultrasound, can help guide and improve your accuracy of your RSBI to apply it to your patient specifically. Thank you for your attention. Thank you.

RSBI

ventilator liberation

respiratory rate

tidal volume

weaning

predictive value

specificity

measurement

clinical indicators