Good afternoon, everyone. Thank you so much for joining us in this very interesting talk about sepsis cardiomyopathy, everything from diagnosis to interventions to outcomes in this group of patients. My name is Siddharth Dugar. I'm a staff physician at Cleveland Clinic. And just to go over, make sure that we are going to disclose all of our relevant financial relationships. The CME will open after noon tomorrow. And with that, I think we can start and everybody has a cell phone. There is an audio response system. So please do scan the QR code and would love to have some interaction going over during this session. I think we can start with Dr. Mike Lenspa. He is a staff physician in Intermountain Critical Care Echocardiography Service. And he will be talking about what modalities can be used to diagnose sepsis cardiomyopathy. Thank you so much. Aloha. Thank you for your attention and being here on a one o'clock session. My name is Mike Lenspa. I have no financial disclosures relevant to this presentation. So what is septic cardiomyopathy? One challenge with that is that we probably are lumping together several different distinct disease pathways into one common syndrome. On the left is a septic heart with a profoundly decreased contractility. Let's see right here. Okay, good, I can indicate these. And then in the center is a septic heart with a low afterload and a hyperdynamic function despite having abnormal strain. There is abnormal function. On the right, there's a septic heart with excess catecholamines that's causing a Takotsubo pattern. All three of these fall under the rubric of septic myocardial dysfunction or septic cardiomyopathy, depending on how you wanna call it. Although they're all very different presentations. One challenge with this is there is no clear definition. This is an outstanding review by Dr. Sarah Beasley on septic cardiomyopathy, which is probably the best review out there. It categorizes all of the different definitions that have been done, and essentially there's no consensus. Even in the multiple studies that have looked at a single component, they can't even agree on the threshold. For example, ejection fraction, 40% versus 50%, or fractional area change, or strain thresholds. None of these have achieved any clear threshold for what constitutes septic cardiomyopathy. As you'll see, there's also non-echocardiographic assessments, including troponin, BNP, SCVO2, or SVO2, as well as tachycardia even. So how should we diagnose septic cardiomyopathy? One thought might be that it might just be ventricular impairment. Now, when septic cardiomyopathy was first categorized or described in the 1980s, it was really described more as an acutely depressed LV ejection fraction, and there was also some ventricular dilation noted. The data in those patients suggested that patients who survived tended to have large end-diastolic and end-systolic volumes, and those who didn't survive tended to be normal. And there was a thought that perhaps this was some trend, and in hindsight, it appears that it's more likely that this was just a recognition that some patients were underfilled, and some patients were not. And then so we say, well, perhaps this is really an issue of cardiac output or oxygen delivery. Perhaps we should be using a cardiac output monitor, or perhaps we should be looking at the cellular level to try to see whether or not there is a primary myocardial injury. I'll point out that biomarkers have limited use. About 88% of septic patients have some elevation intraponin. 70% have elevation in BNP. These are commonly elevated in patients without myocardial dysfunction. And some have proposed that these elevations likely occur from increased membrane permeability rather than ischemia. One can have tissue edema with sepsis, and the heart is no more immune to that than any other organ. Therefore, it's not useful to discriminate against septic myocardial dysfunction, despite the fact that these are commonly elevated. And I'll point out a couple of studies that show that they're commonly elevated. This was a huge population study of 14,000 patients where they had both sepsis and troponin elevation, and the peak troponin elevation was associated with increased risk of atherosclerotic cardiovascular disease. Despite that, there's still plenty of patients who have these elevations who did not end up having any abnormality. And mortality does not necessarily indicate increased risk for having myocardial dysfunction. Similarly, BNP, this is a systematic review of 36 studies and about 3,600 patients. And there was a strong association with BNP and mortality. Again, there was not a clear association that BNP categorized patients with septic cardiomyopathy versus those who did not. So let's go to the meat of how we diagnose this, which is echocardiography. I'm gonna point out a lot of problems with echocardiography. And I think if you were to take one thing away from my talk, I'm hoping it's gonna be this slide. So the study cohorts that are typically used when we're talking about septic cardiomyopathy are typically done on patients fairly late after the onset of sepsis. We're talking about sometimes even up to a week after the onset of sepsis. This leads to a lot of problems. There are some problems with survivor or immortal time bias. There may be some problems also with selection bias. You also use, most of these studies used clinically obtained echoes rather than a prospective rigorous research protocol. And then again, which patients are getting echoes? Well, these are different than all of the patients with sepsis. So there's probably some ascertainment bias as well. And then the last and probably most important is that the loading conditions end up having a huge effect on a lot of our echocardiographic parameters. And so these can be conflated with problems of contractility. And everybody has loading condition problems. So in patients with septic shock, about anywhere from a third to two thirds of them end up being mechanically ventilated depending on your patient population. About a third of them with ARDS and about two thirds with vasopressors. All of these affect all the measures that we end up looking for. And as you can see, this was a study of about 67 patients. Cardiac function changes over time. So depending on when we do the assessment, we're gonna have different findings. As you can see in the left, there is a group of patients who had a normal ejection fraction the entire time. In the center, we have people who started off with decreased ejection fraction that improved. And then on the right, we see a group of people who had a normal ejection fraction and then worsened. All three of these are patients who present with sepsis. Depending on where we would have done our assessment, we might have taken the center or we might have taken the right panel and said those people have septic cardiomyopathy. If we looked at it longitudinally, we would have recognized that it would have been close to double the number. So ejection fraction is probably our main assessment echocardiographically. And the problem is it's highly dependent on load, preload or afterload. And this is a case of someone who had essentially hyperdynamic hypovolemia. And you can see that the end-diastolic and end-systolic chambers are fairly small. This person didn't have any fluid in the tank. And this is a person who would have had a high ejection fraction and would have been misdiagnosed by the older definitions. So what else can we use? Well, there is another assessment and that's called strain. Now, if you're not familiar with this, this can seem pretty intimidating, but strain is basically if I draw that green line over the LV and I look at how long that line is during systole and diastole, how much does it shorten during systole? The percent that it shortens is strain. So if it is a very big percentage or a big negative number, that's a good strain, all right? And that's it. And we can do this at different sections of the heart. And this seems to be potentially better associated with survival. And in the outpatient cardiac literature, this seems to be better at detecting subtle myocardial dysfunction that would have been missed by ejection fraction. And perhaps there's a hint of that in sepsis, although that still is yet to be validated. This is a decent sized meta-analysis showing that strain was associated with survival where EF was not. And we'll get more into that at the last talk here. So what about diastolic function? Well, there's potentially a way that we can look at some of the diastolic measures to try to correlate this with filling pressures. And so this is a very famous graph done in 1997, correlating the E to E prime, which is the X-axis, to the pulmonary capillary wedge pressure. And you see a fairly strong correlation here. How we measure this is we do a pulse wave Doppler over the mitral annulus, and we basically are assessing inflow over the mitral valve during diastole. And as you can see, there's an early and an atrial component and the early component is a function of a pressure gradient between the left atrium and the left ventricle. It's really how much is that left ventricle sucking? The left ventricle is expanding rapidly, blood is rushing in. We also want to counter that with how much it's actually expanding. So we also look at the mitral annular velocity. This is called the E prime. We use tissue Doppler over the medial mitral annulus or the lateral mitral annulus or the average. And that tells you how rapidly the annulus moves away from the apex, which is a surrogate for how much or how rapidly the LV is expanding. So now if we take a ratio of those, we're looking at a pressure gradient, which is our velocity or E divided by a expansion gradient or change in volume, which is our E prime. And so this is essentially a measure of how stiff the ventricle is. And this also seems to be decently correlated with mortality but as mentioned earlier, when we talked about that correlation with pulmonary capillary wedge pressure, it seems that it's really correlated with preload. And as you can see, the patients who ended up having a lower diastolic function were more likely to have received less fluid. And this was a, this found out in three separate studies that all demonstrated that patients who ended up having lower E to E prime tended to have less fluid receipt. What about the right ventricle? Well, here's a patient with acute core pulmonary, right? You see this massive RV that's distended. You see septal bowing into the, you know, into a concave LV. Well, there's a lot of different ways we can measure RV dysfunction. And because the RV is a complex shape, we can't really use ejection fraction with standard echo unless you're using three-dimensional echo. So what we look at here is a fractional area change, which is just essentially a two-dimensional ejection fraction. And we see how much the RV changes over time. And here's a couple of comparisons. A normal would be about 35% or more. Here we have TAPSE, which I'm sure many of you are familiar with. And this is basically an M mode through the tricuspid annulus. And we are seeing the amount of displacement the tricuspid annulus does during systole. And normal would be about 1.7 centimeters or more. Last would be the S prime, which would be the speed that the annulus is moving. So instead of doing the M mode through that same direction, we would use a tissue Doppler. And normally we would expect this to be about 10 or 12. Now we see RV dysfunction is very common in sepsis. These are two separate studies, each of about 400 patients. And they had similar findings. We see that about half the patients had RV dysfunction. And RV dysfunction was defined according to American Society of Echo standards, which also included the fractional area change, included RV size as well as function. And you can see that there's not a neat separation here. Amongst patients who had RV dysfunction, not every single parameter was abnormal. Because the RV is thin-walled, it's unclear how much of these parameters constitute intrinsic RV myocardial dysfunction and how much is just due to preload. The way to think of the RV is preload responsive and afterload intolerant. And we see that people who have RV dysfunction tend to have worse outcomes, but this is probably conflated with the fact that these patients tend to have worse lung disease. And this is probably one of the best ways to illustrate this. This is the same patient after getting put on PEEP. So this was a patient that was part of the ROSE study where we initially were a low PEEP protocol and we ended up to comply with the study going to a higher PEEP protocol. And you can see the change in the RV size as well as the vena cava that all occurred from just increasing the pressure on the vent. Nothing else had changed. So if you want to summarize all of this here, I would say septic cardiomyopathy is common. There's a lot of conflicting definitions with no consensus. Currently I would say biomarkers are not useful enough to be discriminatory. And all of the abnormalities we talked about with ECHO can be due to loading states. But the other underscore here is that you don't need a full ECHO to make these diagnoses. You can do this with stuff that you can learn in really a one or two day course. Thank you very much. Thank you so much, Mike. Next we have Deepa. Deepa Ghotur, she comes to us from Houston Methodist Academic Institute. And she will be talking about the different phenotypes of sepsis cardiomyopathy that we see routinely in our patients. Good afternoon. So I'll be talking about the phenotypes of sepsis-induced cardiomyopathy. I have no relevant disclosure relevant to this topic. And I really broadly want to cover just two main objectives. One is that there are variations in echocardiographic manifestations of sepsis-induced cardiomyopathy. And then we are going to talk about the classification performance of different clusters of septic cardiomyopathy phenotypes. I request that you are able to scan this QR code, because there is going to be an audience response coming up next. I'll wait for everyone to put away their phones. Fantastic. OK. I guess this question is mainly to check the temperature of the audience here. I'm going to wait till we get a few more votes, maybe 50. Fantastic. All right. I think we are talking to the right audience. Fantastic. Again, there's no right or wrong answer, just kind of understanding the audience here. So as Dr. Lanspa just described, sepsis cardiomyopathy is not just a binary approach. It's not about whether there is or there isn't, whether the patient is flu-responsive or not. Multiple measurements kind of go into quantifying or diagnosing sepsis-induced cardiomyopathy. And on top of it, there are many combined mechanisms, as was described earlier as well. And so how do you actually look at the data, or how do you make sense of all of this data? So in order to visualize all the patients, it's important to do what's called a clustering approach. It's a method of biostatistical analysis to cluster patients with commonalities and identify multiple clusters in sepsis-induced cardiomyopathy. This also helps in prognostic enrichment as well. And this study by Jerry et al did just that. What they did, this is actually a retrospective study of combined data of two prospective studies, the hemosepsis study and then the hemoPRED study. What they did was they included patients with septic shock. The hemosepsis study excluded patients with prior chronic heart failure. Sorry, I lost my slides. So a retrospective analysis of these two prospective studies, and they combined the data. And the hemosepsis study, this was performed in over 15 ICUs in France. And they combined all these data and used the clustering analysis to find five major clusters. So what were the components that went into these clusters? So they took multiple echocardiographic parameters. The first one was the LV ejection fraction calculated by the Simpsons method, as well as fractional area change. And again, all these measurements were performed by experts, by trans echocardiography in these ICUs. So one way of measuring the LV function was EF and fractional area change. Next, they looked at, or they combined the data on LV diastolic function. And they took the maximal E wave velocity in early diastole across the mitral valve, as well as they took the E prime, which is the tissue Doppler of the lateral mitral annulus. They also combined the data on how the RV function is doing. So they looked at the RV and LV and diastolic area ratio. Again, all of these measurements were performed by TEE within the first 12 hours of patients being admitted with septic shock. Fluid responsiveness was measured using SVC collapsibility index. And finally, the LV stroke volume was measured, and also cardiac index was measured by calculating the velocity time integral across the LVOT and combining that with the area of the LVOT, which gave them the stroke volume, as well as combined with heart rate and BMI cardiac index. Not only did they take echo parameters, they also combined hemodynamic parameters into this big data. And some of the hemodynamic parameters were heart rate, diastolic systolic blood pressure, mean arterial pressure, lactate, ABG, et cetera. So what they did was they combined all of these principal components, and they did a hierarchical clustering modeling. They also did a K-means analysis to reduce or to denoise these clusters. And as you can see, what they found is, for example, in one of the clusters, this is a group with LV failure. They found those patients to have low EF, low LV fractional area change, as well as lower aortic VTI, which is stroke volume. And another example in cluster 5, they saw patients with higher SVC collapsibility that goes along more with patients who are still hypovolemic. And they also looked at the RV, LV, and diastolic area. So there was one cluster, cluster number four, that showed that there was RV failure. And I'm going to talk a little bit about the RV, LV dilation versus failure, as well. And this is the other cluster with hypervolemic cluster. And they also found that patients with diastolic dysfunction were uniformly distributed across all clusters, except for the first cluster, which is well-resuscitated cluster. So what they did was, based on what they identified, they identified three principal components that would categorize these patients to fall into these very specific clusters, well-resuscitated, LV systolic dysfunction, hyperkinetic profile, RV failure, and persistent hypovolemia. And as you can see, some of the cutoff values are not necessarily like the cutoff values that are by expert opinion, right? These are kind of unbiased and statistical modeling to help them classify into these very specific clusters. They took these three major components, principal components of each of these clusters, and they ran the ROC curves. And as you can see, they all perform very well, meaning that if these patients had these three specific components, how well you would be able to categorize them into these specific phenotype. And the performance was excellent, as you can see from the area under the curve. OK, time for the next ARS. OK, fantastic. And you're all probably absolutely right. I think hyperkinetic profile is probably one of the most common. But if you think about it, it's not really as common, or it's not really more common than the others. Kind of an equal distribution of anywhere between 16% to 20% across all of these groups. And it's important to note that the parameters that went into clustering these patients into these specific phenotypes, I mean, of course, there are additional parameters that kind of go into it. And it's also important to note that the expert opinion, I mean, the expert opinion, as is shown here, like the different phenotypes, it's more of a biased in the sense that when you run the clustering analysis, it's more unbiased and kind of algorithmic approach as well. What about the RV? Are there more to RV dysfunction? Absolutely. So this study was done by Tretalia, and they looked at RV phenotypes in patients with ARDS. And most of these patients had sepsis as a major cause. And they had over 800 patients, and they did a latent class analysis, a method of clustering analysis. And the patients received or had transthoracic echocardiography performed within seven days of their diagnosis of ARDS And as you can see, they're really patients with higher SOFA score, organ failure score as well. And similarly, what they also did was they found three principal components as well, categorizing them into four different classes. And what are these classes? So the first one is the preserved RV. Majority of these patients, 43% of these patients, had preserved RV. And they also identified, they phenotyped into four clusters, and they were able to identify the RV dilation as a separate subgroup where they have preserved RV function as well as RV failure. And you will see later that RV failure patients had more poor outcome. And the incidence is listed on the left-hand side of the chart. And these were the three variable models. I mean, these three cutoff values. And again, these cutoff values were generated by the latent class analysis and not necessarily from expert opinion. I'm going to skip this slide. And finally, the RV phenotypes, again, there are additional components that go into categorizing patients, whether they have RV failure or not. And there could be more subclassification. And this is, again, an expert opinion. And more prognostic studies are needed to prove this. There are, however, several limitations of these clustering approaches. As you heard me say that patients with diastolic dysfunction were spread across the last four clusters, there is some between-cluster overlap happening as well. And when this overlap happens, it gets a little difficult to practice using that in clinical practice as well. There are also other unmeasured factors that were not in the study. For example, the global longitudinal strain, right? I mean, that was not used in these studies as well. And also, these studies were performed at one time point. So the echo, the TEE was performed at one time point and not a serial measurement. So we do not know how these patients transition from one phenotype if and how they transition from one phenotype to the other. There could also be intra-observer and inter-observer reproducibility issues. And again, all these studies excluded patients with pre-existing cardiac dysfunction. So that was not studied as well. So in summary, phenotyping with statistical clustering approaches can identify unique hemodynamic groups. And there are five major sub-phenotypes of sepsis-induced cardiomyopathy and four for RV dysfunction, especially in patients with ARDS. And there are few limitations inherent to phenotype classification. Thank you. Thank you so much, Deepa. That was excellent. So next, we are going to talk about, first of all, everybody still with us? Aloha. So again, we are talking about all those things. But how does it matter? The only thing that matters in critical care is if it is going to affect the mortality. Is it going to cost me more? Is it going to increase the length of stay of my patients? Or is it going to cause some long-term outcomes? That's what I'm here for. The meat of the lecture. Just to go back, my name is Siddhartha Dugar. I'm a staff physician at Cleveland Clinic and also the director of Point of Care Ultrasound. I got a chest sonocyte ultrasound grant. But I don't have any conflict with what I'm presenting currently. So again, we'll start with the past. The past is a prologue. So this is the first study that looked at sepsis cardiomyopathy. This was done in 1984 by Parker and Parillo. Again, the landmark study when we talk about sepsis cardiomyopathy. And they did a pretty good study where they looked at 16 patients, sorry, 20 patients, and they did echo every day. And they're like, what is happening with these patients? And what they found was, I think this will work better, is that as the days of sepsis happen, the patient become hyperdynamic. But the interesting part was when they differentiated the non-survivors from the survivors, the non-survivors actually had a reduction in their systolic function, their LVEF was down. And only during the recovery phase did we see the LVEF go back to normal. While the patients who did not survive actually had a hyperdynamic circulation. So this was kind of an eye opener. Everybody thought that if your heart is going down, you may be doing worse. But what this study showed is your heart going down is maybe a cardioprotective mechanism. And as you recover, the heart recovers with that. Interestingly, they found that the stroke volume did not go down. What happened was as patient developed sepsis, the heart started dilating, it had more volume, so it did not have to push harder. So it maintained the stroke volume, while LVEF, which is again, as Dr. Lanspa mentioned, is load dependent, did show that it decreased, but while maintaining the cardiac output. So a really nice paper to start talking about sepsis cardiomyopathy. Now, the other study that I would like to point out is, again, it cannot be done in human. This is impossible, but this was done in dogs. What they did was they took dogs, they did echo every day, and they tried to look at this LV systolic function, why LVEF, and then they gave them the peritonitis, and they found that most of the dogs, when they develop sepsis, their EF actually dropped. Also, how much the EF dropped totally depended on the amount of toxin that was given to the dog. And as they recovered, the heart function recovered as well. Again, cannot be done in humans, but this tells that as component of sepsis cardiomyopathy happens in every patient. And as they recover, the heart recovers as well. The same thing on EF here is S is survivor, and S is non-survivors. And the same thing, when they develop sepsis, their cardiac function decreased, both the LV and the RV. As they survived, the cardiac function went back to normal. While patients were non-survivors, their cardiac function did not change at all, and it stayed the same till they died. So this is a study that again Dr. Lanspa shared as well. So LVEF, we agree, it's a very crude method, but it's the method that is most commonly available. That's a number that I hear every time there is an echo that is ordered. When they did a systematic review and meta-analysis of close to 800 patients with a mortality close to 32 percent. So this were patients who were really sepsis and septic shock. They found that LVEF could not predict who's going to survive and who is going to die. Or the other way to think about it, LVEF did not play a role in how your patient did with sepsis and septic shock. But the global longitudinal strain, which is a more sensitive mark and can detect very subtle LV systolic dysfunction, showed that the more negative it is, which in turn means the more robust your cardiac function is, the better your outcome. The problem with that is the SMD which is standard mean difference. The best way to understand what standard mean differences is if it is more than 0.5 or 0.8, it's pretty good. If it is 0.5 to 0.8, it means a difference between the people who survived and did not survive is not that much. This is 0.25, if you see here, 0.26. So the difference, mean difference in the global longitudinal strain was 1.3 between people who survived and did not survive. What is the clinical implication of that? Again, we say that GLS improved outcomes or we can predict who is going to survive or not. But again, the clinical implication is a patient with an EF of 55 percent to 50 percent. Yeah, 50 percent will die more, but again, the difference is not truly clinically relevant. So then we just put the LV systolic dysfunction apart. LV systolic dysfunction does not play a role in outcomes of patient. How about diastolic dysfunction? So there are four components what we use for assessment of diastolic dysfunction. But this study by Lendisberg, which looked at the four parameters and they found the parameter that is the best in terms of predicting outcomes in sepsis patient was actually E prime. This is how robust the relaxation of LV is. If you can see this Kepler-Meyer survival curve, irrespective of what your EF is, if your E prime is less than eight, your chances of survival is way lower than if your E prime is more than eight. So it's telling that if your heart is not relaxing appropriately, your chances of surviving in sepsis is low. Makes sense. Because these are the patients who are older, they have more cardiac disease, so they have a high risk of death irrespective of their developed sepsis or not. The same thing, when we do E over E prime, which is a marker of left atrial pressure or filling pressure, there is a really nice grading as your E over E prime or your filling pressure goes high, the mortality goes high. These are the patients when you resuscitate, they develop pulmonary edema, they go on the vent very quickly, they stay on the vent for a longer period of time. Makes sense. But if you really want to phenotype it, it's the E prime that tells you which patients are going to do better and do worse. But then we look at the systematic review. This is another systematic review that was published, I think in 2016, looking at diastolic dysfunction. So again, people who have no diastolic dysfunction, they survive more than people who have diastolic dysfunction. Now this is where the true difference shows up. This is risk difference. So the chances of you surviving with a sepsis or septic shock is 1.82 times higher if you don't have a diastolic dysfunction. That's a very strong statement to make. The other thing is, all the studies are showing no diastolic dysfunction is better. The problem is, people who have diastolic dysfunction are the patients who are aged, the patients who have hypertension, the patients who have diabetes, the patients who have previous heart disease. So are we just using parameters to find the patient population who are going to do worse? And again, they looked at the LVF2, did not make any difference. So this study, I know Mike also shared, but this is another study looking at RV dysfunction now. So we have moved from LV dysfunction, systolic dysfunction, we found there was no difference. Then we moved to diastolic dysfunction, there may be a difference, but now we are going to the RV part. And the reason I'm pointing out this study is, as Mike said, they use five different parameters to assess if the patient had RV dysfunction or not. And not all of the parameters are present every time. Not all the parameters are abnormal every time. But the most two parameter that they used was RV enlargement and TAP-C. So TAP-C less than 17 or 16.5 and RV that is enlarged. And they found that people who had RV dysfunction actually had worse outcome. While if you have RV and LV dysfunction, you don't have worse outcome, which is very interesting. So isolated RV is much worse than RV and LV dysfunction, which is equal to having no dysfunction. Interesting study. So this is by Dr. Lanspa here, where they use different parameters. They did not go with TAP-C, sorry, they did not go with RV enlargement. They went with fractional area change and TAP-C. The way they defined LV systolic dysfunction was using strain and LVEF. And again, there is a significant amount of overlap that we see in different phenotypes or different forms of cardiomyopathy. But the odds ratio with RV dysfunction stands out. It's 3.37, which means if you have RV dysfunction compared to non-RV dysfunction, your chances of dying with sepsis is three times higher. That's a huge increase in their mortality with a very nice confidence interval and a p-value. But every study here is looking at different components of the heart. The Cherry study that Dr. Deepa went over in detail, they're like, you know, there is going to be a significant overlap, but why don't we use machine or statistical analysis to find out different phenotypes? And they found a different prevalence. But look at the mortality. It is still the LV systolic dysfunction that stands out with 50% mortality at the end of their ICU stay and followed by RV. So why are we getting this mixed result? Sometimes LV systolic dysfunction is bad. Sometimes it's not bad. Sometimes diastolic is bad. Sometimes it's not bad. It may be because we are defining it wrong. So this is a study that we published. And what we found is it's not, one is not worse than other. It's a U-shaped curve. As you can see, this is LVEF on the X-axis and in-hospital mortality on the Y-axis. And as your EF gets, you become more hyperdynamic, your mortality is higher. You become more hypodynamic, your mortality is higher. The middle at 55% EF is where your outcome is the best. Now you can imagine, if I have a study patient population that looks like this, I will show hyperdynamic is bad. If I have a study population that is like this, I will say the hyperdynamic is bad. But all we are trying to do is compare one to another. But when you look at linearly, it's just a U-shaped curve. And the same thing, it doesn't matter if it's sepsis, low intensity sepsis, septic shock, or high intensity septic shock. These are the patients who have heart failure and they develop sepsis and they unfortunately die. While in other groups, the hyperdynamic part plays a bigger role as the severity of septic shock increases. Again, is it the heart that is becoming hyperdynamic? Are they vasodilated? Is it too much vasopressors we are acting on this? Or it was just that we did not resuscitate these patients enough? There are so many factors that are playing when we are looking at LVEF. After that, the question is, what is worse? New onset heart failure or pre-existing heart failure? So this is a systematic review that we published a year ago that was looking at, when do you see a worse outcome? Is it sepsis-induced cardiomyopathy or non-sepsis-induced cardiomyopathy? The non-sepsis-induced cardiomyopathy are the patients who have pre-existing heart disease. And we found that patient who have sepsis-induced cardiomyopathy, their outcome is much worse than people who don't have sepsis-induced cardiomyopathy. But this is short-term data. What if my patient has sepsis cardiomyopathy? What happens to them long-term? There are not much work that has been done, but I think this is a really good study that was published by the Intermountain Group by Sarah Beasley. And it's a complex model, but I will try to go over it. This is the LV global longitudinal strain. So minus 10% is your heart is not working that great. Minus 30 is your heart is hyperdynamic. We are looking at history of heart cardiac disease here. These patients have no pre-existing history of cardiac disease, while this do. And what they were looking at, not how it affected mortality. I think we have answered the mortality question with multiple studies. They're like, what if you survive ICU? You go home. What are the chances of you developing major adverse cardiovascular events, like stroke, myocardial infarction, AFib? And the interesting part about this study is it doesn't matter what your global longitudinal strain is. If you have no pre-existing of heart disease, you are fine. So the thing is, what happens with sepsis stays with sepsis. It's only when your history of heart disease, that's when the stakes become higher. Now, here also, we are seeing a U-shaped curve. So if you are hyperdynamic or impaired, the chances of you developing a major cardiovascular event is pretty high. So is the sepsis cardiomyopathy we are seeing more of a marker of how your heart is able to sustain the stress? And going to tell us, maybe the mortality, this admission, we don't know. But we know with this study, that the long-term outcome is going to be very compromised. So not only looking at ICU mortality, but looking at this patient's longer is going to help us truly understand the burden of this disease process. With that, I will end my lecture. Thank you so much. And now we have Dr. Sato. He is an intensivist here in Hawaii. He's really enjoying himself after leaving Cleveland Clinic. But he's going to talk about how do we change management based on what the underlying cardiac function is. So thank you so much, Ryota. Hi, good afternoon. My name is Ryota Sato, working as an intensivist at the Queens Medical Center in Hawaii. I work both in medical ICU and CICU, and I have nothing to disclose. So today's objectives is reviewing LB dysfunction that requires special caution. So when we think about LB dysfunction, we tend to think about LB systolic dysfunction. But as we've been learning, there are two populations that we have to pay attention to. And also the treatment option for LB systolic dysfunction and LB hyperdynamic function. And then we review the current evidence about use of ECMO or other mechanical circulatory support in the setting of sepsis-induced cardiogenic shock. So this is actually a scheme that Dr. Dugar shared. So our group actually investigated the association between LBEF and the mortality in the setting of sepsis and septic shock. And what we found was U-shaped CARB. And as you can see, anybody above LBEF 75%, actually 70%, and then EF below 25%, you see actually higher mortality. And these two populations are some population that you have to pay attention to. This graph is actually adjusted graph. So what that means is we adjusted for something like basal pressure dosage. We adjusted for blood pressures. We adjusted for basic patient's characteristics. But after eliminating all the factors, we also adjusted fluid amount, but still this relationship remained. So now, because we are focusing on these two populations, I'm gonna start from this LB severe systolic dysfunction. So for severe LB systolic dysfunction, everybody, you know, in general, when we talk about how to manage the hemodynamics, tend to focus on oxygen delivery. But is it really important? All of us actually have seen this formula when we took a board exam, you know, DO2 equal, you know, this formula. But is this really reasonable? I'd probably challenge you guys, you know, we know higher hemoglobin is not necessarily helpful in the setting of critically ill patients. But based on this study, to increase oxygen delivery, are we transfusing to increase oxygen delivery? Or we know higher oxygen, we don't have to achieve SpO2 of 100%, but are we increasing the higher oxygen level to achieve oxygen delivery in the setting of substance? Probably not. You know, everybody's okay with transfusion threshold of hemoglobin seven, or maybe SpO2 is maintained around 92%, 93%. So to me, it doesn't really make sense to try to achieve oxygen delivery just based on the cardiac output. So this is really bothering me a lot. So when you think about what's going on or why this is happening, so this is actually the study done by Dr. Gattinoni. This is a fascinating study. So they did a post hoc analysis of obvious trial. And, you know, they found lactic acid and SpO2, it's actually U-shaped relation. And the most of the septic patients actually had SpO2 above 70%. And very few people have SpO2 less than 70%. So they concluded that, you know, the lactic acid elevation in the setting of sepsis is not because of oxygen delivery or transport, but mostly it's from oxygen utilization. So when you get to see the lactic acid elevation, or whenever people are not doing good, of course you want to make sure blood pressure is maintained, but at the same time, you also want to make sure they're using oxygen. So just going back to the clinical scenario, or clinical guidelines, surviving sepsis campaign guidelines recommend adding dobutamine on top of norepinephrine or epinephrine monotherapy in the setting of myocardial dysfunction with reduced perfusion despite adequate resuscitation. So, you know, this is very vague statement, but in my interpretation, this suggests that they're okay with using inotropes in the setting of cardiogenic shock. So what that means is patient has sepsis and cardiogenic shock mixed in a shock. So how would you know patient has cardiogenic shock then? Or then are you going to use inotropes if you think in a patient has cardiogenic shock? That's the biggest question. And then again, we evaluated, you know, how many percentage, well, what's the prevalence of actual sepsis and, you know, cardiogenic shock, mixed shock in the setting of, you know, sepsis? And we found, you know, prevalence is about 5%, or a little bit less than 5%. So it's a small percentage, but given the fact that sepsis is the most common entity or disease that led the patient to come to the ICU, 5% is not negligible. One thing came up to my mind is PA catheter. Maybe PA catheter can tell us if patient is in cardiogenic shock, but previous RCTs actually showed no benefit with PAC. But one thing you want to be cautious is these studies were not including cardiogenic shock too much. I mean, minority had a cardiogenic shock, but majority of these patients are not cardiogenic shock. And in the cardiology field, actually, PAC is coming back now. You know, a lot of studies based on the retrospective studies showing some benefit with, you know, PA catheter with the cardiogenic shock. I wish I could have a paper here now, by now, but we actually investigating the use of PAC in the setting of sepsis and cardiogenic shock context. And what we found was we divided sepsis, you know, a population into two groups. One is sepsis without cardiogenic shock, and group two is sepsis with cardiogenic shock. When we adjusted for the variables, you know, sepsis without cardiogenic shock had no mortality effect from PA catheter. However, on the other hand, sepsis with cardiogenic shock had significant mortality benefit from PA catheter. I hope, you know, we can publish this paper in next couple of days, a couple months, sorry, but we'll see, we'll see. But in the cardiogenic shock population, there are, you know, decent amount of studies showing some benefit of PA catheter. So based on that, PA catheter may be something you could consider in the setting of sepsis and cardiogenic shock. So now, going back to the question, then you make a diagnosis of sepsis and cardiogenic shock. You think it's sepsis-induced cardiogenic shock. Are we using inotropes? There are very limited number of studies. The first study I'm aware of is the study from Milkman, which is published in 2013. So basically, this is a retrospective observational study, and they included 420 patients, and they adjusted for these variables. So this is pretty reasonable variables. They adjusted for some of the hemodynamic markers. They adjusted for CBRK markers. What they found was inotropes was associated with higher 90-day mortality, and dobutamine and epinephrine were specifically associated with a higher mortality in this group. You could say, well, this is still a retrospective study. You know, there must have been some sort of selection bias, stuff like that. Yes, it is true, but this is something to consider. And then our group did another study, similar study, including 417 patients. And we adjusted for LBEF, E-prime, CBRK scores, backgrounds, and we adjusted as much as we can. And what we found was epinephrine and dobutamine was still associated with higher mortality. And this was actually time-dependent and dose-dependent relationship, meaning the longer you're on inotropes, the more people died. The higher dose you're on, people died more. So in addition to this, we also found epinephrine, dobutamine, and merinone are associated with higher risk of AFib vs. RBR. And AFib vs. RBR is known to be associated with worse outcome. Finally, we also had a different study from MIMIC-3 database, basically did a similar study using propensity score matching analysis. They also found epinephrine, dobutamine were all associated with higher mortality. And again, all of these are retrospective studies and you could still blame the study design or selection bias. However, at least, you know, these are the best outcome that we have. We don't have a prospective RCTs in terms of use of inotropes in the setting of cardiogenic or septic cardiogenic shock. There are some studies, RCTs using inotropes without cardiogenic shock, not showing any benefit, but we don't know what to do with the cardiogenic shock. So here's my summary for the septic, sepsis-induced cardiogenic shock. Because cardiogenic shock has some benefit from PAC, I think probably PAC will be beneficial in the setting of sepsis and cardiogenic shock. And in the limited data, even though all of these are retrospective, inotropes can be associated with worse outcome. You wanna be very cautious about why you are using it, what you wanna fix with the inotropes. And again, if you're trying to improve oxygen delivery, then you don't want to fix the number. You wanna treat the patient. So most of the time, when you are starting inotropes, you wanna see some of the markers. For example, you don't wanna treat the SCPO2 only. You wanna see a whole picture. So you have to use inotropes very cautiously. So now, this is not a common practice, but because it's out there, I also wanna mention the use of ECMO in the setting of sepsis-induced cardiogenic shock. So this is probably the most, well, the biggest evidence that we have in terms of use of ECMO in the setting of sepsis-induced cardiogenic shocks. So this is a Maritime Center International cohort study. So this is basically the retrospective analysis of prospective data. And then what they did was they included 82 patients who were on ECMO and 130 patients from other huge cohort from the prospective trials. And then they matched these people, and then they used a propensity score matching, and they compared these two groups. So what they found was VA ECMO was associated with a significant improvement of survival. But what you wanna be cautious is, EF of ECMO group was 17%. Index is 1.5, which is extremely low. But then the matched populations index is 2.2, which is sort of acceptable. EF is 27%. But as I mentioned, when you see the U-shaped curve, 27% is, yes, it's associated with worse outcome, but it's not quite there yet. So I'm not really sure if they can completely match the population. One interesting thing is similar of the same group, Dr. Virchow actually did about 10 years ago, he published this single center case series. Sorry, this is not international, but case series from French. So they included 14 patients with septic shock with severe myocardial dysfunction, and they thought it's sepsis-induced cardiomyopathy. They initiated VA ECMO, and then interestingly, 10 out of 14 survived. Then the EF was 16%. SVR is high. So what this is telling us is, okay, they had a septic shock. However, this is pretty much cardiogenic shock. So authors concluded VA ECMO is associated with better outcomes. But what we can tell is, well, if it's a typical septic shock, basal plastic shock, probably VA ECMO is not a good idea. But what this is telling us is, if shock components is mainly cardiogenic shock, then, and then you also think it's a sepsis-induced cardiomyopathy, then you could think about VA ECMO. But again, it's not standard of care yet. So there are a lot of argument to be done. And finally, I said this is the mechanical circuitry support. We also published some studies about use of mechanical circuitry support using the national data. What we found was, balloon pump, impella, these are probably okay to use if you are sure it's cardiogenic shock. But again, it's a retrospective data. So we need a perspective data. But to initiate mechanical circuitry support, I would say EF index both needs to be extremely low. It's not going to be EF of 40%, 45%, even 30%. I don't know if it's low enough, but if it's 10%, 15%, index is 1.5, 1.6, then I'd probably think it's solid cardiogenic shock. And then SBR is high. So that means it's a solid cardiogenic shock. Potentially indicated if it's pure cardiogenic shock. Five minutes, five minutes, okay. And then if it's basal plasic shock, if heart is moving relatively okay, we probably should not use mechanical circuitry support. Then moving on to the hyperdynamic LV systolic function. So we don't have much data in terms of what to do. You know, hyperdynamic LV systolic function could be from basal plasia, could be from hyperbulimia, but could be from sympathetic overtone. As I showed you, a U-shaped CARP was after adjusting the fluid and also basal pressure dosage. So even after eliminating these factors, still hyperdynamic condition is worse. So there are some clinical trials showing some benefit in terms of shortening of the time to reversal of shock status with hydrocortisone. So based on this, you may want to think about lowering threshold to start the steroids. The other thing is, I think we talked about this evidence in the morning session, but when, you know, we compared basal pressing response based on LVEF, people who had higher LVEF had actually better response with the basal pressing and actually better mortality. So if you're seeing hyperdynamic LV, maybe lowering threshold, you know, to start the basal pressing may be better. So you want to consider to start basal pressing maybe five or 10. This is not proven concept yet, but this is something we can consider in the absence of solid data. And then finally, I'm going to talk about the use of beta blocker. I know this is very controversial, but at least, you know, all the intensivists needs to be aware of this. So we investigated, you know, the beta blocker, ultra-shorting, short-acting beta blockers using Esimolol, Lantiolol. Basically, we included all seven studies. All of these are RCTs, and we did a meta-analysis. So one thing to note is all of the patients are non-compensatory tachycardia, meaning, you know, this study was done after initial resuscitation. They were, you know, fluid resuscitated, and then they were made sure they're not just tachycardic because of hypovolemia. So the investigators used, in most of the studies, Esimolol versus Lantiolol, and then compared with the placebo. Interestingly, heart rate and WBC were lower in the beta blocker group. And then MAP, norepinephrine dosage, cardiac index were same between beta blocker and control group. But more interestingly, stroke volume index was significantly higher in the beta blocker group. So what that means is when it's tachycardia, by slowing the heart rate, you have more feeling time, and then you actually increase stroke volume. Then, even though you decrease the heart rate, when it comes to the cardiac output, it was still maintained. Now, we know there are some population that could develop hypotension from Esimolol. I confess that I see a lot of patients who actually had a hypotension from Esimolol. But we still have to figure out who is gonna really benefit from using beta blocker. So now here's a summary of my presentation. If EF is 30%, 34%, 45%, without evidence of cardiogenic shock, I would suggest just a regular management. If EF is severely low, you could use inotropes, but you wanna be cautious about use of inotropes. And then if you're not really sure what's going on, PAC, PA catheter may be helpful. And then to consider mechanical obstructive support, you wanna have pure cardiogenic shock rather than mixed cardiogenic shock and septic shock, almost pure cardiogenic shock. For hyperdynamic LBC solid function, of course you wanna ensure fluid status. You wanna make sure they're resuscitated. You may consider starting basal pressing a little bit earlier than usual. You may consider starting hydrocortisone a little bit earlier than usual. And then ultra short-acting beta blocker. I wouldn't recommend this necessarily, but you might wanna watch out what kind of study is gonna come out next few years. Thank you. Thank you so much, esteemed panel, for a great presentation.

sepsis cardiomyopathy

myocardial dysfunction

diagnostic criteria

echocardiography

biomarkers

phenotyping

outcomes

inotropes

mechanical circulatory support

individualized management