All right, good afternoon, my name is Dan Hassoumani. I'm a third year pediatrics resident at Indiana University Riley Hospital for Children. And today I'll be talking to you about seizures in children with cardiac disease on extracorporeal membrane oxygenation. I think I clicked this to, good, okay. I have nothing to disclose. In the pediatric population, seizures while in ECMO are common, as high as 18 to 23% in two recent publications from 2018. Seizures while in ECMO can lead to worse both short-term and long-term outcomes. Many of the seizures that pediatric patients have while in ECMO are electrographic in nature only. And cardiac surgery has been found to be a risk factor for developing seizures. Our specific aims, we sought to report the frequency of seizures in status epilepticus in pediatric patients on ECMO for cardiac indications. And also to identify risk factors for seizures amongst pediatric patients on ECMO for cardiac indications. This is a retrospective cohort study between January 2014 through December of 2017 at a single center tertiary care children's hospital. Inclusion criteria were any children less than 18 years of age who were placed on ECMO for cardiac indications. Exclusion criteria were any of those children who were on ECMO for less than six hours. Continuous neurotelemetry for the first 48 hours following ECMO initiation. Or 24 hours after the last seizure, and then as clinically indicated, is the standard of practice at our institution. Here are some general characteristics of our patient population. A large majority of our patients had congenital heart disease and had cardiac surgery prior to being placed on ECMO. 26% were cannulated via eCPR. The median length of ECMO was 96 hours. And 67% of our patients survived to hospital discharge. In our patient population, 18% of patients develop seizures. 56% of these were electrographic in nature only. And of the patients who developed seizures, 38% also developed status epilepticus. The median time from ECMO initiation to first seizure was 40 hours with a range of 21 to 79 hours. Risk factors associated with seizures. So of all the risk factors that we explored, only cardiac arrest and eCPR were associated with seizures. Other notable variables that were not associated with seizures included age at ECMO cannulation, duration of cardiopulmonary bypass, pH or CO2, OR versus ICU cannulation, neck versus central cannulation, or any specific cardiac anatomy. This is a graph of cardiac arrest duration with, on the y-axis has cardiac arrest duration, and on the x-axis we have patients without seizures compared to patients who did have seizures. And while cardiac arrest duration was not directly associated with seizures, it was close to statistically significant. And there does appear to be a clinically important difference. So we examined this variable a little bit more closely. To further investigate, we dichotomized patients into those who arrested for greater than 50 minutes compared to those who arrested for less than or equal to 50 minutes. We did a sensitivity analysis to pick the best cutoff time, and that was at 50 minutes. The relative risk of seizures for patients who had a cardiac arrest lasting longer than 50 minutes was 3.3. And those results were statistically significant with a p-value of 0.014. Norwood procedure was not a risk factor for seizures in our cohort. However, previous studies have identified Norwood physiology as a risk factor for post-operative seizures. Amongst the patients who did not have a cardiac arrest but developed seizures, three of the five of those patients had a Norwood surgery. When comparing non-arrest patients, Norwood procedure was associated with seizures. So 60% of those who developed seizures without a cardiac arrest had a Norwood surgery, compared to 13% of those without seizures and without a cardiac arrest. Taking these results into consideration, 14 out of the 16 patients who developed seizures either had a cardiac arrest or had Norwood procedure. If we were to only use video EEG to monitor patients with cardiac arrest or patients with a Norwood surgery, we would have only monitored 44 out of 88 patients, and we would have had a sensitivity of nearly 90% while still having an adequate specificity for capturing seizures. And our outcomes. So seizures were associated with longer ICU length of stay and lower survival to discharge. Seizures were not associated with length of overall hospital stay or length of ECMO duration regardless of survival. In conclusion, seizures are common in pediatric cardiac patients on ECMO, about one in five in our study. Patients who develop seizures tended to do so early with a median time of 40 hours following ECMO initiation. The majority of seizures were electrographic in nature only. And seizures were common in patients who had cardiac arrest or had a Norwood procedure. Therefore, we recommend monitoring all cardiac patients for at least the first 72 hours following ECMO initiation. However, in centers with more limited video EEG capability, we would recommend monitoring at least the most at-risk patients, specifically those with a cardiac arrest or Norwood surgery. I'd like to thank my mentors, Dr. Matt Friedman and Dr. Chris Astropietro, and I will do my best to try to answer any of your questions. Thank you. Clearly, there's a great group of individuals. I was supposed to moderate this, and I was a minute late, and the clock started. So great work, guys. Thank you for the excellent talk. We'll be happy to take questions. If not, I have a quick question for you, if you don't mind. You know, obviously, this being a very critically ill population, there are many confounders that go into this, be it medication specifically, antibiotics, sedatives, anxiolytics, all of which may influence the seizure activity. How did you account for these, or how did you try to tease the signal out? Because these are clearly very important findings, but they obviously have downstream effects in terms of costs and resources. So I don't think that we had a specific, in our analysis, we didn't specifically look at other medications that patients were on with regards to specific sedatives to see if that was leading to or decreasing the amount of seizures. So it'll be something that we would have to look at further. Other questions from colleagues in the audience? Thank you. Next up, we have David Hutchison. He is a pharmacist, clinically active at the University of Rochester Medical Center in Rochester, New York. And he's here to talk to us about the risk factors for thrombosis in infants receiving treatment-dosed enoxaparin following cardiac surgery. Welcome. Thank you. Good afternoon, everyone. I'd just like to take a moment to discuss a project that we worked on over the past year regarding the use of enoxaparin to prevent clots in our kids with congenital heart defects. And I have nothing to disclose on this presentation. So everyone in the room is probably aware of the detrimental effects that can occur when children, especially in the ICU setting, develop a VTE. Now when you have a specific population, such as those with congenital heart defects, the ramifications and deleterious effects can be seen, which lead to increased intensive care, mechanical ventilation, infection, paradoxical cerebrovascular accidents, and long-term consequences may include loss of access, which can prevent future catheterization surgeries and the potential for a heart transplant. Now within this special population, the range of VTE could be as high as 25-33% has been reported in the literature. So to combat this problem, the physicians in the pharmacy department at the University of Rochester Medical Center collaborated to initiate a guideline in which we use enoxaparin at treatment doses to prevent these catheter-related clots, specifically in pediatric patients. Now in order to enroll in this specific protocol, the child needs to be less than 150 days of age and be undergoing a complex congenital heart disease. And a lot of these children are going to be ones with single ventricle physiology, typically undergoing a Norwood procedure. Now to prevent the central line clot, we would initiate enoxaparin at 1.5 to 2 milligrams per kilogram subcutaneously every 12 hours. Other patients that can be enrolled in this protocol are those who have a central line and have a congenital heart disease with additional anomalies or organ dysfunction and or thrombotic risk factors. And obviously, if there's a prolonged use of a central line is anticipated. So this was a retrospective observational study in which we looked at these patients between 2012 and 2017. Now in order to be enrolled in the observation study, they needed to have a central line in place and have undergone the enoxaparin protocol that I just described. Throughout the patient's hospital stay, our typical practice is actually to get a series of imaging just to see actually if they're when a clot does present or upon admission maybe from an outside hospital to see if there's a clot already present and then also prior to discharge. So all of the infants were followed clinically throughout their hospital admission with fetus ultrasounds, echoes, and cardiac angiograms. So in this study from the five-year period, we screened 194 patients. You can see that of those 194, we had 42 that were excluded from the study with the most common reason being a non-complex congenital heart defect. We had also excluded patients where we observed a clot was observed before the initiation of the protocol. They were excluded and ultimately require systemic anticoagulation and then a series of patients that went to undergo surgery who actually were put on ECMO right after the surgery were excluded. So in our trial of these patients, we had 19 or 12.5% were diagnosed with a VTE during the observation period and we had 133 or 87.5% with no VTE observed. When you look at the most common sites of where the patients had the VTE, the inferior vena cava was the most frequent at 31.6%. Additional sites included the right femoral, right IJ, superior vena cava, the left femoral, and the right atrium, atrial septum. When you look at the demographics of these children, you can see that the median age was 8 days and predominantly these kids were single ventricle undergoing Norwood SANO procedure. You can also see that there was a statistically significant difference between the VTE group and the no VTE group in terms of the STAT 5, which is typically the patients that have the most complex congenital heart defects and have the highest mortality. And Norwood procedure is among one of the categories in the STAT 5. One of the risk factors that we obviously look for in development of a VTE is cardiac catheterization. So a lot of times these kids may come in from an outside hospital and undergo cardiac catheterization in order to determine their exact anatomy. So you can see there is no difference, any form of preoperative cardiac catheterization between the two groups. But when you look at the difference in postoperative, so it may have a surgery have occurred and the surgeons need to undergo a catheterization just to see a function of the shunt or any other types of problems that may arise. And there was statistically significance between those two groups. In terms of operative details, there is no difference in those undergoing cardiopulmonary bypass, aortic cross clamp, deep hypothermic circulation arrest. But there was a statistically significant difference between the VTE group and the no VTE group in terms of delayed chest closure. When you look at the complications, not surprisingly, the children that developed the VTE essentially had longer hospital stays. They also had a higher incidence of cardiac arrest. And because of that, most likely they warranted longer days on the ventilation. Of note, when you look at bleeding, there was only of all the patients that we screened that there was only two incidences of bleeding. And that was really defined by a mediastinal fluid drainage greater than 7 mls per kilogram per hour for two consecutive hours or 84 mls per kilogram for a 24-hour period or the need to re-explored for bleeding. It's of note that right blood cell transfusion is a risk factor for the development of clots. So we actually looked at the total number of exposures of RBC among the groups. And not surprisingly, you saw a increase in total number of red blood cell donor exposures in the OR, in the post-op. And then when you take into account all donor exposures when it comes to platelet therapy, FFP, cryoprecipitate, and red blood cells, those are all higher in the VTE group. Another risk factor that we looked into was the central lines that each of these children received in terms of placement before going to the OR and then the placement after. So you have median values of the number of lines of noted to be statistically significant and higher in the VTE group in terms of the number and the duration of central lines. When you look at when our initiation of anoxaparin therapy occurred, typically our goal is going to be on post-op day one when the cardiologist on service when the feels that there's minimal mediastinal fluid, that it's serosanguinous, and the platelets are greater than 50,000. So about on post-op day one, you had about a little over 50% start to receive anoxaparin therapy. On day two, that number rose to about 70%. And then on post-op day three, approximately 84% of the children were started on the anoxaparin therapy. There was no statistical difference between groups, between the VTE group in terms of dose of anoxaparin. And that range is from 1.5 to 2. And that could be just the preference because of the guideline. There is that range. It could be provider preference and choosing. It also could be attributed to any renal dysfunction that the child may have. But there was no difference between the two. There was, not surprisingly, a longer treatment of anoxaparin therapy in the VTE group. When you look at the initial anti-TEN-A monitoring, so we have a clinical pharmacist that works closely with these patients who's monitoring the anti-TEN-A levels. And now these levels are typically drawn four hours after the third dose. And when you look at the initial anti-TEN-A, you had lower levels or no statistical difference between the two groups. But you can see that lower levels obtained and they were not therapeutic. So I think that is room that we may not actually be given higher enough doses right in the beginning. It also can be attributed to different factors that the patient, being in ICU setting, can affect the pharmacokinetic differences between the two. But essentially, they were not statistically significant. The average anti-TEN-A level over the entire course of the treatment for the children was determined. And you can see that patients in the no VTE group fell within the therapeutic range of 0.5 to 1. And that value is actually 0.6 compared to 0.37 in the VTE group. And those were median values. When you look at the time to therapy or time to goal of the anti-TEN-A, there was no difference between the VTE group, which is the dotted line, compared to the no VTE group, which was the bold line. This curve was conducted from a survival analysis log-rank curve. And we're looking at, you can see the y-axis was the freedom from the venous thrombosis. So essentially, the first 10 days that the child was on therapy, there was no clots noted. But by day 10 to 30 is when we first started to see the visualization and the occurrence of these VTEs to occur. So this may be attributed to another central line being placed into the child. And then now you start getting into the longer hospital stays. A multivariate analysis was conducted. So despite the anaxoparent therapy, factors independently associated with development of VTE were total number of red blood cell transfusions, which was not surprising, and the mean anti-TEN-A levels. We feel our study strengths really emphasize the importance of higher initial doses of anaxoparent therapy in these children, with the protocol ranging from 1 and 1 half to 2 milligrams per kilogram per dose every 12 hours, and maybe actually increasing it closer to 2 in all of these patients, targeting therapeutic anti-TEN-A levels in that thromboprophylaxis mentality. The initiation of anaxoparent therapy at post-op day one. So we realized that this may not be feasible in all centers. So I think it's really important that each center look at their bleeding rates and do what's best in terms of, obviously, for the patient and the high risk of bleeding that can occur. The population of patients that are at highest risk for these central lines are these cardiac children. We have close monitoring and are able to do dose adjustments of the anti-TEN-A by a clinical pharmacist that's rounding every day with this population. And all the subjects had imaging studies screening for VTE prior to discharge. So in conclusion, this was the first study describing treatment dose of anaxoparent for central line VTE thromboprophylaxis in infants following cardiac surgery. The lower anti-TEN-A levels were associated with post-operative VTE and higher initial doses to achieve a therapeutic levels. Anti-TEN-A levels and red blood cell transfusions were independently associated with the VTE. And anaxoparent targeting therapeutic anti-TEN-A should be considered to reduce post-operative central line VTE with low incidence of bleeding. With that, I'd like to acknowledge my study colleagues, Dr. Sierra Stahlberg, Eileen Talley, Dr. Michaela Brown, Amy Jarula, Drs. Schwartz, Alferez, and Dr. Shillette. And with that, I'd like to open the floor to any questions. Thank you for an excellent presentation. Yes. I'm just going to repeat that question for the audience's sake. A question here is, did the team do any pharmacogenomic testing to guide therapy in these patients? There's no pharmacogenetic testing. Other question from... I have one question for you. This looks like a very important change. It's a before-after kind of a study that your team's implemented. And clearly, your results are promising within the limitations of excessive bleeding and such. Is there... Could there be just an element of selection bias, and therefore, is there a move to maybe study these prospectively and compare them in an observational fashion? Yeah, exactly. So our center is currently enrolled in a trial right now that is prospective in nature. And I believe the investigator is at New Haven in Yale. So there is an ongoing prospective trial. Fantastic. Excellent. Thank you so much. Thank you. Thank you. Our next presentation is by Dr. Linda Klein. She is the Associate Dean of Nursing at the Mayo Clinic in Rochester, Minnesota. And she will be talking to us, a description of the pharmacogenomic testing in a cohort of cardiovascular ICU surgical patients. Welcome. Thank you. Good afternoon, everyone. We're so pleased that some of you found us, even though we did not make it into the on-site program. So you are to be congratulated for coming to our session this afternoon. So I'm really pleased to be here. And I would like to express my thanks to SCCM for the Gold Star Abstract Award. We are very humbled that we received one of the top scores on our abstract. We received funding for this work from the Mayo Clinic Center for Individualized Medicine and also from internal research funds. So just by show of hands, since we're a small group, how many of you are familiar about pharmacogenomic testing? OK, a couple. Would you say your centers use pharmacogenomic testing to guide therapy? Show of hands, yes. So one. OK. So I am just going to breeze through some of these background slides. So really, the goal with pharmacogenomics testing is to identify the right drug and the right dose based on a patient's gene panel. So we are looking for variants or changes in genes that affect the effectiveness of a medication, the dose of a medication, and also the likelihood of side effects and toxicities. Currently, pharmacogenomics is used most frequently to individualize cancer treatments, heart disease therapy regimens, as well as antidepressant therapy regimens. So very briefly, PGX testing identifies genotypes and phenotypes. So the genotype is based on the individual's collection of genes. And the metabolizing pathways of an individual are identified as ultra-rapid, rapid, extensive, intermediate, poor, or an effective metabolizer. And again, this is to determine effective versus toxic doses. So in the setting of the intensive care unit, there really has been very little application of pharmacogenomics testing in the ICU, most assuredly because most ICU admissions are unplanned. There has to be ability for rapid turnaround of pharmacogenomics testing results. It's very important to keep in mind that this type of testing is to guide decision-making and management of a patient, and also including the patient's clinical parameters. it is not a magic bullet. There also has to be knowledge of pharmacogenomics among the ICU team. Cost of doing this type of test is important, and there really are not any or very few prospective studies looking at pharmacogenomics in the ICU. So to overcome some of these limitations, our team took on a single-center prospective descriptive cross-sectional study based at Mayo Clinic in Rochester, Minnesota. Our goals were to determine whether or not it was feasible to obtain the pharmacogenomics testing in a convenient sample of patients, could we describe phenotypes in the metabolizing pathways in these patients, and whether or not these identified phenotypes may influence ICU treatment. So we recruited a convenient sample of 100 patients with a planned cardiovascular ICU stay. Our goal was to determine whether or not we could get these results into the patient's EHR while they were in the intensive care unit. So patients over the age of 18 who provided their own informed consent with a planned CVICU stay were recruited. We added the pharmacogenomics testing to the pre-op labs, and we had a 90 percent consent rate over a very short period of four months. So there actually was quite a bit of patient interest in what their genes might look like. So overall, our sample consisted predominantly of mid-60-year-old white males. You can see on the slide here the mean Apache 3 scores. A majority of these patients were undergoing valve repair or valve replacement, fewer coronary artery bypass graft surgery, median very short ICU stays, although there was quite a large range in ICU stays, and very short periods of mechanical ventilation as well. So I would like to spend a little bit of time explaining the focused pharmacogenomics panel that was developed at Mayo Clinic in the personalized genomics lab. So this is a PCR-based assay that amplifies the DNA segment that contains the polymorphisms. So this focused pharmacogenomics panel identifies 10 genes, and those are listed here. Along with those 10 genes, this panel is specific to 55 general medications for which there is strong evidence in the literature with a practice guideline or from a practice society that a dose should be considered to be increased, decreased, or perhaps select a different medication. However, keep in mind that this panel is not specific to the plethora of medications administered to patients in the ICU. So what did our feasibility look like? We had a very easy time enrolling patients, about a 90% consent rate. I will probably never see those types of consent rates again in my clinical research career. And again, we really wanted these results. Could they be available while the patient was in the ICU? So overall, the median time to posting of the results from the time that the lab was run in the focused precision pharmacogenomics lab was around 52 hours median time. However, this includes weekends and holidays. 28% of patients had results onto their chart from the time the lab was drawn to within 36 hours. So those patients that were in the ICU longer or had their lab added three days before surgery, the pharmacogenomics results were posted to the electronic health record. So there were a number of issues that led to lag time in the availability of the results to the chart. The sample maybe wasn't delivered to the lab in time. These assays were run once a day. So if the cutoff time was 1PM and the lab arrived at 130, it had to wait until the next day. The blood maybe wasn't collected. There may have been an issue with equipment calibration or the patient may have needed additional sequencing. So this is the piece that we found very, very interesting. In the 100 patients, 98% of the sample had identified actionable or potentially actionable variants. Most commonly, 82% of the sample had actionable variants for warfarin. Down the line here, interestingly, clopidogrel was 33%, 25% for statins, a mainstay in the cardiovascular surgical population, and 12% for haloperidol. So very, very interesting results there. So what do these results of a feasibility descriptive study mean for cardiovascular ICU patient care? So what we learned from the single center prospective study is that you can do pharmacogenomics testing in the CVICU in non-emergent patient situations. However, the data science session this morning did have on the slide, if any of you were at the data science session this morning, omics-based care and omics results. You will be seeing those much more frequently in EHRs. So more to come on that hopefully in the future. So a majority of our sample had one or more actionable or potentially actionable identified results. And these results really do have implications for management of patient's medication therapy. Vessel patency is very important in the CVICU, as is rhythm management, pain management, and statin therapy. So what do these preliminary results mean for practice? So if you take nothing else home with you today, our point is that pharmacogenomics testing is a piece of information. It goes along with all the other labs that patients have, vital signs, hemodynamic monitoring, all those types of things. So it's really another tool in our toolbox to improve patient care. Maybe in the future, we might have point of care testing for pharmacogenomics, but that does not exist at this time. There is great consideration that needs to be given to these phenotypes of metabolizing pathways. Is there an absence of metabolizing, a reduction of metabolizing, or is a patient going to have an increase of a medication effect that may enhance side effects or lead to toxicity? So lots of considerations for clinical care of these patients. So you also need to consider the strength of the evidence that's available in the literature as far as when do you make specific medication recommendations, how do you make dosing adjustments, and again, that keen consideration for side effects and toxicity. I see this area as wide open for ICU research. The science is building in this area. If you think about the Precision Health Initiatives at the National Institutes of Health, the All of Us Initiative, so I think we're going to be seeing much more attention to all things omics, including the ICU. So the rigor of the current knowledge needs to be carefully considered. How this pharmacogenomics testing contributes to patient outcomes currently is not known. Whether or not this type of testing could contribute to improved symptom management, early pain, nausea, vomiting among surgical ICU patients is not known. The cost of pharmacogenomics testing is coming down quite dramatically, but whether or not the cost-benefit ratio of something like that adding one more test to a patient's panel, that is also not known. And we really don't know that much about knowledge among our critical care team, whether or not individuals have sufficient knowledge, skill, and comfort with pharmacogenomics to use this additional source of information. So we feel that innovative teaching strategies are needed to engage clinicians across all disciplines, so they are comfortable with pharmacogenomics. So this slide gives you a list of some of the best resources. If you're interested in learning a little bit more about how some of these actionable or potentially actionable recommendations are made, these are the top authoritative resources. So that's all I have. Thank you for your attention. Thank you. Questions from members of the audience? I have a question for you. Sure. Obviously, this is a great pilot study, and by the sound of things, your team is seeking to build on this experience and knowledge. Is there a consideration for extending it beyond the cardiac surgical ICU into either into the other ICUs or even cardiovascular patients? Because the issues you bring up are very germane to cardiology at large. Correct. Correct. You are correct. So the idea is thought about there are several patients in the Mayo Health System that have pharmacogenomics results on their chart. And now with Mayo being all on one EHR, that information is easy to find. But that's just a single center, I would say. But most prominently, I totally agree, particularly with the anticoagulant management and the stat management really has huge implications for post-ICU care. So more to come. Fantastic. Thank you. Thank you so much. Next up, we have Dr. Andrew Greenwald from the Columbia University Medical Center in New York, who is here to discuss with us molecular mechanisms of hypotension in sepsis. Welcome. Hi. Thank you all for coming this afternoon. My name is Andrew Greenwald. I'm presenting some of the work I did with my mentor, Peter Yim, on hypotension in sepsis. So a bit of background. Of course, sepsis has quite the high mortality, estimated between 30% and 50% of hospitalized patients suffer from the infliction. And about 30% to 50% of those patients die. This is despite decades of really well-done, well-intentioned trials aimed at different care bundles and packages to improve mortality. However, it's easy to say that this is rather inadequate progress in the area. Something that is also of quite an interest to us and to our group is the use of benzodiazepines in sepsis. It's well-known and well-documented that the acute use of benzodiazepines for sedation in the septic patient leads to worsened mortality. And as well as patients who chronically take benzodiazepines and people who suffer from chronic alcohol abuse have worsened mortality in sepsis. And this data is largely driven from studies that show that alternative sedation strategies, mostly based on Presidex and Fentanyl, show some better mortality in the sedation of the septic patient. We are anesthesiologists as well in our practice. Something that we know clinically is that in the septic patient, propofol creates quite the worsened vasodilatation during induction of anesthesia for a septic patient. And the direct mechanism of that is somewhat unknown and of great interest to us. The GABA-A receptor is the main target of benzodiazepines. It's a heteropentameric ligand-gated chloride channel. And the function of that depends on the resting membrane potential for chloride. In the cell in the resting state, activation of the GABA-A receptor results in a depolarization of the cell. And in the activated cell, activation of the GABA-A receptor results in hyperpolarization of the cell. In sepsis, the GABA-A receptor is broadly overexpressed throughout many tissues. This is thought to have a considerable contribution to the increased mortality in sepsis. The mechanism that which the GABA-A receptor is upregulated in sepsis is thought to be via interleukin-1-beta signaling. In this paper here, a group demonstrated in neurons the mechanism by which interleukin-1-beta drives cell surface expression of GABA-A receptor signaling quite elegantly. The effect of the GABA-A receptor on the vasculature is a bit unclear as well. GABA-A is thought to be classically a neurotransmitter. But as the GABA-A receptor is expressed on many tissues throughout the body, it can have many effects all throughout. It's thought that any effect of GABA-A activation signaling on the vasculature was thought to be from a decrease in central adrenergic signaling. And further, there is activation of GABA-A receptors on leukocytes is associated with some immune suppression. However, a lot of the GABA-A effect on the vasculature is somewhat unknown and a lingering question for us. The hypothesis for our study was that if GABA-A activation can cause hypotension via a direct effect on vascular smooth muscle, and that this effect would be worsened in the septic state. To probe our hypothesis, we sought to also determine how interleukin-1 beta signaling drives GABA-A expression on vascular smooth muscle cells. So we grew human omental artery smooth muscle cells in cell culture and treated them with interleukin-1 beta. And the figure here on the left-hand column is our control cells that were not treated with interleukin-1 beta at 5 and at 7 days. And on the right is our cells that were treated with interleukin-1 beta at 5 and at 7 days. What's projecting a bit poorly is the immunostaining that we did. We then treated the cells with an anti-GABA-A alpha-5 subunit antibody, which immunofluoresces in green here. What we were able to demonstrate is that our interleukin-1 beta treatment, as in neurons as previously described, also increases cell surface expression of the GABA-A receptor. In this experiment, we did not treat our cells with any detergent. So any expression of any antibody activity is solely expressed on the cell surface. To further probe our hypothesis, we turned to what's called a Flipper, or a fluorescent light plate reader assay, which is a high throughput assay that uses a membrane potential reporter dye to determine any cell membrane potential change from any ligand activation. So more directly to probe our hypothesis about GABA-A effect on the vascular smooth muscle, this is the result of some of our Flipper work. In figure A, along the x-axis is the time in our assay. At about 15 seconds there is the injection of our, excuse me, 10 seconds was the injection of our test compound. And in the y-axis is the membrane potential change as reported by our reporter system. What you can see is that potassium chloride generates a pretty significant membrane potential depolarization of the membrane potential, as does propofol. And then what we also looked to see if we could hyperpolarize our cells as well, and we used something, a compound called NS1619, which is a nonspecific cation channel ligand. On figure B, in the x-axis, are the various compounds that we tested. And then on the y-axis is the quantitative membrane potential change that we see. What we were able to demonstrate is that there is a statistically significant increase in, or statistically significant membrane depolarization with propofol, as well as propofol and bicuculline, along with potassium chloride, and then a hyperpolarization with NS1619. We then further sought to determine the effect of calcium channel agonists on vascular smooth muscle activity in the context of GABA-A activation. So in the figure demonstrated here on the left, figure C, we were looking for reporting of intracellular calcium. So time is noted on the x-axis, and then the reporter effect in the y-axis. You could denote at 50 seconds, endothelin is injected into our cells. And then we see concurrently a nice rise in the intracellular calcium potential, most strongly in our DMSO, just our control test. And then that's an effect that's a bit attenuated by our GABA ligands, propofol and mucemol. And the figure on the right, figure D, again in the x-axis, we have various concentrations of our GABA ligands, along with endothelin, and then the reported change in intracellular calcium, as reported there. What's important to note here is that this does not seem to be what we notice to be a dose-dependent related effect. Whether it's at 1 tenth micromolar, 1 micromolar, or 10 micromolar, we still do see a statistically significant decrease in intracellular calcium release in response to endothelin in cells pretreated with a GABA agonist. The next question we sought to ask ourselves is, how can we make this more clinically relevant? Mucemol is not something that's used clinically, endothelin is also not a clinically used vasopressor. So we decided to focus more on norepinephrine, norepinephrine being well described as the primary vasopressor of choice in septic shock, with a considerable amount of research to support this supposition. And this slide, we see here the norepinephrine-induced increase in intracellular calcium in human-cultured vascular smooth muscle cells. Along the x-axis is the various compounds that we test-injected, HBSS is Hanks-Buffert's saline solution, just a sort of electrochemically neutral solution to dissolve our test compounds in. And then on the y-axis, or excuse me, then ending is norepinephrine, mucemol as a laboratory GABA agonist, and propofol being one of our test compounds. On the y-axis is the intracellular calcium released as reported by the Flipper membrane activation unit. What we do see is that injection of norepinephrine does cause a statistically significant increase in intracellular calcium, as expected. And then this effect is attenuated in the third column by treatment with mucemol. And again, it is decreased by treatment with propofol as well. Our next thought was that if GABA A activation can decrease norepinephrine-mediated calcium release, this must be something that can be augmented in the septic state. So we treated our cells with interleukin-1 beta, and then exposed them to norepinephrine as well. So on the x-axis on this figure is our two different time points, blue being our control, and then in red are time point treatments of interleukin-1 beta. We see at one hour between our control and our interleukin-1 beta is that there is not a decrease in norepinephrine-mediated calcium release. However, at three days of treatment with interleukin-1 beta, we do see a decremental decrease of calcium released in response to norepinephrine signaling. The next question we had was that if we are able to knock down the effect of the intracellular calcium release in response to norepinephrine in the septic state, is this something that can be enhanced with GABA A signaling? And this would go to what we observed clinically in that patients who are septic have an enhanced vasodilatory response to GABA A ligands. So again, on our x-axis, we have our various injection compounds. The blue being our treatment groups that are not treated with interleukin-1 beta, and the red group being the groups that are treated with three days of interleukin-1 beta. In the leftmost pair of figures, there is no GABA A ligand injection, and we do see about a 40% decremental decrease in our intracellular calcium. On our y-axis here is the percent of baseline intracellular calcium released. The baseline is set to be our control group that is no interleukin-1 beta, and then our standard dose of norepinephrine. So with just three days of interleukin-1 beta, we decrease the amount of calcium that's secreted in response to norepinephrine stimulation. Then later with mucimol, our laboratory GABA ligand, and with propofol, we do see a decremental decrease in calcium secreted in response to norepinephrine stimulation. However, it's not statistically significant. This is something that we attribute to in the septic state. Our hypothesis is that so much of the calcium release has already been knocked down by interleukin-1 beta treatment, and this is not something that we can affect with our GABA ligands as classically. So the conclusions that we have are in our human vascular smooth muscle cells, endothelin induced calcium elevation is attenuated by a GABA-A ligand. However, norepinephrine induced calcium release is not something that's attenuated by a GABA-A ligand. Prolonged exposure to interleukin-1 beta did increase the cell surface expression of our GABA-A alpha-5 subunit on the plasma membrane, which was something that we translated from work in neurons. And our preliminary studies do demonstrate a trend towards reduction in norepinephrine induced calcium release after prolonged interleukin-1 beta, but nothing definitive as of yet. Some of the future directions that we're looking to go towards is to try and determine how come we can see this effect in endothelin but not in norepinephrine? Is there something different in their secondary messenger system or the like? The other big question that we have is we're simulating sepsis solely by treating cells with interleukin-1 beta, which may not be the perfect model to simulate what we see clinically. The presentation of sepsis from a urinary source or from an intrapulmonary source are two very different patient phenotypes, but we're kind of grouping them together just with interleukin-1 beta treatment here, which kind of leads. And of course, sepsis does have a very heterogeneous disease phenotype as well, and we are kind of focusing on this small, narrow laboratory model. One of our questions is the appropriateness of that as well and how we can optimize that. Thank you for your excellent presentation. We'll take questions from the audience at this time. Thank you. Thank you to all our excellent presenters, and thank you to members of the audience for being here this late despite being in Orlando. Appreciate everybody's time and attention, and have a good rest of your evening. . Thanks for watching! Thank you for watching. Thank you for watching. Thank you. Good afternoon. My name is Jeremiah Dubey. I'm a critical care pharmacist at UC Davis Medical Center. And I'll be the moderator for this session. Really quickly, we're moving the second session as scheduled to the first, while we fix the slides for the first. Nothing to see over there. I would say move to the center of the row, but it seems a little silly. Sit wherever you like. The exit doors are behind you. Please silence your cell phone. And you can get CE for this talk through the app. Speakers will disclose their conflicts. You may take photographs, but please do not publish them. And if you would like, you can share or tweet at hashtag CCC49. The first speaker is Uptal Balala. He will be presenting on pre-arrest, is that right, pre-arrest echocardiographic changes. Could they predict cardiac arrest in hospitalized children? Thank you for this wonderful opportunity. Good afternoon, everyone. I know it's almost close to end of the day, but thanks for coming and listening to our talk. I'm honored and privileged to be invited for the STAR Research Award presentation. First of all, I would like to acknowledge my students who have put in a lot of hard work on this project, Charleston West, who just came in, and Jose Augustin Gutierrez Gomez, a student from University of UDEM in Mexico, Monterrey. And they put in a lot of work on analyzing EKG together, and so thanks to them. I do not have any disclosure to report. Literature related to adult sudden cardiac death suggests that pre-arrest changes in EKG variables can be used to predict cardiac arrest. There is no similar data on pre-arrest changes in the EKG variables in children prior to in-hospital cardiac arrest. So in this particular study, the adult intensivists were able to demonstrate certain EKG changes before the arrest. But in this study, they looked at the EKG changes 24 hours prior to arrest. Sinus pauses and AV block were commonly seen prior to PEA and bradycystole caused by respiratory events, which suggests that hypoxia can cause impaired nodal function prior to leading to cardiac arrest. RR and QRS prolongation were primarily found in those which were related to multi-organ dysfunction, and that suggested myocardial conduction delay caused primarily by sequelae of severe multi-organ disease, including hypotension. These findings agree with experimental data showing that hypoxia in guinea pig and rabbit hearts display delayed conduction through sinus node and AV node. In this particular study, certain specific EKG changes were reported prior to in-hospital cardiac arrest in adult. Again, there is some data which suggests that pre-arrest changes in the EKG variables can be used to predict cardiac arrest in adults, but there is no similar data on pre-arrest changes in EKG variables in children prior to in-hospital cardiac arrest. So our hypothesis was specific pre-arrest EKG changes do occur prior to in-hospital cardiac arrest in children. Patient characteristics and central monitor data on pre-arrest EKG variables were collected for pediatric in-hospital cardiac arrest occurring between December 2017 and December 2019 at our Children's Hospital of San Antonio. Pre-arrest cardiac rhythm and EKG parameters were evaluated in 30-second epochs during five minutes prior to initiation of chest compressions. EKG intervals which we reviewed were RR, PR, QRS, QT, QTC, and J-point and ST-point positions. So again, this is just a clip from the central monitor which was downloaded pre-arrest. And then we looked at specific intervals at 30-second epochs in five-minute time frame before the chest compressions were initiated. The initial and peak EKG intervals and the initial and nearer values of J-point and ST and T-wave positions within five minutes prior to arrest were compared using two-tailed student T-test with a significant P-value of less than or equal to 0.05. A total of 335 EKG tracings were reviewed from 37 in-hospital cardiac arrest among 29 children between this time frame. The median age was five months. The documented cardiac arrest rhythms were PEA in 21, bradycardia in 12, asystole in three, and ventricular fibrillation in one. So what were the findings? We saw that RR interval increased from 0.52 to 1.38 median value, PR interval increased from 0.12 to 0.16 median, QRS complex also showed increase from 0.08 to 0.12, and QT and QTC and QT and QTC also showed some change, as you can see, and there was depression of the J-point and the ST segment from the initial time point, which was five minutes before the compressions were begun. So there was a significant change in all EKG variables except J-point and ST segment measurements within five minutes prior to the arrest. RR interval prolongation was the most frequent and earliest EKG change prior to arrest. So we wanted to look at if RR interval change was significant, was it just because they were becoming bradycardic? And so we wanted to look at what happened first, the change in the RR interval or they were simultaneous with the heart rate significantly going down. And so we looked at the median time between the peak RR interval and drop in heart rate below 60, which would alert the team to initiate chest compression or do something. And so we found that it was 30 second with IQR 0 to 90 seconds. These were our outcomes. So we found that 88% of our patients had ROSC and survival to ICU discharge. The median time to ROSC was 7 minute 30 seconds. We found that 27%, almost one third of our patients had hypokalemia within 24 hours prior to the event. And only 2.7% cases had hyperkalemia within that time frame. Hypokalcemia was seen in 8.1% of cases 24 hours prior to the event. And what was very significant was that almost half of the patients were exposed to dexamethasone within 24 hours prior to the event. Again, we looked at these things because we wanted to make sure that these were not the major confounders. EKG trends can be identified prior to in-hospital cardiac arrest in children. These EKG changes could be potentially used in creating innovative cardiac arrest alert system for monitors. What are some of the future directions? We would like to evaluate EKG intervals at two other time points, that is at one hour and six hours prior to arrest in same patient in order to get some control information. And we would like to correlate these changes with any dyselectrolytemia or exposure to medications which are known to cause rhythm changes and bradycardia. We would like to capture and analyze more cardiac arrest events to look for further trends. So we are actually working with PeddiResQ multi-center group, which is the pediatric group which is looking at quality of resuscitation across the nation and even across the globe. And so we are looking at capturing this particular part of the data from all those centers. We would also like to look at the heart rate variability to predict any in-hospital cardiac arrest. Because again, when we saw RR interval change as one of the commonest and most prominent change, we believe that heart rate variability could be one of the important components which could help us predict in-hospital cardiac arrest. And we would like to create alert system in monitor based on the intervals. But of course, it will need some case control study and much larger patient population to look at, which we are already working on. That ends my talk. Thank you so much. And I'm happy to welcome questions. I have a quick question. How do you account for the association with the use of dexamethamidine in the cardiac arrest that you saw? So it's a very good question. I think that is the reason we would like to get some control information in the same patient, meaning one hour and six hours prior to arrest, whether some of these EKG interval changes which we saw, were they related to medication or electrolyte imbalance? But at least from what data we have, we know that five minutes prior to compressions were begun, they had particular EKG interval measurements. And then as they started deteriorating because of either hypoxemia and or hypotension or combination, there was significant change in the interval. So I think that's something very important. But yes, we would like to look at control data. Are there other questions? OK. Thank you. Thank you. Our next speaker is Lin Liang from Virginia Commonwealth University. He will be presenting on the effects of NLRP3, lymphosome blockade and post-resuscitation cerebral function. Thank you. Good afternoon, ladies and gentlemen. I'm Dr. Liang. I'm presenting this slide for Guanghui Zheng. He is the author of this study. And we all come from the Weill Institute of Emergency and Critical Research and St. Joseph Memorial Hospital. It's my great honor to be here to present our study with you. So let's get started. And the study was performed at Weill Institute of Emergency and Critical Care Research. And the protocol approved by the IACUC of Virginia Commonwealth University. We disclosed that we do not have any potential conflicts of interest. And use of MCC950 is investigation only. No relevant financial relationships can exist. According to our variety of studies, the sudden cardiac arrest is still one of the leading cause of death and remains a major public health issue. Despite the major advance in medical science and techniques in the past decade, the survival rate in the medical science and in patients with cardiac arrest remains low. And most of their knowledgeable functions are poor, even after successful resuscitation. Several gaps in our knowledge of mechanism on cardiac arrest are in the way. Among them, the global ischemia reperfusion injury of the brain, between cardiac arrest and CPR, is a significant contributor to post-cardiac arrest neurological dysfunction. And also, mobility and mortality of these patients. Growing evidence suggests that inflammatory response with overproductions of IR1-beta contributes to the cerebral ischemia reperfusion injury. Furthermore, IR1-beta signaling is regulated by the NRP3 inflammasome. The NRP3 inflammasome is activated by assembling of NRP3, ASC, and caspase-1. Then the assembled complex cleave, pro-cytokine IR1-beta, IR18, to the activated form. These cytokines act as effectors, leading to one type of cell death, pyrosis. Studies have shown that the NRP3 inflammasome is activated after regional cerebral ischemia reperfusion is solved, and in turn, exacerbate the ischemia reperfusion injury. However, there's no evidence regarding the role of NRP3 inflammasome blockade with MCC950 in the global ischemia reperfusion injury of the cardiac arrest and CPR. In this context, we sought to investigate the effects of the selective NRP3 inflammasome inhibitor, MCC950, on post-resuscitation plasma IR1-beta, cerebral microcirculation, brain edemia, 24-hour survival rate, and neurological functions in the random model of cardiac arrest and CPR. After animal preparation, 36 male SDRs were randomized into three groups, one MCC950 group, two control group, and three same group. Each group was subdivided into a 6-hour non-survival group and a 24-hour survival group. Cardiac arrest and ventricular vibration was induced through a guide-wide advance from the right jungle wing into the right ventricle. After 6 minutes on 3DVF and 8 minutes CPR, the rest in the MCC950 group received one dose of MCC IP injections immediately after arrest. Instead, vehicle injections with water, sterile water, was given in the control group. The rest in the same group underwent the same surgical procedures without VF and CPR. In the non-survival group, the cerebral microcirculation was mentioned at around 6 hours. Then the rest was sacrificed and the brain was obtained for the brain water content measurement to assess cerebral edemia. Capillomere survival plot and neurological deficit score were used for evaluating the neurological outcomes in survival rates. In addition, the plasma R1 beta was mentioned for evaluating the systemic inflammatory response and NSC for representing the severity of brain injury using the ELISA method, according to manufacturer's instructions. Here are what we found. Figure 1A are representative graphs of cerebral microcirculations in these three groups. From the top to the bottom is same MCC950 and control. And from the figure, we can find that the density of micro vessels in the control group was significantly decreased after cardiac arrest and CPR when compared to the same. And use of the MCC950 can preserve the density of micro vessels, which was similar with the same. In addition, the quantitative index of microcirculation, including the preferred vessel density PVD and microcircular flow index MFI was found to decrease after cardiac arrest and CPR in the control group when compared to the same group. And MCC950 interference can restore the PVD and MFI to their normal level. And figure 2 here show that the brain water contained in the control group was increased significantly after cardiac arrest and CPR in the control group when compared to the same, which indicate the brain edema at rough six hour. The good news is MCC950 can prevent the brain from edema. We also find that the NSC, which indicate the severity of brain injury, was decreased, was increased significantly after cardiac arrest and CPR in the control group, both at rough six hours and 24 hours. We further address whether the protective, the MCC950 can prevent the brain from edema. MCC950 can inhibit the increase of NSC. We further address whether the protective role of MCC950 is through the inhibition of NRP3 inflammasome. We clarify our hypothesis that the increase of IL-1 beta after cardiac arrest and CPR in the control group, both at rough six hours and 24 hours, was blocked by MCC950. Furthermore, the 24-hour survival rate of the ROS in the MCC950 group was significantly improved. All of the six rats in the MCC950 group survived, and only one rat in the uncontrolled group survived at ROS 24-hour. In addition, the neurological damage score was also improved significantly in the MCC950 group when compared to control. From the video, we can see that the only survival rate in the control group was had poor neurological performance, and the rats in the MCC950 group has much better neurological outcome than overall performance. From the video, we can see that you can take food and water by itself, just like the rats in the same group. In conclusion, selective inhibition of the NRP3 inflammasome with MCC950 reduced systemic level of IL-1 beta, preserves cerebral microcirculation, mitigates cerebral edema, improves 24-hour survival rate, and neurological functions in our rat model of cardiac arrest and CPR. Therefore, our research findings suggest that MCC950 may be a promising therapeutic candidate for lessening post-resuscitation cerebral ischemia reperfusion injury. That's all. Thanks for listening. And any questions? Thomas Trebek, Saffer Center for Resuscitation Research at the University of Pittsburgh. You reported systemic levels of IL-1 beta. Did you look at cerebral levels of IL-1 beta at probably different brain regions, or were you focused on the systemic effects of the drug? You mean that we… Actually, we mentioned the plasma IL-1 beta. Sorry, can you repeat the question again? He's asking about systemic inflammation parameters versus the brain. We only looked at wet-dry-wet ratio in the brain and only looked at systemic IL-1 beta. Actually, we did not mention the IL-1 beta in the brain. Thank you. Can you speak to future directions for this research? Okay, thanks for the advice. We are working on our job to… further working on the IL-1 beta and also the NRP3 inflammation with MCC950. We selected in the translation medicine opinion, we are focused on the methods of giving the drug, and also we are starting the combination of MCC950 with other drugs to explore the effects of the inflammatory response. Yeah, thank you. Thank you, Dr. Ling. Thank you. Our next speaker is Brandon Michael Henry from Cincinnati Children's. He will be presenting Outcomes of Repeated Runs of Extracorporeal Membrane Oxygenation Support in Adults, Analysis of the ELSO Registry. Good evening, and thank you for the opportunity to present our study results with you today. I have no disclosures, so I'll dive right in. Cardiopulmonary recovery and readiness to decannulate from ECMO is not objective or success guaranteed. So, if a patient deteriorates post-decannulation, a subsequent run of ECMO may be considered. However, repeated runs of ECMO have difficulties, one of which, for example, recannulation. Outcomes of multiple runs of ECMO have been reported to have variable success in pediatric patients. However, data is limited in survival benefit in adults remains unclear. So, we investigated survival rates in adult patients receiving multiple runs of ECMO during the same hospitalization, and we aim to identify which factors influence mortality and survival in order to enable for better patient selection in the future. We collected our data from the Extracorporeal Life Support Organization Registry, also known as ELSO. ELSO is a prospective collection of cases from institutions around the world of ECMO runs. We took all adults greater than 18 years of age receiving two or more runs of ECMO during the same hospitalization for any support type. We then compared outcomes and complication rates based on the type of support, and then evaluated the factors influencing outcomes in multivariate logistic regression. So here we have a flow chart that shows our study selection, our patient selection and patient identification. Of the 51,061 adult runs of ECMO in the registry as of January 2019, 2,288 of those, or 4.5%, were part of multiple ECMO runs. That equated to 1,144 patients. 1,041 of those had exactly two runs of ECMO during the same hospitalization. And 103 had three or more runs of ECMO during the same hospitalization. So here we see patient demographics and characteristics broken down by survival, with patients who died on the left and those who were discharged alive on the right. First thing you'll notice is that patients who were discharged alive were a bit younger. They had a median age of 47.7 compared to those who died who were 54.8. That intrigued us, so we broke down the data into age groups. You see that patients who were then 18 to 30 had a survival rate of 51.7%, and it dropped thereafter. So we can see that with increasing age, survival decreases. And when you reach age 51 and above, survival was all the way down to 30.9%. And we'll come back to that when we get to the multivariate analysis. When we looked at mode of ECMO support on the first run, we see that patients who received VV ECMO had the highest rate of survival at 52%, whereas patients who received VA ECMO had much lower survival at 33%. There were no other differences in factors such as weight, sex, or hours on ECMO. So we'll jump into survival. So here we have, on the left, is the number of runs that a patient received, broken into two runs or three or more runs. The second column has the support type on the first run, number of patients, survival to discharge for those patients. And all the way on the right, we have the ELSA registry survival to discharge, which basically we can equate to a single-run patient. So for a patient who received pulmonary support on the first run, regardless of what they received in the second run, had a survival to discharge of 51.7%, about a 7% difference from the ELSA registry data. In patients who underwent a first run of cardiac support, it dropped a bit to about 33.7%. In a patient who received first-run ECPR, the survival rate was 32%, which is surprising. It's actually 3% higher than what we consider the single-run ELSA registry data. When we look at three runs, again, you'll notice that pulmonary patients who received pulmonary support in the first run, regardless of what they received on their other runs, had a survival to discharge of 37.2%, which was the highest for that group, but dropped to 27.7% and 15.4% for ECPR. So here we have survival to discharge by decade of ECMO support. You can see that the number of patients on the bottom there increased from 27 in the 90s to 82 in the 2000s, and over 1,000 in the past decade. So that made up the majority of the cases. Much of that is probably attributed to the fact that ECMO has become much more utilized in adult patients. Over the past decade, however, as you'll see that there was really no change in survival. It went from 37% in the 90s to only 40.1% in the past decade. OK, so here we'll break down survival data into a bit more detail. Here we have survival to discharge for two ECMO runs by type of support on the first run and second run. So we have the first run column and the second run column all the way on the right. So we can follow, then, a sample patient. So a patient who went from pulmonary support to pulmonary support on a second run had a survival to discharge of 55.9%. And that's the highest rate of survival we have in the entire cohort of patients. When a patient goes from pulmonary to cardiac, you can see that it drops. So it goes down to 37.5% for a patient going from first-run pulmonary to second-run cardiac. And same with ECPR. It drops again to 22.2%. But consistently, we saw throughout the cohorts that patients who received a second run of pulmonary support or a final run of pulmonary support when we get the three-run patients had the best survival. So if we look at cardiac patients, patients who had heart failure and went from cardiac support to cardiac support again, they only had a survival rate of 31%. However, patients who went under cardiac support and then needed pulmonary support as a second run had a better survival at 10% higher. And that's consistent, as you see throughout the data set, that pulmonary support in the second run offers the best survival. Here we see the same data, survival of the discharge for three or more ECMO runs by type of support on the first run and the last run. So this is for the three-run patients, the 103. So it's a bit limited data set. But it's broken down in a similar way. So we have what they had on the first run and then what they had on the last run, regardless of what they had in between. So patients who had pulmonary support on the first run and pulmonary support on the last run had the best survival at 44.4%. We see a similar trend when we get to cardiac patients. We see a similar trend when we get to cardiac patients, is that the first-run cardiac group who went on to need a final run of pulmonary support having a 33.3% survival as compared to 27.6% for the cardiac-to-cardiac. So next, we want to know what factors we should look for in selecting candidates for multiple runs of ECMO. So here we have first-run variables predicting death on the final run. So for first-run pulmonary support, we see that age is associated with mortality. It's a very small odds ratio, but when you break it down to the fact that for every one-year increase in age is associated with a 3% increase in odds, mortality becomes a bit significant. An inotrope requirement was also associated with first-run inotrope requirements associated with mortality on the final run. For eCPR, again, we notice the age trend, but hemorrhagic complications really stood out here with an odds ratio of 116. So that was a pretty big factor associated with mortality. For first-run cardiac support, there was nothing other than age, which is, as you can see, a trend throughout the first-run variables. And interestingly for us, there was no first-run cardiac or pulmonary support ECMO complication that was associated with mortality on the final run in multivariate analysis. So here we have now final-run variables predicting death on the final run. So these are complications of the final run that then indeed produced mortality. We can see that for pulmonary support, again, we have the age factor coming in, but also the metabolic complications, neurologic complications, pulmonary complications. For first-run cardiac support, we have the infectious metabolic complications that come into play. And for the eCPR, we have, again, we can point out those hemorrhagic complications endowing very bad fortune on those patients. So in conclusion, adults who deteriorate post-decannulation and require an additional run for pulmonary support are the ideal candidates for multi-run ECMO. Outcomes in patients requiring repeated cardiac or eCPR support runs, especially in older patients, are poor. No first-run pulmonary or cardiac support complication decreases odds of final-run survival. So those we probably should not take into account when screening candidates for additional runs of ECMO. And we'd advise that more than two runs of ECMO for non-pulmonary indications should be undertaken with caution, especially in the setting of cardiopulmonary resuscitation. Thank you, and I'm happy to take any questions. I have one quick question. It looked like your neuromuscular blockade had a protective effect. Yeah. Do you think that was because there was colinearity with the pulmonary indication for ECMO? It's certainly a possibility. It's something that we think we have to look at a bit more. So we're gonna take the data set as it is now and then look at specific groups of patients. So we're gonna take the pulmonary group, analyze them from a separate cohort from the cardiac to cardiac group, because the patients are heterogeneous. They're not greatly able, we can't really compare them exactly. So hopefully that will deal with some of that colinearity like you noticed in that data set. Yes. I have one kind of technical question with the data gathering. So were all of the separate, like a VV and VA runs temporally separate, or were some of those reported as separate runs for like we started with VV and then added an A-limb or started with VA and then added a V-limb for north-southing? So are you asking in terms of? Like did they consider those two separate runs or were all of these decannulation? No, they're all decannulation in a new run. So the question was, how did your referee change from VA to VV ECMO? Yeah, changes are noted. If we go back to the slide, we have like a conversion and those are particular factors there. So those are noted. They're not considered to be part of the same or under a multi-run. So this is patient, but we definitely decannulated and then had to put back on ECMO. Thank you. Thank you. And our last speaker was also our first speaker. Uptal Balala is going to now present on extracorporeal cardiopulmonary resuscitation in the pediatric cardiac population. Once again, thank you for the invitation. Again, it's an honor to be here for STAR Research Award presentation. So we'll talk about eCPR in the pediatric cardiac population, which is our analysis from the virtual pediatric system database. And again, thanks to my collaborators, Harvey Lassa, Ravi Thyagarajan, and Daniel Cuffey. We have put in a lot of good efforts together and we have some good data here to share. I do not have any financial disclosure or relationship to report. eCPR has been utilized as a successful rescue modality for failed conventional CPR after pediatric in-hospital cardiac arrest among children with congenital and acquired heart disease for over 30 years. In North America, majority of eCPR occurs in cardiac ICU. Now let's look at eCPR utilization in cardiac ICUs. So again, this report from Jeffrey Alton and co-authors on epidemiology and outcomes of cardiac arrest in pediatric cardiac ICUs showed that almost one third of cardiac arrest deployed eCPR in the cardiac ICUs. And 38.6% survived to hospital discharge. What about unit versus patient factors? So Harvey Lassa from Texas Children's Hospital and co-authors published the report from Get With The Guideline Registry on eCPR in pediatric in-hospital cardiac arrest and they showed that majority eCPR were in cardiac patients, 79%. But what was not reported and what is not known is was there any unit characteristic cardiac ICU versus mixed or PICU which were related to the outcomes. Also Tia Raymond from Dallas, she reported from National Registry of Cardiopulmonary Resuscitation on outcomes among neonates, infants and children after eCPR and 3.8% eCPR use was reported from the registry. And there was a significantly improved survival after eCPR among cardiac patients as compared to non-cardiac patients. So we have some existing literature which provides limited data on impact of both patient and ICU level characteristics on outcomes after eCPR in the pediatric cardiac population. So additional studies are needed to stratify the heterogeneous cardiac population that is cardiac surgical versus cardiac medical in order to fully understand the risk factors for survival. So leveraging standardized cardiac and mixed pediatric ICU clinical data from virtual pediatric system or VPS, we aim to understand the impact of patient and hospital level characteristics on outcomes of eCPR in pediatric cardiac surgical and cardiac medical population. So we hypothesize that ICU type, that is CICU versus mixed or PICU is associated with eCPR survival and favorable neurologic outcomes of medical and surgical cardiac patients resuscitated with eCPR in the VPS database. Our inclusion criteria were ICU encounters which were reported in the VPS between January 1st, 2010 to 31st, December, 2018. And all the encounters which were submitted to VPS in this timeframe which had ECMO or cardiopulmonary bypass as a procedure reported within the two hours of initiation of CPR. We excluded trauma cases and those in which ECMO occurred more than two hours after the initiation of CPR. Our primary outcome was survival to hospital discharge and the secondary outcomes were favorable neurologic outcome that is pediatric cerebral performance category, PCPC, of one, two, three or no change from the admission baseline and this was based on the previous report from Robert Berg and co-authors. Characteristics and outcomes were compared by diagnosis using quantile regression, chi-square test and Fisher's exact test with a significant P-value less than or equal to 0.05. Independent and multivariable mixed effect logistic regression accounting for correlation at patient and hospital level were assessed for association between characteristics or cardiac groups and survival as well as favorable PCPC. So we had more than a million ICU encounters reported in VPS between 2010 to 2018 amongst 7,334,756 patients. Of these, 12,931 cardiac arrest events were reported in VPS amongst 11,853 patients and of these, only 671 eCPR events were identified based on our inclusion criteria so the eCPR rate within the VPS database during that time frame was 0.07% only. Majority were cardiac surgical as compared to cardiac medical patients and as you can see, majority are patients who are less than one year age and the children comprised 15% of patients and adolescent, only 6% of patients. There were 52% patients which were male and the congenital heart disease was seen in 84% of the patients. What were our outcomes? So survival to hospital discharge was reported in 51% of cases and favorable neurologic outcome was reported in 48% of survivors. Of course, the limitation we had was amongst 31 cases, the PCPC was missing. So this is some demographic information and some outcome information. So median weight was 4.2 kilo, median CPR duration prior to ECMO was 45 minute which is pretty similar to what has been reported by Harvey Lassa and co-authors previously. Median ECMO duration was 90.7 hours, median mechanical ventilation duration was 13.8 days and median ICU length of stay was 21.2 days. So this is a comparative table between the patients who are managed in cardiac ICU versus those who are managed in mixed PICU. As you can see, there were no significant difference with patient level characteristics, that is weight, age, gender, race amongst the two groups. As far as the CPR duration prior to ECMO, the patients who are managed in cardiac ICU had significantly longer CPR duration before going on ECMO as compared to those who are managed in mixed or PICU. And then the other characteristics such as ECMO duration, mechanical ventilator days and physical length of stay were not significantly different between the patients who are managed in cardiac ICU versus mixed ICU. Similarly, the survival and favorable neurologic outcome were not significantly different between the two groups. When we look at the survivors versus non-survivors, as you can see, the patient level characteristics except the race, the weight, age, gender did not have any significant difference between the two groups. The white race had better survival as compared to other races. Looking at the ICU level characteristics, so again, as we can see, cardiac surgical patients had better survival as compared to cardiac medical patients, but the other ICU characteristics such as ICU type, ICU bed strength, academic status, and whether the procedure was done weekend versus weekdays, and the CPR duration prior to ECMO were not significantly different between the two groups. When we look at the favorable neurologic outcome, so we can see that again, white race was associated with favorable neurologic outcome, but the other patient characteristics did not have any impact on the neurologic outcome. When we look at the ICU characteristics, again, as we can see that the cardiac surgical patients had favorable neurologic outcome as compared to cardiac medical patients, and those patients who had ECPR done during the weekdays had better neurologic outcome as compared to those who had ECPR done during the weekends, and the CPR duration was also significantly better in those who had favorable neurologic outcome. Multi-variable logistic regression showed that survival to hospital discharge was significantly associated with illness category, cardiac surgical versus cardiac medical, while the ICU type and ICU bed capacity were not. So the surgical cardiac patients were more likely to survive to hospital discharge compared to medical cardiac patients. White race and ECPR timing non-weekend were only factors significantly associated with favorable neurologic outcome, and you can see the odds ratio for the two. So to summarize, survival to discharge was not affected by age, race, gender, ICU bed capacity, ECPR timing, and academic status. Favorable neurologic outcome also was not affected by age, race, gender, illness category, ICU type, and ICU bed strength. Patient-level factors potentially outweigh the organizational factors. Of course, we had some limitations here, so again, this was the study, the report based on the large database with inherent limitations. The way we defined ECPR cohort could also potentially challenge the true ECPR population, and we did not have information on CPR quality, post-resuscitation care, or complications. So we conclude that the overall incidence of ECPR among critically ill children with cardiac disease in VPS database is low, with significant variability in utilization across ICUs of varying size, type, and academic standing. Patient-level characteristics, surgical were more than medical, impact survival to hospital discharge and neurologic outcome after ECPR more than the ICU-type characteristics. Thank you. You come to, yeah, sure. Hi. Thank you for the presentation. A quick question. What was the cause of death of the patient who did not make it out of the hospital? Did you analyze what did they die of? That's a good question. We don't have that granular data, but we know the majority are related to neurologic, meaning brain death. Is it the complication of the ECPR itself or post ECLS management? One second. I might have that slide. I don't have that supplemental slide, but a significant proportion of patients who died had the PCVC category of six, meaning brain death. So again, we know that because of the way CPR is delivered and how the ECMO is deployed, they are at higher risk of neurologic injury. And one of the commonest cause of death in these patient populations is brain death. We have the data, so we should be able to look at it more granularly and provide that information. Thank you. Good question. Thank you. I have a couple of questions, but it might just be one. You might be able to answer it in one part. Why was the rate of ECPR utilization so low, and what are the barriers to utilization? It's a very good question. I think you have to understand that VPS database is predominantly from the mixed or PICUs as compared to the large cardiac ICUs. And so some of it might be related to the type of database we have utilized to capture the information. But honestly, this is the first report where we have actually looked at all different types of ICUs. Majority of the literature is from the cardiac ICU or the patients who had cardiac disease. So we believe that the low rate of ECPR could be related to the database from where we have captured the information. And in general, in pediatric, ECPR is still evolving, especially in the in-hospital cardiac arrest setting. Thank you. Thank you for coming.

benzodiazepines

mortality rates

septic patients

sedation

GABA-A receptor

hypotension

interleukin-1 beta

smooth muscle cells

propofol

calcium release

ECMO

pediatric cardiac patients

patient characteristics

cardiac surgery

neurological outcomes