This webcast hosted and supported by the Journal Club Critical Care Medicine series. In today's webcast we feature an article from Critical Care Medicine. My name is Tamas Zagmeni and I'm a Professor of Intensive Care at Cardiff University in the United Kingdom. I will be moderating today's webcast. Thank you for joining us. Just a few housekeeping items before we get started. First, during the presentation you will have the opportunity to participate in interactive polls. When you see a poll, simply click the bubble next to your choice. Second, there will be a Q&A session at the end of the conclusion of the presentation. To submit questions throughout the presentation, type in the question box located on your control planner. Third, if you have a comment to share during the presentations, you may use the question box for that as well. And finally, everyone joining us for today's webcast will receive a follow-up email that will include an evaluation. Please take five minutes to complete it. Your feedback is greatly appreciated. Please note that this presentation is for educational purposes only. The material presented is intended to represent an approach, view, statement, or opinion of the presenter, which may be helpful to others. The views and opinions expressed herein are those of the presenter and do not necessarily reflect the opinions or views of SCCM. SCCM does not recommend or endorse any specific test, physician, product, procedure, opinion, or other information that may be mentioned. And now I would like to introduce today's presenter. Shaurya Taran is a clinical associate in critical care at the Toronto Western Hospital and a graduate student in clinical epidemiology with the Institute for Health Policy Management and Evaluation at the University of Toronto. His clinical interest is neurocritical care, and he will be pursuing a neurocritical fellowship at Harvard University beginning of next year. Dr. Taran's research focuses on practices of mechanical ventilation in patients with acute brain and spinal cord injuries. He is a Philipson scholar with the Elliot Philipson Clinical Scientist Training Program and is also a member of the Clinical Investigator Program at the University of Toronto. Thank you, Dr. Taran, for joining us today. I will now turn the presentation over to you. Thank you, Tomas, Jill, Rafaela, and the podcast organizers today for inviting me to talk on this study. Today I'll be discussing a recent publication on discordances between factors associated with withholding extubation and extubation failure after a successful spontaneous breathing trial in the intensive care unit. Before I begin, I'd also like to give a huge thanks to all of the co-authors and co-investigators on this project, including Dr. Angerman, Dr. Pinto, Dr. Ferraro, and Dr. Andre Amaral, who is the study supervisor. Again, a lot of this presentation and certainly these results would not have been possible without high-level input from all those individuals. So again, a major thank you to all involved. So without further ado, let's get started. And just for me to get a sense of who's listening today, I was hoping to ask this first polling question. If you could please kindly indicate if you're a nurse in attendance today, a physician, an advanced practice provider or pharmacist, or a PT, OT, or respiratory therapist, or if you're a student. Again, just helps me get a sense of who's listening today. Okay, perfect. So the polls have come through and it looks like the majority of individuals today are advanced practice providers, 56% are pharmacists and physicians at 44%. So fantastic. Thanks for taking the time for that poll. Okay, so just in terms of setting the context a bit, everyday clinicians in the intensive care unit have to make important decisions about who might be ready to be extubated. And this is actually no small matter because the stakes are relatively high. On the one hand, there's the risk of a patient having extubation failure, which is typically defined as the need for reintubation within 48 to 72 hours or some other preset specified period after the initial extubation attempt. And on the other hand, there's also the possibility that you're unnecessarily delaying extubation in a patient who might otherwise be ready to be liberated. There are attendant risks associated with receiving unnecessary mechanical ventilation. And again, you're exposing patients to those risks. Now we know from lots of elegant studies that extubation failure seems to be an independent risk factor for excess mortality. So even after adjusting for illness characteristics, comorbidities and so forth, something about the act of failing extubation itself seems to predispose to heightened mortality. It's postulated that in a patient who might not be ready to breathe on their own, there's time for additional organ failures to develop in the post-extubation period. And this has actually been elegantly demonstrated by increases in SOFA scores within that fragile population. And there are other kind of speculations as to why this might be as well. We also know that delaying extubation is associated with increased ventilator associated complications, including pneumonias, higher hospital costs and longer intensive care unit lengths of stay. This has again been very elegantly demonstrated by many studies. Historically and traditionally, factors involved in the extubation process have typically been characterized with respect to extubation failure. Again, it's an important outcome, one that we hope to avoid for our patients. And so it's no surprise then that this has been the majority of research. Less attention, I would say, has been paid to the outcome of withholding extubation, but I hope you'll agree that in clinical practice, both outcomes should be considered simultaneously to inform the highest quality decision-making. Now, I'm going to spend a few minutes on this slide here, so please bear with me. I think that some of these concepts can be a little complicated to grasp on first attempt, and hopefully this visual schema will help clarify some of those concepts. So imagine the two outcomes that clinicians care about, extubation outcome and timing of extubation as being on two axes. This is purely a theoretical construct, but hopefully it becomes clearer as I go through this slide. So on the horizontal axis, we have failed versus successful extubation. And on the vertical, we have prompt versus delayed extubation. Extubation would be considered prompt if the patient is liberated very soon after they meet weaning readiness criteria, and it would be considered delayed if there's a clinically relevant time lag between meeting weaning criteria and the extubation actually happening. Now, we can place patients in this grid according to which level, so to speak, of these two outcomes they experience. So in the first case, for example, the patient has met weaning readiness criteria, the clinician recognizes this promptly, and extubation is also performed promptly. The patient also does not go on to require reintubation, so we would consider that as a successful extubation. That is, again, the hope and goal for all patients, but of course, as we know, not all patients fit within that portion of the grid. In the next case, extubation is performed with a delay, speculating as to why this might be. Perhaps there are risk factors for extubation failure present, the clinician recognizes them and wants to give time for them to be optimized, and yet nevertheless, the patient still fails extubation. Again, theoretical construct, but I don't think it's a stretch to say that in a lot of cases, where patients fall between these two categories is often dictated by their illness characteristics, which we as clinicians don't have much control over. So a patient might be young, robust, and otherwise healthy, and fall within that green category. They might otherwise be frail, elderly, and have lots of comorbidities, which places them in the red category. Where there might, by contrast, be more of an opportunity for meaningful changes in clinical trajectory are in these orange boxes. So let's consider the bottom right box first. In this case, the patient is meeting weaning readiness criteria, but extubation is delayed. However, when the extubation actually happens, it's successful. So in this case, the opportunity for change might be that the extubation could have happened more expediently. On the other hand, in this top left scenario, extubation is being performed quickly, but the extubation attempt was a fail. In this case, again, speculating as to why this might be, perhaps there was misrecognition of risk factors for extubation failure that could have warranted earlier optimization. So in other words, these orange boxes represent opportunities to change a patient's clinical trajectory. Again, a bit of a simplification, but for the sake of the rest of the slides and the discussion, I think it's reasonable to adopt. So how does this schema actually help? Consider it from the perspective of any given factor. Let's take age or sex, for example, as relevant factors. Now consider the direction of association of these individual factors with each of the two outcomes of interest. When the associations go in the same direction, in other words, when they're concordant, then for that given factor, a patient may fall in the red or green boxes. Whereas when the associations go in opposite directions, i.e. they're discordant, then that factor would place the patient in the orange boxes. So if you can find orange factors, again, to use the color analogy, you could potentially change your approach to extubation in a clinically meaningful way. And if you find factors that are concordant, i.e. red or green, then you could be reassured that the clinical gestalt you're using to drive your extubation decisions is based in good logic. So regardless of the direction, there are meaningful implications. This translates, again, thank you for entertaining me with sticking through that talk. Again, I think it's important to clarify these concepts up front because they can get a bit complicated if not addressed in advance. So the study objective here was to identify clinical factors associated with withholding extubation and to determine whether this association was concordant, i.e. arrows aligning, or discordant, i.e. arrows in different directions with extubation failure. So let's also pause very quickly and define some important study definitions. The first of those being a spontaneous breathing trial. This is a test of the patient's readiness to breathe unsupported by the ventilator, and it's determined by standardized pass or fail criteria. There are lots of nuances to how SBTs are performed, whether it's with pressure support, how much pressure support, whether it's with a T-piece, is it 30 minutes, an hour, or two hours. We're going to gloss over some of those nuances, and I will say that all of the sites that were included within the study use standardized SBT criteria. The second definition to be cognizant of is that of withheld extubation, which is no extubation performed within 24 hours of a first successful SBT. That would be a clinically relevant time lag. Extubation failure we defined as the need for unplanned reintubation within 48 hours of extubation, which is consistent, again, with commonly reported timeframes. And another important additional definition is that of discordant factor. That factor is associated with one of the two outcome levels of interest, but either has no association with, or an association in the opposite direction to the other, the so-called orange factor. So, in order to undergo a spontaneous breathing trial, at least across the sites that were included within the study, the patient has to first meet certain eligibility criteria, and those are indicated in the left box here. There has to be spontaneous respiratory effort, that's a given. The patient's partial pressure of oxygen to the fraction of inspired oxygen concentration has to be greater than or equal to 150, and their FiO2 has to be less than or equal to 0.5. If no arterial blood gases are available, then the SpO2 has to be greater than 92%. They should be on no more than a positive end-expiratory pressure of 10 centimeters of water, and on no more or higher than one vasoactive drug at 0.2 micrograms per kilogram per minute of norepinephrine or equivalent. In head-injured patients, there should be no concern for elevated intracranial pressure, and no need for active carbon dioxide control. So, once patients meet those eligibility criteria, typically in Toronto it's a respiratory therapist, or at least in the sites considered in the study, the respiratory therapist will perform the spontaneous breathing trial. And during and after the trial, they will be documenting any presence of respiratory distress, any desaturations lasting for more than two minutes that are deeper or below 88%, and any major fluctuations in heart rate or mean arterial pressure in either the upward or downward direction by more than 25%. In patients who do not have any of these criteria, then the SPT is considered a success, and what happens thereafter I will get into on a subsequent slide. So, again, before proceeding, I wanted to get a sense of the following. So, among patients with a successful spontaneous breathing trial, what proportion are also successfully extubated without the need for reintubation? Is it 40% to 50%, 50% to 60%, 60% to 70%, 80% to 90%, or 90% to 100%? We'll give a moment for responses to come through. Okay, so, 43% of participants have indicated C, 43% D, and 14% have indicated B. So, the literature would suggest that between 80 to 90% of individuals are successfully extubated without the need for reintubation among those that have passed a spontaneous breathing trial. So, the correct answer here would be D. And interestingly, this makes the spontaneous breathing trial the gold standard test, so to speak, for determining a patient's ability to breathe unsupported by the ventilator. It outperforms clinical gestalt. This has been shown in multiple studies. And interestingly, also outperforms a patient's own sense of whether they're ready to be liberated or not. So, correct answer to this question is D. So, moving on, this schema outlines the extubation process as it occurs in the sites included in the study. So, patients are identified as eligible for a spontaneous breathing trial. Again, typically, it's by the respiratory therapist. The SPT is performed with a maximum of 8 centimetres of water of pressure support and a maximum of 5 centimetres of water of positive end expiratory pressure. In fact, a lot of the sites are doing SPTs on 010. And at the end or the conclusion of the spontaneous breathing trial, which is typically lasting about 30 minutes across the sites, the SPT is considered a success or a fail, again, depending on the criteria outlined on the previous slide. Where it is a fail, the SPT is discontinued and prior ventilatory settings are resumed with careful optimisation of the patient's condition to hopefully allow weaning to continue. When it's successful, the patient is transitioned onto minimal ventilatory support. And the primary care team discusses the decision to extubate with interdisciplinary input. The respiratory therapist here has a major role in determining the safety and appropriateness of extubation for that patient. The final decision is up to the attending critical care physician. But of course, this is almost always an interdisciplinary decision in large part. So, in terms of the methods to describe them a bit, we built a cohort of mechanically ventilated adult patients using the intensive care observational registry or I-Corps for short. I-Corps includes data collected from eight intensive care units across six hospitals in Toronto, Ontario, Canada. And the data captured spans between the years 2014 and 2019. I-Corps has furnished data for a lot of studies over the past couple of years. And in Toronto, we're very proud of this database. It's a very rich repository of useful information, especially when it comes to mechanical ventilation. The inclusion criteria for the study were all adult patients aged 18 years or older with a first successful spontaneous breathing trial within the first 28 days of initial ICU admission. So, if the patient had been readmitted, only the first admission was considered. Our outcomes, the primary outcome was composed of three mutually exclusive levels, with those being withholding extubation, outcome one, successful extubation, outcome two, and extubation failure, outcome three. To think about these graphically, imagine you have a patient with a successful spontaneous breathing trial. Thereafter, one of two things can happen. Either they're promptly extubated, again, using the definitions previously mentioned, or extubation is withheld. And in that latter case, that would be outcome one. Of the group that is promptly extubated, again, one of two things can happen. Either the extubation is successful, which would constitute outcome two, or it ends in failure, which would constitute outcome three. It's important, some of you may be looking at the schema and realising that of these patients who are not promptly extubated, some may actually go on to have extubation success. But again, I want to emphasise that those would not be considered in addition to those on this portion of the graph here, who were successfully extubated after a prompt recognition, because as I will explain in the modelling strategy, outcomes must be mutually exclusive and exhaustive. So, if the patient had had an extubation performed without delay, and it was successful here, again, combining those with this outcome would have led to overlap with this red box here. So, again, this was not done. Hopefully, that portion will become a bit clearer when I describe the modelling strategy. So, for context, we also assessed additional outcomes, including the length of stay in the ICU, tracheostomy insertion during the index admission, and all-cause mortality within 24 hours of extubation. So, in terms of the analytic plan, we fitted a multivariable multinomial logistic regression model, specifically to identify factors associated with withholding extubation and extubation failure. I will describe this in detail on the next slide. But before doing that, we also performed multiple sensitivity analyses to assess the robustness of these findings, excluding patients that we thought might foreseeably behave differently, or that clinicians might approach the calculus of extubation decision-making differently for. And those populations included those with traumatic brain injury, acute neurologic conditions, acute cardiovascular disease, and so forth. So, let's spend a brief minute talking about multinomial logistic regression, because this is a modelling strategy that may not be familiar to everyone. This is a form of regression analysis that permits two or more categories or levels of the outcome variable to be considered simultaneously. If two outcome categories are being considered, then that's basic binomial logistic regression. It's really where you start getting into three or more categories, where it becomes multinomial. This is valuable, especially with the study setup I've described for us, because it enables associations to be drawn simultaneously between multiple independent variables in each level of the outcome variable. Similar to binomial logistic regression, the outcome variable itself is categorical, whereas the independent variables, they can be dichotomous, they can be categorical, they can be continuous. There are multiple criteria that must be satisfied before one undertakes this analysis, important ones being that it assumes independence among the dependent variables. And again, belonging to categories must be mutually exclusive, so there can be no overlap in the Venn diagram, so to speak, in terms of where the patient or outcome occurs. If that's happening, then again, you cannot be using this modelling method. Additional things such as multicollinearity, linear relationships between outcomes and predictors, all of those pre-assumptions must also be checked for before applying this modelling strategy. I'm happy to speak more in the question and answer period if questions do come up on the modelling strategy, but for now, I want to spend a brief amount of time just talking about the results here. So overall, we included 9,910 patients in the overall cohort. The mean age was approximately 60 years, and 38.8% of those patients were female. Of the patients included, 62.2% were extubated promptly, so within 24 hours of a first successful SBT, and 37.8% were extubated beyond 24 hours or not extubated at all. Among patients who were promptly extubated, so that's this first group here, 12.9% failed the extubation. This flowchart describes where and how patients were excluded and for what reasons. So overall, there were 16,297 adult patients captured within I-Corps between 2014 and 2019. And after applying various exclusions listed here, the majority of which related to lack of documentation on spontaneous breathing trials or lack of documentation around whether it was successful or not. Some patients were excluded because they already came in with a tracheostomy. Again, after applying those exclusions, we were left with 9,910 patients in the overall cohort. So of those 9,000 unchanged patients, 6,167 were extubated promptly, 3,743 with a delay. And that right portion constitutes outcome one withheld extubation. Again, of this group, 797 patients were re-intubated within 48 hours of extubation, which leads us to our failed extubation proportion of 12.9%. Again, one of our pre-specified outcomes. And successful extubation was observed within this group in 87.1% of cases. So this is the table one in the study, and I'll walk you through it briefly here. Focusing first on a patient's demographic characteristics, you can see that the average age was about 60 years, and that prior cardiovascular disease was one of the more common comorbidities seen. The next portion here goes through the illness criteria, or excuse me, the illness characteristics on admission. You'll see that the Apache three score outlined there, and the usual breakdown, or a relatively common breakdown of admission indications, such as trauma, respiratory failure, cardiovascular disease. Important also to note that about 50% of these patients came from the OR as a postoperative admission. We also felt it clinically meaningful to look at characteristics on the spontaneous breathing trial day itself. A lot happens on that particular day that may guide a clinician's decision to extubate or not. And we were interested in looking at whether any sedative use, opioid use, antipsychotic use, renal replacement therapy use on SBT day, and other variables informed the decision making in any way. So I'm going to walk you hopefully as slowly and clearly as possible through the following tables, which encapsulate most of the information presented, or the results presented with our study. So we'll focus first on the association among individual factors with our first outcome of withheld extubation. Wherever there is a red arrow, that is meant to signify an increased association, and a green arrow signifies reduced association. So again, focusing on withheld extubations, you can see that age and female sex are both significantly associated with withheld extubation, whereas none of the other demographic factors are. Looking now at illness characteristics on ICU admission, you can see that a higher illness severity as quantified by the APACHE-3 score is associated significantly with withheld extubation, as is a diagnosis of acute neurologic condition with one of the higher odds ratios from this study of 2.06. And now looking in the opposite direction, reduced association with withheld extubation was seen with active malignancy, prior cardiovascular disease, a higher initial GCS, higher hematocrit, and a diagnosis of acute cardiovascular condition, in addition, finally, to elective and emergency surgery. So I partitioned this table into two slides. This is the second half of those slides. We're continuing again with red, meaning increased association, and green, meaning reduced association. We see that time to SBT is significantly associated with withheld extubation, with an odds of 1.15. A higher SAS or agitation score on SBT day is associated with withheld extubation, any sedation use, any opioid use, isolation precautions on SBT day, and also the use of renal replacement therapy day on day of SBT. Reduced association, on the other hand, is seen with mobilization on the day prior to SBT. So same table, but now looking at the second outcome of failed extubation. Again, red, meaning increased association, and green, reduced association. And of the demographic factors, you'll see that none demonstrate a significant association with this particular outcome. When it comes to illness characteristics, a higher Apache 3 score demonstrates significance, higher GCS on admission, a diagnosis of acute hypoxemic respiratory failure. This one's important, and I'll speak more to this in the discussion, as well as a diagnosis of acute cardiovascular disease and acute neurologic condition. On the other hand, reduced association is seen with prior cardiovascular disease, higher hematocrit, and elective surgery. So the final stretch of these tables, again, continuing in the same way as we have been before, looking at SBT criteria, time to SBT is associated, perhaps not surprisingly, the higher length of time it takes to get an SBT. There's an association with increased risk of failure, SAS that's higher on SBT day, any sedation use, and isolation precautions on SBT day. All of these are significantly associated. Let's zone in on the orange factors. I've spent a lot of time hopefully convincing you that this is really where the interesting aspect of the study is. And these are where the directions of association, again, to remind you, go in opposite directions. So focusing on the bottom right corner of this quadrant, we noted a higher association with withheld extubation, but not a higher association with extubation failure when it came to increasing age, female sex, and the use of renal replacement therapy on SBT day. I'll get into what the possible implications of this could be in the discussion. Similarly, a higher association with extubation failure was seen, but no association with withheld extubation for the conditions listed here, those including acute hypoxemic respiratory failure, acute cardiovascular condition, and a higher initial GCS. For context, again, we looked at additional outcomes. This table in particular comes from the study supplement. And you will note that it's divided by patients who were extubated with a delay versus those who were extubated promptly, delay on the left, promptly on the right. And you can see at a glance that ICU length of stay is higher when extubation is delayed compared when extubation is delayed compared to when it's done promptly. This is concordant with lots of other studies. Proportion of patients receiving tracheostomy is also higher at 13.6% when extubation was delayed compared to when it happened quickly. On the other hand, death within 24 hours of SBT was similar at 1.1% and 1.2% respectively. For the sake of brevity, I've left out many of the sensitivity analyses from this slide deck. I'd invite those of you who are interested in looking at the numerical outcomes to look in particular at table E5 in the supplement. To cut to the chase, there were no substantial changes in the strength or direction of associations, excluding each of the pre-specified patient of the pre-specified patient groups. The odds ratios for individual variables changed marginally, but not, again, meaningfully in either direction. So our sensitivity analyses, again, confirmed what the main analysis demonstrated. So what does this actually mean and what are the takeaways? Well, I'll start by summarizing what we found first, which is that in this cohort study of almost 10,000 patients in eight ICUs in Toronto, withholding extubation was relatively common, occurring in 37.8% of cases, and certain factors had an expectant-concordant association between the two outcomes, whereas several factors showed a discordant association. This has the potential to guide decision-making in important ways. To help understand that a bit, I want to refer to this figure from our study, which basically demonstrates that for the first three factors, there is, again, a significant association in the same direction for APACHE score time test PT and acute neurologic conditions with respect to withholding extubation and extubation failure. Clinicians, when these conditions are present, might be reassured that their clinical gestalt, that the patient is just maybe not quite ready for extubation, needs a bit more time, seems to be borne out by our results. And perhaps there's appropriate recognition of these patients and appropriate time devoted to their optimization. So clinicians can, again, feel reasonably reassured that when these conditions are present, the strategies they're already perhaps subconsciously employing to guide their decision making are logical and robust. For additional factors, for example, those listed in the middle text here, there's no association with withholding extubation, or excuse me, there's an association with withholding extubation, but no association with failed extubation. So perhaps these might not play a role in increasing the risk of failure, and perhaps they might help to reassure clinicians. And finally, for the last set of listed factors, there's no association or a reduced association of withholding extubation, again, with cardiovascular conditions that are acute, initial GCS, acute respiratory failure, but there is an association with failed extubation. And again, in the right context, perhaps they may alert clinicians to factors that they're not taking into account or maybe marginalizing in their decision calculus. Of course, there are limitations here. This is observational data. It's prospectively collected, but it's a retrospective study. They're subject to residual confounding. We can discuss at length how these implications might be modified by those limitations. But I think the general take-home principle that, for some factors, associations are concordant, for others, they're discordant, is a valuable bottom-line message to try to emphasize. And I think, for myself and our study co-authors, that was, again, the message we were trying to emphasize. In terms of patient group-specific observations, we noted in our cohort that acute neurologic conditions were associated with both withheld extubations and extubation failure. This is consistent with other studies. As was mentioned at the beginning of this podcast, I have a personal interest in neurocritical care and practices in mechanical ventilation in this population. And a lot of my ongoing work, at least at the preliminary level, is confirming what this current study found, which is always nice when results among studies are in concordance with each other. We also noticed that patients with active malignancy had a lower likelihood of withheld extubation compared to other conditions we looked at. Perhaps, again, speculating as to why this might be. Clinicians are more cognizant of the fragile nature of these patients and don't want to expose them to the risks of ongoing mechanical ventilation. Perhaps there may be more use of time-limited trials of mechanical ventilation, again, in this fragile population that we could speculate on end. But I suppose those are two important factors that may explain this finding. I've alluded to some of these limitations already. But to discuss briefly others that haven't come up, of course, as we all know, clinically, there are factors other than a successful SBT that often drive the decision to extubate. And I would argue that sometimes the decision really comes down to these logistic factors. We were at the mercy of the data available to us and could not account for many of these and other relevant factors in our analysis. But suffice it to say, they do exist. And clinically, I'm sure we can all remember times where an extubation was delayed because it was getting late in the day, or the patient had to go for an MRI, and we all felt safer doing it when they came back or waiting until the next morning. Interestingly, a lot of other studies that have been published in this space have found that logistic factors often drive the decision to withhold extubation in patients with a successful SBT. So again, the decision to extubate is not exclusively dictated by a successful SBT. Life would be too simple if that were the case. We, similarly, were not able to account for cough strength, secretion burden, or estimated risk of post-extubation stridor. Again, with my interest in the neural population, there seems to be important signals for each of these things with respect to final extubation outcome. This data is not captured within the database, but would be interesting to look at in the future. And weaning across all sites, as is important to mention, was carried out by respiratory therapists and by weaning protocols. And so, our results may not apply in settings where RT availability is limited or where protocols are not used to guide extubation decisions. And finally, as I've already alluded to, this is observational research. The associations are by no means causal, rather hypothesis generating. So, to summarize what I think are the main points here, many patients deemed ready for extubation are not liberated within 24 hours. Some characteristics associated with withholding extubation are not associated with a higher risk of extubation failure. And the converse is also true. This may inform decision-making because, in the following ways, again, it may support decisions that clinicians are already making, perhaps subconsciously, or it may cue clinicians to consider certain factors they might be, again, marginalizing or not paying much attention to before they extubate a patient. And with that, I'd like to conclude this presentation, but also want to briefly acknowledge the Acute and Intensive Care Outcomes Research Network at the Sunnybrook Hospital for assistance with the development of the study. Again, a lot of the methodology employed herein was made more robust, I think, with high-level input from multiple very smart people, and we are all, as co-authors, indebted to them. And again, a final major shout-out to all co-authors who made this study possible. These are the references cited, and I'm happy to take any questions at this time. Thank you for listening. Thank you very much for the fantastic presentation and for the really interesting data that we have seen. There are a couple of questions coming, and the first one is going back to, probably, to the first poll. What do you consider as an acceptable level of reintubation rate? You said that about 80 to 90% of the extubations are successful. Thank you for that question. That's very important. I guess a knee-jerk answer to that might be the lowest rate is always the best, but I would say that's probably the wrong answer. And the reason is, if your rate of reintubation gets too low, for example, if it's in the single digits, you're probably waiting too long to perform the extubation, and you're probably subjecting a reasonably large proportion of those patients to unnecessary mechanical ventilation with all of its attendant risks. The opposite is also, I would say, true. If the extubation rates are getting too high, then maybe your approach is a bit too aggressive. Some of those patients could be optimized a bit more carefully. And so, as with so many things we do in critical care, it's about finding the right balance and using risks and benefits on either side to guide your strategy. In terms of what proportion might be considered acceptable, it's hard to say, but I would start by saying that probably around 10 to 20% is a reasonable place. And again, if it's getting any lower than that or much lower than that, certainly if it's in the low single digits, you're probably waiting too long to extubate patients. And if it's getting much higher, you're probably doing it too aggressively. So, somewhere around 10 to 20%, I think is reasonable. Also important, again, with the caveat that not all patient populations behave in the same way. We all practice in very different environments. Some of us, you know, post-surgical patients, for example, who might be otherwise young and healthy, trauma centers and so forth, these rates may all look very different. We're in exclusively medical ICUs where patients might be older, frailer, lots of comorbidities. Again, you might accept a slightly higher rate of reintubation, recognizing that you can't keep people intubated forever. So, I would say probably 10 to 20% as a simple answer, but it's, of course, always more nuanced than that. And I think this is what I see in the UK practice as well. So, it's reassuring that we are. I saw slides when you were talking about the limitations and the possible logistical limitations. Do you have any data on when those SBTs were performed during the day? And if there is a data, is there any merit in looking at whether that has got effect on the extubation or pre-forwarding extubation? Thank you for that question. I think it's very important. I would say that, to my knowledge, I don't think the timing of when the spontaneous breathing trial was performed is recorded in the case report form that is used to guide the data extraction within I-Corps. From personal experience, having recently been a trainee at the five academic sites at the University of Toronto, a lot of these SBTs are performed early in the day by respiratory therapists on patients who meet waiting readiness criteria. And that information, in a large proportion of cases, is already available to the rounding team by the time rounds take place. It would be nice to see if later SBTs might perhaps predispose to extubation with a delay. I would probably chalk that up to a logistic factor that might be the eventual decider on when extubation actually happens. It would be, again, very interesting to look at. Unfortunately, I don't think we have the data available, at least within this dataset, to tease out the nuances of that specific question. Thank you very much. I'm quite envious of that setup that you have the SBT performed by the time the rounds are started. I've certainly heard that one before from colleagues at other practice sites. I think we're very lucky to have fantastic respiratory therapists who are very, very vigilant about identifying patients early on for us. And in the majority of cases, again, just from personal experience, we make these decisions routinely before noon or even before rounds or during our pre-rounds. And this is all kind of tidied up and nicely taken care of before day's end. Not always the case, of course. Logistic factors, patients away for scans and procedures will take priority sometimes, but whenever it's possible, we try to do it early. Thank you. Do you have any data on what kind of respiratory support the patients had after the extubation, whether that was delayed or not delayed? Whether you have used simple face masks or high-flow nasal oxygen or CPAP? Is there any information available? I-Corps does capture data on use of non-invasive ventilation and high-flow nasal cannulae after extubation. We did not look at the proportion of the use of those things in this particular study. We didn't look at prophylactic, again, use of those modalities in an effort to reduce re-intubation risk. Again, there's lots of high-quality data suggesting that extubation in the right patient populations to one or more of those modalities in combination may reduce re-intubation risk. Very, very important outcome, but again, was not looked at in the study. I don't think I can speak to what proportion of the patients we included had been extubated to high-flow or non-invasive, but I'm happy to look into that information if available and potentially disseminate it in the future. Thank you. Can we be certain that the decision-making for those with delayed and failed extubation is correct? Is it possible that the delayed extubation is also a potential cause in leading to failure? That is a fantastic question. I certainly think it is possible, right? With every delay that happens every day that extubation is delayed, patients are receiving more sedation, they're potentially being immobilized less, they're accumulating risk factors that may ultimately predispose to an eventual failed extubation. So, I certainly think those two outcomes may go hand in hand, and one may dictate or inform the risk of the other. That is a wonderful observation and certainly one that probably the majority of studies looking into this would also support. Thank you. And one of the final questions is mine. I do look at and I work on frailty in intensive care. And my question is that, do you have any kind of frailty data or comorbidity data beyond the normal ICU scores? And if you do have, is there any concordance or discordance between these factors and failed extubation or delayed extubation? Also, a very important question. To my knowledge, again, I don't think I-Corps captures baseline frailty within its CRF, but there are surrogates that, again, you can potentially use to infer frailty. There is data available on patients' motor strength within the ICU, their ability to participate with physiotherapy and occupational therapy. Again, baseline frailty before ICU admission, though, I'm not sure if that exists. Certainly, we didn't look at it within the context of the current study. I will say, though, that if memory serves correctly, there was a study within the past couple of years in intensive care medicine by some Canadian colleagues in Ottawa that suggested that duration of mechanical ventilation is higher, proportion of patients receiving tracheostomy is higher, mortality is higher among patients that had higher clinical frailty scores. Again, that is not data that comes from this study, just to be very clear about that. But certainly, Tomas, you as an expert in this field could probably speak more to this than I could. Not something, again, that we looked at within this particular study. Thank you very much, and thanks for the presentation. This concludes our Q&A session. Thank you very much, Dr. Teran, for presenting this important data, and thank you for the audience for attending. Again, everyone who joined us for today's webcast will receive a follow-up email that will include an evaluation, and I would like to ask you to take five minutes to complete it. Your feedback is really much appreciated. And on a final note, please join us for our next Journal Club Critical Care Medicine on Thursday, January 27th. And this concludes our presentation today. Thank you, and goodbye.

study

factors

withholding extubation

extubation failure

critically ill patients

Toronto

results

extubation decision-making