Hello, my name is Scott Weiss and I'm a pediatric intensivist at the Children's Hospital of Philadelphia and it's my pleasure to present this lecture entitled Pathophysiology of Pediatric Sepsis. I have no relevant disclosures related to what I'm going to talk about other than I'll say that I do research on mitochondrial dysfunction, which I will mention during this presentation. By the end of this session, you will hopefully learn about the epidemiology of pediatric sepsis, an overview of the pathophysiology of sepsis, including common clinical features, and appreciate the different hemodynamic subtypes, including warm versus cold shock, thrombocytopenia associated with multiple organ dysfunction, and sepsis, HLH, MAS overlap. So sepsis is really a spectrum of pathophysiologic response to infection that includes different levels of increased inflammation, abnormal perfusion, and organ dysfunction. Rather than being a diagnosis, it's really a syndrome that involves an underlying infection. And this syndrome is based on a presence of several nonspecific clinical and laboratory parameters. In and of itself, sepsis is not a diagnosis. It's not enough to recognize that sepsis is present. One must go further to identify the source of the problem, which for sepsis is the underlying infection. So in 2005, there was a publication of consensus criteria to help uniform the terminology that researchers and clinicians were using. And this very much followed the prior adult concept of the spectrum of sepsis, starting with systemic inflammatory response syndrome, or SIRS, which is two or more abnormalities in temperature, heart rate, respiratory rate, or blood cell count. And for pediatric SIRS, one of these abnormalities must be either temperature or white blood cell count. Sepsis is the presence of SIRS with infection. And severe sepsis is the presence of sepsis plus organ dysfunction. Septic shock is when that organ dysfunction includes cardiovascular dysfunction. In practice, severe sepsis and septic shock is generally how we use these terms at the bedside. And the more updated definitions from Sepsis 3 have eliminated the terminology of severe sepsis and just used the phrase sepsis to refer to infection with organ dysfunction and septic shock with cardiovascular dysfunction. It's important to note that in pediatrics, shock and cardiovascular dysfunction does not necessarily have to be hypotension, but is rather defined by the presence of abnormal perfusion, which can be hypotension, but could also be hyperlacteamia, acidosis, oliguria, delayed cat refill, or other signs or symptoms of abnormal perfusion. So as I mentioned in Sepsis 3, the adult group has redefined adult sepsis as life-threatening organ dysfunction caused by a dysregulated host response to infection, essentially redefining severe sepsis as sepsis. This takes away the confusion of calling SIRS plus infection sepsis in the absence of organ dysfunction, and rather redefine sepsis as only having infection plus organ dysfunction. And then septic shock was expanded to not only include circulatory dysfunction, but also cellular and metabolic abnormalities, which was operationalized as an elevated lactate, which for pediatric septic shock is already included in the definition of abnormal perfusion. So if we think about the pathway from uncomplicated infection to death, we see that patients at the time can develop systemic inflammatory response syndrome, which can progress to varying degrees of organ dysfunction. And then if that organ dysfunction is cardiovascular, that represents shock. Sepsis 3 defines septic shock as the presence of cardiovascular dysfunction with infection, whereas pediatric sepsis still uses the slightly older nomenclature where sepsis is defined as infection plus SIRS, and severe sepsis is when organ dysfunction develops. So when we talk about infection, we often mean bacterial infection, but sepsis can be triggered by viruses, fungi, rickettsia, or parasites. In addition, the infection is often not able to be definitively identified, and so the presence of a clinical syndrome with a high probability of infection should suffice for thinking about sepsis or septic shock. These include presence of pus in sterile body fluid, a clinical finding of an acute abdomen, cough, hypoxemia, and opacity in chest x-ray that might signify pneumonia, or certain rashes that are highly correlated with the presence of invasive infections. Notably, depending on the study, only about 30 to 60% of children who have sepsis or septic shock ultimately have a positive blood culture. So the presence of bacteremia is helpful. The absence of bacteremia does not exclude the diagnosis of sepsis. And then in terms of the epidemiology of the bacterial infections that tend to cause sepsis in children, there's about an equal incidence of gram-positive and gram-negative bacterial infections. So gram-positive species are most often staph and strep, often group A strep or strep pneumoniae, or group B strep and neonates. And then in the gram-negative infections are often the enterobacteriaceae, such as E. coli and Klebsiella, and then as well as Pseudomonas, particularly in hospital-acquired infections. The site of infection hasn't changed much over time. This was a study that looked at three data sets from three different time points, 1995, 2000, and 2005, for children with sepsis and found that the most common sites were respiratory primary bacteremia, genitourinary, and abdominal. Of course, it's important to point out this bar at the end here, which showed an increase in the cases of sepsis in which there was an unspecified site of infection. These are most often culture-negative sepsis events, which probably reflects a higher degree of suspicion and diagnosis of sepsis, even when a specific infection cannot be identified but is highly suspected. Definitions of organ dysfunction have been published. These are criteria that were published, again, in the Pediatric Consensus Conference from 2005. They're somewhat cumbersome to remember. I think clinically, it's most important to recognize that there are six organ dysfunctions that need to be attended to, cardiovascular, respiratory, neurologic, hematologic, renal, and hepatic. The specific dichotomous cut points that are used are probably less important, but recognizing that your diagnostic evaluation must focus on identifying dysfunction in these six organ systems is important. Oftentimes more than one organ dysfunction is present in sepsis. This is termed multiple organ dysfunction syndrome, or MODS, and this is typically present early in children with sepsis. This study, which is a little bit older now but has been replicated in more recent studies, about 90% of children who develop MODS do so within the first 48 hours of PICU admission. That's not to say children can't develop worsening or additional organ dysfunctions. They certainly can, but in sepsis, at least, much of this organ dysfunction is present early on when children are first admitted to the ICU. In addition, the accumulation of the number of organ systems that are dysfunctioning is the primary risk factor for death in pediatric sepsis. You can see data here from the Sprout study, which is a point prevalence study of 128 sites across 26 countries worldwide, showed that as the number of organ dysfunctions accumulated over time, the risk of hospital mortality increased in a stepwise fashion. Overall, the outcomes after pediatric sepsis vary substantially depending on the population studied. The overall mortality, including community and hospital-acquired sepsis, is roughly around 3% to 6% in children, but if you're sick enough to get into the ICU, and again, this varies depending on the entry criteria for the ICU and the country studied because of different thresholds and different number of ICU bed availability, PICU mortality is higher for the sicker cohort, upwards of 25%. For those who survive, for children who survive sepsis, functional disability has been reported in up to one quarter, and some studies slightly higher proportion of patients, and so this is new functional disability, including problems with physical function, emotional, social functioning as well, things like attention deficits, difficulty concentrating, problems at school and social networks, PTSD, and so on. Infection after having an episode of sepsis is also fairly common, so studies report upwards of 20% to 30% of children getting readmitted, and about half of these are for new infections, suggesting that children who have sepsis are susceptible to recurrent infections within the ensuing months after recovering. When we look at when children die from sepsis, there was a study published looking at two deaths from sepsis in two sites, CHOP and Nationwide Children's Hospital, and this study found essentially two time periods when children tend to die from sepsis. The first is a peak in children who come in with refractory shock or develop cardiac arrest soon after presentation, and they tend to die in the first 48 hours, and then there's a second peak of children that's really less of a peak and more of a sort of drawn-out period of time over one to several weeks, where children settle into more of a subacute to chronic persistent MODS phenotype, and oftentimes pass away as life-sustaining therapies are withdrawn due to multiple reasons, including the presence of sepsis, but often informed by underlying comorbid conditions. Risk factors for death in pediatric sepsis include a presence of comorbid conditions, including cancer and prematurity, hospital-acquired versus community-acquired sepsis, the number of organ dysfunctions, presence of worsening fluid overload, and extremes of age. For children that's less than one year of age. So moving on to the pathophysiology of sepsis, there are multiple overlapping and interconnected systems that combine to ultimately provide the phenotype of sepsis and septic shock. These include innate and adaptive immunity, various cytokines and chemokines, coagulation abnormalities, and activation and dysfunction of the endothelium, issues with macrocirculation and microcirculation, and mitochondria and metabolic problems as well. We're going to go through these individually now. Your innate immune system is your first line of defense. These are immune factors that are present prior to pathogen exposure. They're generally not enhanced in that they don't have memory and they're not clonally expanded but rather provide a consistent response, whether this is the first exposure to a pathogen or a current exposure. It's generally indiscriminate in that it does not have specificity like the adaptive immune response. There's no immunological memory. It's primitive in that it's more evolutionarily conserved than the adaptive immune system, which has evolved to a greater degree over time. So the components of the innate immune system include physical and surface barriers such as skin and mucosal surfaces, and then phagocytic cells including neutrophils, monocytes, macrophages which circulate and are evident in tissues, and then more organ-specific cells like macrophages present in the lung called alveolar macrophages, Cooper cells in the liver, and microglia. Eosinophils are also part of the innate immune response, more specific to parasitic infections, and the natural killer cells play an important role as well. The endothelium creates a barrier at the vascular interface, and so the endothelial lining is part of the innate immune system, although we'll talk about endothelial activation and dysfunction separately. And then there are also non-cellular mechanisms and molecules. In particular, in sepsis, nitric oxide. So nitric oxide is formed by the combination of L-arginine with oxygen to release nitric oxide plus citrulline as its metabolic byproduct. So there are several enzymes that produce nitric oxide through this reaction. There are constitutively active nitric oxide synthases, including endothelial and neural NOS, and these produce low levels of nitric oxide and are important in various roles of normal physiology. However, in sepsis, you get a massive upregulation of inducible NOS, and that leads to an overproduction of nitric oxide relative to the normal homeostatic state. And so the factors that induce upregulation of NOS include cytokines, bacteria, hyperglycemia. This results in a prolonged and sustained nitric oxide production, and because nitric oxide is heavily involved in vasodilation, this is one contributor to vasoplegic or distributive shock. Upregulation of nitric oxide and overproduction can also lead to oxidative stress with the formation of reactive oxygen species, which can lead to cellular injury. So several studies have shown an upregulation of nitric oxide production in sepsis. Here's a study from Hector Long's group in children showing that nitric oxide is measured using byproducts of nitric oxide metabolism, nitrite and nitrate, is much higher in the first three days of septic shock compared to non-septic controls. However, this upregulation of nitric oxide is not all harmful. There's an important signaling and antimicrobial role for this upregulation of nitric oxide, and that's evident when attempts have been made to completely inhibit the inducible NOS enzyme. So in this randomized controlled trial of adult sepsis, when a molecule was given that inhibited inducible NOS, there was an increase in mortality or a decrease in survival that you can see here. The patient that received placebo fared better than if you received this inhibitor, suggesting that while potentially problematic, there's also a beneficial role for the upregulation of nitric oxide in infection. PAMPs and DAMPs are also part of the innate immune system. So PAMPs are pathogen-associated molecular patterns and DAMPs are danger-associated molecular patterns. Essentially, PAMPs come from exogenous sources and DAMPs come from endogenous sources. But these are molecules that stimulate the innate immune response. LPS are generally conserved. They're conserved elements on pathogens that are present on microbes. These include things like lipopolysaccharide, or LPS, lipotochoic acid on gram-positive organisms, peptidoglycans, mannase, and bacterial DNA. These tend to trigger receptors, including toll-like receptors, mannose receptors, non-receptors. For example, LPS triggers toll-like receptor 4, and lipotochoic acid triggers toll-like receptor 2 from gram-negative and gram-positive organisms, respectively. Examples of DAMPs are molecules that are released from injured cells, and these really are there to warn the body of potential danger and injury, so that we can recruit an immune response in reparative mechanisms. When released in high amounts and spread throughout the body, they can cause a systemic inflammatory response syndrome. These include things like various heat shock proteins, uric acid, high-mobility group Box 1, and mitochondrial DNA, which has important homology to bacterial DNA. It itself can act as a DAMP to trigger the surge response. Moving on to adaptive immunity. Adaptive immunity is, as it says, it's adaptive in that it changes in response to specific pathogens, and it's specific in that different elements of the adaptive immune response will be amplified depending on the pathogen. It provides long-term protection against infections, and it's more specific than the innate immune response. It's triggered in response to certain pathogens, and over time, hours to days, it increases in the magnitude of the response, its potency, and its efficiency, particularly with repeated exposures because upon initial pathogen exposure, it induces memory, which is housed in certain cells like memory T cells and so on, and memory B plasma cells such that a more specific immune response could be recruited more quickly upon subsequent exposure to a specific pathogen. The components of the adaptive immune system include antibodies and lymphocytes, in particular T and B lymphocytes, certain antigen-presenting cells, the major histocompatibility molecules that help trigger T and B cells, and the complement system, and especially the classical pathway. There are several different subsets of T lymphocytes, helper T lymphocytes or CD4-positive cells, neurotoxic T lymphocytes or CD8-positive cells, and regulatory T lymphocytes or suppressor T cells, all function to coordinate the adaptive immune response. In particular, CD4-positive helper T cells are stimulated when an antigen-presenting cell such as a monocyte, a macrophage, or a dendritic cell presents an antigen in connection with the major histocompatibility complex II. That leads to a series of signal transduction events that upregulate cytokine production in order to activate, differentiate, and proliferate other cells so it self-amplifies the immune response and those cytokines also recruit and activate various phagocytic cells which help to eradicate the infection. Cytotoxic T cells or CD8-positive cells are involved specifically in the lysis of cells. So all cells have, or most cells have MHC I, major histocompatibility complex I, and those cells will produce antigen in response to a certain infection such as a viral infection. Those CD8-positive cells will recognize the antigen MHC I complex and that will induce those cells leading to, again, eradication of that intracellular infection. And lastly, regulatory T lymphocytes, also known as Tregs, formerly kind of referred to more broadly as suppressor T cells, these somewhat counteract the effect of the CD4 and CD8-positive cells to keep the overall immune response in check by allowing for modulation and suppression of immunity. But up-regulation of these cells to an abnormal extent can actually induce a hypoimmune response or immune paralyzed state which may put, which may hinder the ability to clear an infection or put patients at risk for development of secondary hospital-acquired infections. The ability to reverse this immunoparalytic state, which tends to occur later in the septic response, sort of day three and later, is an important area of active research. Moving on to cytokines and chemokines, so broadly this is the class of proteins that modulate and regulate all aspects of the innate and acquired immune response and inflammation. In general, secretion is relatively brief and self-limited, but when ongoing production is in process, you can have a sustained inflammatory response. There are multiple sources of each cytokine and these cytokines often have broad effects which leads to sort of the redundancy in the inflammatory response that we see in sepsis. It's a slightly slower response in that most of these proteins are not pre-made and available for ready secretion, but rather require active transcription and translation prior to synthesis. So there are dozens, if not more, cytokines that can be involved in a complex response such as sepsis, but there are some key ones that we should be familiar with. So early on in the course of sepsis, TNF-alpha and IL-1 are very important, particularly in bacterial infections where they help to stimulate the innate immune response. In general, these cytokines peak early and then often will return to lower levels or even to normal levels despite an ongoing propagation of the septic response. So they're important in inducing the septic response, but they're less important in sustaining the septic response. IL-6, on the other hand, is often produced in response to TNF-alpha and IL-1 and drives the acute phase response and stimulates more of the adaptive immune response, in particular B lymphocytes, and this tends to be a cytokine that's more easily measured because it tends to be sustained over time. Interferons, in particular interferon-alpha and beta, are especially important in the viral response, often secreted by fibroblasts and K-cells, and important in activating macrophages. IL-8 is often secreted by neutrophils and is often an important homing signal for leukocyte trafficking, so we refer to it as more of a chemokine because instead of inducing other cells to do things, it tends to bring cells to the area of infection. IL-8, again, is often a sustained cytokine that can be measured over time. And then IL-2 is produced by the adaptive immune cells, T-cells in particular. It helps to stimulate the cytotoxic T-cells, including NK-cells and B lymphocytes, to produce antibodies and help eradicate infected cells. There are also anti-inflammatory cytokines, which help to regulate the immune response and restore immune homeostasis. These are often secreted either concurrently or slightly later than some of the early pro-inflammatory cytokines, and three of the most important are TGF-beta, IL-4, and IL-10. Many of these cytokines and DAMPs and PAMPs will stimulate various signal transduction pathways that can be quite complex and overlapping, but there are several to generally be aware of as the ON switches. And so three of the more common ones that are relevant to sepsis include the JAK-STAT pathway, which tends to be triggered by interferons, and particularly interferon gamma. The NF-kappaB pathway, which generally is stimulated by toll-like receptors, in this case LPS-stimulating toll-like receptor 4, and the MAP-kinase AP1 pathway, which here is shown illustrated as stimulated by TNF. All of these things ultimately have the downstream effect of increasing transcription of various pro-inflammatory genes, and so referred to as the ON switches because they stimulate the pro-inflammatory response. In kind, there are also several OFF switches, which work interstitially to either reduce or modulate the effect of these inflammatory pathways and help to down-regulate cytokine and hemokine production, and there are various phosphatases, inhibitors, and inactivators that work to counteract these pathways so they're not interstitially overactive. The balance of the ON and OFF switches plays an important role in whether there's more of a pro- or anti-inflammatory phenotype over time, and the evolution of that balance over time in sepsis. In addition, there's been increasing recognition of the importance of various checkpoint proteins. When an antigen-presenting cell, such as a monocyte or a dendritic cell, takes on a pathogen and is able to break it down to present antigens on major histocompatibility complexes to, in this case, the T cell receptor on a CB4-positive helper cell, that interaction alone is insufficient to activate the lymphocyte. You need additional co-stimulatory molecules that also need to be present to get the full downstream response, and so these include things like CD40, activating CD40 ligand, and B7-1 and 2. There are checkpoint proteins, illustrated here as CTLA-4 and PD-1, which help to reduce the likelihood that lymphocytes are activated by antigen-presenting cells. And so the overall level of stimulation of a lymphocytic response is modulated by the presence of these checkpoint proteins that help to minimize the activation or the overactivation of lymphocytes. These proteins, however, such as CTLA-4 and PD-1, can also be upregulated in a way that may be pathogenic in and of itself. So, in particular, over time, it's increasingly recognized that overexpression of PD-1, CTLA-4, and others can lead to inability for the immune response to effectively clear an infection or may lead to a hypoimmune or immune paralyzed state such that patients are at risk for developing secondary infections while they're in the hospital. And so, there are ongoing trials looking at inhibition of some of these checkpoint proteins to try to restore the normal immune response and move away from an immunosuppressed state, either in the subacute phase or sepsis, or even later on, several weeks after sepsis has resolved. And so, what we see in sepsis is ultimately a balance between the systemic inflammatory response syndrome and the compensatory anti-inflammatory response syndrome, often referred to as CARS. And so, if we look at different patterns of the immune status in sepsis, we can see in the red bar that oftentimes, particularly with certain infections, in this case, illustrated with meningococcemia, we get a robust pro-inflammatory response that rapidly comes into check within the first couple days of sepsis with effective treatment. We have other cases where we have a less robust pro-inflammatory response, but that evolves into more of a hypoimmune response. And then, there are other patients with a less robust inflammatory response and more of a marked immune-suppressed inflammatory response. And so, the failure to return to homeostasis, either from the pro- or anti-inflammatory response, and to get progression, in this case, illustrated by the hypoimmune response, can lead to death. We can also have cases where we have abnormal persistence of the pro-inflammatory response that can similarly lead to death. So, it's this failure to return to a level of homeostasis within the immune system that leads to poor outcomes. Okay, so moving away from the immune response itself and into abnormalities within the coagulation response. So, sepsis induces a imbalance between anti-coagulation and pro-coagulation factors. And most of the time in sepsis, you get a relatively, a relative weighting towards a pro-coagulant physiology. And so, you get up-regulation of platelet activation and aggregation, an increase in tissue factor, phospholipids that can induce clotting, and an increase in plasminogen activation inhibitor, all of which can promote the formation of microvascular thrombi, and impair end-organ perfusion, and lead to organ dysfunction. At the same time, the factors that tend to balance out the pro-coagulation factors are all relatively suppressed. So, tissue factor pathway inhibitor, APC, protein S, DPA, antithrombin 3 and ADAMS-13, all are factors that help to keep pro-coagulation in check. And these generally tend to be either consumed or down-regulated, which leads to an overall pro-thrombotic milieu that leads to the microvascular thromboses and alterations in microcirculation. We'll talk about the downstream effects of that altered microcirculation in a little bit. So, the endothelium is very much activated in infection and may actually become overactive or dysfunctional in sepsis. And so, it's important that the endothelium up-regulate certain adhesion molecules, such as E-selectin and ICAMs, and participate in the secretion of cytokines when there's a local infection, so that it can help recruit the immune response to that area to eradicate and heal from the infection. It also participates in the more pro-coagulant milieu by reducing anticoagulant thrombomodulin and tissue factor pathway inhibitor, and so the endothelium becomes relatively more sticky. Again, that helps to slow the flow of blood and trap leukocytes in the area of infection. There's also an increase in the endothelium permeability by alterations in gap junction proteins that facilitates leukocyte trafficking out of the vascular space and into the interstitial space that actually allows for access to infectious pathogens. And it's important to recognize that what we typically call capillary leak more likely occurs at the venial post-capillary site rather than at the capillaries themselves. So again, pictorially, this activated endothelium recruits phagocytes and assists with tethering, rolling, adhesion, and ultimately extravasation with upregulation of cytokines and hemokines that feed forward and promote this response such that we get immune cells in the areas of the interstitium where the pathogens tend to be present. Of course, when this happens on a systemic level and you get massive activation of the endothelium, then you've got post-capillary leak and extravasation of immune cells and inflammation in areas of the body that aren't actively infected. And so you get inflammation that's counterproductive. In addition, it's important to recognize that between the endothelial cells and the vessel lumen is a layer called the glycocalyx. The glycocalyx is an active layer of multiple antimicrobial peptides and anticoagulant proteins. You get denuding of that glycocalyx. And so you get, again, that contributes to this endothelial dysfunction. And then you get abnormalities in various junctional proteins, which leads to alterations in the ability of the endothelium to keep fluid within the vascular space and instead leaks out, usually through these gap junctions, less often through the cells themselves, but through these gap junctions into the interstitial space. Okay, moving on to macrovascular dysfunction. So this is what's typically referred to as cardiovascular dysfunction or shock. And in septic shock, it's actually, although classically described as distributive shock with vasoplegia, it's actually a combination of multiple different shock types. In children, we often see hypovolemia and cardiogenic shock contributing to the full spectrum of abnormal cardiovascular dysfunction in addition to vasogenic abnormalities. So when we talk about shock, septic or otherwise, we're talking about an imbalance in oxygen delivery and oxygen demand, or more aptly termed, substrate delivery and substrate demand. So it's not just oxygen that cells need, but it's historically been referred to as oxygen. I think it's convenient to think about it in the terms of oxygen delivery and demand. So when we think about oxygen delivery, this is determined mainly by two general factors, cardiac output and the content of oxygen in the arterial blood. And so the content of oxygen in the arterial blood can be determined by the equations seen there, primarily dependent on hemoglobin and hemoglobin saturation. And so when we think about that equation, we can think about the factors that contribute to a decrease in oxygen delivery during macrovascular shock. And these include low cardiac output, which means a low heart rate or low stroke volume, a low hemoglobin, low oxygen content, either oxygen saturation or oxygen dissolved in the blood, and then microvascular shunting, which we'll talk about as a separate topic in a minute. And it's important to recognize that, at least in the early stages of sepsis, the inability to sufficiently deliver oxygen and other substrates to the tissues is what ultimately triggers organ dysfunction and organ failure. But we must also consider the other side of the equation. So it's not just oxygen delivery, but it's also oxygen consumption. So that's referred to as VO2, and the equation is cardiac output times the difference between oxygen content in arterial and venous blood with a fudge factor of 10. And the extraction of oxygen is the ratio of oxygen consumption to that of what's delivered. This can be estimated using saturations, but is more appropriately calculated using oxygen contents. The normal level of oxygen extraction is about 25 to 30%, which is why our central venous oxygen saturation is around 70 to 75% when measured from the SVC. And we can increase oxygen extraction ratio when oxygen delivery is reduced when oxygen delivery is reduced in order to keep oxygen consumption constant. And so that is measured physiologically by a decrease in the central venous oxygen saturation. And so we can look at this pictorially as well. So this is a graph that depicts the relationship between oxygen consumption on the y-axis and oxygen delivery on the x-axis. And we can see as oxygen delivery is reduced, as septic shock evolves, we have a relatively constant oxygen consumption. And so on the right-hand part of that graph, we are in the area of supply-independent oxygen consumption where we can actually increase oxygen extraction in order, the percentage of oxygen extraction while oxygen delivery is reduced in order to maintain cellular functions with a constant oxygen consumption rate. And we measure that again as a decrease in SVO2. However, beyond a certain critical point, we enter a phase of supply-dependent oxygen consumption whereby oxygen consumption falls if oxygen delivery is further reduced. And this is typically indicated by hyperlactatemia, which signifies a transition to anaerobic metabolism because oxygen is now a limiting factor. And the goal of critical illness and in sepsis resuscitation is to move folks from the left side of this graph to the right side. And so why do we see lactate produced? Well, this is a very simple overview of metabolism where we see glycolysis leading to the production of pyruvate, which is normally in the presence of oxygen shuttled into the mitochondria for aerobic respiration and ATP production. When oxygen is a limiting factor, then that pyruvate is alternatively metabolized by LVH to lactate. And the lactate spillover is seen in the blood leading to hyperlactatemia and lactic acidosis as a signifier of oxygen delivery dependent decrease in oxygen consumption. So we know that the presence of lactate is a risk factor for mortality. In this study by Halden Scott and colleagues in Denver, you can see that in children who did not have normalization of lactate over time, they had a higher degree of organ dysfunction at 48 hours. So the ability to move children from the left side of that pyrograph to the right side where their oxygen consumption is no longer limited by oxygen delivery is an important therapeutic goal. Moreover, macrovascular dysfunction is not solely indicated by hypotension. Again, it should really be abnormal perfusion. That is our signifier of macrovascular dysfunction and the presence of shock. And why is that? Well, in children, you can see that as there's hypovolemia, that due to vasogenic leak or dehydration, we can see that vascular resistance can increase, which can help to maintain blood pressure. And even while cardiac output has fallen, blood pressure has been stabilized at a reasonable level until quite late, when we can no longer use vascular resistance to compensate. And so hypotension tends to be a late finding, particularly in infants and very young children. So because of the ability to maintain an elevated and increased SVR, the clinical presentation of children with septic shock can vary. So adult sepsis is classically described as distributive shock with a high cardiac output and a low SVR state, often presenting as warm shock. So with warm extremity, bounding pulses, low diastolic blood pressure. Whereas in pediatric sepsis, because of physiologic differences in the ability to maintain SVR, we see more of a varied appearance. So this study using PA catheters showed that there was a relatively higher incidence in children of this low cardiac output, high SVR state, presenting often with cool extremities. Diminished pulses, delayed CAP refill. And in fact, that pattern of findings was associated with a higher mortality than if you presented more classically with warm shock. So the features of warm and cold shock are shown here, and the physical exam findings can differ. Although new evidence suggests that physical exam findings alone may not adequately represent the degree to which children have myocardial dysfunction or their SVR state. And so this is a relatively crude way to distinguish physiology, and probably should be quickly augmented by more invasive hemodynamic assessments such as cardiac ultrasound and arterial blood pressure monitoring. But you can see in this study by Joe Brilli and colleagues in England that patients tend to cluster into two groups. Those who have cold shock with low cardiac index, high SVR, and those with warm shock with high cardiac index, low SVR. And when they looked at the typical demographics of these patients, community-acquired shock tend to be more likely to present as cold shock, whereas hospital-acquired shock in patients who've had central venous catheters and tended to be older, more adolescents, often cancer patients, tend to present with a more classic adult distributive shock picture. So again, as I mentioned, the ability to differentiate warm versus cold shock solely on the use of clinical exam findings is somewhat limited. This was a study out of India where children were classified as likely to need more fluid, inotropes, or pressors based on clinical exam findings alone. And you can see 80% of the patients were thought to be needing more fluid. Only a few needing inotropes or vasopressors. But when you added in cardiac ultrasound, then you can see that patients were largely seen to not need fluid but were more likely to have myocardial dysfunction needing inotropes, findings that were less well-appreciated based on clinical exam findings alone. All right, so moving into microcirculatory dysfunction. So even if one can restore signs of perfusion, such as cap refill, urine output, mental status, and even blood pressure, there may still be dysfunction at the microcirculatory level. So what does that mean exactly? So here we take normal oxygen delivery in mLs per minute going through three different small blood vessels with an oxygen extraction by the interstitium here between the blood vessels. And you can see we're delivering 240 mLs a minute of flow. And oxygen is removed at a constant rate. And so in the venous blood, we end up with less oxygen. And 24 divided by 60 gives us 60% SVO2 remaining in the venous blood. So that's under normal conditions. So when you're in a low flow state, but it's homogeneous, so this would be in a setting of say hypovolemic shock, the issue is that cardiac output is reduced. And so we're reducing oxygen delivery, but in a uniform manner. So if there's preservation of oxygen utilization at the tissue level, you can increase oxygen extraction ratio. And ultimately that results in a much lower SVO2 of 20%, but an even distribution of oxygen extraction. However, if we have microcirculatory dysfunction from vessel dropout, so that's derivative from disruption of vessels or clotting of vessels from being in a procoagulant state, even if flow is normal. So in this case, total flow 240. So you get dilation of normal vessels to accommodate for that increased flow. The tissues around those vessels still have the same oxygen requirement and metabolic demand. So they still only extract what they need similar to the normal state here. And so SVO2 actually goes up because overall extraction is less whereas these areas receive no flow and therefore there's this relative ischemia due to lack of oxygen in these adjacent tissues here. And that leads to cellular stress in those areas and ultimately organ dysfunction and organ failure despite having a super physiologic SVO2 measurement. And that's microcirculatory shunting, much harder to treat because most of our therapies to date target the macrocirculation as opposed to the microcirculation. And so this is a pictorial representation. You can see in the normal picture here using sublingual microscopy, nice robust network of blood vessels and septic shock. However, you can see a decrease in functional capillary density which drop out of vessels throughout various areas of the tissue. And ultimately you can quantify functional capillary density which tends to go up in survivors and down in non-survivors as an indicator that microcirculatory dysfunction when persistent is associated with poor outcomes. Another marker of microcirculatory abnormalities which is more readily attainable than sublingual microscopy is the PVACO2 gap. So you can measure partial pressure of CO2 in venous and arterial blood. Normally, that difference should be low, about five or six millimeters of mercury. However, when you see an increase in that gap such that the amount of CO2 in the venous blood is much higher than that in the arterial blood, that either indicates that there's anaerobic metabolism with less bicarbonate buffer so that CO2 is less effectively buffered in the venous blood or that CO2 is accumulating in poorly perfused tissue and that's registered as a relative increase in the venous CO2 levels. So if you have an arterial line, you measure the peripheral venous and peripheral arterial CO2 levels and that gap is higher than six. That's associated with poor outcomes. Lastly, we'll look at the mitochondria. So mitochondria, an extraordinarily complex organelle. But basically, if you recall, it's a double-membraned structure with an outer and inner membrane. In the inner membrane is embedded the complexes in the electron transport chain which serially pass electrons while pumping protons into the intermembrane space which are ultimately brought back in through ATP synthase to produce ATP while reducing oxygen to water. When there's dysfunction of the electron transport chain or the inability to shuttle pyruvate into the mitochondria or an insufficient amount of oxygen, this process cannot work effectively and so you get a decrease in aerobic ATP production. You get a relative increase in reactive oxygen and nitrogen species production because of offloading of electrons from the chain. And you get an opening in the mitochondrial permeability transport core which leads to release of certain proteins, particularly caspases which can induce cell death. So the inability for mitochondria to utilize oxygen even when oxygen delivery is able to be restored again leads to an increase in accumulation of pyruvate and an increase in lactate, not necessarily because oxygen is insufficient but because the mitochondria can't utilize that oxygen effectively to produce ATP and so that's called dysoxia or cytopathic hypoxia. So in the last few slides, just sort of thinking about how this pathophysiology comes together to give you the clinical picture of sepsis. So what does a child with sepsis look like? Well, they have signs of infection and inflammation, alterations in perfusion and laboratory measurements of organ dysfunction. However, I mentioned in the very beginning the nonspecific nature of those criteria and so lots of other things can look like sepsis. So trauma, burns, cardiopulmonary bypass, autoimmune and rheumatologic diseases such as vasculitis, metabolic disorders, particularly in infants, adrenal insufficiency and pancreatitis all can masquerade as potential sepsis. And lastly, you can't forget about unusual pathogens. So it's not just bacteria that cause sepsis although we think about that most commonly. But thinking about fungi, particularly in hospitalized patients who have been exposed to prior courses of antibiotics and are on TPN, certain viral pathogens, particularly influenza and SARS-CoV-2 can induce a septic state. And then rickettsial infections such as Rocky Mountain spotted fever, particularly if there's petechial rash and hyponatremia, hypotomegaly and other tick-borne illnesses such as ehrlichiosis. Some children with sepsis have more of an immune-paralyzed phenotype, a thrombocytopenia multi-organ failure phenotype or sequential multi-organ failure phenotypes. So I'm not gonna get too much into the details of these with the exception of to speak about TAMOF and immune paralysis in the next two slides. So TAMOF is a subtype of sepsis that's defined as the presence of thrombocytopenia, AKI, shock and modern multi-organ dysfunction syndrome. It's generally caused by low ADAMS-13 which is an enzyme that cleaves von Willebrand's factor into smaller multimers. And so the low level of ADAMS-13 leads to an increase in ultra-large von Willebrand factor multimers which tend to cause micro-circulatory thromboses and leads to micro-circulatory dysfunction. So you can help to restore ADAMS-13 levels with plasma exchange. And so in this proof-of-principle trial in patients who received plasma exchange, you can see their organ dysfunction scores indicated by PVOD decreased over time as opposed to a sustained increase in those who did not get plasma exchange. And then lastly, immunoparalysis. So a lot of this work comes out of Mark Hall's group at Nationwide Children's. In this study, they looked at two indicators of immune response. So TNF-alpha produced ex vivo by cells exposed to LPS. And then the number of HLA-DR, percent of HLA-DR expression on monocytes as an indicator of immunologically competent monocytes. And you can see the non-survivors in the gray boxes tended to cluster where these two markers were low. Interestingly, you can restore TNF response by giving GM-CSF. And so you can see patients who got GM-CSF compared to placebo had an increase in TNF-alpha production. And also had a reduction in the likelihood of secondary infection. So the bars here are the patients who received placebo and the white bars, which there really are none because there were no infections in the patients who received GM-CSF. So suggesting that you can not only reverse the immune paralyzed phenotype, but also reduce the likelihood of secondary infection. And then hyperferritinemic sepsis is a overlap state with HLH and MAS. And tends to be a more of a final common pathway that is commonly found in patients with some of these sub-phenotypes. In this study, 11% of patients had this MOS phenotype, MAS overlap phenotype. And that's important because studies have shown that this highly pro-inflammatory state is responsive to immune suppression as opposed to immune supplementation that we saw on the last slide. You can give immune suppression with IVIG, methylplasma change, and anakinra, which is an IL-1 receptor antagonist to help knock down a sustained pro-inflammatory response and improve outcomes. So in conclusion, sepsis is relatively common in children, particularly in the children who are critically ill in the intensive care unit. Cardiovascular and respiratory dysfunction are the most common, but accumulating number of additional organ dysfunctions leading to progressive MODS is the most important risk factor for death in children. Sepsis starts with an infection that leads to a very complex and redundant host response that involves overlapping and interacting effects from the innate and adaptive immune system leading to inflammation, upregulation of coagulopathy, endothelial activation, and even dysfunction, and then microcirculatory and metabolic abnormalities. For children, unlike adults, more often we'll present with cold shock, with myocardial dysfunction and elevated SBR, with a compensated blood pressure, at least initially, other signs of abnormal perfusion that needed to be recognized for early recognition of sepsis. And then there are certain clinical subtypes of sepsis that seem to have consistent and distinct pathophysiological mechanisms that may be amenable to targeted therapies. So thank you very much for your time. Really appreciate the opportunity to speak about the complex nature of sepsis pathophysiology in children.

pediatric sepsis

pathophysiology

inflammation

organ dysfunction

septic shock

bacteremia

multiple organ dysfunction syndrome

mortality rates

risk factors

immune response