Hello, everyone. Welcome to the SCCM Virtual Year in Review Part 2. Thank you, Dr. Riley, for sharing your insights on COVID-19. And now I'm going to share my insights on sepsis. My name is Lakshmi Santosh, and I am an Assistant Professor of Pulmonary and Critical Care at UCSF, University of California, San Francisco. I'm also the Associate Fellowship Program Director, and you can find me on social media at LakshmiMD. I have no conflicts of interest to disclose. And my alternate title is, A Break from COVID, Back to Bread and Butter. Who knows? We'll find out, we hope. So no conflicts to disclose. Thank you to the SCCM. So when we think about, you know, how do we measure a year and the year interview in sepsis, there is just so much great literature that came out in the world of sepsis. And so this is not going to be an exhaustive journal club where we dive into the literature and each specific table. This is more going to be like a tour, a big picture overview of the year and review in sepsis, kind of the big highlights, the big papers that you should know about, and that hopefully you'll use the rest of the SCCM virtual conference to check out further, more in-detail sessions on some of these milestone papers. So we are going to take the following outline. We're going to, of course, talk about our new surviving sepsis guidelines with a big focus on what's new and what's changed. So we're not going to go through all the recommendations, but really highlight the new recommendations. We're going to talk about the different studies on the perennial question in critical care, which are the best IV fluids to use. We're going to talk a little bit about policy implications and big data, machine learning and sepsis, as that's a hot topic. And we're also going to talk about what happens to our patients sort of beyond their sepsis hospitalization as intention is increasingly paid to post-intensive care syndrome or PICS. So like I mentioned, we have this landmark milestone of the new surviving sepsis guidelines with 93 new recommendations that came out in October of last year. And so we're going to give you the TLDR version, the too long, didn't read version, summarize things for you. But we encourage you to check out the document. So very big picture takeaway is we want to keep doing some of the great critical care evidence-based practice that we've been doing with focusing on early antibiotics, ideally within the first hour. We're going to keep using norepinephrine as our first line vasopressor of choice. We're going to keep addressing goals of care early in critical illness. The QSOFA, which was very hot a couple of years ago is a little bit de-emphasized now. We are definitely not going to be using starches or gelatins as study after study have shown that they do not help. And hopefully the nail is finally in the coffin of vitamin C, which again, we have another study, many studies, most recently the AXE trial showing that it does not help. We're going to start talking about balanced crystal woods. And we'll dive more about that in our section on IV fluids. We're also going to keep doing high flow nasal cannula and sepsis with hypoxemic respiratory failure, high flow is preferred to non-invasive. And then the recent trials, you know, adrenals approaches, et cetera, talking about the use of steroids in refractory shock when you're on multiple vasopressors. So let's dive into some of these high level recommendations in a little bit more detail. So the surviving sepsis guidelines and the campaign thought some of these recs are just very critically important and warrant their own infographic poster to disseminate widely within your hospital. So they need to, the first one is about antibiotics timing and they say, you know, whether shock is present or absent, if you think sepsis is definitely happening or probably happening, we need to administer antimicrobials immediately, ideally within the first hour to get that early mortality benefit. And that even if it's possible, if it's quite high on your diagnosis, thinking about administering these antibiotics immediately, again, if you're in shock, it's easy to just give those early antibiotics, broad spectrum antibiotics. And even if shock is not present, you need to think about early antibiotic therapy, but also if shock is not present rapidly reassessing to see, could this be a sepsis mimic? Could this be non-infectious contributors to sepsis? And that you have a little bit more time if the patient is not in shock to kind of distinguish, am I in sepsis or a sepsis mimic? And they talk about, you know, doing it history, of course, physical exam, relevant labs, further data gathering to make that definitive assessment. Is this sepsis that I'm dealing with, or is this a sepsis mimic or a non-infectious cause? The second big kind of infographic that they put up that they wanted to emphasize was about vasopressors. And again, we're reaffirming and strongly recommending norepi is the first line vasopressor, targeting a map of 65. And that, you know, some recommendations that were more recommended, not as strongly where things like considering an A-line for arterial monitoring of pressures, a new recommendation saying that peripheral vasopressors, you know, would be, should be strongly considered. It should be considered if you are unable to have central access just yet. And they know in the asterisks here, you know, especially make sure all the usual precautions that this is a large bore, well-situated, closely serially monitored IV that you're doing peripheral vasopressors in, that you have metrics after which you are going to need to switch to central access. This is a temporary thing. This is not indefinite, et cetera. And that, you know, if the map is uncontrolled, despite norepinephrine, we're going to add vaso. And that if we think that there's concurrent cardiac dysfunction, or if there's persistent shock, despite, you know, volume resuscitation, thinking about whether we need to add, you know, either some inotropic support with dobutamine or switching to epinephrine. So these are kind of our tried and true recommendations, just reemphasizing these with very few tweaks from the recommendations that we've followed in the past. And so who specifically kind of got upgraded or downgraded that they really know? So they have a nice table in the surviving sepsis guidelines that actually talk about explicitly who got upgraded, who got downgraded. And so one big downgrade was the infamous 30 cc's per kg fluid resuscitation recommendation. And, you know, especially since that was tied often to incentive payments and performance metrics, many frontline providers, you know, clinicians in the emergency room and urgent care and critical care felt like this blanket recommendation of 30 cc's per kg was not always appropriate, not always appropriate for people who had heart failure, kidney failure, liver failure, etc. And so this is less of a strong recommendation that's now been downgraded, who got upgraded. So the recommendations for using steroids and refractory shock, and specifically, they call out a dose of IV hydrocortisone 200, if you're in persistent shock, despite multiple vasopressors. And then of course, the recommendation again, for crystalloids being significantly stronger than colloid, gelatins, starches, that was also called out specifically as well. So this is kind of a quick summary of who got upgraded and who got downgraded in the latest guidelines. So switching gears, because we've already been talking about fluids, the perennial question comes up of which fluids and we had a bunch more trials that, you know, it didn't clarify things completely. And so there's more for us to study. So the first trial that we're going to talk about is the basics trial, of course. And I'm going to use this format for my slides where kind of the PICO format where we're going to talk about our population, what was the intervention, what was the control group, and what was the outcome study. So the basics trial looked at 75 Brazilian ICUs patients in there, and they did a two by two well-designed trial, it was an RCT of balanced solution, like plasmolyte versus normal saline. And they also looked at slow versus fast infusion rates. So they were looking at again, plasmolyte or balanced crystalline solution versus normal saline. And their primary outcome was 90 day survival, 90 day mortality, and they found no difference actually between saline and plasmolyte. And the when they did look at the second part of that analysis of the rate of infusion, which addresses the question of, you know, how quickly should we resuscitate, and they looked at, you know, again, these 10,500 plus patients, randomizing them to 333, which was slow, and 999, which was fast 999. And they found that, you know, again, no significant difference, which should be somewhat more reassuring that if you need to bolus wide open at 999, you know, there's not harm there. So I think that arm was actually more reassuring provided reassurance to practice that once you need once you think someone is fluid responsive, you need that to give them a bolus wide open, there's no harm for giving it acutely at that fast infusion rate of 999. And so the next study that came out soon after was the plus study. And the plus study looked at 5037 patients in 53 ICUs in Australia and New Zealand, again, looking at plasma light versus normal saline and 90 day mortality. And this was a trial that affirmed the basics trial and that there were no significant mortality differences, and no significant real outcome differences. You see here the Kaplan Meyer curve with almost complete overlap between the saline group and the P light group, you know, overlapping curves, and you see the maximum creatinine increase in creatinine receipt of renal replacement therapy, also not significantly different either. I think the main big kind of critique of this trial that people have discussed is that the majority of the plasma light group, as is done in most institutions, are receiving a lot of medications that are still running in normal saline. Again, if you have septic critically ill patients, they're getting a lot of medications like antibiotics and other continuous infusions that are usually prepared in normal saline. So the question was, did this perhaps attenuate the benefit scene of the plasma light group. And so, you know, we have prior literature, you know, the SMART trial and others, which had shown us that the preference for for balanced crystallide solutions like plasma light, and then these two trials plus and basics, raising the question of, does it really matter? Are we back to that therapeutic nihilism and critical care of it, that it actually doesn't matter, volume is volume, as long as you're using a crystalloid. And so I think further studies and meta-analyses are on the way to really decide this. I think for now, at our institution, the practice is still to favor, you know, based on the earlier trials to favor plasma light as a balanced crystalloid solution. And that, you know, time will tell about how these benefits hold up, or if it does matter or not. But I think that this argument of whether this benefit was attenuated is an important one. And we do have prior literature with the SMART trial and others showing the benefits of balanced crystalloids. So I'm not changing my practice quite yet to abandon balanced crystalloids, but we'll see, we're watching that data closely. And the other thing to know is that fluids are, of course, just one component of our resuscitation. I like this graphic from Jenser and colleagues in CHEST, where they talked about, you know, early shock, severe shock, refractory shock, that we have a bunch of tools in our toolbox, right? And fluid resuscitation is just one of those many tools that we're using. In addition to vasopressors, of course, we're thinking about the whole path of physiology and identifying and treating hypovolemia with fluids. But also we're treating, you know, most of these patients have severe acidemia, many of them are on continuous renal replacement therapy. If they're hypocalcemic, we're giving them calcium, we're thinking about how to combine our vasopressors, you know, do we need to add vaso or epinephrine? We're thinking about our adjunctive agents, as mentioned, like hydrocortisone. We're thinking about do they need rescue therapies like methylene blue, things like that. And so, you know, fluids are just one tool in our toolbox that we have to remember. So don't get too nihilistic. And, you know, another graphic from the same article, just showing us that, that we have to think about what people's cardiac output is, and their fluid responsiveness and really assess those in a two by two session, what is the cardiac output? What is the fluid responsiveness? To think about, of course, just like we teach our med students on the first day, you know, what type of shock are we in? And how do we best treat that? And so that's just a quick couple of moments thinking about our perennial question of which fluid and so the jury is still out. Next, moving on to this concept of big data and sepsis, there was a lot of trials and controversy about the role of big data and machines, man versus machine when it came to sepsis. And so this is a graph from my institution that is often heralded as that, you know, our sepsis mortality, once we started the electronic sepsis surveillance, we use Epic as our EHR in the emergency room and on the inpatient units, you know, our adult sepsis mortality actually dropped. And so this is a graph that we see in a lot of our QI meetings, etc, showing about the benefits of electronic sepsis surveillance with tools like our EHR Epic. And at the same time, we recognize the limitations, right, we have to resist that lactobolus reflex and always click reorder whenever we get a call from the bedside nurse saying the lactate is slightly elevated. And so how do we kind of reconcile these and this has been in the news a lot recently, you know, there is this stat news investigation with Epic saying that the algorithms for which they're detecting sepsis are under a corporate firewall, and they're not necessarily accurate. And there was another article that came out just a few weeks ago, talking about how subtle shifts in patient data over time can actually significantly change the outcome, which undermines patient safety. And so in the literature, the big trial that came out that made these news news waves was this trial in JAMA Internal Medicine by Wong and colleagues, which actually looked at a very robust data set of 38,000 hospitalizations at Michigan, and found that the prevalence of sepsis in these patients doing, you know, chart review was about 7%. And Epic missed 67% of those cases. And there were though, sepsis alerts, the familiar sepsis alerts triggered for 7000 patients about, but the positive predictive value is only about 12%. So low positive predictive value, lots of false positives. And that, I think, jives experientially, what many of us see is that we just click through the sepsis alerts, because they're firing kind of indiscriminately, you know, on 7000 patients, when it's actually only 2500 patients. And so that tracks with with my experience, at least. And there's been a number of trials that have come out since then a lot of them from Dr. Prescott's group, Dr. Haley Prescott's groups, you know, this one looked at what are the encounters that people have with the healthcare system in the week before sepsis, I really like this paper, because they basically talk about how most patients who come in later with sepsis or septic shock have seen a healthcare provider, whether it's ER, urgent care, outpatient clinic, you know, primary care in the preceding week. So we have a lot of opportunities actually to catch that sepsis and intervene early. Nice paper. This is another paper from colleagues at Washington University in St. Louis looking at sepsis prediction tools, using AI machine learning for the general ward setting. So how do we kind of detect this predict this before these patients even hit our ICUs. And another paper from Dr. Prescott's group and annals of ATS looking at, you know, different methods to define baseline organ function when you're looking at sepsis. So, you know, even though there were these terrible headlines about epic and EHR algorithms, you know, not being sensitive enough, there's a lot of great research from fantastic investigators, critical care investigators, who are really trying to change this and make things more precise and actually try to detect sepsis in patients before they clinically present. So it's an exciting area to watch, I think, in the years ahead. And lastly, to round out our year in review, thinking about the ongoing literature about what happens for our patients beyond sepsis. And, you know, as we know, there's about 1.4 million sepsis survivors every year. And these patients are not, quote, back to normal, they're left not only with physical challenges, but also mental health challenges, cognitive challenges, emotional challenges. And I think there's more awareness of this now, perhaps with COVID and long COVID. But for our bread and butter, you know, core critical care patient population, this has been a huge issue for for decades. And, you know, I like this graphic from from from nature and it reviews nephrology showing how there's this early mortality, of course, in sepsis from the systemic inflammation. But there's also this delayed mortality and long term mortality related to, you know, long lasting immune dysregulation, perhaps increasing susceptibility to things like cardiovascular disease and cancer. And again, this is also being aggressively studied and researched during COVID. But we know that even our sepsis survivors not from COVID are experiencing, you know, long term sequelae as well. And so, you know, in COVID, we know well that this great figure from Nalbandani and colleagues in nature about the chronic symptoms that happen post acute illness, post critical illness, and in many cases, even without being hospitalized, everything from general symptoms to pulmonary symptoms, neurological cardiac, you know, increased susceptibility to blood clots, renal injury. And that's why I think you've seen this last pandemic year or two, a proliferation of not just post ICU clinics, and post COVID clinics, but more centers that are linking them explicitly together. Because I think what we found that our PICS framework, which actually looks systematically at mental health, pulmonary health, cognitive health and physical health, and focus on family caregiver health is very translatable from PICS to COVID. And so this study with the Cairo group of post critical illness clinics, run by Dr. donation colleagues, looked at, you know, all these new clinics that have sprung up with post ICU and post COVID clinics really around the US. And they found, you know, again, if you've anecdotally been hearing about a lot more clinics popping up recently, that's true. And I think that's because you see here, this huge number of ICU recovery clinics that started in 2020. And I think many institutions like ours at UCSF, they wanted ish to do a PICS clinic for many years, but there was a lot of barriers. And then when COVID hit, you know, navigating those barriers creatively to create these longitudinal and sustainable ICU recovery clinics was very helpful. And so I hope that in this half hour, we talked about, you know, the new surviving sepsis guidelines, took a whirlwind tour, talked about the new literature about sepsis and septic shock, fluid management, we talked about big data in sepsis, and what are the latest tools show us, and a little bit about what happens beyond sepsis, and how do we take care of patients when they're surviving critical illness and septic shock. And so if you thought your favorite article wasn't quite here, addressed by me nor Dr. Riley, you know, again, it's too bad we couldn't meet in Puerto Rico, but we'd love to hear from you. Feel free to email me or find me on social media. Thank you so much for listening. And thank you for your time. Enjoy the rest of the virtual conference.

sepsis guidelines

early administration of antibiotics

norepinephrine as first-line vasopressor

intravenous fluids

big data

long-term effects of sepsis