SCCM Resource Library-Star Research Presentations: Pediatric Pulmonary

Star Research Presentations: Pediatric Pulmonary - 2020

Video Transcription

Good afternoon, everybody. I am not Dr. Rezwan. I am Dr. Erin Holfish. I'm one of the pediatric intensivists from Stony Brook Children's Hospital.  And I'll be presenting today are the effect of protocolization of status asthmaticus and treatment of the pediatric intensive care unit. I would like to thank Dr. Rezwan, who  is sitting right over there, as one of our main contributors to this project. She'll be joining our ranks as one of the pediatric critical care fellows  at Montefiore in a couple of months. So I'm sure she's looking forward to what we're going to be doing as pediatric intensivists. This is the rest of my team that's  helped out with our research as well. I have no financial disclosures. So as we all know, especially at this time of year in the wintertime and any time the season changed,  status asthmaticus or severe asthma exacerbations is one of the most common causes of admission to the pediatric intensive care unit.  And daily costs of a PICU stay plus an entire hospital stay can be in the thousands to hundreds of thousands of dollars. This can result of significant time off  of work, of school, of daycare, and significant time for the families that are having to pay for all of the time spent in the hospital. And children reported in these young years  can miss about 10 and 1 half million days of school of asthma and have a significant reduction in their productivity. There is more than 25 million people that are affected,  and numbers are increasing yearly, leading to increase in pediatric hospitalizations. And there's been a higher increased lifetime prevalence of asthma.  1 in 11 children have asthma, and more than 1 in 5 go to the ED per year for asthma-related complications. There have been many studies out there in the general inpatient care areas  that have talked about albuterol weaning protocols that have shown a reduction in variability of practice, time to interventions, general hospital length of stay,  as well as costs and reduction in beta agonist use without increasing readmission rates. However, those have not yet trickled into the pediatric intensive care world  where we have a larger variability of practice, which can lead to a larger variability in our admissions and a large variability in our PICU length of stay.  Prior to enacting our new protocol, we did not have any formalized criteria or standard scoring system for admissions to our pediatric intensive care unit.  If the ED called us and said, we need a kid that needs continuous albuterol, as in our unit, continuous albuterol is the only place that albuterol  is administered, they had to come to the PICU. And it was provider-dependent on the dosing that was started, as well as what they were weaned to and how quickly they were weaned.  There were asthma assessments that were made. It was about every four hours. However, any decision to change dosing outside of adverse reactions to the kids were customarily  performed at the end of the current dosage, which could be a 10-hour length of time between the time the albuterol started to when it was finished before we changed  any of our titrations. At the same time, any transition off of the continuous was made to nebulized intermittent dosing, which was usually about 2.5 milligrams.  And the transfer out of the PICU was done once the child was receiving albuterol only every two hours and moving closer to using it every three hours.  So we hypothesized that a protocol-driven approach to the treatment of status asthmaticus would decrease our PICU length of stay, decrease our hospital length of stay,  as well as decrease cumulative dose of albuterol in the PICU. We also hypothesized that it would lead to a decreased administration of adjuvant medication, as well as  a decrease in advanced respiratory support needed. So as you can see, we had a major undertaking in the workgroup that went about revising our protocol,  all the way from our pediatric pulmonology group and nurse practitioners to people like myself as the intensivist, all the way down to our IT department  that helped us build our new models within our electronic medical system, so everybody that was involved. We spent an entire year just revising our protocol  and using a, until we found one and adapted our previously validated pediatric asthma scoring for our institution. And this was an RT and RN-driven protocol.  And then we spent several months with a multidisciplinary training committee until we had more than 95% of everybody that would be using this protocol, intensivists,  hospitalists, residents, RT, nursing, all of them were trained on this using our multi-model and case scenarios. And then our goal live date was December 1st of 2017,  so right in time for respiratory season, which was a really good time to switch everything. So again, this was our pediatric asthma score that we used.  We did age-defined norms for accepted respiratory rates, oxygen saturations, auscultation, retractions, and as well as work of breathing.  And based on all of that, they were given a composite score and mild, moderate, severe, or very severe. And I know this slide is small and busy,  but it is to illustrate that we have different pathways based on the number that was assigned and weight-based pathways for our very severe continuous  albuterol with the dosing that was needed, as well as any adjuvant medications. And every hour, there was an assessment made by our respiratory therapist or our bedside  RN that was done on alternating hours. This is a sample progression through our pathway, starting through the severe, very severe category,  and illustrates how we wound up transitioning them once they reached the moderate pathway to intermittent MDIs and then going through until they were able to be discharged.  So this is an IRB retrospective study that was done from December of 2016 to December of 2018. We have a 12-bed ICU staffed 24-7 by pediatric intensivists, second and third year  pediatric or ED residents, bedside nursing, and shared respiratory therapists with the general inpatient unit. Patients had to be between 2 and 18 years of age  and their primary diagnosis of status as maticus or transfer and admitted to our PICU or transferred to our ICU within four hours admission to the floor.  Generalized exclusion criteria are those that are admitted with any comorbid conditions or neuromuscular disorders, immune deficiencies, that would confound the ICU.  So in terms of our results, our demographics, we identified 39 in the pre-intervention and 44 in the post-intervention. Besides having a little bit more males  in the post-intervention group and a little bit more Caucasians in the post-intervention group, we saw no significant difference between the two groups,  including their asthma history, their matched pediatric asthma scoring, or their history of tobacco exposure.  What we did find that was significant was a decrease in length of stay, both in the pediatric ICU length of stay and their hospital length of stay  for their general between the two groups. We did find there was a decrease in the amount of albuterol per kilo that was given to the intervention group, but this  wasn't found to be statistically significant. And there wasn't a difference between the adjuvant respiratory support required between the two groups.  So what does this mean? Protocolization of status asthmaticus treatment in the PICU led to a reduction in PICU length of stay by almost 17 hours and almost 24-hour reduction  in length of stay for their entire hospitalization. With a relative decrease in cumulative albuterol dosing during their PICU stay, without any difference in need  for advanced respiratory support or adjuvant medication. Again, this was a single set or retrospective study. And we also, at the same time, did an entire global revision  to our asthma protocol, including the treatment that the child received in the emergency department. And that could have confounded some of the data that we did analyze.  We're in the process right now of going through and also looking at adherence to protocol with our deviation notes. So I do not have that to discuss at this time.  So in conclusion, creation of a protocolized de-escalation pathway for status asthmaticus in the PICU is feasible and able to reduce PICU and hospital length of stay.  A shared scoring system between the RN and RT on alternating hours is effective without overtasking one department. This also allows for clear goals and numerical assessments  of how patients are progressing with their illness and can lead to estimated length of remaining time needed in the ICU, as well as better resource utilization.  However, further multi-centered, high-powered studies are needed. Thank you. Thanks so much for the great talk. Any questions from the audience?  Nice study, Aaron. This is Bob Parker. Length of stay. Was the length of stay calculated using the time that the person actually left the unit or left the hospital  or from the time that the order was written to transfer out of the unit and to the discharge order written? The length of stay was calculated  from the times in the electronic medical record, so the time that the transfer order was placed in terms of when the admission order was placed in the ED  to the transfer order being placed in the pediatric intensive care unit, and then from their entire length of stay until the discharge order was placed either in the PICU  or the outpatient. Were you able to determine whether there was a significant difference from the time of the decision to transfer to the time that the order was placed in the EMR?  I know this was retrospective, so that might not be available. I'm sorry. So you're wondering, is there a difference in the time to the order being?  Yes, so sometimes, so it's 10 o'clock in the morning. I said, this person's good to go. And the unit's busy. You don't get around to writing the order until 1 o'clock  or 3 o'clock or whenever. Same thing, probably more variability on discharge. But when you're looking at length of stay as one of your outcomes, the more precise  you can time when the decision was made, that may actually be a more meaningful. Right, and so as it is a retrospective study,  I can't go back and say when was the decision made versus when the order was placed. But maybe the decision was made more quickly with all of the data that was available  as we saw a reduction in the length of stay overall with both. Gotcha, thanks.  Thank you very much. Thank you.  All right, this brings us to our next presenter. It's Dr. Anthony Sochei from Johns Hopkins All Children's Hospital. He'll be presenting his work on stress ulcer prophylaxis  for children admitted to the ICU for status asthmaticus. Anthony, thank you. Thank you, everybody, for being here. And then also, SCCM, for allowing us to present our data.  I'm Anthony Sochei. I'm one of the intensivists just down the I-4 corridor in St. Petersburg, Florida. And I'm going to present some data about our use of stress ulcer prophylaxis  in this specific population of children rather than the global PICU population, which a lot of our recent literature has been focused on, or not. Let me see. Indeed.  I can do it without the slides, too.  So.  I don't have any disclosures. If I owned any stock in ranitidine or Lansoprazole, I'm pretty sure they would take it away after I was done.  If you haven't been looking at the schedule for SCCM already, then there actually was a pro and con debate this morning on stress ulcer prophylaxis ICU study  and implementation of it and whether it's appropriate or not. And there have been several publications, most recently in Pete's Critical Care in 2019, kind  of assessing from a Canadian group what different types of prophylaxis are used in what populations. And was there any risk reduction in bleeding? And those studies are interesting,  but also fail to do one thing, which is focus down on a homogeneous population rather than every single ICU patient that we take care of, which is highly variable.  The disease severity is very different. And potentially, their mucosal cytoprotection at the level of the gut is also different in those different disease states.  And as a result of corticosteroid exposure, children with asthma are routinely placed on stress ulcer prophylaxis. And this is a process that is happening  at every children's hospital across the country with some variance. And the concept here is that you have some type of hypoxemia or acidosis in combination with a low gastric pH  and impaired cytoprotection. And then you give them some exposure to a glucocorticoid. And then this results in gastritis or ulceration at some place that has a vessel.  And then you have bleeding. And if that bleeding is bad enough, then you get hypotensive. You need boluses, vasoactives, and potentially a blood transfusion.  If those things happen, then we call that clinically important bleeding in the literature. This is so you can squint really hard. But this is a forest plot from meta-analysis,  one of many meta-analyses. This is specific to pediatrics looking at various H2 blockers compared to placebo versus no treatment.  But on the other side is also a funnel plot or a cone plot that looks at various different things that are used to treat or prevent gastritis and stress ulcer bleeding.  And the bottom line is that, yes, if you put them all together on the bottom, H2 blockers may veer from crossing over to the insignificant area. And if you look at the cone plot,  almost nothing actually becomes significant there is out of the cone. And therefore, it's very hard to suggest that we are applying this therapy  with any actual clinical benefit to the child. And yet we apply it in large numbers. And so my research group decided, hey, we would take a look at that.  And we would start by just looking back at our own data. And so we took two years of asthma data. And I'll step back one moment and say, in 2019,  there was a recent publication, again, in Pediatric Critical Care Medicine, that looked at who was applying these therapies and to what populations.  And it's more common in those who don't receive enteral nutrition and kids who get systemic corticosteroid exposure or NSAIDs. And there was no difference in the percent  in GI bleeding outcomes. And some often cited literature in support of stress ulcer prophylaxis are summarized in some meta-analyses from the early 2000s  that actually assessed neonates in post-natal steroid exposure, which is not a 12-year-old asthmatic, but on steroids. And so we actually published this publication,  retrospective data for two years, that just looked at every single asthmatic we admitted, whether or not they received stress ulcer prophylaxis  and did any of them have any bleeding events or gastritis. And as it turns out, not a single one of them had gastrointestinal bleeding.  Two of them had gastritis that's self-resolved, that didn't require any type of therapy, including an H2 blocker or a BPI. And this was incredibly hard to publish. Why?  Because it was a negative study. And because of that, it's also somewhat underpowered, since it's such a rare event. And so we decided that's a gap that we  can try to fill by using our registry data that's available, that many of our institutions pay money to be part of. And so we looked through at the FIS registry for 10 years,  and children 5 to 18 years of age admitted to the PICU without some other reason to be on a PPRH2 blocker, like eosinophilic esophagitis or GERD.  And they didn't have some of their major medical comorbidity that makes them more than just an asthmatic. They had to have the presence of stress ulcer prophylaxis  on day of service number one. The primary outcome was either GI bleeding defined as the code. But as you know, we don't always code for the things we take care of.  And so we also look for surrogate markers, including if, unless you had a chylothorax, if you're on an octreotide infusion, it's very likely that you may have a GI bleed.  If you were given sucralfate, blood transfusions during your hospital stay. We also looked at the incidence of gastritis and also  adverse events associated with stress ulcer prophylaxis, including C. diff, associated diarrhea, thrombocytopenia, et cetera.  I doubt any 17-year-old's got necrotizing enterocolitis, but we looked for it anyway.  And when we collected those data, we initially had 25,000 children identified from 52 children's hospitals that reported these data actually pretty reasonably well.  And 3,000 of them were excluded for the criteria we just mentioned. And 7,000 of them got stress ulcer prophylaxis. So as it turns out, a majority of the patients  with status asthmatic actually do not receive stress ulcer prophylaxis in these institutions. If this graph goes from least percentage of prescribing stress ulcer prophylaxis  to the other end where it's the highest percentage, you can end the bar graph show the number of total cases in that institution during that time period.  On number five, if you look at, has the highest number of volume but has the least amount of prescribing. But there are several large volume centers to the right  that actually prescribe stress ulcer prophylaxis with high frequency. But there's a ton of variability.  We also looked by year. And you can see that this is, again, blue being no stress ulcer prophylaxis, pinkish being stress ulcer prophylaxis,  that it's increased from the 20 percentage range in the 2010 period to almost 40% in the last three years.  Most of the time we're using H2 blockers, which is consistent with literature that suggests that H2 blockers may be more safe than PPIs. Although, again, depending on which group  of population you look at. And from 2016 on, we actually have NHLBI classification data in our billing. And so at least there was no major differences  in the severity of illness when you use that index for these asthmatics.  One other way of classifying severity of illness in these asthmatics is what therapies they applied, we applied to them. And if we look at aminophilin,  it was more frequent in the stress ulcer prophylaxis group, or terbutaline was statistically not different but maybe a little bit greater.  The frequency of ventilation and non-invasive ventilation, interestingly, was on the opposite end. And we had a high percentage in this population  of intubation and mechanical and non-invasive ventilation, which is actually interesting because the reported rates are actually much lower than this, 7% or so.  Bottom line, here's our primary outcome. Exceedingly rare. Just like in our single center study, there were 12 children who met any criteria that we used to define bleeding  who had status asthmaticus from 5 to 17 years of age without any other comorbidities. And there were six kids in each group. And that difference was not observably  or statistically different.  The incidence of gastritis was also extremely rare. We didn't have anyone on octreotide. We didn't have anyone who had a blood transfusion. We didn't have anyone develop any  of the stress ulcer prophylaxis adverse events. The mortality rate was low, which is good. And the length of stay was about three days and no different between the two cohorts.  Drawing conclusions from registry data, you always need to be super careful. But with that said, in this specific poll,  when these patients, it's extremely rare for an asthmatic to have severe GI bleeding, regardless of the presence of stress ulcer prophylaxis.  And perhaps we should be looking for other risk factors. For example, an existing coagulopathy, which in our single center study, we had children with  that who we didn't apply stress ulcer prophylaxis for. Kids who had neurologic injury or trauma or the existence of a tube that's physically there  and irritating the bowel and the gastric lumen. Another thing just to mention briefly is that we admit children, for example,  for rheumatologic flares, where we give four to eight times of the dose of steroids and don't put them on stress ulcer prophylaxis. Cancer protocols have steroids as part of them.  We don't put them on stress ulcer prophylaxis. And another great example is IBD. I mean, a disease characterized by ulcers in the GI tract,  there is no recommendation from our GI colleagues about using stress ulcer prophylaxis in several studies that contraindicate its use for mucking  around with the flora inside the gut. So I think there's controversy on the use of this for a reason. I hope that these data maybe prompt some collaboration  and some prospective study and trial where we can randomize people into receiving and not receiving it and see if we have any events.  But our data would suggest it's so rare, we may not be able to detect that difference, even if we had many setters collaborating together. I appreciate your time.  And open it up for any questions, comments, and any feedback. Thank you very much. Thank you, Dr. Sochei. We're now open for questions.  Hello. Very nice talk. I do have some issue with your study, though, the way it was looked at. You mentioned the limitations of a retrospective study.  And you come to a lot of conclusions related to that without any real data. In other words, you say there's no gastritis, but you didn't do biopsies.  There's no evidence that there was biopsies performed. So I think that conclusion is overestimated. And you say sometimes severe bleeding, which is what I'm suggesting.  You have really no evidence of severe GI bleeding, but you don't actually have evidence of no GI bleeding. There's evidence to suggest that over 25% of anybody who goes  in all comers into the PICU are going to have some evidence of GI bleed. So what you're suggesting there is that you're bleeding in these specific patients is actually less  than the cohorts of all comers into the PICU, which might be true. I'm not saying no. But then again, you're not having a lot of evidence to prove or suggest your conclusions there.  So I think a better conclusion for your study, which I think has its value. Don't think I'm bashing on your study. But massive GI bleeding. And of course, I think you can correlate  that with the other pathology you were suggesting. So duration of high-dose steroids would also come into, you know, a factor here. And I think if you were to prolong  or expand your denominator there by including other cases where the duration of steroid use might change a little bit, at least your results. And I think that has real importance  because you're not limiting what you're finding specifically in the PICU patients because you can extrapolate that to the outpatient environment,  at least the hospitalized environment. So my recommendation, I guess, would be to take that into heart and see if you can expand on your population under study here.  Thank you very much. Yeah. When we do go to eventually, you know, write this up and submit it, it will obviously be slightly toned down as far as very careful consideration  for what conclusions you can be drawn. I agree with your sentiment that there are data on, if you have children and you scope every single one of them,  almost 70% of them have ulcers when they present to the ICU. And whether you have bleeding that is obvious or massive is less frequent. Macroscopic versus microscopic bleeding  versus clinically important or hemodynamically unstable bleeding. When we did our single center study and looked at those data,  their exposure to H2 blockers or PPIs were less than 24 hours. Why? Because most of them got switched to enteral nutrition.  And so it's very common, at least in our institution then, and it's something that's difficult to tease out using registry data or retrospective data,  is also the indications for when and how long we apply these therapies. The duration of steroids for most asthmatics is about a five-day course.  But I don't have that data about what they received as an outpatient, et cetera, to account for it. Precisely. So like I say, your database would be beneficial  to expand that and possibly include other pathology. Your limitation, your limiting it to asthma, I assume, is just for short course steroid use, which itself has value.  Don't think it doesn't. But again, like I say, you can extrapolate from whatever data you have to what we're doing in the private, excuse me, the outpatient environment,  but also to the hospital room outside of the ICU. Because like you know, the majority of these cases are put on therapy without going to the ICU, and it's just a more severe element  that you're pulling out. So I think by the way you've limited your population, you've really taken a lot of that value out. But then again, it's a short course of steroids.  I mean, we can talk in circles. I think a lot of the peers in critical care may feel that including a hospitalist patient who may not have the same severity of illness,  may not result in the same, quote, degree of stress to cause stress ulcers. I agree with you. It's whether or not you account for that variable  in the analysis up front, or you exclude them, and therefore I don't have to account for it. So I can't comment on hospitalist patients. I can't comment on outpatient children.  But I can comment on kids who are admitted to the PICU who met our inclusion and exclusion criteria. Right. And you're basing a lot of that on.  I'm so sorry, but I think we're going to have to cut short this and perhaps have the questions afterwards. We're running out of time, and we're getting to our next speaker.  All right. Thank you, guys. Thank you.  Thank you for a stimulating talk. Our next speaker will be Dr. Rowan from the Riley Hospital for Children speaking on Submission 11 Factors Associated  with Failure of Noninvasive Ventilation in Children After Stem Cell Transplant. Dr. Rowan. Thank you all for being here. I am going to talk to you today about noninvasive use  in children that have had a stem cell transplant. This study was funded in part by the St. Baldrick's Foundation. So there's mixed data about how useful noninvasive ventilation  is in the adult population. In 2001, the New England Journal published a promising study where they looked at early use of noninvasive compared to supplemental oxygen.  And they found that in this small cohort, there was a decreased rate of intubation and mortality. But when this was looked at later in a larger multicenter cohort  of almost 400 adult immunocompromised patients, there was no difference in the intubation rate or the mortality. There's no data like this in pediatrics.  And there's one fairly large retrospective cohort study that looked at all oncologic diagnosis in children and the use of noninvasive. And they found a failure rate of 25%.  But for our interest, this population didn't have very many stem cell transplants. And our subgroup of PLEASI, the stem cell transplant subgroup of PLEASI,  really became interested in this topic after we published a large cohort of invasively ventilated patients. And we found in the invasively ventilated patients,  41% had received a trial of noninvasive prior to getting intubated. And those that did had a twofold increase in the odds of PICU mortality.  And so the problem with this database was that we didn't have any patients that were successfully treated with noninvasive. So we couldn't make any strong conclusions.  So we sought to take a look at this. And we aimed to investigate the use of noninvasive ventilation in all children that had received an allogeneic stem cell transplant.  And also aiming to identify factors that were associated with failure. We had 11 pediatric centers that contributed data. We included children and young adults up to the age of 21.  They all had received allogeneic stem cell transplant. We defined noninvasive for this purpose as BiPAP or CPAP. We did not include high flow nasal cannula.  We defined failure as needing to be intubated or having an arrest prior to intubation. We excluded children that were treated with noninvasive only in a post-extubation course.  And you had to have a full code status. You can't have any code care limitations to be included in the study. Our cohort had 153 patients with a median age of 11.  Forty-six percent were female. And the median time post-transplant to noninvasive use was 46 days. While most of the patients were started on their noninvasive  in the PICU, a quarter, almost a quarter of them were started outside of the PICU. And this is predominantly on the transplant floor. The primary mode that was started was BiPAP.  And of our cohort, 63% failed their noninvasive ventilation. We also looked at transplant variables. And none of these were significant.  The variables that we looked at demographically or the transplant variables were associated with failure. So aiming to look at some early predictors of failure,  we examined the start of ventilation. And we looked at if they'd received a course of high flow, what their SATs and FIO2 were at the start.  And none of these variables were associated with NIV failure. But what we did find is that those who failed noninvasive spent a longer time on oxygen prior  to initiation of noninvasive ventilation. We looked at starting settings for the BiPAP or CPAP and cause of respiratory distress and none of those were associated with failure.  Not surprisingly, CVVH and vasoactive agent use was associated with failing your noninvasive. We then examined multiple time points with vital signs and respiratory parameters.  So we looked at BiPAP, we looked at the noninvasive settings. We looked at FIO2 saturations, blood pressure, heart rate and respiratory rate. And what came out of this analysis was  that tachypnea was consistently associated with failure of noninvasive at multiple time points. We focused on hour four because we wanted to aim  for a relatively early prediction model. And at this time point also gave us the most, the largest set of data points for all the patients.  And we can see that as your tachypnea worsened, your odds for respiratory rate, or your odds for intubation also increased. But we really wanted  to improve the clinical applicability of this model. And so since the age group was not different between our failure and our success,  and most of the children included in this cohort because it is a transplant cohort is school age, we focused on one cut point. And we chose 40 based on a sensitivity analysis  with an ROC curve. And as you can see, the kids that are displayed in red that failed have a higher proportion of patient,  have a higher proportion that were having a respiratory rate in the 40s and some up to 100. So we pulled respiratory rate over 40, vasoactive agents, age,  use of renal replacement therapy, and O2 use, supplemental oxygen use prior to noninvasive into a multivariate model. And when we pulled all these into a multivariate model,  we can see that tachypnea over 40 and vasoactive agent use were still significantly associated with odds of failing the noninvasive. And the area under the curve for this model  to predict NIV failure was good at .82. Not surprisingly, if you were successfully treated with noninvasive, you had a longer length of NIV use,  but also a shorter length of PICU stay.  When we look at survival, as you would expect, those that are successfully treated with noninvasive had a better PICU survival.  But what was interesting is that we see when we look all the way out to hospital discharge, their survival drops off even in the kids that were successfully treated  with noninvasive. The cardiac arrest data created a pretty big cause for concern when we pulled this out. So 14 patients out of the 153 had a cardiac arrest during  intubation or prior to intubation. When we looked at this, 11 of the children arrested during intubation. So 10% of the intubations  in this cohort were complicated by a cardiac arrest, which is an extremely high number. And then of those children that arrest, only 2 or 18% actually were able to be recovered  and survived to PICU discharge. There were three children that arrested before we ever even attempted to intubate them, and none of those children survived to PICU discharge.  So we weren't really, we were surprised to find this, and we started to look at some associations with cardiac arrest. And encouragingly, the same variables pulled  out of the kids with cardiac arrest. So we looked at just the kids that were intubated, and we looked at those who arrested and those who did not. And on univariate analysis,  we're seeing that there's more vasoactive use and more tachypnea in the kids that arrest. And also important, especially from a transplant perspective,  the kids that arrest are closer to their transplant, so likely not engrafted yet, more immunocompromised.  So from this data, the cardiac arrest data should make us all concerned, and really as a profession should make us consider if we need to have earlier intubation in these children.  And if we're going to continue to use noninvasive in this population, then objective prediction modeling at who was going to be successfully treated and help to aid  in timing of intubation would be really important. And it seems from this study that tachypnea and vasoactive agent use would be important variables  to include in a prediction model. I'd like to thank, I had a ton of collaboration with the 11 centers contributing. My collaborators are all listed up there.  And thank you for your attention.  Thank you. Thank you for a great presentation. Are there any questions?  So the question was about data and position, if they were like supine or upright. I don't have that data. We just thought I had enough to ask the question.  We weren't expecting to see such a dramatic difference from what I would consider a general PICU, cardiac arrest intubation.  Hey, Courtney. That was great. Do you have questions, do you have data on how long it took a patient to fail for the ones who failed?  I do, but I haven't pulled that into the analysis. When we look at the failure rate, the median time of length of non-invasive is .2 days, so it's not taking a ton of time.  I have to do the calculation in hours. Is that four hours? Yeah, it's roughly four or five hours. Yeah, roughly four hours.  It is a little challenging because there was a larger cohort started outside of the unit where the vital signs are not taken as frequently, which is why we didn't have a lot  of one-hour and two-hour data, I think.  The other thing is that in most other studies that have done this in undifferentiated PICU populations, they've also found heart rate and FiO2 at two or four hours to be predictive.  I don't know if you have that data or if that just fell out. It, on univariate analysis, was not associated in any way we looked at it,  and I think that that is because it is so abnormal all the time. Like, they were all abnormal.  Nice job, Courtney.  Heidi Flora, University of Michigan. Can you just clarify for me, these are patients who are on vasoactive while they're on NIV, yes?  Yeah, so we looked at vasoactive use on NIV up to the day of intubation. And what about that data related to CRRT? Are they on CRRT also before they're intubated?  So about half of the CRRT. There's not as many on CRRT are on, but there are a few that are put on later.  So if, are you able to also look at the combination of multi-organ system issue, number of organ systems involved, then predicting failure and or arrest?  So I don't have all the organ system failure, but that is my next study, I think. Because I really want to know that.  I will say, even when we pull CVVH out of the model, if we just pull only the tachypnea  and the vasoactive, the area under the curve is .78. It doesn't really change. Those are the, those seem to be the drivers of failure.  Thanks.  Next presentation, John McCluskey from Hopkins. Why did these patients arrest?  Were they hypox, profoundly hypoxic, difficulties with intubation,  drugs that were used for intubation? Did you look at all those parameters? Really important variables that we did not have the foresight to collect.  My suspicion, given when we look at where they were started at on their non-invasive, that there's probably some delay in care because there's an overwhelming proportion of arrest  of children that were started outside of the PICU.  But it's such a small number, it's hard to know what to say with it. Right. And again, that, you were saying to Nauda, it was just .2 days.  I mean, why would you even try, if they're that bad, why even, why go to negative? Why don't you just go ahead and intubate them right now because you're going to get in trouble?  I 100% agree. I think this is our old perception of don't intubate these kids because they all die. So, we try to not intubate them and then they die.  Right. And we work with oncologists who says when they go to the PICU and get intubated, the bone marrow transplant patients, they die.  But that's, you're right, but not all of them, okay, not all of them. I don't know if we are self-fulfilling the prophecy. Yes.  Great talk, Courtney. I just, to piggyback on what John was saying about collecting some data on the patients who did arrest, it's a small number.  And I'm sure your collaborators would be willing to go back and get some of those data points for you. Yeah, I think it'll probably be important to make a stronger paper.  Thank you for your talk.  Did you notice any difference if the patient was engrafted and had a good normal white cell count? Or what duration of post-transplant was that?  The median white blood cell count was not different between failure and not failure.  The only time the engraftment piece really came into play is the cardiac arrest because there is that big difference. But timing post-transplant was the same between the two groups.  And it was over 40 days.  So, most of those patients at 40 days should be engrafting. And the white cell count, the median was two. So, they are not immunocompetent but engrafted at that point.  Thanks. Just a quick question. Do you think fluid balance or the degree of renal failure may have been a factor in influencing this?  So, we pulled in, it's really challenging to look at fluid balance because of the way the fluid balance is collected on the transplant floor  and the variability, but we pulled in percent change in weight from admission to the start of noninvasive and it didn't, it wasn't associated with failure.  Thank you.  Thank you so much.  Thanks so much. So, our next speaker will be Dr. Yeha from the Children's Hospital of Philadelphia speaking on the implementation of a ventilator weaning pathway,  how it improves outcomes in pediatric ARDS.  Thanks everybody. I'm going to be talking about the implementation of a vent weaning pathway at Children's Hospital of Philadelphia. These are my co-authors.  So, the rationale behind this study is that like there haven't been any interventions which have been proven to improve outcomes in pediatric ARDS.  And this is a tough field to do outcomes research and or trials in because we have lower mortality. And so, we end up using a lot of composite outcomes.  And the most common one is probably ventilator-free days at 28 days.  And so, this is a composite between mortality and ventilator days and survivors censored at 28 days.  So, a few years ago, we implemented the CHOP ventilator weaning pathway.  And this was a multidisciplinary QI intervention with doctors, nurses, RTs, QI, analytics people to like, you know, to track our progress.  And this was initially a QI initiative.  And the pathway, briefly, we'll go into a little bit more detail about the pathway is patients were eligible when they were on a mean airway pressure of less  than 18, unconventional, not paralyzed. If you try to do a vent weaning pathway in paralyzed subjects, you're going to miss.  The most common modes of ventilation that we used at CHOP were pressure control and PRVC in the SIMV mode.  And there was a built-in extubation readiness test as part of this pathway.  And this was implemented in May 2016.  This is the pathway, okay. I don't expect you guys to read it. If you can read it, that's awesome. I'm happy to show you these slides afterwards so you can look at it in detail.  I will call your attention to, if you can't read it, then what these, there's a separate thing for pressure control.  There's a separate thing for PRVC in terms of what you're weaning. But the important thing to look at is that this was an RT-driven pathway in which the RTs had  to reach out to the, most of the frontline ordering clinician, which was either a nurse practitioner or an advanced practice provider, a resident or a fellow.  And so they had to actually, they were doing assessments on the ventilator periodically anyway.  So I think the thing that we operationalized, which is probably the main reason why this pathway may have worked,  is that we required assessments every two to four hours. And so keeping in mind that the RTs were doing these ventilator assessments  in two to four hour increments anyway, now we just have them look at the patient and say, like, looks weanable, versus I wouldn't wean that.  The extubation readiness test was a drop, once you reach certain PEEP and pressure support and FIO2 criteria, then you were dropped to minimal settings.  And then if you were, you were kept on these minimal settings for two hours. And if you passed, you passed. And the team was notified that, like, hey, man,  your patient passed an extubation readiness test at this time. And it was compared to, like, if, like we should be set up to extubate  or for whatever reason, like either there was a procedure or something or a good reason to leave the child intubated for that, at that moment, then the patient would return  to their original settings, but the team was at least notified that they had, in fact, passed an extubation readiness test.  So what we hypothesized as a secondary analysis of an ARDS data set, which was prospectively collected, we hypothesized that the vent weaning pathway at the time  of introduction was associated with a shorter duration of ventilation in ARDS survivors and shorter PICU length of stay.  So this was a, whoops, this was a prospective cohort of intubated ARDS patients that we'd been following since 2011. The dates of inclusion for this study are 2011 to 2019  with implementation of the pathway in 2016. So about, I don't know, what is that, five years before, four years after?  So, three years after. So five years before, three years after. So, like, it gave us some time to, like, actually get a few trends  and get some numbers to be able to do this.  We limited the analysis to survivors because we reasoned that this was a target population for this.  And patients who were, like, getting worse on the ventilator or, like, were otherwise made DNR or, like, were planned elective withdrawal care, like,  they were never entered into this pathway. They were never eligible for this pathway. We also ended up excluding extreme outliers, which was patients with greater  than three standard deviation or more than 100 days on the ventilator. The exposure was pre and post pathway implementation.  And the outcomes we looked at were ventilator days and then secondarily PICU days. And our balancing metric was reintubation rate.  So the analysis we chose was interrupted time series analysis. And this accounts for temporal trends. And we personally adjusted, we secondarily adjusted for autocorrelation.  That is to say that, like, May and June are more correlated with each other than May and February.  The measurements that we did were median ventilator days per PICU quarter over three months. And this is, the reason we did this is a lot of you guys have seen QI graphs  in which people are presenting monthly averages or things like that. But this is a diagnosis which occurs at maybe, like, 80 times a year in the CHOP ICU.  If you limit it to survivors, that's about 60 times a year. So that means if you were to do a monthly average, you would end up with, like, something like five patients a year.  And we all know that it's not going to be evenly scattered at five patients a year. It's going to be, like, some months are going to have two patients.  Some months are going to have 10 patients. So that's no way to do math. So we ended up picking quarters or three-month chunks, and we did the median value in an attempt  to get at the central tendency of what are, of what, to be able to tell whether or not we were making a change in the ventilator duration of survivors.  And when you do it this way, then we had something like an average of 14 subjects per quarter. Okay? So we could get some pretty consistent medians and averages if we needed to.  And every time increment had at least 10 subjects represented. We hypothesized that there would be a change in level. That is to say, fewer ventilator days and fewer PICU days.  Okay? And we conducted sensitivity analyses where we adjusted for pre-specified confounders. And then after we constructed our Table 1, we found what was different between the pre  and the post, and we separately adjusted for those. So when I say that, like, we hypothesized that there would be a change in level,  then there's different ways that interrupted time series analysis can look. And depending on what your intervention is, and best practice is probably  to pre-specify which direction and which of these graphs you actually think that your intervention is going to cause.  So all of these are different ways that a intervention, if it's going to have an effect, can have an effect.  So in the first box, in the upper left, for example, there's a change in what I call level. In the second box, there's a change in trajectory.  In the third box, there's a change in level and trajectory. In the fourth box, in the bottom left, there's a delayed change in trajectory.  In the fifth box, there's a change in level, which goes back to baseline at some point. And then in the last box, there's a change in trajectory, which results in a new level.  So it's helpful if you pre-specify which one you think is going to happen with your implementation or your pathway. We hypothesized this one, that there would be a change in level.  So why interrupt a time series analysis? And the answer is that time matters. Okay? Like, temporal trends absolutely matter. And if you think they don't, here's an example.  This is the unemployment rate for the last, I don't know, what is that, 15 years. And this is when Bush was elected, this is when Obama was elected,  and this is when Trump was elected. Now, if you have people trying to do naive comparisons of unemployment rate  between these three presidential periods and assign causality to it, and believe me, people do this all the fucking time, okay?  And if they were to ignore temporality, you would end up with a grossly erroneous conclusion.  These are the results of our analysis.  In the pre- and the post-intervention period, there are 585 subjects. There were seven of them who were excluded for greater than 100 days of mechanical ventilation.  So there weren't a huge contributor to the extreme outliers.  The post-intervention period, those patients are a little bit older. There's a slight distribution, a different distribution of how they ended up with ARDS.  And there's about a half as common a use of alternative ventilator modes, such as the oscillator or APRV. This was the pathway uptake.  We very rapidly got to about 60, 80% uptake, okay, with the implementation, which is great for, you know, great for us and great for the purposes of this analysis.  This is what happened with ventilator days. So we saw an immediate and sustained reduction in about three and a half ventilator days.  There was about a two-day reduction in ICU length of stay, which was not statistically significant.  These results were robust to outliers, excuse me, to robust to sensitivity analysis. So like, so the unadjusted analysis shows a lessening by 3.6 ventilator days  in after pathway implementation. And in every sensitivity analysis, depending on what, it didn't matter what we adjusted for,  there was a consistent about three to four days shorter of duration of mechanical ventilation in survivors. And they were all significant.  This was not the case with PICU length of stay, okay, in which the, there was a consistent effect size of about two to three days, but it was almost none of them were significant.  Reintubation rates, importantly, were not increased.  So we concluded from this that ventilator weaning pathways are associated with about three fewer ventilator days in survivors with ARDS.  There's no clear difference in PICU length of stay and no difference in reintubation rates. So we think this has a few implications. So one, okay, we are arguing causality here.  We are saying that like the pathway actually improves outcomes in ARDS by shortening duration of ventilation by three days.  Three days shorter on the ventilator is comparable effect size as clinical trials are powered for, that power for ventilator free days.  The second implication of this is that the institution of a pathway will affect the power calculation of a composite outcome like ventilator free days.  Reminding everybody, ventilator free days is a combination of mortality and ventilator duration in survivors.  If you set up your trial to be powered for like 14 ventilator free days in baseline and then your intervention will raise you up to 16 and you base that calculation  on observational data, but as part of your trial, if you were to introduce a ventilator weaning protocol because you're protocolizing everything in a trial,  then you are going to improve ventilator duration even though you didn't mean to. So what ends up happening is that this ends up affecting the power calculation of your trial.  And so this should be thought about and potentially adjusted for in the design phase of your trial. If you were introducing a protocol to a bunch of places  that didn't necessarily have a protocol before.  So the limitations are numerous. Single center, there's no control group.  The best way to have done this study would have been to do like half the unit with a ventilator weaning protocol and the other one, not even the same unit  because then there would have been bleed over of, you know, attendings because it's all the same attending group. But like if ideally this would have had a control group,  but that wasn't feasible with this design. There's always the possibility of unmeasured co-interventions around 2016. And so like the assignment of causality  to a single ventilator weaning pathway may be incorrect.  And the lack of difference in PICU length of stay warrants investigation. Although it's numerically smaller, it was not statistically smaller.  So one thing which may have been happening is we were trading invasive ventilator days for BiPAP days, but ultimately did not end  up changing the kid's PICU length of stay in a meaningful way.  Thank you.  Thank you for a wonderful and PG-13 presentation. Do you have any, does anyone have any questions?  Hi, Nader. Great talk. Quick question. Did you find that the unit or the providers were able to notice and feel they were extubating their patient sooner?  And if so, was there like a change in psychology of the provider, you know, of the attending and say, oh yeah, this kid's ready because the ERT was ready to go?  It was remarkably variable. Thank you, Terry. It was remarkably variable. Until we actually presented the initial pathway data after about a couple of years,  most of the the most common tenor of things was that even though it was being ordered 60 to 80 percent of the time, the most common tenor of things was a complaint  that this pathway probably wasn't working. And it actually wasn't until the first time we presented the QI data in a QI setting  and said that like, no, you're getting patients off the ventilator faster. And then we actually got buy in. And so like it actually, but it took it took a while.  And because initially this this is a secondary analysis. So like we did not design this study to like prove to people.  And like we didn't have pre-specified hypotheses until later on. This was a QI initiative. So we went through the like the standard PDSA cycles of like trying to get things better  and things like that. The pathway itself didn't change too much. There were minor iterations to the extubation readiness test, but the bulk of the pathway didn't really change.  And so the what's what's interesting is that like the RTs all felt from the get go that this was working, but the ordering clinicians actually took a little bit longer  and actually it their their tone didn't really change until after they saw data.  Thanks a lot for the excellent presentation. My question, did you look at how many patients in both groups has been put on noninvasive ventilation after extubation?  And did you consider that as a failure? So we did not consider extubation to NIV as a failure. And we also do not have data on in this database about who got NIV.  NIV, this was the database that this analysis was performed in was an ARDS database that was focused on intubated ARDS.  And so we actually have very little post-extubation data other than mortality. If somebody got extubated and died, we would have that.  But we didn't actually do a ton of there wasn't a focus on the initial database of post-extubation support. And so we have all the data because it's a single center.  And so I could do it, okay, I just haven't done it. Yeah. But it's a good question. It's an important question.  And I suspect it's the mechanism for why there's not a major difference in length of stay.  Yes. Similar question, but the length of the noninvasive after extubation, you didn't check it?  It's a great question. We will chase it down. I just have not yet chased it down.  Excuse me. How did you define a day?  And did you look at time of day of extubation to get at the issue of what we all do on the long-term kids where I'm not going to extubate them in the middle of the night?  You're entirely correct. Throughout this entire time period, just a little bit of context before I get to that. Throughout this entire time period, we had in-house attendings.  The second part of your question, the first part of your question, day is calendar day. Okay. And so it could be anywhere from 48 hours to 24 plus 1, 24 hours plus 1 minute.  So like, so yes, that is, it's the way things are defined. And that's the relative data we have. And so like that, and that is important to bring up because what that implies is  that the effect size of three days comes with a little bit of, take a look at the 95% confidence intervals.  And just like that's probably a more realistic, the conservative end of it. It may be the more realistic effect size.  With respect to not wanting to extubate a patient, over the course of the post-implantation pathway, as a unit,  I kind of felt that we became more comfortable extubating at odd hours. Because we always had this in-house attending presence.  But now we actually had a respiratory therapist coming at you and saying, this kid passed their extubation readiness test.  And because we're forward thinking like that, we made them NPO. And so it's been six hours. Do you want to, you know, I know it's midnight, but, you know.  And so I think there was, as with a little bit of time, like we actually did become more comfortable extubating at weirder hours.  The fact that, which actually suggests, if anything, that like if attendings did not, were not willing to do that, if faculty were not willing to do that,  that would have biased our results toward the null. And the fact that we saw an effect size actually, to me, suggests that like it's,  the effect would have been even bigger if we counted hourly. Okay. Because like then we would have had really granular data.  And then we would have been able to answer like time of day of extubation. Most extubations still probably happen during light of day.  But there were far more of them which were happening at night. Great talk, Nader. Thanks, man. Is this built into Epic? Are these orders? It is. These are Epic orders. Right. So.  These are Epic orders. Is there best practice advisor to say, let's do this, people? So there's, the way the Epic orders are set right now is  that this is the ventilator weaning pathway is ordered. There's a link to this page which tells you like how to do the weaning pathway.  That is, that prompts the respiratory therapist to do periodic checks. That, and then discuss the findings of those periodic checks with the ordering clinician.  And then they are responsible for manually putting in the new ventilator orders. So like as PEEP goes from 10 to 8, or PEEP goes from 8 to 5, or like rate goes from,  you know, 20 to 15 or whatever, you know. And so there still has to be, there's not a, there's not a best practice advisory built into Epic saying what to do.  It's a prompt to do this. But then the extubation readiness test is also built in as part of it. And the, the prompt after the extubation readiness test pass is a sign that's put  up on the patient's door, okay, that says this patient passed an ERT. And so that's, that's the, that's the workflow we've come up with to let people know  that this kid is ready to be extubated. And, and the goal was to make people feel bad for not extubating. Yeah. Very good. Again, that's quality improvement work. Yeah.  They're believers and non-believers, and they've got to keep hammering them. And then the other thing, one last question. Why are you looking at time for discharge from the PICU?  I don't know why people worry about that. Because, you know, or should you look at time there be, they're ready to be discharged from the PICU.  Because a lot of times you have the same problem that we do. Kids are sitting in the PICU because there's no beds on the floor. Because my metric is days and their calendar days,  I suspect that that particular issue is probably a little bit washed out in, in this, for my particular analysis.  I'm, I'm so sorry, but I think we're out of time for our next speaker. But thank you again. These were great questions. Please keep it going, and it's been very engaging.  So our next speaker is Dr. Umar Salimi, who is going to be talking about his work on recombinant  angiopoietin 1 and how it protects against LPS-induced inflammation in the developing murine lung. Hi, everyone. Thank you for attending my talk. My name is Umar Salimi.  I'm a pediatric critical care fellow at the Children's Mercy Hospital in Kansas City. And today I'm going to tell you a little bit about the scientific implications of angiopoietin  signaling in neonatal sepsis-induced lung injury. I have no conflicts of interest to disclose. Bronchopulmonary dysplasia, as we know, is a disease of impaired pulmonary parenchymal  airway and vascular maldevelopment in the injured developing lung. With the advent of lung protective ventilator strategies, the contribution of postnatal  sepsis to BPD has increasingly been recognized. We have focused on the pulmonary endothelium in sepsis and BPD.  Specifically vascular growth factors have key roles in regulating both endothelial permeability and developmental angiogenesis in the lung.  Our proposed model of sepsis-induced lung injury in the neonate focuses on pulmonary endothelial cell, or EC, activation.  Specifically in sepsis, bacterial products such as lipopolysaccharide, or LPS, are released and bind toll-like receptors in the lung.  Downstream, pro-inflammatory transcription factors, such as NF-kappa B, are activated. And this leads to the elaboration of cytokines and endothelial adhesion molecules, such as  ICAM-1. ICAM-1 will facilitate the infiltration of neutrophils into the alveolar space.  The consequences are lung inflammation, acute lung injury, and long-term parenchymal remodeling.  We have sought to explore these relationships and bench models, including animals and isolated  endothelial cells, in order to better characterize the relationship between sepsis and long-term chronic neonatal lung injury.  Interestingly, in this model, we have seen an upregulation of angiopoietin-2, a known biomarker of human disease.  In fact, angiopoietin-2 is reliably increased in bloodstream samples of patients across the age spectrum in both pediatric sepsis and ARDS.  The relationships between angiopoietins, a vascular growth factor family, have been well characterized in sepsis.  Specifically, angiopoietin-1, angiopoietin-2, and TI-2 are involved. During typical angiopoietin signaling, angiopoietin-1 binds the TI-2 receptor and activates it by  phosphorylating it. This leads downstream to homeostatic angiogenesis, endothelial cell adhesion by junctional complexes  and enhanced barrier function, as well as endothelial cell survival. Meanwhile, angiopoietin-2 is relatively low at the cell surface and does not bind TI-2  while it is sequestered in the cell as well. This promotes the release of angiopoietin-2 from the cell to the cell surface so that  it outcompetes angiopoietin-1 for TI-2, dephosphorylating the receptor and leading to endothelial immune activation.  This in turn promotes further inflammation during sepsis, vascular leak, and interestingly, angiopoietin-2 also promotes its own gene expression, propagating this cycle.  There has been a relative lack of study of the angiopoietin pathway mechanistically in pediatric and neonatal disease.  Therefore, we sought to understand how angiopoietin-TI-2 axis dysregulation in sepsis affects lung  development, as well as whether angiopoietin-1 supplementation can be beneficial in neonatal sepsis-induced lung injury.  We hypothesized that recombinant angiopoietin-1 will protect against sepsis-induced ALI in  long-term alveolar remodeling through sustained homeostatic angiopoietin-TI-2 axis signaling.  Our first objective was to understand whether recombinant angiopoietin-1 suppresses LPS-induced acute lung injury in alveolar remodeling in vivo.  And to do this, we used a mouse model of six-day-old wild-type mice. This is a convenient model to study lung development as mice are born with... prior  to the secondary phase of alveolar septation, meaning that the alveoli have still a long way to go in formation.  Within this model, we used recombinant angiopoietin-1 protein, which we had to work to understand  how dosing would best optimize study of lung injury, as it had never been dosed in neonatal mice before.  So we extrapolated dosing from prior adult animal model studies, as well as used a few different dosing strategies.  IP delivery was best in our 3-gram neonatal mice in order to best... best deliver this compound systemically.  We gave the recombinant angiopoietin-1 as a pretreatment to lipopolysaccharide. LPS was administered at a dose which we knew reliably produced alveolar simplification  in the long term. Tests were taken at 24 hours for further study. We first wanted to understand whether recombinant angiopoietin-1 led to increased angiopoietin-1  activity in the lung, and this... and whether intraperitoneal administration would be a reliable way to deliver angiopoietin-1.  Therefore, we went back to the interaction of angi-1 with the TIE2 receptor, knowing that angiopoietin-1 phosphorylates TIE2, we performed immunofluorescence of TIE2...  phosphorylated TIE2 specifically. The image shown depicts phosphorylated TIE2 in green at the cell surface.  Blue indicates nuclei of any lung cell, and red is an endothelial cell-specific antigen.  LPS induced a robust dephosphorylation of TIE2, while recombinant angiopoietin-1 pretreatment  actually produced an increased green signal indicating that TIE2 was more strongly phosphorylated in animals pretreated with angiopoietin-1 despite receiving LPS.  To understand whether angiopoietin-1 affects acute lung inflammation, we then measured lung inflammatory mediators.  The graph depicts a gene expression by PCR of interest... the adhesion molecule ICAM-1 and cytokines IL-1 beta and keratinocyte-derived cytokine, the mirin homolog of IL-8.  Looking at the red, we see that LPS induces robust levels of all inflammatory mediators, while the orange indicates approximately a 30 to 60% reduction of ICAM-1, IL-1 beta,  and KC in angiopoietin-1 pretreated subjects. The western blot to the side of the image, it looks at NF-kappa B activity within our  model, looking specifically at the NF-kappa B P65 subunit, measuring protein levels of activated NF-kappa B. Phosphorylation of P65 is a key step to NF-kappa B activation prior  to its translocation to the nucleus. So we see that there's increased phosphorylated P65 in LPS-treated animals, and this rise  is prevented essentially by recombinant angiopoietin-1 pretreatment. Looking at lung injury, we assessed apoptosis by the tunnel stain method.  Green indicates fragmented DNA, which signifies apoptosis. LPS induces a robust apoptotic response through TLR signaling.  This is well characterized in the literature, and we see that recombinant angiopoietin-1 pretreatment actually prevents apoptosis almost to the degree of control.  Cellular neutrophil infiltration and increased protein levels are a salient feature of acute lung injury.  Therefore, we performed BALs in our newborn mice by tracheal cannulation and the installation  of approximately 250 microliters of saline, recovering about 150 to 200 microliters with each pass. And first, we characterized total cell counts and cell differentials.  As expected, the total cell counts were robustly increased with LPS while the percentage of neutrophils was also increased.  Recombinant angiopoietin-1 decreased total cell counts and also preserved the alveolar macrophage population.  Protein levels were also reduced with angiopoietin-1 in LPS-treated subjects. And at 24 hours, we already see effects on acute lung remodeling, looking specifically  at metalloproteinase-mediated elastin fiber degradation. We see that LPS induces MMP9, metalloproteinase-9, in the lung of our mouse subjects, which correlates  with elastin fiber degradation by staining. Recombinant angiopoietin-1 preserves elastin fiber continuity and density as well as suppresses levels of MMP9.  And in the long term, at day of life 15, we see that lung morphology is also preserved with angiopoietin-1.  LPS created a state of alveolar simplification, both qualitatively and quantitatively shown  here with reduced alveolar counts in LPS-treated subjects that were increased in recombinant angiopoietin-1 pretreated subjects.  Our final objective was to determine whether angiopoietin-1 directly antagonizes angiopoietin-2.  And to do this, we worked in culture and endothelial cells, specifically human pulmonary microvascular  endothelial cells, a fetal cell line which we stimulated with LPS, angiopoietin-1, or angiopoietin-2.  LPS alone induced a robust gene expression response of angiopoietin-2 that was suppressed with the addition of angiopoietin-1.  When angiopoietin-2 alone was given without LPS, we saw increased levels of inflammatory  mediators, IL-1 beta and IL-8, which could be suppressed with the addition of angiopoietin-1.  In addition, ICAM-1 was increased in angiopoietin-2 treated cells, which again was suppressed  with the addition of angiopoietin-1, indicating increased immune activation directly with angiopoietin-2 alone that could be overcome with angiopoietin-1 co-treatment.  So to summarize what we've discussed, in a mouse model of neonatal sepsis-induced lung injury, recombinant angiopoietin-1 decreased in key inflammatory mediators by measurements  through Western blots and PCRs, suppressed apoptosis by the tunnel stain method, reduced alveolar capillary permeability as assessed through alveolar neutrophil and protein counts,  and reduced acute lung remodeling as well as alveolar simplification in the longer term. Mechanistically, angiopoietin-1 suppressed angiopoietin-2-mediated inflammation in human  pulmonary endothelial cells in vitro. In conclusion, recombinant angiopoietin-1 protects against LPS-induced acute lung injury  and alveolar simplification in newborn mice by antagonizing angiopoietin-2. We speculate that angiopoietin signaling during human sepsis can be therapeutically modified  in neonatal acute lung injury and bronchopulmonary dysplasia. Future directions of our work will look at using recombinant angiopoietin-1 as more of  a rescue treatment given after LPS to better replicate a clinical scenario. In addition, we wonder whether recombinant angiopoietin-1 can cause any harm.  It is known to promote endothelial cell proliferation and could potentially lead to pulmonary vascular hypertrophy.  So looking at measures of pulmonary hypertension in our animal models will be important. And as angiopoietin-2 all died, it's been difficult to study angiopoietin-2 due to the  lack of a reliable animal model as angiopoietin-2 knock out my sexually developed calothorax, which in itself starts to answer the question.  I want to acknowledge my lab group at the Children's Mercy Hospital, neonatology, as well as the Division of Pediatric Critical Care.  Thank you all for listening to my research, and I'll be happy to take any questions at this time. Thank you, Dr. Salimi, for a very nice presentation of your work.  I think we have time for maybe one very quick short question of less than 30 seconds. Great talk.  Are you planning to do this in a live bacteria model, like either CLP or intratracheal pneumonia? No, that was not in our current plans, but that's a very interesting approach.  The reason I ask is just like LPS remains a sterile model, and there's things that immune cell endothelial interactions do which are helpful.  So it's possible that in the setting of LPS in your model, the endothelium die and become leaky because it makes it easier for neutrophils to influx and kill the bacteria.  So it would be a more convincing translational therapy if you could show me that blocking that pathway with angiopoietin 1 was not causing a bunch of bacteria to accumulate and then  kill the patient with intact endothelium, but a overproliferation of bacteria either in the lung or in systemic circulation.  So that's why there's been a lack of translation between preclinical and clinical stuff, and  I do wonder whether the addition of a live bacteria model may change some of your conclusions or strengthen them. Yeah. Thank you for your recommendation. I appreciate that.  Thank you.  Okay. Our next speaker will be Dr. Mamari from UT Southwestern presenting on persistent tachycardia  and comorbid pneumonia as being associated with noninvasive positive pressure ventilation failure and bronchiolitis. Dr. Mamari. Hi, everyone.  Thank you for joining me and thank you for the introduction. My name is Nia. This is some of the data that I accumulated in fellowship, and now this was at Children's  National and now I'm at UT Southwestern. And so I'll be discussing persistent tachycardia and comorbid pneumonia and how those are associated with NIPPV failure in bronchiolitis.  So as we all know, acute bronchiolitis is one of the most common causes for admission to the PICU, and there is a wide spectrum of modalities that we use for respiratory  supportive care. And the start of that is high-flow nasal cannula, continuous positive airway pressure, and then  we can move on to bi-level positive airway pressure, which here I'll refer to as noninvasive positive pressure ventilation.  And then there is intubation and, of course, things beyond intubation. And so NIPPV can be delivered through a variety of interfaces.  Some can be nasal configurations or full-face configurations, and the interfaces can vary by size, fit, comfort, or availability in your institution or your region.  Another interface that can deliver bi-level airway pressure is the REM cannula. And so this is an interface that was developed in the NICU.  And in premature neonates, it reduced the extubation failure and reduced the need for re-intubation.  And its use sort of crossed over to the PICU, and while it's subjectively more comfortable  than a tight-fitting BiPAP mask, its use has also grown in the setting of acute respiratory failure due to bronchiolitis. And so this is the patient population I'm looking at.  So the knowledge gap is that in this patient population, the factors associated with NIPPV failure leading to subsequent intubation in these infants with bronchiolitis are not  well-defined. And so our hypothesis is that some of the factors can be physiometric variables like heart rate and or respiratory rate before and after NIPPV initiation.  Those are associated with NIPPV failure and subsequent intubation. So accordingly, we sought to determine if there is an association that exists between  these physiometric variables and with failure of NIPPV. And we were looking at pre- and post-physiometric variables.  So this was a retrospective cohort study from 2013 to 2017 in a large PICU in a large academic  children's hospital in infants under one year of age with bronchiolitis on whom NIPPV was applied via the RAM cannula.  And so our exclusion criteria were critical congenital heart disease, sepsis, neuromuscular disease, and tracheostomy dependence.  The independent variables that we explored were demographics, comorbidities, NIPPV settings, and physiometric data.  So again, heart rate, respiratory rate, the differences between the two before and after.  And we looked at hourly vital signs, so hourly heart rate and respiratory rate up to 10 hours  before and up to 10 hours after as data were available or up until intubation if that occurred before 10 hours.  And our primary outcome was failure of NIPPV, which to us meant subsequent intubation. So we identified 138 patients.  And a third of those failed NIPPV and were subsequently intubated. And that's shown in orange, that proportion in orange.  And those kids who failed, they spent about a median of six hours on NIPPV and then were intubated.  And once they were intubated, they spent about a week on invasive mechanical ventilation. So for our first results table, we looked at demographics and we compared the success  group and the failure group and found no differences in demographics between those groups. We looked at age, sex, weight, race, ethnicity, and comorbidities.  And again, we found no differences. Most of our, the median of our cohort was around three months of age and about 5.5 kilos.  When we looked at our infectious pathogens and the distribution of viral pathogens was not different between the success group and the failure group.  But when we looked at superimposed or comorbid pneumonia, we found that that was more prevalent in the failure group than the success group.  And so about 53% of patients who failed NIPPV had pneumonia versus 15% in the success group.  Next we looked at the NIPPV settings and again found no difference between the success group and the failure group. And our, the PIP, the median PIP in both groups was about 30.  The PEEP was between 8 to 10. The ventilator rate was about 30 and the FiO2 was 40 to 50%. And again, no differences between the two.  When we looked at medication use, we looked at antibiotics inhaled versus systemic steroids or bronchodilators.  Again, found no differences between the success group and the failure group. And these are all unadjusted differences and so they, these are obtained by simple logistic regression.  So moving on to our physiometric data, first we looked at the respiratory rate. So again, we looked at this hourly up to 10 hours before and 10 hours after NIPPV initiation.  And we have the success group in green and the failure group in orange. And we looked at before NIPPV initiation and after.  And we found that in both groups, the success group and the failure group, the respiratory rate significantly decreased after NIPPV initiation.  And the decrease was similar in both groups. So the decrease was down by 2 to 4 breaths per minute or 5 to 9% from the pre-NIPPV values.  In a similar way, we looked at the heart rate. And so again, we looked at pre- and post-NIPPV initiation and the success group is in green and the failure group is in orange.  And we found that the heart rate decreased after NIPPV initiation only in the success group, but not in the failure group.  And we found that the median heart rate in the failure group after NIPPV initiation was significantly higher than the median heart rate in the success group.  The decrease in the success group in the heart rate was about 6 beats per minute or about by 4%. While in the failure group, there was no decrease or even a slight increase.  And so we combined these data and now adjusted them in a multivariable logistic regression. And so again, we still found that comorbid bacterial pneumonia and persistent tachycardia  or median heart rate after NIPPV initiation were still independent variables associated with failure of NIPPV.  And so they were still variables that differentiated those who failed from those who succeeded.  And so our conclusions were that the factors that are associated with NIPPV failure and subsequent intubation among these infants with bronchiolitis in the PICU on whom REM  cannula was tried and delivered NIPPV were comorbid pneumonia and persistently elevated heart rate after NIPPV initiation.  And that respiratory rate decreased after NIPPV initiation through the REM cannula. So in the future, to further confirm these results, we'd like to perform some prospective  observational studies of NIPPV and comparison studies of these interfaces and other interfaces and their impact on intubation rates and other outcomes.  And so ideally, we have the overarching goal of optimizing NIPPV. And so this would be finding the right interface with the right settings on the right patient  at the right time in their course of illness.  This is everyone I'd like to acknowledge who's very helpful in this project. Thank you. Thank you. Are there questions? None in this institution.  It was less than five. Yeah, less than five patients. Thanks. Thanks, Mia, for presentation. How did you define pneumonia radiographically or BAL?  And if those kids didn't get intubated, how did you know they're not infected? Yes. So the way we defined pneumonia was mainly through ICD-9 and 10 codes, which I know are fraught  with their own limitations. When we could, we would try to we would confirm the diagnosis of pneumonia if there was a respiratory culture  through an ET tube, that would confirm the ICD-9 or 10 code. But of course, in the patients who were intubated, we don't have luxury of an ET tube respiratory culture.  So essentially, ICD-9 and 10 codes. How did you define the failure of noninvasive support? Need for intubation. So if you were intubated. My unit, the number one need  for the younger staff is tachycardia. So they intubate. So self-fulfilling. What criteria? Is that gas failure? I mean, you should intubate a broncholytic  and the CO2 goes from 70 to 120. If it's true, obstructed lung disease. So it's very subjective. It is subjective. And I'm sure that there is variability  in the indications for intubation. And I don't have the data for what indications the deciding provider had to decide to intubate. So we just know the patient was on NIPPV  and then the patient was intubated. And we tried to extrapolate from there.  Hi. I was just wondering about your experience with the Ram cannula. So my experience with using Ram cannula as a noninvasive positive pressure, it's almost, I mean,  they very rarely trigger. And so it's almost kind of like a noninvasive APRV where they're just kind of breathing around. Is your experience different than that?  Because I've wondered what exactly are we doing with the Ram cannula? And I don't know if anybody has raging success stories. I don't know that we have raging success stories.  But the settings that we've used in this institution are, I would think, on the high side. So it's like 30 over 10 with a rate of 30.  So you're pushing a delta P of 20, 30 times a minute. And so while our settings included that you could trigger. I mean, so we had the servo eye  and there was that little purple mark. And so it could sense the flow trigger and trigger. But it was already going at 30 times per minute.  And so I know that we are decreasing respiratory rate on the Ram cannula. And I know that in general, noninvasive has led to reduction in intubation rates.  But in terms of how are we actually changing the, the lung mechanics, more remains to be seen. So. I apologize. And with that, I think we've run out of time.  But Dr. Nishizaki, perhaps after, afterwards, yeah. Thank you. Thank you, Dr. Yamari. So our next presenter is going to talk about characteristics  and outcomes of children with necrotizing pneumonia. And this is Dr. Minette Ness-Kochenwala.  Hi, my name is Minette Ness-Kochenwala. I'm one of the Pediatric Critical Care Fellows at Nicholas Children's Hospital. And I'll be talking a little bit about necrotizing pneumonia.  So to start with, necrotizing pneumonia is a relatively rare diagnosis. However, it is a severe complication of pneumonia. It accounts for about 3.7%  of all community-acquired pneumonias. And although its clinical features are similar to uncomplicated pneumonias, they tend to have a higher incidence of pleural involvement.  And they also tend to have, be sicker patients overall. There is an increased risk of respiratory failure and shock in these patients. However, the current data  on necrotizing pneumonia is really centered primarily on single-center retrospective studies and also expert opinion. So essentially, although this is a rare diagnosis,  it has severe complications. And there is a significant need for national data on this diagnosis. The methods of this study, it was a retrospective study  of the KID database in 2016. It included about 6 million non-neonatal weighted discharges. We then filtered this database for the ICD-10 code of lung necrosis.  And our N for that was about 746 patients with lung necrosis. Children with lung necrosis alone were then compared to patients with lung necrosis  who also had a co-diagnostic code for bacterial isolation. Whether that was in the blood, endotracheal sample, or other method. And then sample weighting was used on these patients  in order to produce national estimates. For our data analysis, we did the chi-squared test, as well as the Mann-Whitney U test for data analysis.  We also did a univariate analysis in order to assess variables that impacted mortality. Any of these variables that had a P value of less than .1 were then included  into a multivariable regression analysis. Any of the categorical data I'm presenting will be presented as P values with odds ratios for 95% confidence intervals.  And any of the continuous variables will be presented as medians with interquartile ranges. As far as my results go, I first want to talk a little about some epidemiologic data.  These are the age groups for patients with lung necrosis. As you can see, the highest age group is between two and five years, accounting for about 35% of the patients  with lung necrosis, which is relatively significant. And this data is significantly different than the age groups within the database in general.  I also looked at the race and ethnicity of patients with lung necrosis. And as you can see, this data set is also significantly  different than the distribution for race and ethnicity in the general database for all discharges as well. As far as the regional distribution  in the western region, it was about 24%. The Midwest, it was 21%. The Northeast, it was 18%. And the South was about approximately 37%. And then I looked at a little bit  at the complication rates in lung necrosis. We know that they have a higher complication rate, so I wanted to look at exactly which, what was the frequency  of these specific complications. As you can see, chest tubes were very commonly used in approximately half of the patients with lung necrosis, which is a very high percentage.  You can also look at mechanical ventilation was employed in about a quarter, pyrethrox in about a quarter. And the overall mortality for patients  with lung necrosis was 4.1%, which is much higher than the usual mortality in PICU patients. I then looked at the presence of a chest tube and how that impacted these patients.  So the patients that had lung necrosis and a chest tube had a significantly increased length in stay, and this was statistically significant.  If the chest tube was present, their median length of stay was approximately 14 days, whereas if they did not have a chest tube present, their median length of stay was 10 days.  The presence of a chest tube also significantly increased hospital charges, and this, again, was statistically significant. The average or median length of, median charges was  about 150,000 with a chest tube present, as opposed to about 85,000 without a chest tube present. And then we looked a little bit at infectious etiology,  which is something that has been explored in these single center trials. So about 40% of patients with lung necrosis also had a concurrent co-diagnosis  that indicated that they had a bacterial isolation. The most common bacterial etiologies, the first was Strep pneumo at about 12%, and the second was Staph aureus at about 10%.  And this is consistent with other studies that have shown that Strep pneumo is usually the most common, followed by Staph aureus.  Interestingly, we found that Pseudomonas was present in approximately 4% of these patients, and Mycoplasma in approximately 3% of these patients.  And then we looked a little bit at the patients with lung necrosis with bacterial isolation, as opposed to the patients with lung necrosis  where they were unable to get bacterial isolation. The patients that had lung necrosis and bacterial isolation had a higher complication rate,  and these are some of the examples of complications that were higher in the patients with bacterial isolation and lung necrosis.  For example, chest tube was significantly higher in those with bacterial isolation, mechanical ventilation, and as you can tell, mortality is also significantly higher  at 6% as opposed to 2.8% as far as those without bacterial isolation. So if you are able to isolate a bacteria, it does indicate that they will have a more severe course  with increased complications. And then we looked at what specific variables would be associated with an increased risk in mortality with patients with lung necrosis.  So for that, we did a univariate analysis. So in the presence of these variables, mortality was increased. As you can see, septic shock has a relatively high odds ratio of about 16.  So when septic shock was present, the mortality was 16 times more likely, as well as the presence of any or all, any of the chronic complex conditions.  And as you can see, we also looked at mechanical ventilation, pneumothorax, and bacterial ideology. After that, we took any of the variables  in the univariate analysis with a P value of less than .1, and we added them to the multivariable regression analysis for mortality.  When we do that, did that only two variables significantly increase mortality in lung necrosis. The first one was septic shock with an odds ratio of 7.8,  and the second one was the presence of any complex chronic condition. Bacterial isolation in the multivariable regression analysis did not significantly  increase the risk of mortality in the multivariate regression. As far as complex chronic conditions, I found this pretty interesting. About 28% of patients  with lung necrosis had a complex chronic condition. So about a quarter of those patients. And the presence of a complex chronic condition was associated  with a significantly higher mortality rate, as we saw in the multivariable regression analysis. Those without a complex chronic condition, the mortality rate was 1.1%, whereas those  with a complex chronic condition, the mortality rate was 11.9%, which is a very significant difference. The other thing is that the bacterial ideologies in these patients  with complex chronic conditions was also significantly different. Staph aureus was the most common bacterial ideology in these patients, with a frequency of 13.4%.  Pseudomonas was the second most common, with a frequency of 12.8%. And then strep pneumo was the third most common with a frequency of 11.8%. So a relatively different frequency  of bacterial ideologies in these patients as well. As far as limitations for this study, it was a retrospective study. It is also a database study.  So it's based on physician coding, and it is limited by both the diagnostic codes and, you know, whether the physicians actually coded the correct diagnosis.  And as well, there is no data on any long-term outcomes. The data entry stops at discharge. As far as conclusions go for our study, we found that the need  for a chest tube was associated with a significantly increased length of stay and hospital charges, that those patients with bacterial isolation  and lung necrosis had higher complication rates, and the presence of septic shock or a complex chronic condition significantly increased the risk of mortality in these patients.  These are some of my references. And that's all I have. Thank you. Thank you. I'm open for questions.  Thank you. I have a quick question. You mentioned some very interesting geographic variation across the different regions. Do you speculate why that might be?  And it might be something worth reporting in your paper, actually. Yeah. Actually, you know, and I guess one thing I didn't have time  to mention is that the KID database does have a slightly higher incidence of reporting of the hospitals in the south than those other regions. So that might account for some of it.  And, but I'll look into it a little bit more. Yeah. Thank you.  Thank you. Great. And our anchor for the afternoon is Dr. Anna Murray from the University of Texas Southwestern presenting on routine morning chest radiography usage  in a large quaternary pediatric ICU.  Hi, my name is Srabia Annamarie, I'm a PICU fellow at UT Southwestern Children's Health in Dallas, and I'm going to talk to you today about decreasing routine chest x-ray use in  the PICU. Daily AM chest x-rays are routinely obtained in many PICUs, but it's still largely unclear  how many of these chest x-rays lead to a change in clinical patient care, because it's a topic that is relatively understudied in the PICU world.  As we move into an era of increased care utilization, there are, at times, harm to the interventions that we provide. Routine AM chest x-rays can have significant disadvantages.  They can lead to increased radiation exposure, worsened quality of stay, including loss of sleep and increased delirium.  They can increase the work burden for hospital personnel, often requiring a nurse and a respiratory  therapist at bedside to obtain the routine x-ray, increased need for portable chest x-ray machines and x-ray technicians.  This can all translate into increased patient costs and increased costs for the hospital as well.  The adult world has successfully implemented protocols for decreasing routine chest x-ray usage without compromising patient care.  However, the largest pediatric study looking at routine chest x-ray use was done in 2001.  The Quasny study looked at 512 chest x-rays, which led to a change in management in about 45%. Most of those interventions were in young children with acute cardiopulmonary issues.  So we sought to look at routine chest x-ray use in our PICU. Our aims were twofold. Number one, to elucidate the reasons and rationales for obtaining a routine AM chest x-ray in  our PICU. And number two, to determine if these chest x-rays led to a change in management. We started off with a pilot review of chest x-ray use in our PICU over a non-consecutive  two-week period. During that time, 143 chest x-rays were done. And out of those, 74 were routine AM chest x-rays. But only 18 led to a change in management.  So armed with that data, we decided to study routine chest x-ray use in our PICU by doing  a single-center retrospective chart review in our 44-bed quaternary care ICU over a one-year period from July 2018 to July 2019.  We obtained a sampling of about 60 chest x-rays from each month. Patients with routine AM chest x-rays were identified through an Epic-generated report.  We excluded all admission and PRN chest x-rays. And we determined if routine chest x-rays led to a change in management by looking at  changes in orders and respiratory settings, like orders for endotracheal tube adjustment, provider notes, like attending notes, resident notes, RT notes, nursing notes, and chest  x-ray radiology interpretation for every chest x-ray. Statistical analysis was done to compare the change in management group with the no change in management group.  We collected data on patient demographics, routine chest x-ray indications and characteristics, and interventions as a result of the x-ray.  So we looked at any tubes or lines that were adjusted, ventilator changes, respiratory therapy changes, changes in fluid and diuretic management, antibiotic use, and any tests  or procedures that were done as a result of this x-ray. Out of the 714 chest x-rays that we reviewed, we found that the patients that were included  were 49% male with a median age of four years and a median ICU length of stay of 21 days. The top five primary and secondary diagnosis by disease group were included, with the most  common disease group being respiratory. Chest x-rays were ordered in our cohort for multiple reasons, the most common being disease progression and endotracheal tube location.  At times, one chest x-ray was often ordered for multiple reasons. So when we looked at the demographics by group, we found that the change in management group  and the no change in management group were similar in all categories, except for gender  and number of days in the ICU before the chest x-ray was obtained, with the change in management  group more likely to be male and have a fewer number of days spent in the ICU before the chest x-ray was obtained.  There was no difference in diagnosis by disease group for the five most common diagnoses. When we looked at chest x-rays by level of respiratory support, we found that patients  who were on room air, oxygen binasal cannula, or with chronic respiratory failure at baseline at the time the chest x-ray was obtained, no chest x-ray led to a change in management.  When we looked at patients who were on high flow nasal cannula, non-invasive ventilation or mechanical ventilation, we found that the chest x-ray resulted in a change about 11  to 15% of the time. Interestingly, for patients with a tracheostomy, regardless of their level of respiratory support,  the chest x-ray led to a change in management only 4% of the time. The most common intervention as a result of a chest x-ray, a routine AM chest x-ray was  endotracheal tube adjustment. The majority of interventions, I'm sorry. The majority of interventions were in our invasive mechanical ventilation cohort and  few interventions happened in our non-invasive group and our high flow groups. We noted that 12% of those with endotracheal tubes had their endotracheal tube adjusted  after a routine AM chest x-ray. We also noted that the median age of patients with their endotracheal tubes adjusted was  one year, much lower than the four years for patients who did not have their endotracheal tubes adjusted.  The next thing that we did was to look at factors that were associated with a change in management.  Patients that had a change in management were more likely to receive a sedation bolus within 30 minutes of their routine AM chest x-ray and were more likely to have additional chest  x-rays that had been done in the preceding 24-hour period. In conclusion, in our data set, the majority of routine AM chest x-rays, 86% to be exact,  did not lead to a change in management and it may be unnecessary to do routine AM chest x-rays in non-intubated patients.  And last, in patients who were intubated, endotracheal tubes were only adjusted 12% of the time. There were limitations to our study.  It was a single center study and patients in the cardiovascular intensive care unit were not included.  There were multiple decision makers that were determining changes to patient management and there was potential for variability in data collection given the risk of incomplete  documentation in the medical record and the presence of two data gatherers. We did have high inter-rater reliability between our two data gatherers with scores of 86% and 90%.  The next steps for our study are to first enrich our data set by adding more chest x-rays,  do a multivariate analysis looking for factors such as age and diagnosis that could be associated  with a change in management, second, create guidelines for routine AM chest x-ray ordering in our PICU, and third, do a prospective study looking at the effect of those guidelines.  Here are my references. I'd like to thank the following people, Dr. Darnell, Dr. Morrison, Dr. Sheehan, my mentors,  Dr. Green, my program director, and Denise Roden and Ricky Surgdon who were instrumental in data collection and data analysis. Thank you. I'm now open for questions and comments.  Great. Thank you. You're welcome. Please.  So there is a program underway with our respiratory therapists to work on doing that, but that hasn't been fully implemented yet.  Do you think it's worth looking at the size of the patient as a risk factor in determining what steps forward should be made? Yes, I do.  I do think that we will see that smaller patients will likely have more changes done, and we're going to do that in the multivariate analysis. Great. Thank you so much.  Thank you, everybody. Thanks. I think both of us would agree with thank you to all the presenters and their mentors  for wonderful presentations, and without you, this wouldn't be possible. Thank you again, and we'll see you at the reception.  Reception, and thank you for being a great audience with lots of good questions. Thank you.

Video Summary

Dr. Yeha conducted a study at the Children's Hospital of Philadelphia to assess the impact of a ventilator weaning pathway on outcomes in pediatric ARDS. The pathway, implemented as a quality improvement initiative, involved specific criteria for eligibility and readiness tests. The study analyzed data from a prospective cohort of intubated ARDS patients before and after the implementation of the pathway. The results showed a significant reduction in ventilator days after the pathway was implemented, indicating its effectiveness in shortening the duration of mechanical ventilation. However, there was no significant difference in PICU days or reintubation rates. The study suggests that the pathway could improve outcomes in pediatric ARDS but does not have a substantial impact on PICU length of stay or reintubation rates. The study emphasizes the importance of evidence-based interventions in pediatric critical care.<br /><br />In another study, the use of routine morning chest x-rays in the PICU was evaluated. The study found that while chest x-rays were commonly ordered for various reasons, only a small percentage led to a change in clinical management. Specifically, chest x-rays obtained in patients on room air, oxygen nasal cannula, or with chronic respiratory failure did not result in any changes in management. However, chest x-rays obtained in patients on high-flow nasal cannula, non-invasive ventilation, or mechanical ventilation led to changes in management about 11-15% of the time, with the most common change being adjustments to the endotracheal tube. The study suggests that routine chest x-rays may not be necessary in non-intubated patients and further research is needed to establish guidelines for their ordering in the PICU.

Asset Subtitle

Pediatrics, Pulmonary, 2020

Asset Caption

The Society of Critical Care Medicine's Critical Care Congress features internationally renowned faculty and content sessions highlighting the most up-to-date, evidence-based developments in critical care medicine. This is a presentation from the 2020 Critical Care Congress held February 16-19, 2020, in Orlando, Florida, USA.

9-The Effect of Protocolization of Status Asthmaticus Treatment in a Pediatric Intensive Care Unit

Erin Hulfish

10-Stress Ulcer Prophylaxis for Children Admitted to the ICU for Status Asthmaticus Appears Unwarranted

Anthony A Sochet

11-Factors Associated With Failure of Noninvasive Ventilation in Children After Stem Cell Transplant

Courtney Rowan

12-Implementation of a Ventilator Weaning Pathway Improves Outcomes in Pediatric ARDS

Nadir Yehya

13-Recombinant Angiopoietin 1 Protects Against LPS-Induced Inflammation in the Developing Murine Lung

Umar Salimi

14-Persistent Tachycardia and Comorbid Pneumonia Are Associated With NIPPV Failure in Bronchiolitis

Mia Maamari

15-Characteristics and Outcomes of Children With Necrotizing Pneumonia

Manette Ness-Cochinwala

16-Routine Morning Chest Radiography Usage in a Large Quaternary Pediatric ICU

Sravya Ennamuri

Meta Tag

Content Type Presentation

Knowledge Area Pediatrics

Knowledge Area Pulmonary

Knowledge Level Advanced

Membership Level Select

Membership Level Professional

Membership Level Associate

Tag Asthma

Tag Pneumonia

Year 2020

Keywords

Dr. Yeha

study

Children's Hospital of Philadelphia

ventilator weaning pathway

pediatric ARDS

quality improvement initiative

prospective cohort

intubated ARDS patients

mechanical ventilation

ventilator days

PICU days

reintubation rates

evidence-based interventions

routine morning chest x-rays

clinical management

	Society of Critical Care Medicine 500 Midway Drive Mount Prospect, IL 60056 USA Phone: +1 847 827-6888 Fax: +1 847 439-7226 Email: support@sccm.org	Contact Us About SCCM Newsroom Advertising & Sponsorship DONATE	MySCCM LearnICU Patients & Families Surviving Sepsis Campaign Critical Care Societies Collaborative	GET OUR NEWSLETTER
© Society of Critical Care Medicine. All rights reserved. \| Privacy Statement \| Terms & Conditions The Society of Critical Care Medicine, SCCM, and Critical Care Congress are registered trademarks of the Society of Critical Care Medicine.