Hello, my name is Scott Weiss and I'm a pediatric intensivist at the Children's Hospital of Philadelphia and this talk will be about clinical guidelines for the management of pediatric sepsis. I have several disclosures related to my research in sepsis, none of which are particularly relevant other than to say that I am an author on the most recent Pediatric Surviving Sepsis Campaign Guidelines published in February of 2020. By the end of this session, the goal is for you to learn about early recognition of pediatric sepsis and how screening algorithms can help to facilitate improvement in early recognition and the management of children with sepsis and septic shock, including how to apply clinical guidelines at the bedside. So sepsis exists somewhere on the spectrum between uncomplicated infection and death. And sometimes it can be unclear as to exactly what it is you're treating. So for example, if a patient presents over here towards the right end of this graphic with clear evidence of shock or organ dysfunction, it's probably evident to all providers that you're treating sepsis. What about children who present over here? Are you preventing sepsis or have you successfully recognized sepsis early in its early phases so as to avoid progression and more severe symptoms? There are two documents that have really guided the early recognition and management of pediatric sepsis over the last 25 to 30 years. The first is the American College of Critical Care Medicine clinical practice parameters and more recently, the Pediatric Surviving Sepsis Campaign Guidelines. Here you can see the most recent ACCM and Pediatric Surviving Sepsis Campaign Guidelines, their algorithms compared. And we're going to go through these in sort of a stepwise fashion throughout this talk. But there are notable similarities between these algorithms focusing and emphasizing on the same key features. So before one can treat sepsis, you must recognize it and diagnose it, which again, when you're on the right side of that prior graphic and you have clear evidence of shock and organ dysfunction, may be relatively easy to do, but a little bit harder as you move leftward in patients who have more subtle symptoms or early symptoms. So it can be quite a challenge. In addition, sepsis is not really one syndrome or disease, but rather a similar set of clinical and laboratory manifestations of multiple different processes, including viral infections, variety of different bacterial infections, and so on. And not only do the patients present with different microbiological etiologies, but the host response can be very different, even to the same infection between different patients. And then, of course, patients present at all different time points, from early until late in the stage of sepsis. And then we do not have a gold standard diagnostic test. And so we rely on a combination of rather nonspecific diagnostic clinical and laboratory criteria to ultimately arrive at a suspicion and diagnosis of sepsis. So this is a picture of Rory Staunton. Many of you might be familiar with him. He was a young boy, previously healthy, who presented with early symptoms of sepsis after sustaining a rather small cut on his arm that got super infected and then progressed. And he presented with systemic symptoms that were not necessarily easily related to his cutaneous infection, and ultimately progressed rather rapidly, represented, and unfortunately passed away. His legacy is that he has called to arms the need for better ways to recognize sepsis in its earliest stages. And so building on the prior ACCM guidelines that had called for methods to be put in place to aid early recognition of sepsis, the most recent guidelines, the Pediatric Surviving Sepsis Campaign Guidelines published in February of 2020, more explicitly stated a suggestion of implementing a systematic screen to assist with the timely recognition of septic shock and other sepsis-associated organ dysfunction in children who present acutely unwell to an acute care setting. And so the basis for this suggestion was that several studies have demonstrated that a routine systematic screening process can help to facilitate earlier recognition of sepsis, which can lead to more timely therapies and ultimately improve patient outcomes. So as one early example of this, Texas Children put into place both a system to aid in the early recognition and hasten therapies for children who present to the emergency department with sepsis, and they showed a dramatic improvement in time to administration of IV fluids, time to administration of the first and the third IV fluid bowls, as well as time of administration of antibiotics. But when you think about a large ED system such as Texas Children or Children's Hospital Philadelphia or many of the other institutions where we work, most of the patients who present with a fever and other infectious symptoms will not have sepsis. And so ultimately, what we end up looking for is a needle in a haystack. And to do that, it's helpful to have a tool to help hone in on that needle. And so several hospital systems have now developed and published screening algorithms that are implemented widely on most, if not all, patients who present in order to aid in clinician early, clinician's recognition of the early signs of potential sepsis. And so what criteria are included in these screening algorithms is the number of patients And so what criteria are included in these screening algorithms is somewhat, or the best criteria to include in these screening algorithms is still somewhat unknown. But to date, most of the screening algorithms published have relied on three different sets of criteria, which come from the 2005 International Pediatric Sepsis Consensus definitions, which includes SIRS, presence of infection, either suspected or diagnosed, and presence of shock or other organ dysfunction. So as an example, this is the sepsis screen published, I think, still currently at use at Le Bonheur Children's Hospital. And you can see it's somewhat complicated here in that there are a lot of different elements that one must consider when applying this screening algorithm to an individual child, which, if you scale this up to tens of thousands of children who present over the course of a year, most of whom don't have sepsis, you can see it's an incredible burden to apply, although it can still be helpful. But you can see that this, the details are not important, but the criteria used, again, are based on this concept of SIRS, suspected infection, and organ dysfunction. So whether or not SIRS are helpful is not entirely clear. So this was a very large study of adult patients from Australia and New Zealand, presented with infection plus organ failure, so would essentially have sepsis. And you can see that there was a group of patients, not the majority, but a substantial minority, who did not meet SIRS criteria, with either zero or one SIRS element, and one SIRS element met, and still had substantial mortality related to infection plus organ failure, not that much different from those patients who had SIRS criteria. In children, there's less data that provides similar sort of evidence, but this nice study from Hull and Scott at Denver Children's looked at all children who presented to their ED and were ultimately admitted to the ICU. Now, this wasn't just focused on sepsis patients, but nonetheless, you can see that the presence of having two or more SIRS criteria had fairly low sensitivity and only modest specificity for needing ICU admission, suggesting that SIRS, meeting SIRS may not be, meeting SIRS criteria may not be helpful for screening for sepsis, and the absence of meeting SIRS criteria may falsely lead you away to a concern for sepsis. So, other tools have been developed and implemented. This is one that has been promoted by the AAP as their pediatric septic shock collaborative, and used in a different set of criteria to sort of focus in on key elements that may be suspicious for sepsis. And so, instead of starting with SIRS criteria, suspicion of infection as evidenced by fever or other symptoms was a key initial element of the screen, and then there was screening for SIRS criteria, or evidence of altered perfusion, even if SIRS was not present. And then they added on to this high-risk criteria to take into account variability in individual patient risk for possibly having sepsis. So, what about lactate? So, it's very clear that having an elevated lactate is a risk factor for both prolonged length of stay and increased mortality. However, most children who have an elevated lactate ultimately do well, and some children who have a low lactate can progress quite quickly and ultimately have poor outcomes. So, it's also not absolutely sensitive or specific for adverse outcomes or the presence of sepsis. In addition, lactate is most often used to confirm suspicion of sepsis rather than to screen for it, whereas suspicion of sepsis is probably more correctly based on abnormalities and perfusion that can be detected clinically, such as delayed cap refill, low urine output, altered mental status, and so on. And so, the use of lactate to screen for the presence of sepsis in children is not yet recommended. So, other screening algorithms. So, this is one using the CHOP-ED with some success over several years. Again, tries to simplify even further that AAP tool to look for the presence of fever as reported at triage or on the initial set of vital signs along with the presence of tachycardia or hypotension. If any of those are present, then the triage nurse does a quick assessment for perfusion, altered mental status, or high-risk conditions. And if any of those are present, the patient is rapidly triaged for a sepsis huddle so that the patient can be further evaluated quickly and the decision can be made to begin or initiate further evaluation or treatment of sepsis or to stand down if the patient otherwise seems to be well. So, with this tool to detect sepsis as defined as admitted to the ICU with severe sepsis or septic shock, the alert alone had fairly good sensitivity and specificity. When you add physician judgment for those who did not fire the screen, there was even higher sensitivity and specificity. But you'll notice the positive predictive value is only about one in four, meaning that three out of four patients who fire the screen ultimately, as part of the huddle, stand down and don't get treated rapidly for sepsis. This doesn't mean they don't get treated for infection, but there's no need to deliver the same degree or bring the same degree of resources and speed to that patient's bedside and also avoids over-treating cases where there may be an alternative explanation for the patient's symptoms other than infection. So, with implementation of this screening algorithm, you can see that the number of missed sepsis cases at CHOP went from around 17% to less than 5% with sustained improvement over a long period of time. So, a nice example of where a systematic algorithm that's relatively simple can really help to aid in early recognition and decision making. So, there is, of course, the challenge of hospital-acquired sepsis, where patients are not admitted to the hospital with sepsis but develop this over time. And from the Sprout study and many other studies, including in adults, we know that the risk of mortality if you develop sepsis in the hospital is much higher than if you present in an acquired fashion. And so, having screening algorithms that are active for inpatients is also a point of active research for children. So, moving on from screening and early recognition to how does one begin initial resuscitation. So, there are two key overriding principles when one embarks on treating a patient for sepsis or septic shock. The first is to restore oxygen and other substrate delivery with the goal to improve end-organ perfusion. And this should be targeted to specific improvements in key features that are readily measurable at the bedside. The second is to identify and treat the underlying infection. And so, this includes a combination of antimicrobials and appropriate source control maneuvers such as drainage of abscesses and removal of infected hardware that may be contributing to the patient's septic symptoms. So, for almost two decades, early goal-directed therapy was the overriding principle driving the initial resuscitation of patients with sepsis. And this really stemmed in part from an article published in the New England Journal of Medicine in 2001 by Manny Rivers and colleagues where after developing and applying a physiologically based protocol to adult patients with sepsis that included giving oxygen, giving fluid, giving vasoactives, and giving blood within the first few hours of presentation, they were able to decrease hospital mortality from 46 percent to 31 percent. Virtually overnight, this changed practice from standard of care to early goal-directed therapy and really changed what was defined as standard of care. There was a similar article from South America that showed improvement using a similar algorithm in children reducing mortality from 39 percent to 12 percent with an early goal-directed therapy algorithm. So, have there been any changes since then? So, over the last few years, there were three large randomized trials that tested whether early goal-directed therapy, as envisioned by the protocol implemented by Manny Rivers, was or is still necessary. And so, these included the PROCESS, ARISE, and PROMISE studies, which were done as separate trials, but sort of in a collaborative fashion to cover patients broadly throughout the world. And all three trials, you can see for PROCESS, ARISE, and PROMISE, showed no difference in mortality for patients, adult patients who were treated with early goal-directed therapy versus usual care. However, if you looked at the usual care arms in these studies, you can see that they were very different than what was done prior to that 2001 publication in that almost all patients got aggressive goal-directed therapy. What was different is that the Manny Rivers protocol required universal measurement of central venous oxygen saturation and CVP, and probably that is not necessary in all patients. It may only be helpful in those patients who don't respond to early therapy. And so, exposing too many patients to unnecessary measurements and procedures may actually lose some of the benefit and potentially even creates harm. So, moving into the pediatric algorithm, we're coming back to these. So, again, we can see that in the 2014 ACCM guidelines, in the first hour of initial resuscitation, really prioritized the things we've been talking about. So, oxygen, IV fluids, antibiotics, and then moving quickly to vasoactive blood and steroids, and we're going to go through all of those. In the more recent Surviving Sepsis Campaign guidelines, again, in the first hour or so of resuscitation, the goals are to obtain IV access, collect blood, start antibiotics, administer fluid, and then vasoactive is for fluid refractory shock. So, slightly different layout, but again, similar fundamental principles. Okay, so if fluid is the cornerstone for initial resuscitation of abnormal perfusion and shock, which fluid is recommended? So, there are several options. One can be a variety of different crystalloids, colloids, or blood products. And so, the best choice has been a point of study for several years now. And so, crystalloid fluids, which are cheap or readily available, easily administered, highly compatible with other medications, are often the first-line choice therapy. And so, which crystalloid fluid to give is still not entirely clear. So, if we look at the two most commonly used crystalloid fluids, saline and lactated ringers, or Hartman's in areas outside of North America, we can see that the major difference here is that LR Hartman's fluids are overall more balanced in that they have less chloride and a non-chloride buffer that helps to limit the acidic pH. And so, this makes this fluid more similar to the composition of our human plasma. Whereas, saline provides an important chloride load. And that chloride load has potential adverse physiologic effects. And so, this was a study of adult sepsis. It was a retrospective observational study. It was very large and ultimately was able to show that as the percentage of total fluid administered for sepsis resuscitation was more and more balanced, that is, increasing use of LR, there was a stepwise decrease in mortality compared to exclusive resuscitation with 0.9% saline. And then, two large randomized trials, both done at Vanderbilt University, one in critically ill and one in non-critically ill adults, not just sepsis, but all patients, was able to show that random allocation to the arm that included balanced fluid, which was laxative ringers and a little bit of plasma light, resulted in a reduction in adverse kidney events, as well as a reduction in hospital death. For children, there are no large randomized studies, but there have been two large observational studies. So, the first, shown on the left, included patients who presented with concerns for septic shock to largely community hospital EDs. And this was a paper published by the CHOP group and showed really no difference in patients who received normal saline or LR-based resuscitation. You can see the mortality curve here. However, a group at Emory looked at patients who presented to specialized PICUs, so academic children's hospitals, and showed that if you resuscitated with only balanced fluids, that there was a slight reduction in mortality and need for renal replacement therapy, suggesting some benefit for resuscitation with balanced fluids. And why crystalloids versus colloids in the first place? So, this study is called the SAFE study. This was a landmark trial that was published in Australia and New Zealand. It compared saline to albumin for adults who required ICU, many of whom had sepsis, but this included patients who also had fluid requirements for reasons other than sepsis and ultimately showed no difference in probability of survival. However, only saline was the comparator here. There was no comparison with buffered crystalloids such as lactated ringers or Hartman solution. Nonetheless, even when compared to saline, in those patients who specifically were being resuscitated with fluids because of sepsis, there's a suggestion that colloid resuscitation may actually be better than saline resuscitation. So, if you look at this subgroup analysis, you can see that there was an indication, although not statistically significant, that those patients who preferentially received albumin had a lower mortality. That was also suggested in a more recent study where adults with sepsis, if they received colloid resuscitation, had a trend towards a improved mortality or lower mortality with albumin versus crystalloids. And this study was all crystalloids, although the majority was stalin administration. So, if one's going to choose to give a type of fluid, one must also consider how much fluid to give. And for many years, there was a concern that patients weren't receiving sufficient amounts of fluid. So, Joe Carcillo out of the University of Pittsburgh published a very nice study in JAMA in 1991, again observational, but showed that children who had septic shock, who received more than 40 mLs per kilo of fluid in the early phases of resuscitation were more likely to survive than if you received less than 40 mLs per kilo of fluid. However, there's been increasing attention to perhaps unnecessary amount of over-resuscitation with fluids, and concern that using too much fluid potentially could swing the pendulum towards an increased risk of mortality. And so, this is, on the right, one example of such a study that showed a stepwise increase in risk of mortality with progressive fluid overload. Now, a lot of this fluid overload is attributable to ongoing fluid administration, not just the initial bolus resuscitation in the first hour or two. And so, there are important differences when considering when to move from an aggressive fluid replacement strategy to a more restrictive fluid minimization strategy. But we also must think, can fluid be harmful in other ways? So, while I mentioned the fluid overload studies largely include fluid administered over several days, the FEAST trial examined early fluid resuscitation in a large group of children who presented with infection and altered perfusion, or fever and altered perfusion, living in sub-Saharan Africa. So, slightly different patient population than typically is cared for in North American PICUs, but nonetheless an important study because they found a rather surprising result. So, they tested over the first 48 hours, difference in mortality between three groups, those given albumin or saline resuscitation up front, and those treated with maintenance fluids without an initial fluid bolus. And what they found was that the group that did not receive a fluid bolus actually had a lower 48-hour mortality rate than either those patients resuscitated with albumin or saline. And this raised a lot of concerns about the possibility, at least in some patients, that initial regressive fluid resuscitation could be harmful. In particular, those patients at risk for malnutrition or severe anemia from malaria and other infections that may predispose to anemia. So, in totality, the review of this data, as interpreted by the Surviving Steps campaign, came up with the following recommendations. So, in healthcare systems in which intensive care is accessible, either at the institution where the child is located or easily via transfer to another center, the suggestion is to administer fluid bolus therapy up to 40 to 60 mLs per kilo of fluid over the first hour, but titrate this fluid carefully in a goal-directed manner to markers of cardiac output, and restrict or discontinue fluid resuscitation if signs of fluid overload develop. Conversely, in healthcare systems in which intensive care is not accessible, either locally or via transfer, the recommendation, largely based on the results of the FEAST trial, is to avoid bolus fluid unless the patient is hypotensive, in which case fluid up to 40 mLs per kilo could be cautiously administered over the first hour, again titrated to clinical markers of cardiac output. And then, in terms of which fluid to use, there is a suggestion that browns or buffered crystalloids, such as laxative ringers or plasmalite, may have an overall more beneficial effect than saline, although there's an acknowledgement that additional high-quality data are really necessary to fully test whether that is the right way to go. So, while one is resuscitating to restore oxygen or another substrate delivery, one must consider how best to eradicate the inciting infection. So, the first step of this is to administer antibiotics or other antimicrobials, depending on the likely source of infection. So, why is there a high priority on early antimicrobial administration? Well, this is a very important observational study out of Canada, looking at adults, showing that with each hour delay from onset of hypotension related to sepsis, there was an increase in the odds of mortality. In children, much less robust data, but two studies, one from a single center at CHOP, showed a similar trend in a stepwise increase in mortality with each hour delay in administration of antibiotics, although it was not statistically significant until delays beyond three hours. And a study out of New York State, also in children, showed that while overall administration of early antibiotics as part of an early bundle of care improved mortality, when you look specifically at the difference in mortality between administration of antibiotics within one hour or in three hours, there was no numerical difference related to antibiotics alone. And so these data suggest that time to antibiotic administration is important, but unclear as to exactly whether there's truly a benefit to one hour versus a little bit behind. When examining more of the adult studies, since the initial study published by Dr. Kumar that I showed earlier, you can see that the totality of data does indeed favor early antibiotic administration within one hour for patients who have suspicion of septic shock or hypotensive shock, but that nobody's been able to replicate quite the impressive degree of benefit that was demonstrated in the Kumar paper. Nonetheless, all the data together strongly suggest a need to attend to speed when administering antibiotics for these patients. In terms of how to administer antibiotics quickly, so having a antibiogram is helpful so that clinicians are not having to think about which antibiotics to give for each individual patient, and so here's one example from CHOP. In addition, it's important to recognize that giving certain antibiotics, such as a beta-lactam or cephalosporin, can be given quickly, often as an IV push, so that they can be given as part of ongoing fluid and other elements of resuscitation without tying up IV access for 15 minutes or longer to administer slowly over a pump. And so the most recent recommendations for children as a surviving sepsis campaign in terms of early antibiotics are that if a child presents with septic shock as recognized clinically by abnormal perfusion, there is a strong recommendation to administer antimicrobials as soon as possible and ideally within one hour of recognition. However, in children who present with a suspicion of sepsis but do not have clinical evidence of shock, there's a weaker suggestion that antimicrobial therapy should be also started as soon as possible, but allowing for time for appropriate evaluation in order to confirm one's initial suspicion that sepsis is likely present, but there's a warning to not wait beyond three hours of initial recognition. And then, of course, if children develop shock during the course of the evaluation, they should be administered antibiotics immediately. So that's the first half of this algorithm for the surviving sepsis campaign, where patients with septic shock and suspected sepsis should receive the same initial resuscitation with those six key steps listed there. However, there is a greater need for attending to time in patients who already have shock and a little bit more wiggle room for those children who have suspected sepsis, some of whom will ultimately have that diagnosis ruled out, and others will have that suspicion confirmed. And again, if patients cross over to developing shock during the evaluation, they should be treated immediately. So it may not be sufficient to administer antimicrobials alone for some children, and one must consider and look for the presence of infectious sources that may not be amenable solely to antimicrobial therapy. And these include several intra-abdominal processes that require surgery, evacuation of empyemas, removal of infected hardware, consideration of endocarditis that might need specific antimicrobials or, in rare cases, surgery, and then other nitises of infection that may require drainage. In addition, one must consider not only bacterial infections, but certain viral infections as well as fungal infections, and in certain places, rickettsial and other type infections. For fungi, patients at high risk include those with immunodeficiencies who are on immunosuppressants, presence of a central line, receipt of recent broad-spectrum antibiotics and TPN. Specifically, if patients have prolonged fever, hypoglycemia, or thrombocytopenia, one must at least consider the possibility of fungal infection as the cause of sepsis, because these antifungals are not commonly prescribed as routine antimicrobials, but in some patients may be important to consider or include in addition to antibiotics. So what if fluid alone and early antimicrobial therapy are insufficient and patients continue to share signs of abnormal perfusion or ongoing organ dysfunction? And so once patients are administered 40 to 60 mLs per kilo of fluid and continue to exhibit concerning signs, they could be called fluid refractory. And so in these patients, initiation of a vasoactive infusion is very important. And so which vasoactive therapy to start with? So in adults, there was a randomized controlled trial that compared two of the more commonly used first-line vasoactive agents, dopamine and norepinephrine. And what they found was that patients who received dopamine had a slightly lower probability of survival compared to norepinephrine. A large portion of the differential risk in mortality was related to the increased risk of arrhythmias, in particular, supraventricular arrhythmias in patients who received dopamine compared to the norepinephrine group. The generalizability of these findings to children where arrhythmias are much less common is not entirely clear. However, dopamine, which has often been used as the initial empiric first-line therapy in children with septic shock, has many off-target effects, including immune suppression and various endocrinologic effects, which may contribute to adverse outcomes in spite of being able to restore hemodynamics and improve blood pressure. In fact, there have been two randomized controlled trials that have directly compared dopamine to epinephrine head-to-head in children with septic shock, one in South America and one in India. And both of these studies demonstrated that epinephrine had a survival benefit compared to initial therapy with dopamine. But ultimately, one should really titrate vasoactive choice to the patient's physiology. And so patients, children in particular, young children in particular, will often present with cold shock where they have primary vasoconstriction and myocardial dysfunction, whereas some patients, particularly older patients, teenagers, will present with the more classic findings of warm shock with vasodilatation and hyperdynamic cardiac output. And in these cases, there may be benefit to a different therapeutic strategy in terms of initial vasopressor selection. So those with warm shock may benefit from additional vasopressor therapy, and those with cold shock may benefit from initial inotropic support with epinephrine. And also, especially at higher doses, is a potent vasoconstrictor, and so it could be beneficial for warm shock as well. But if one can understand the physiology of the patient, then there may be benefit to a more selective approach, although this has not been formally tested in children with septic shock in a research study. One particular challenge is whether or not we can reliably differentiate warm versus cold shock using clinical findings alone. So capillary refill, temperature extremity, quality of peripheral pulses, and so on. And so this study by Dr. Ranjit and colleagues in India showed that when patients were assessed only by clinical findings alone, there was thought that most patients needed fluid. However, when echo and lung ultrasounds were added, there was a higher proportion of myocardial dysfunction that had previously gone unrecognized, suggesting that patients may need inotropes. And so one must be cautious when they use physical signs alone to whether or not that is entirely accurate to titrate vasoactive therapies. And so incorporating additional measurements beyond clinical signs, such as assessment of cardiac function by bedside ultrasound, may be very helpful to understand the patient's physiology in order to better determine how to initiate and then titrate vasoactive therapy. So again, the surviving sepsis campaign recommendations suggest that epinephrine or norepinephrine, rather than dopamine, are appropriate first-line agents in children with septic shock. Again, there's never been a head-to-head comparison of epinephrine and norepinephrine that's sufficient to determine whether or not one of these two medications is superior. But ideally, if one could try to understand the patient's physiology, epinephrine for cold shock and norepinephrine for warm shock would be preferential. In terms of when to start vasoactive infusions, so when patients meet criteria for fluid refractory shock after about 40 to 60 mLs per culiter of fluid, or sooner if they develop signs of fluid overload or have evidence of monocardial dysfunctioning on additional testing. And then, because many of these patients present without an insight to central venous catheter and obtaining central access can be a challenge in many settings, especially in young children, these vasoactive agents can be administered temporarily through a peripheral vein if central access is not accessible, but central access should be pursued if these medications are going to be continued for any period of time beyond the first hour or two of resuscitation. What about additional vasoactive agents that may be thought about as patients progress in their course and get admitted to more of an intensive care setting? One option for a non-catecholamine vasoactive agent is vasopressin, and the rationale here is that when you measure vasopressin levels in adults and children with sepsis, they tend to be low, suggesting a vasopressin deficiency in some patients. In the VAST study pictured here, there was a direct comparison of increasing titration of norepinephrine versus addition of vasopressin as a catecholamine-sparing agent in adults with septic shock, and there was a slight trend towards a higher probability of survival in patients who received vasopressin as a catecholamine-sparing agent. But this did not compare vasopressin to norepinephrine, but rather vasopressin as a catecholamine-sparing agent rather than as an alternative to norepinephrine. However, what's less clear is whether this translates to an improvement in outcomes, and so it's still unknown whether adding an inodilator such as milrinone can improve patient outcomes in addition to its physiological benefits. And so what about those patients who not only don't respond to fluid, but also continue to have abnormalities in perfusion or hypotension despite titration of catecholamines and other vasoactives? So in these patients, one must reconsider the diagnosis about whether septic shock is the most appropriate diagnosis for that patient. So for example, in neonates, the possibility of obstructive congenital heart disease needs to be considered. In older patients, cardiogenic shocks such as myocarditis or increasingly MIS-C from coronavirus from SARS-CoV-2 virus needs to be considered. And then possibility of developing things that may occur with septic shock such as pneumothorax or tamponade, anaphylaxis, hemorrhage, and so on should all be considered. One should also attend to normalization of calcium, particularly in infants, as well as glucose concentration because hypocalcemia and hypoglycemia often occur in critically ill patients and can contribute to a shocked state. And then finally, whether or not additional adjunctive therapies such as stressors, corticosteroids, whether to administer those patients with fluid and catecholamine-resistant shock remains controversial. And so steroids in particular require a little attention. So in the mid-2000s, there were really two large trials, both done in adults, that helped to inform the decision about whether to start corticosteroids in patients who had fluid and catecholamine refractory shock. So the first study shown on the left was done in France and showed that there was an increased probability of survival with use of stress dose hydrocortisone in addition to administration of flugocortisone. A larger study, multinational, shown on the right, which used hydrocortisone monotherapy without flugocortisone showed no mortality benefit for patients who received steroids versus placebo. And so that led to some controversy into whether or not corticosteroids were truly beneficial in these patients. So 15 or so years later, we have two new studies, again, both done in adults. The first shown on the left, also out of France, again showed a benefit to corticosteroid therapy. And the study on the right, multinational study, again showed that there was no benefit to, no mortality benefit to steroids versus placebo for these patients. However, there was evidence that the group that received corticosteroids had an accelerated recovery in that they had a shorter duration of shock, decreased time on the ventilator, and a shorter ICU length of stay. And there may, in fact, be host factors that underlie some of the differences in the likelihood of benefit from receiving corticosteroids. So this was work done by Hector Wong and others, looking at children with septic shock, in which they were able to divide patients into different endotypes, A and B, and then risk stratified them based on a set of biomarkers called Persevere, called the Persevere model, and showed that if you were in endotype B and high risk by the Persevere model, and you received steroids, that your mortality was much lower. However, if you were in endotype A and high risk, there was no benefit to receiving corticosteroids. Interestingly, the patients who were in endotype B had evidence of increased glucocorticoid receptor signaling. And so there may be a subgroup of patients who are more likely to respond to corticosteroid therapy and actually do benefit, whereas in other patients, they are unlikely to benefit, and therefore are only likely to experience potential harms related to steroid therapy. And so looking at patients as individual subgroups rather than as just large heterogeneous groups of septic patients may be important in future studies. So current recommendations for corticosteroids from the Surviving Sepsis Campaign suggest that you should use corticosteroids against hydrocortisone if one can achieve hemodynamic stability with fluids and vasopressors alone. This is consistent with the adult recommendation. And if you have fluid refractory and catecholamine-resistant shock, then there's a suggestion that either administration of stress-josed hydrocortisone or not using stress-josed hydrocortisone is equally appropriate in the absence of pediatric-specific data and given the controversial and conflicting adult data. And so ultimately, one could choose to use steroids or not use steroids for those patients who have fluid refractory or catecholamine-resistant shock. There's also an ongoing trial now testing whether or not hydrocortisone therapy is beneficial for children with septic shock. So what about concurrent respiratory support? So pneumonia is the most common cause of sepsis in children and adults. And so one must attend to treating hypoxemia. And so certainly patients who have oxygen saturation less than 90 or 92% require supplemental oxygen delivery in order to reverse hypoxemia and improve oxygen delivery. In addition, for patients who are in shock with low perfusion state, if they are having increased work of breathing, then that can seriously compromise cardiac output to other organ systems. Because if patients are working hard to breathe, over 20% of their cardiac output may be shunted to their respiratory muscles. And so intubating, sedating, or otherwise supporting their breathing to reduce their work of breathing may allow for cardiac output to be better distributed to other organ systems. If one is going to intubate a patient either for respiratory distress and hypoxemia or because of persistent shock, then induction with ketamine is preferred over etomidate. Etomidate does allow for hemodynamic stability but has adverse effects, particularly on the hypothalamic pituitary axis and causing adrenal insufficiency. And so there is some evidence that etomidate may be associated with increased harm. And so ketamine is currently recommended as the preferred agent. In addition, supporting oxygen delivery with ensuring adequate hemoglobin is likely to be helpful, although it's not entirely clear what the optimal hemoglobin goal should be for children with septic shock. So if children are stabilized, then hemoglobin above 7 appears to be acceptable. But if children are having evidence of ongoing altered perfusion or shock, then a higher hemoglobin closer to 10 has traditionally been preferred, although not entirely clear whether or not an absolute goal of 10 is necessary. And so this was the TRIPICU study that compared restrictive versus liberal transfusion practices aimed at different hemoglobin thresholds, sort of transfusing at either hemoglobin less than 7 or less than 9.5, and showed no difference in the rate of new or progressive multi-organ dysfunction. However, this was in children who had already been stabilized, not in children who had ongoing resuscitation needs. A study that addressed the latter was done in Scandinavia in adult septic shock, comparing two hemoglobin thresholds and showed that there was no mortality benefit to targeting a hemoglobin level of 10 or higher in the early phases of resuscitation. However, this study only included adults and did not include children. So what about other adjunctive therapies beyond initial resuscitation? There are several, and I'm just going to highlight these briefly. So I mentioned already the need to limit excessive fluid administration, particularly beyond the early phases of resuscitation. And so one strategy to minimize fluid overload is to administer diuretics or even renal replacement therapy to help to reduce the percent fluid overload in children once they're stabilized. Other adjunctive therapies that have demonstrated some evidence of efficacy, including the use of plasma exchange for thrombocytopenia-associated multiple organ failure, IVIG for patients with toxic shock syndrome may provide some benefits, although there's conflicting data. For patients who have the more recent syndrome of MIS-C related to SARS-CoV-2, IVIG does seem to be effective in helping to reduce the cardiovascular dysfunction. And then immunomodulatory therapy using IVIG, methylprednisolone, and or anakinra, which is an IL-2 receptor antagonist, for the small subset of patients who might have sepsis, MAS, or HLH overlap, has shown some benefit. And then finally, the use of rescue ECMO therapy with either VA support for ongoing shock or VV support for patients with persistent respiratory failure has been shown to be beneficial in both neonates and older children with sepsis, with survival rates that exceed 50%. And while comparisons to similar groups of patients who did not receive ECMO is difficult, the historical mortality rates for patients who have ongoing shock, severe respiratory failure, and hypoxemia, and multi-organ failure, who would be candidates for ECMO, is much higher than 50%. And then there's been increasing interest over the last several years in antioxidant therapy, and so one regimen that's gained interest is the addition of ascorbic acid or vitamin C and thiamine to stress dose hydrocortisone. And so this has shown some potential for efficacy in observational studies from adults, and there are several interventional trials testing HAT therapy now in adults. A recent pediatric study from a single center, again, observational and retrospective, but using propensity matching, suggested that there may be benefits for this triple therapy compared to hydrocortisone alone. At the current time, this therapy cannot be recommended for routine use because of still lack of high-quality data. So bringing everything back together, again, this is the resuscitation algorithm promoted by the Surviving Sepsis Campaign, which, again, prioritizes six key steps as part of the initial resuscitation of children with septic shock and suspected sepsis, including IV access, blood culture, empiric antibiotics, measuring lactate, not so much to screen for sepsis, but more to help to guide ongoing resuscitation, fluid administration with the caveat of limiting this therapy in low-resource settings where ICU care is not available, and then early initiation of vasoactive agents if shock persists despite fluid therapy. And then moving down as you resuscitate patients, providing adequate respiratory support, titration of fluids and vasoactives, advanced hemodynamic monitoring, such as use of cardiac ultrasound to understand if myocardial dysfunction is present, and then other means of infectious source control beyond just antimicrobials. And so why are these protocols relevant? Well, there have been several studies now that have shown that when there's a bundle of care that's provided in a systematic and consistent manner, and that's with attention to time, that outcomes can be improved. So on the left is a study out of New York State, which showed that a bundle consisting of early antibiotics, blood culture, and intravenous fluid bolus together was associated with improvement in mortality, whereas the individual elements in and of themselves were not statistically associated with mortality. And then at the Children's Hospital of Philadelphia, there's been a sepsis pathway in place in the emergency department for several years, and in a multivariable analysis, the patients who were treated on that protocol had a greater odds of being free from organ failure on day two of illness, 48 hours after presentation, compared to patients who were treated for sepsis but not on the protocol. So in summary, screening in a systematic fashion is an important aid to help clinicians with early sepsis recognition and to begin treatment in a timely manner. Early sepsis resuscitation should focus on fluid and vasoactive support to ensure adequate oxygen and substrate delivery, with attention to timely antimicrobial therapy and infectious source control. As sepsis itself is not sufficient as a diagnosis, one must understand the infection that's causing sepsis and seek to mitigate that. And then one must actively test for dysfunction and support other organ systems, including the respiratory system, and consider its interplay with other organ systems like cardiopulmonary interactions that lead to persistent shock in the setting of increased work of breathing. And then finally, the use of guideline bundles, while they're not perfect and require frequent iteration as new data becomes available, clearly demonstrate that there's improvement in outcomes when a collective set of therapies are done in a consistent manner such that they can be administered quickly and timely and efficiently. So thank you very much. I appreciate your time, and I, well, would normally take questions, but I suppose not in this format, but thank you very much and good luck.

pediatric sepsis

clinical guidelines

diagnosis challenges

screening algorithms

fluid resuscitation

early administration of antimicrobials

vasoactive therapy

additional therapies

improved patient outcomes