Folks, it's a pleasure to join this group. I am charged this morning, and I was actually given a charge, to address a topic that I think is both timely, for which I think there's sufficient controversy that we're left with the opportunity here to be a little more exploratory in terms of where the guidelines end and where opportunities for both current care and future studies exist. And then I'm going to take you on a little thought experiment with a couple new data slides at the end just to spice it up. I do want to reiterate that the approach I'm going to take today in no way is intended to contrast specifically with existing survivor sepsis campaign guidelines, of which I have served on that committee in the past. And so I really am here focusing on the questions that pertain to hemodynamic resuscitation in the immediate post-initial environment. And so I do want to reiterate, and I'll show you data to perhaps underline, how very important the work that we do together on maintaining good adherence to survivor sepsis campaign guidelines continues to be as the evidence evolves. My appointments, as I say, University of Colorado and Denver Health. I have funding to my institution, both from the NIH, the Department of Defense, and industry from Baxter Religious and Work that I'll speak about today. There is a single prevailing challenge that we have, even in an era where we see overall decreasing mortality. And that is this problem that we see, which is almost insurmountable relating to the progression and persistence of sepsis organ failure, even in the face of establishing circulatory effectiveness. And it almost is that we have reached a plateau when it comes to our capacity to move the needle on this issue. On the one hand, and I take you down here to the cellular level and perhaps subcellular level, we know that sepsis is characterized both by diffused microcirculatory and macrocirculatory vasodilatation linked with mitochondrial dysoxia, a failure of oxygen extraction and metabolism. And we know that in combination between cytokines, altered microvascular hemodynamics, transvascular leak occurs so that the intravascular space becomes relatively underfilled. Inflammatory cells, diaperties, and migrate. And the consequence of relative circuitry underfilling without resuscitation is microcircuitry incompetence, mitochondrial dysfunction, tissue hypoxia, et cetera. But in a correlative fashion, once we initiate fluid and vasopressor therapy, we may be compounding the situation in ways that are very complicated. Because once we have established circulatory filling or effectiveness, we appreciate that the dynamic nature of the circulation, which is pulsatile flow, non-static flow, can further worsen bio-injury of the endothelium. Now we're very comfortable with this idea in the airway. We think about ventilator-induced lung injury in our discipline very commonly. But indisputably, we can show both in the microdish, in the flow cell, in small animals and humans, that cyclic overdistension of an endothelial surface and the vasculature will amplify endothelial valutrauma. And we know that that will further worsen leak, worsen tissue edema, et cetera. So we're caught in this complex kinetic space where, despite optimizing macrocirculatory resuscitation, we've plateaued in a way and have this persistent evidence of organ dysfunction. And the arguments that I'm going to make to you today are somewhat theoretical, largely based on evidence, and essentially are aimed to address this issue of underpressure. Is it sufficient to just optimize the macrocirculation through effective arterial filling through fluids and presses, while also, and I'll touch on this, mitigating venous congestion, if we don't also optimize microcirculatory flow? And it's this coherence function, as it's been referred to, between the macrocirculation and the microcirculation, which we could think of as being macrocirculatory fluid responsiveness, resulting in microcirculatory flow responsiveness that we see is substantially disorganized, incoordinate in the septic diathesis. And I'm going to make the argument that, while we have made a lot of progress over here, our work continues now to reconnect the two components of circulatory coherence. And the way I'm going to do that is to argue, first of all, that we need to resolve some key questions in the field. And that has to do with, how do we get to macrocirculatory resuscitation before we think about the microcirculation, and then optimizing the idea of coherence, optimal flow, and reserve between the two? Well, it's timely to appreciate that when it comes to being under pressure in sepsis resuscitation, in shock resuscitation specifically, the arguments about restrictive or liberal fluids I think is almost in a sort of penultimate or ultimate phase. We've argued that, on the one hand, restrictive fluid strategies can reduce overall fluid balance and leakage, allows for early vasopressors as a prioritization, may prevent pathologic edema as we saw develop in our example slide over here. And we certainly know that there is a strong association with higher net volume balance and excess mortality, which may not be causal, but is strongly associated. On the counterbalance, there's good evidence to suggest that if you don't liberally resuscitate the circulation and augment preload, you end up using a lot more vasopressors. And so if you optimize filling, you can decrease pressure use, decrease microcirculatory dysfunction. And broadly speaking, this remains the current empiric approach. Large amounts of data suggest that, and if I took a survey here, many of your patients within the first 24 hours receive up to 5 liters of crystalloid fluid in the initial phase of resuscitation. So if you put that together prior to yesterday morning, what that meant was that large numbers of prospective randomized trials that compared a more restrictive fluid approach compared to either a more liberal approach or just usual care. Many of the times, usual care is more liberal and less presses as a priority. In about 1,200 patients meta-analyzed in each arm, the point estimate, including the very large Mayhoff classic trial published last year focusing predominantly on ICU-level patients randomized to one of those two approaches, was somewhat favorable of a more restrictive approach, but not statistically so. And so we would argue that before yesterday morning, there was reasonable equipoise with regard to which of these strategies, when you look at mortality and organ dysfunction, resulted in improved outcome. And why did I say yesterday morning? Because prior to yesterday morning, we had strong data in the ICU patients from classic that showed no difference, even though you could administer far less fluid in a restrictive strategy than a liberal strategy or usual care strategy, to see a difference in 90-day mortality, adverse events, either individually or collectively. Yesterday morning, we released the results of the Clovis trial, a large new multi-center randomized prospective study that included patients prior to the ICU as well as in the ICU. I'm not going to present the entire data. It's out there available for you to look at. But I want to contextualize the findings of the crystalloid liberal or vasopressor early study as they pertain to today's discussion. In this study, we evaluated about 780 patients in either an arm that prioritized presses or pressocentric or liberal fluocentric early on in the course of the presentation with undifferentiated shock. These patients were then determined to have probable or possible sepsis and were randomized. And once they were randomized, they had a systolic blood pressure in the 90s, a white count around 14, a modestly elevated lactate. They weren't the sickest patients right at the beginning. These were their first numbers. About 10% were ventilated right de novo, but almost 30% ultimately required intubation. Pneumonia constituted almost a quarter of the cases. And when we look at patients like this, you'll accept that this is highly reflective of the patients that we see at our centers. So in terms of what happens to these patients, and I want to reiterate that my discussion today is not in any way meant to address, challenge, or contrast with the notion of a 30 mL per kilo crystalloid bolus because, indeed, both groups of patients in the study received around 2,000 mLs of crystalloid prior to randomization. But the big differences in this investigation was in the early periods after randomization where a fluid liberal strategy resulted in significantly higher volumes of a crystalloid administration, both at six hours through 24 hours, and if you add back in the pre-randomization group preserved all the way through. Vasopressors between the arms were given both more frequently earlier, less than two hours in the pressor group, and for longer durations, almost 10 hours in the restrictive group, showing that there was significant separation for our ability to determine effect in the study. And as you may have learned yesterday with us, the primary outcome of 90-day mortality was importantly, importantly neutral, no difference between it, but almost 15%. Now I acknowledge that this is not the sickest group of septic shock patients that may experience mortality double that, but when I speak about a plateau effect where we've optimized macro circulatory resuscitation and still see sepsis organ failure-related death, we appreciate the significant challenge in the area. And just to reiterate that this similarly didn't translate into long-term or patient-sensitive outcomes at the level of organ failure-free support or hospital days, and particularly didn't importantly result in excess morbidity using pressors and lower fluids didn't complicate the progression to renal dysfunction, worsen non-circulatory organ failures, including respiratory failure. But as you can see, the incidence of adverse events that were pre-specified was similar. I'm going to skip through subgroup analysis and encourage you to look at that, but I want to finish up this first part of our discussion, framing the question about macro and micro circulatory coherence as the way to take the next step through optimization, to appreciate that when we use a vasopressor-centric fluid resuscitation, both in the classic study focusing on ICU patients, CLOVAs, including ER and ICU patients combined, they're both effective and safe, but are still associated with high morbidity and mortality burdens. Importantly, fluid bolus administration in both these groups was administered agnostic to whether patients were still fluid responsive. Now, let me reiterate that point. We use a protocol where if the patient's blood pressure remained low, they received more fluid. But as I'll show you in the second part after this morning's discussion, that by no way informs an understanding of whether the patient's circulation is still fluid responsive. And an agnostic approach, while useful to estimate average treatment effect across groups, is certainly not the way we practice personalized medicine at the bedside. There's no doubt that subgroups may be identified, and I'm going to highlight one of them in a minute. But I'm going to suggest that because our studies didn't either mandate, require, or specify any form of hemodynamic monitoring other than blood pressure, we still haven't looked under the hood of why there was no separation or difference other than to say both approaches are safe. So taken together, I think that we can assume that if you use a personalized approach that is prioritized, there's really no significant difference if you use blood pressure as the target. And I want to transition then to a different approach, which is to answer this question, which is, will my patient respond to a fluidic bolus? Now, let me reiterate that in the approaches that we've studied thus far, very appropriately because they capture usual care, we use an agnostic approach. Your blood pressure's low, you get more fluid or more presses. But we know that up to more than half of patients who are fluid non-responders may actually still be in shock. Conversely, patients who have established this effective macrocirculatory blood pressure may still have evidence that they are fluid responsive. I'll show you what that means in a minute. Reiterating that being fluid responsive doesn't mandate more fluid. It's a state of circulation. Before I move forward, I want to just separate one key observation, which is this. Measuring the circulatory preload, you can do that using a fancy VEXA score that I'll show you in a minute, measuring intravascular venous pressures, et cetera, is entirely unpredictive of whether a patient's circulation will be fluid responsive. Now, let me say that because it's not entirely intuitive. People recognize that a high central venous pressure connotes a full venous circulation. But even patients with a high central venous pressure, when their circulation is challenged, may experience augmented flow and stroke volume in a way that connotes a response to circulation. And so if we combine a large number of studies, it turns out that things like central venous pressure, while really good at telling you about the state of the venous circulation, are literally no better than the flip of a coin, area under the receiver operating curve, combined from meta-analysis, than about 50%. So predicting volume responsiveness is an inconsistent relationship between stroke volume and cardiac preload. We know that static measures like blood pressure, CVP, really are uninformative about this. And if you only guide therapy using those measures, can drive volume overload. Resuscitation strategies that target static endpoints, such as MAP and CVP, also result in outcomes no different than usual care. We know this broadly from the triad studies, process, promise, and arise. And fluid responsiveness, we would argue, should be identified before starting volume expansion to treat circulatory insufficiency. And so why do we believe that? We believe that based on an established physiologic principle of stressing or perturbing the venous capacitance system, where either a fluid bolus, or passive leg rays, or even micro boluses are trending, will stress preload, which, if the circulatory preload is low, augments stroke volume. We use an arbitrary but well-validated cutoff of 10% to indicate a fluid response state. But if the circulation is full, based on the Starling-Fulling curve, we appreciate that the same increase in stress volume results in very little increase in stroke volume. Using that paradigm, we appreciate the current strategies, including the Surviving Sepsis Campaign, emphasize that the quality of evidence to use this to use a non-agnostic approach are very weak. And I would argue, however, that since 2021, to underline my colleague's observation, it's time for us to update our guidelines. Why? Because regardless of how you measure stroke volume, invasively or non-invasively, and the work I'm going to show you uses a non-invasive technique. This is the Starling bioreactants technology that uses a non-invasive surface bioreactants approach to detect phase shifts that can be interpolated, similar to the Doppler shift phenomenon, to look at flow across the thorax of blood. You can estimate changes in blood flow, either in response to a passive leg raise, augmented blood flow here, or with a volume bolus that is consistent across cardiac cycles, respiratory cycles, and irregular heartbeats. And these changes, or changes in stroke volume to perturbation, can be very informative about whether there is a fluid response to state. Here's Carlos, our outstanding research coordinator, who consented, under some duress, to be photographed doing a passive leg raise. There's his legs up. We look at the change in stroke volume from baseline to three minutes after challenges can be shortened, actually. And he had a 25% increase in stroke volume. Now, is Carlos fluid responsive? Yeah, he is. Does he have microcirculatory insufficiency? No. Is he septic? Certainly not. I just told him not to drink anything from 6 p.m. the night before. So he demonstrates the effect, and the reward, in addition to much thanks, was one of those grande triple latte with the cream and the frou-frou, and I had a fellow like that one day. Cost me a fortune. And, of course, there's a lot of information. We, with colleagues at KU, have performed a series of experiments. I'm going to touch just on a small randomized pilot trial of 124 patients published a couple years back, where we randomized patients to an approach for early sepsis resuscitation, either using an intervention-based approach, passive leg raise, to measure change in stroke volume and inform further care. So this is not a fluid response agnostic approach. It uses this approach to minimize fluids or usual care. And we enrolled patients very similar to those you've seen in Clovis. We had a protocol that if the patient was enrolled early on after no more than a liter of fluid, had a passive leg raise that showed a 10% increase in stroke volume, the opportunity for a crystalloid bolus was provided, and the patient was then monitored, reassessed in cycle. And as we did in Clovis, there were opportunities for interventions for persistent hyperperfusion. The differences between the two arms were striking. The patients, again, were very, very similar to Clovis, older, pneumonia. As a sepsis diagnosis, we had a sex imbalance that was not significant in adjusted secondary analysis. But the outcomes were really important and interesting to inform future study. A large difference in 72-hour fluid balance, not just 72-hour fluid administration, but balance, a reduction by 2 thirds of the likelihood of progression to kidney injury requiring renal replacement. And this echoed findings from my colleague Heath Latham at KU in a pre-post study. A reduction by half in the need for mechanical ventilation, remarkable if you think about it. And numeric differences without significance in shortening of ICU length of stay, and likely of survival to discharge home. These data have been combined in meta-analysis. And even here, there is still insufficient power to show improvement in mortality, more work to be done in this regard. But you can see that the net trends of approaches that use fluid response guided resuscitation, so-called dynamic resuscitation, including the trials that we've performed, is in an interesting and provocative direction. So when we take the second part and think about it, physiologically informed fluid and vasopressor resuscitation that uses a more dynamic approach leads to lower fluid balance, reduced need for organ failure support. And these dynamic assessments could be improved on top of a new usual care. And I'm going to make an argument to you that it provides a rational approach to guide fluid bolus administration in a pressor-prioritized fluid resuscitation strategy. So now, instead of using a more naive approach or agnostic approach, let's inform the way we think about additional fluid to guide precision dynamic resuscitation. We're in the process of gearing up to do a very large 1,000-patient prospective randomized trial. We're calling it Fresh First. I'm looking for funders, since the DoD decided they weren't going to fund it this cycle. So anybody that's gotten in, help me here. But the protocol essentially will use a similar approach, where we randomize patients into two arms of a study, but here with undifferentiated early shock, use the change in stroke volume once they've had some initial fluids to guide therapy in a non-responsive guidance approach. And then once we've separated these patients in terms of their response to therapy, guide treatment in a fashion where mortality, organ failure support are the primary and secondary outcomes. So where does this leave us? I would argue that we're now in a circumstance where we actually have sufficient information to be ready for the next level. We can be comfortable that it's both safe and effective to use either a more fluid-prioritized or vasopressor-prioritized approach. But if you're going to use a vasopressor-prioritized approach, we have further indication, certainly in early pilot study work, to suggest that a less agnostic informed approach is imperative. That I believe will get us to the point of where we can now then talk about microcirculation, and that's more for another day's discussion. But perhaps it will become along that we can look at whether the microcirculation, particularly below the tongue in sepsis, can be significantly impaired, where you get microcirculatory shunting. We'll be able to understand the role of the endothelial glycocalyx, and as we're doing in grants in our group, measure the shedding of the microvascular endothelial glycocalyx as a change in the perfused boundary region. Recent work that we presented at ATS showed that this was both feasible and correlated with measures in the urine and plasma of biomarkers indicative of endothelial glycocalyx shedding and injury. And a further aspect, ultimately, is that we would then want to move on to the venous side of the circulation, and not only optimize the arterial filling and responsiveness, the macro-microcirculatory coherence, but think about the later phases which we want to avoid, which is venous excess abnormalities. So taken together, early identification, prompt and immediate reestablishment of effective arterial filling, oppressor-prioritized restrictive stabilization, dynamic fluid response-guided crystalloid bolus optimization in a fashion that will then ultimately, perhaps, lead to our ability to cohere the micro- and macrocirculation. I'm truly indebted to a really phenomenal group of home collaborators, the Petal Network, our Fresh Investigating Network, collaborators in the room from Baxter, Doug Hansel, Jen Satchin. Now you know why, and I used to say during COVID, you need to believe. My thanks very much.

hemodynamic resuscitation

sepsis patients

integrated approach

fluid responsiveness

microcirculation

stroke volume