Thanks very much. All right, so we have two talks now on sepsis. First is going to be on the pathophysiology, and the second we'll talk about treatment. And I tried to link them a little bit, so some of the things that we talk about in the first talk will then be addressed somewhat in the second talk. So I guess my only disclosures are that I am going to talk about the Phoenix criteria a little bit. They're maybe a little too new to be on the board exam, although they're certainly making their way into the MOC questions. And then I'm involved with the Surviving Sepsis campaign, which I will talk a lot about in the treatment talk. All right, so first I think it's important to just think about what sepsis is, right? And it is really not, in and of itself, a diagnosis. It's a syndrome that represents a spectrum of pathophysiologic responses to infection. And involved in that is some combination of inflammation, abnormal perfusion, and organ dysfunction to different degrees and with some complex underlying pathophysiology that we'll go into. It is recognized by a set of nonspecific clinical and laboratory parameters, which makes it really challenging because there's a lot of things that it overlaps with. But in and of itself, it's not a diagnosis, right? So it's really important that we take it a step further, right, and understand why a child might have sepsis, right? So you need to identify the source of the problem, and that is, therefore, the infection that is driving the underlying sepsis response. There's some important terminology to be familiar with. So this is the now sort of older set of terminology that was used to define the spectrum of sepsis, and that is still very commonly used in the literature. And at the bedside, to be honest. And it's, I think, important to be aware of. And there may still be some board questions that reflect this older nomenclature now. But the systemic inflammatory response syndrome is two or more abnormalities in temperature, heart rate, respiratory rate, and white blood cell count. In pediatrics to have SERS, one of these must be either abnormal temperature or abnormal white blood cell count, so it's not enough to just be tachycardic and tachypneic. And if then, at least according to this nomenclature, if you had SERS caused by an infection, that was sepsis. If you had the addition of organ dysfunction, that was severe sepsis, and then cardiovascular dysfunction was septic shock. So in children, it's long been recognized that hypotension, particularly in young children, can be a late finding. And so, therefore, hypotension while present or if present is helpful, but you can still have cardiovascular dysfunction in the absence of hypotension. And so it's really the presence of abnormal perfusion, of which there are many different clinical and laboratory markers that really identify as a child with having septic shock. More recently, sepsis 3, which is really an adult set of criteria, but kind of reestablished a new set of nomenclature around what is sepsis and septic shock. And basically, it sort of said the difference between an infection with or without sepsis is not inflammation, but rather the presence of organ dysfunction. And so it took this concept of severe sepsis and sort of changed that to sepsis. And so now, at least as of 2016, with the sepsis 3 criteria, sepsis is life-threatening organ dysfunction caused by a dysregulated host response to infection. The reason it's called sepsis 3, just in case you're wondering, is there were two prior adult task force, and so this was the third iteration, so that's why it's sepsis 3. And then septic shock was also advanced, and so it's not just circulatory dysfunction, but rather also the presence of cellular and metabolic abnormalities, which I think Jerry explained nicely a little bit earlier, and that was operationalized as an elevated lactate. More recently, the Phoenix criteria were published just earlier this year. The definition is the same as sepsis 3, but the operational criteria are different. And so in children, sepsis is defined by having two or more Phoenix sepsis score points, which is basically life-threatening abnormality in respiratory cardiovascular coagulation and or neurologic dysfunction, and then if you have at least one cardiovascular point, then you have septic shock. So I think it's really important that we don't get too lost in the terminology, but it is important to be familiar with them. But for a clinical practice, it's really important to recognize that sepsis is somewhat, it's an arbitrary definition, right, because the reality is this is the spectrum between uncomplicated infection and infection that leads to death. And somewhere along that spectrum, you have inflammation and then organ dysfunction, which itself has a spectrum, and then shock being at the severe end of that. And the older nomenclature that I showed you initially had this dichotomy, where sepsis was the presence or absence of inflammation, and then you had these severe sepsis and septic shock modifiers. Sepsis 3 said, no, no, no, it's not inflammation, but organ dysfunction that delineates sepsis, and so it drew the line there. And then Phoenix now kind of moved it a little bit over and said, well, it's not actually any organ dysfunction, it's organ dysfunction that's severe enough to actually increase the risk of mortality. And so that's kind of how things have progressed over time. When we, again, as I mentioned earlier, when we talk about sepsis, we need to think about the infection. So blood culture's helpful when they're positive, but only positive in about 30 to 60% of septic children. The epidemiology of sepsis, at least in terms of bacterial infection, which is the most common, is pretty equal between gram-positive and gram-negative, with gram-positive organisms, staph and strep, and gram-negatives, mainly E. coli, klebsiella, and pseudomonas, although obviously there are other organisms as well, but those are the most common. And then the most common sites of infection, respiratory, bacteremia, genitourinary, and abdominal. So those are the most common four sources of sepsis. So if you get any questions about epidemiology, that's kind of important to know. And that's been consistent over many different epidemiological studies. The sites of organ dysfunction, which again now are, with both adult and pediatric definitions, the defining feature of sepsis, right, that you have organ dysfunction. In children, we think about, used to be six, now eight organ systems that could be dysfunctional. So cardiovascular, respiratory, coagulation, and neurologic are the Phoenix-defining organ systems. So you need dysfunction in at least one of those organ systems to technically meet those criteria for sepsis. But you can also have organ dysfunction in other systems, such as renal and hepatic, and those are important and increase your risk of mortality as well. And then there are two new systems that Phoenix recognized that include endocrine and immunologic. So endocrine is operationalized as glucose problems, high or low, and immunologic is really meant to be an immunosuppressive state, but is operationalized using neutropenia and lymphopenia. And we'll talk more about the immunosuppressive features of sepsis in a moment. The accumulation of organ dysfunction is the most important risk factor for death in, well, in pediatric sepsis, but also in adult sepsis. And so every study that's ever looked at this, no matter how you define organ dysfunction, what criteria you use, as you accumulate organ systems, your risk of death goes up. This is data from the Sprout study, but it's pretty consistent across every study that's looked at this. In terms of outcomes after pediatric sepsis, the mortality varies dramatically depending on the population. Overall, if you look at all comers, most of whom are not that sick and do well, mortality is about 3% to 6%. But if you start looking solely at patients who make their way to the ICU, it's somewhere between 5% and 25%. About a third have functional disability long-term after sepsis. We know that from the LAP study that Jerry ran. And then readmissions for recurrent infections are also fairly common, so one-fifth to one-third of patients. Children tend to die in a bimodal fashion after sepsis. So there's a group of kids, about 20% to 30%, who die early, within the first 24 hours. These are patients who generally have refractory shock and die typically of cardiac arrest. And then there's a much bigger group of children who die over a longer time period, usually with persistent mods. These are patients who get better somewhat, they stabilize, but they never truly recover. And oftentimes we withdraw life-supporting technology. Risk factors for death include what I said is the most common, or the most important, which is number of organ dysfunctions, things like comorbid conditions, hospital-acquired sepsis over community-acquired fluid overload, and then young age are key risk factors. All right, in terms of the pathophysiology, this is going to be kind of a whirlwind because there's a lot. We could talk a whole day on this, and we have like 12 minutes. So we're going to go a little bit quick. There's a little more detail in the slides than I'm going to go over here, but I did want to hit on a few key points. First is that the pathophysiology of sepsis is complex, and it's redundant. So it's not one thing, and that's why it's been so difficult to treat. But in general, there's an infection which incites all of this, and that leads to several downstream effects. It usually starts out with an innate immune response, often in response to pathogen or danger-associated molecular pathogens. These are conserved antigens that stimulate certain innate immune factors, and we'll show those in a moment. And then the adaptive immune response is largely driven by T cells. Although B cells are involved, T cells seem to be the predominant and most important and certainly the well-characterized adaptive immune component. We'll talk more about that. Then there's activation of complement, which I'm not going to get into today. There is a transition towards a procoagulopathic state. I'll show a little bit about that. And then the endotheliopathy of sepsis, we'll talk about that. And then there's some other components that are downstream of all of that, and that includes microvascular shunt, microthromboses, interstitial edema, tissue hypoxia, and then mitochondrial dysfunction. So we'll touch on, again, some of these things but not all of them. So the innate immune response driven by PAMPs and DAMPs is important and can be kind of harrowing when you look at it. This is a relatively simplified version of all the different possible PAMPs and DAMPs and their sort of signal transduction pathways. But there are a few key ones that the boards tend to ask about. And in particular, those are the PAMPs that activate toll-like receptor 2 and toll-like receptor 4. So toll-like receptor 4 is your classic LPS, so gram-negative endotoxin trigger. And then toll-like receptor 2 is your gram-positive bacteria. So things like lipozoic acid activate toll-like receptor 2. So those are fair questions for the boards. There are some intracellular toll-like receptors that are triggered usually by viruses, but also by mitochondrial DNA. So you heard this morning about mitochondrial DNA being like bacteria, and that's true. So generally when mitochondria become dysfunctional, mitochondrial DNA gets released into the cytosol, and that triggers toll-like receptor 9. And then the viral RNA and DNA tend to trigger toll-like receptor 3, 7, and 8. So 3, 7, and 8 viruses, toll-like receptor 9, mitochondria, toll-like receptor 2, gram-positive bacteria, and toll-like receptor 4 is gram-negative. And then the boards also tend to like high-mobility groupox-1, which tends to be a late damp, and that works through RAGE, which is receptor-advanced glycosylation end products. The soluble form is actually measurable and is associated with the development of ARDS in sepsis. So that's another one you might hear about. So high-mobility groupox-1 and its receptor RAGE or soluble RAGE. All right, on the adaptive immune side, T cells are the main things you want to know about. And this could be very complicated, but the key principles here are the sort of pro-in... The first key principle is the pro-inflammatory milieu that T cells tend to drive. And they do this through two pathways, the Th1 response, which is your sort of classic CD4 and CD8-positive cells that tend to upregulate inflammation. And this is actually adaptive initially, right? So they help to reduce bacterial load. And then the Th17 response, which is named because it's driven by interleukin-17. And this is sort of a secondary pathway that's triggered more by macrophages that induce T cells, cytotoxic T cells, to, again, try to control the inciting infection. T cells also, however, are important with the anti-inflammatory response. And so you have these effector T cells, your CD4 cells, your CD8 cells, which are helpful to stimulate inflammation and control bacteria. And then that tends to give way to a T cell phenotype that's more immunosuppressed. So you have these upregulation of different subsets of T cells, such as T regulator cells or Tregs, as well as myeloid-derived suppressor cells, and a macrophage phenotype that shifts from a pro-inflammatory state, the M0 or M1, into an M2 state. And that all is anti-inflammatory and immune-suppressive. That tends to drive anti-inflammatory pathways through IL-10, IL-4, and TGF-beta. Those are your three key anti-inflammatory cytokines the board tends to ask about. And again, the whole point of that is to try to bring inflammation back towards homeostasis. The problem is if that becomes excessive, then you end up getting immune suppression. In addition, there are other factors that tend to exacerbate this. And so when T cells are activated, they protect themselves from being overstimulated by having checkpoint proteins, such as PD-1 is the sort of most common example, although there are others. So there is upregulation of PD-1 and its associated ligand, PD-L1, as well as loss of antigen-presenting proteins, such as HLA-DR. So the loss of HLA-DR and the upregulation of these checkpoint proteins, such as PD-1, again, further exacerbates this sort of immunosuppressive phenotype. And so while these are, again, adaptive to try to bring back inflammation and get you back to homeostasis once the infection has been treated, in sepsis, these appear to be upregulated in a somewhat uncontrolled fashion and can lead to immune suppression. There's also a third pathway that's not as commonly talked about but is increasingly published on, and I think we'll start to see some questions about this on the boards, and that's the neuroinflammatory reflex and autonomic nervous system dysfunction. And the reason this is important is because it's now measurable. So there's a lot of studies that are published on lack of heart rate variability as a marker of this. And so I think we're going to start to see this more in not only clinical use but also on boards questions. So the concept here is that the efferent arm of the vagus nerve helps to control inflammation, right? So it has a lot of anti-inflammatory effects, and it does that through hormonal release, through acetylcholine increase in various organ systems, as well as direct effects on immune cells, usually in the spleen. And in sepsis, what happens is you get this imbalance, and you get excessive sympathetic and a relative decrease of parasympathetic nervous system activity. And that leads to the loss of heart rate variability, but more importantly, it leads to a very pro-inflammatory environment. And so there's actually some evidence that if you can do vagal nerve stimulation, that you can actually reduce inflammation and control inflammation that way. But that imbalance, I think, is the key thing to know pathophysiologically. And so what we get is this sort of imbalance between inflammation and anti-inflammation in sepsis, where ultimately some patients are very inflamed, and they have this sort of excessive inflammation. And you have other patients that sort of continue to have this anti-inflammatory or immune-suppressive phase that becomes pathogenic, and you don't have this return to homeostasis, right? And those are the patients that tend to have the worst outcomes. I mentioned the altered coagulation. I'm not going to say a whole lot here other than to say that there's an imbalance. And in general, the things that tend to drive coagulation go up, and the things that tend to inhibit coagulation go down. And so you get this pro-coagulatory environment, which tends to propagate microthromboses and leads to microvascular shunt, which I'll talk about at the end. The endothelium is a major player in sepsis pathophysiology. Most people refer to it as endothelial activation. Some people refer to it as dysfunction. There's probably a spectrum where it is activated and helpful to some extent and then eventually becomes dysfunctional. We need that endothelium to become activated in such a way that you can bring your immune system, your circulating white cells, to the source of infection. And so, you know, that's why leaky vessels and things exist, right, so that you can have permeability and leukocyte trafficking. But obviously, if those things become excessive, then they become pathogenic. Other parts of the endothelium that are important to know are the glycocalyx. So these are the glycopeptides that kind of stick off the endothelial cells that serve a really important role in regulating the passage of material through the endothelium. And then you also have, and that's shown here, and then you also have loss of different cell junction proteins. And all of that, again, initially adaptive, but becomes counterproductive if it becomes excessive. All right, so moving away from the molecular pathophysiology to some more of the clinical findings. So you heard about shock earlier today. I think the important thing is that septic shock isn't really a shock type. It's really a combination of three different shock types, to different extents in different people and actually to different extents over time in the same patient. So some patients are more hypovolemic, others have more cardiogenic component, and some are more vasopelagic with a distributed picture. Regardless, that leads to this imbalance between oxygen delivery and oxygen consumption. I'm going to just go over this quickly since you heard about this, but I like this graph a lot. And you might see these terms questioned on the board. So initially, oxygen delivery, even if it decreases, so if you move from right to left side of this graph, oxygen consumption is maintained constant, and that's the supply-independent phase of oxygen consumption. And the key clinical features of this is that you have increased extraction of oxygen from the blood, which results in a lower systemic venous oxygen saturation. And then if you hit this critical point, then any further decrease in oxygen delivery is going to cause a decrease in oxygen consumption. This is the supply-dependent phase of oxygen consumption, and this is indicated by an increase in lactate production, suggesting that you shifted to anaerobic metabolism. This critical point can shift and tends to move rightward on this graph when you have infection, inflammation, and sepsis. And so you can see that supply-dependent phase even at levels where you might not under a different condition. I'm going to skip the lactate stuff since Jerry explained it nicely. This is just a graphical presentation of the fact that hypotension tends to be a late finding, in particular in infants and very young children, because vascular resistance can be maintained in these patients, and so blood pressure is maintained even though cardiac output is falling. So be careful with that. These kids can maintain a blood pressure until they can't, and then it becomes very serious very quickly. The clinical features of sepsis in children differ from adults. We've known this for quite some time from a very nice study done in Pittsburgh in the 1990s with pulmonary artery catheters, when they were still commonly used. And from that study, we learned that the predominant presentation of pediatric shock is actually not distributive but rather more cardiogenic in its presentation with a low cardiac output and a relatively high systemic vascular resistance. And those are the patients that also tend to have the worst mortality, or at least they did then, probably still true now. The clinical features of warm versus cold shock are shown here. I think you're familiar with these. I'm going to talk more about this in the treatment talk about why, while it's important to recognize there's some limitations in how we can actually use these. And I'll come back to that. The microcirculatory dysfunction, so I mentioned this. I think it's important to really walk through why this is important. We focus a lot on the macrocirculatory components of cardiovascular dysfunction when we think about heart rate and blood pressure and urine output and mental status and so on, but less so about the microcirculatory dysfunction, predominantly because it's harder to measure and harder to treat, but it's really important to kind of conceptualize. So if we take a normal situation, so this is a normal tissue bed, where we have an influx of 240 mLs of oxygen per minute into four different arterioles. And so with homogeneous blood flow, which is normal, then 60 of that, 240, gets divided equally amongst the four blood vessels. There's an oxygen consumption. In this example, it's 24, so 24 from each area around each arteriole. And then what comes out of each vessel is 36. And then if you did the math, that's actually an extraction of 40%, which is about normal, so your SVO2 is 60%. When you have low flow state, so this is sort of pure cardiogenic shock, then you deliver less oxygen. So now we're only delivering 120, but it's still equally distributed. We still have the same amount of oxygen consumption because we're in that supply-independent part of our curve. And we can see that our SVO2 is down. So we're delivering less, we're extracting the same, so our extraction ratio is higher, so our SVO2 is lower. The problem with microcirculatory dysfunction in sepsis is you get this heterogeneous flow pattern where some vessels are perhaps clotted off. And so you might be delivering a good amount. Maybe you've worked really hard to resuscitate your patient, so you've improved oxygen delivery, but now you have less arterioles available for that blood to go through. So that 240, instead of being divided by 4, is now divided by 2. You have the same oxygen consumption in those local tissue beds. So now what comes out is that you've extracted less oxygen from that blood overall, and so your SVO2 is actually elevated at 80%. So an elevated SVO2 in sepsis could be that you've done a really good job resuscitating them and improving their oxygen delivery, and you probably have, but it also is a warning sign that perhaps there's some microcirculatory and or mitochondrial dysfunction, which is harder to treat. One proposed way to look at this has been an elevated in the venous arterial CO2 gap. So this is normally you should have, if you measure oxygen, sorry, carbon dioxide in your vein and your artery, the gap should be less than 6. So if it's more than that, that's abnormal. And that could suggest that you've got some microcirculatory dysfunction. And in this study, I think nicely shows that mortality was generally in that area. And then lastly, you'll hear a lot more about these subphenotypes in sepsis. And this is why I mentioned at the very beginning that sepsis is a syndrome, not a diagnosis. And so you need to sort of think about treatment differently than saying we just treat sepsis, right, which I am going to talk about in a moment. But thinking about what particular kind of sepsis you're treating. And so we now know there are patients who are highly inflamed and stay that way. There are other patients that are immune paralyzed and can't recover from that, and they need very different treatments. I'll skip that. So key points to take away. So sepsis-associated cardiovascular and respiratory dysfunction are most common, but there's eight organ systems you need to worry about. MODS is the key risk factor for death. It begins with an infection, but then leads to a very complex and redundant host response. And we need to start thinking about sepsis subphenotypes, because that's really where the future is in terms of treating sepsis. So I will pause there.

sepsis

pathophysiology

organ dysfunction

pediatric sepsis

immune response

clinical approaches