Well, thank you. It's a pleasure to be here, and I thank the organizers for the invitation. This has been a great session so far, and I think my conclusions are going to fit very well with the first two presentations, which is biomarkers are extremely important but by no means can be used alone. They need to be combined with clinical assessment, and I think that we've heard that theme through both of these first presentations. I'm going to give you an abbreviated version of the presentation that's available online in the interest of time, but there's much more information available if you're interested. So I've listed my disclosures, and I wanted to focus on the potential role of biomarkers in sepsis. There are a number of things we've heard about related to diagnosis and prognosis, but there are other things that biomarkers can do. They can help us decide when to use antibiotics, try to tell us if it's a bacterial versus a viral infection, in the case of aspiration, maybe tell us if it's a chemical or bacterial aspiration event, how long to use antibiotics, and that's been very well studied in community-acquired pneumonia and sepsis in the ICU. How intensely to treat. Could we use a biomarker that's telling us we need to add more therapy, and I'll show you one trial that addressed that point. And then we could maybe define the site of care where it's been most specifically studied in community-acquired pneumonia, and probably most importantly, we've heard about this already, can we use biomarkers to define the prognosis of patients with severe infection? Now the reason this talk is entitled Beyond Procalcitonin is that most of the literature has been generated so far for all of these points with procalcitonin, but there are a lot of other potential promising biomarkers. By no means can I review all of them, and I'm going to highlight just a few in a moment. What is the ideal biomarker for sepsis? Ideally we'd like something that's very specific for bacterial infection, so that we know when to use antibiotics, who to treat, and we'd ideally not like it to be affected by prior antibiotic therapy, so if the patient got a dose before you got your hands on them, you might still be able to know that there's a bacterial infection. You'd like it to work in any kind of host, independent of the immune status, which is unlikely to truly happen. You'd certainly like it to have prognostic value, where the findings tell you what's likely to happen, and you'd like to identify infection early in the course of illness so you could intervene. You might intervene sooner than your clinical criteria would tell you. You'd like the results to be available rapidly, ideally from a point of care testing, where the cost isn't too high for routine use, and in many instances you might want to follow it to decide when to stop antibiotics and what the ultimate long-term prognosis would be. So procalcitonin is the first biomarker that's been extensively studied, although C-reactive protein appeared earlier. I don't think there have been as many recent studies in sepsis and pneumonia with CRP as in procalcitonin. This is a hormone kind that's expressed by neuroendocrine cells, but cells like the liver can produce it in response to stress, so it's an acute stress response. It's a reactive production, an acute phase reaction in the presence of bacteria but not with viruses, and importantly, you can get levels in the serum 100,000 times above the level that you might find in a healthy individual, and it rises very rapidly. It actually rises sooner than a rise in C-reactive protein. That's been the interest in procalcitonin, very high levels, rapidly produced, acute phase reactant, earlier than a mediator like C-reactive protein. And in the surviving sepsis guidelines, which we've already heard about earlier, they define two potential roles for procalcitonin. They say it should not be used to decide when to start therapy. That you're going to have to do based on clinical assessment, and it would be a huge clinical mistake to have a low procalcitonin level in a very sick patient and say, nope, I'm not going to start antibiotic therapy. That's not a good idea. But they do say that along with clinical assessment, maybe you could use a biomarker like procalcitonin to define the duration of therapy. This might help us with antimicrobial stewardship, and in the end, I'll talk about how important this may be in treating patients with COVID. Why do we look for new biomarkers? Procalcitonin isn't perfect, and although I don't have nearly enough time to talk about this review, this review highlights the fact that there are over 200 biomarkers being studied for the treatment of sepsis. And again, it's sort of like a shot in the dark to know which ones are ultimately going to be the ones we rely upon in the future. In this article, there's just some of them listed here that are being studied, and by no means do I think that the ones I'm going to highlight next are the only ones to be discussed. But one of the ones that's gotten a lot of interest is pancreatic stone protein, and it's now becoming available as a bedside tool for point-of-care testing. It's an insoluble peptide glycoprotein made in the pancreas and other organs, and it has an intrinsic antibacterial activity. In response to stress and organ damage, levels rise, and it can be a biomarker of sepsis and infection, but probably the key interest in this is it can rise as early as five days prior to the clinical diagnosis of sepsis and sooner than Procal or C-reactive protein. So this is potentially interesting and promising, because if you're monitoring this, you may know when to start antibiotic therapy sooner than with other biomarkers. And in the study that's cited here, in the first 24 hours after admission for sepsis and septic shock, it has better prognostic value for mortality than other biomarkers, and you can have some prognostic value on ICU admission if you combine it with clinical scoring. So I think this is a biomarker to watch. I don't know if ultimately the promise will prove as valuable as with procalcitonin, but certainly one that seems interesting. Another one that's of interest is pentraxone. This is also an acute phase reactant. It's part of the innate immune system, and you get higher levels with severe sepsis than non-severe sepsis, so it may be useful for diagnosing severe sepsis. In a meta-analysis, its value on admission was a prognostic indicator in sepsis, and in 17 studies, it had a higher value in non-survivors than in survivors, and thus it was a good predictor of mortality, and these are the data from that meta-analysis showing that it has some predictive value for mortality. So again, maybe pentraxone will be a future biomarker that we'll be interested in. The other one that I think is particularly interesting is precepsin. Precepsin is measured in the serum. It's measured as an activation of the innate immune system by bacterial infection and not by non-bacterial infection. CD14 is a toll-like receptor that recognizes gram-positive and gram-negative bacteria, and the complex between CD14 and lipopolysaccharides binds to cells, it's cleaved, and it can stimulate cell proliferation. When it's cleaved, soluble CD14 is precepsin, but in addition, when this complex enters the cell, it can also be cleaved and released, and that's also another source of precepsin, and it's been useful because it rises with infection and stays elevated with persistent infection. So it might potentially be useful for de-escalating and escalating antibiotics, meaning when the levels go up and then come down, it may be time to stop therapy, but if the levels stay persistently elevated, you may need to be looking to understand why and maybe adding more antibiotics. And persistently elevated levels also predict adverse outcome and ICU complications of sepsis. And this is just a schematic diagram of precepsin showing the CD14 LPS complex on the surface of the cells and demonstrating its lysis to form soluble CD14, which can also happen internally within the cell. So there are two sources of precepsin. Probably the real answer is going to be no single biomarker, but a combination of multiple biomarkers to diagnose sepsis. This is not a new concept. It's been published now for several years. This was an Italian study of over 700 patients with infection, 120 with non-infection in five Italian ICUs, and they looked at five biomarkers, ProCal, soluble phospholipase A2, precepsin, soluble IL-2 receptor, and Estrem, and infection was best diagnosed with a multiple biomarker score. A single biomarker was not nearly as valuable for predicting sepsis as multiple biomarkers, and that's illustrated here that when you looked at a combined score from all of these biomarkers, you got a much better predictive value for the presence of sepsis than any one biomarker. Now, this may be cumbersome for bedside use, but it's an important point that there may not be a single biomarker that by itself is good enough. Another study looking at a multi-marker approach to predict sepsis was looking at mortality, looked at patients with sepsis and septic shock, and multiple biomarkers were best for predicting 30-day mortality. In this study, they used ProCal, precepsin, Galectin-3, SOFA score, so they used a clinical score as well as biomarkers. And again, they demonstrated that when you used the combination of multiple markers, it was the very best predictor of mortality in this study. There are many things that ProCal has been demonstrated to do that these new biomarkers need to also study, and probably one of the best points here has been it can guide the duration of antibiotic therapy. This was a Dutch study of 15 ICUs throughout Holland where they used an algorithm for serial measurements of procalcitonin, and with procalcitonin guidance, they used less antibiotics, and they had actually an improvement in mortality, which was subsequently investigated in future studies. Serial measurements of procalcitonin have been proposed to be a way to guide escalation of antibiotic therapy. This was an older study of 1,200 ICU patients, randomized to ProCal guidance, escalation of therapy, where if the level rose rather than fell, that might be a clue that you were missing something and you needed to add antibiotics. And this was not only a negative study, but it illustrated harm, that if you followed this algorithm to escalate antibiotic therapy, you increased length of stay and you increased the rate of mechanical ventilation. So this is not a demonstrated value yet of this particular biomarker. It may also help us stratify patients with community-acquired pneumonia. And in this study, if you combined procalcitonin with other measurements of severity, you could identify the need for ICU admission. The combination of multiple clinical predictors of severity with a high procalcitonin led to as many as 40 percent of that group needing ICU care. Now in COVID, biomarkers may play an important role, because in this review, it's very clear that less than 4 percent of COVID patients have bacterial infection on admission. And if we look at these data, what we see is, in this left-hand column in the box, that presence of infection is very low, typically less than 5 percent. The use of antibiotics on admission in COVID is very high, probably as much as 90 percent. So we've got to have a strategy for not using as much antibiotics in COVID patients when they come in as we're currently doing. And biomarkers may be useful. One biomarker that's been studied here is serum amyloid A. This is another acute phase reactant that arises rapidly with infection. It's not specific, and this study didn't talk about timing, but the levels were higher with more severe illness than with milder illness. I've just highlighted a few of the studies here that looked at milder and more severe illness with higher serum amyloid A with more severe illness. But there have been now a few studies that have looked at procalcitonin for stewardship during COVID. This was a pre- and post-antimicrobial stewardship study in COVID-19. If the procal was less than 0.5, that was supposed to be a guide to stop antibiotics. And when this was used as a guide in this study from Singapore, more antibiotic discontinuation was done within four days, a shorter duration of therapy. And interestingly, when they finished IV therapy in the procal-guided group, they didn't go to more therapy with oral therapy. They just stopped. And without procalcitonin guidance, they switched to oral therapy. They were reluctant to stop antibiotic therapy. So that's a potentially valuable use. And this is another interesting combination of clinical and biomarker studies done in Thailand, two studies by the same group, where they used a combination of procalcitonin and the clinical pulmonary infection score, which is a multi-factor score of fever, white count, purulence of secretions, radiographic patterns, and oxygenation. And if patients had a procal less than 0.5 and a CPIS less than 6, they were urged to stop antibiotics or not start them. And when this was studied both prospectively and retrospectively, they led to more responsible antibiotic use. Now, we also have VAP. And VAP is clearly not eliminated. And this was a review of a paper that I commented on that showed that in COVID, we have VAP. We have, interestingly, in this study, what they found was if you compared the COVID period to the pre-COVID period, there was more VAP in the COVID period than in the pre-COVID period. But not just in COVID patients. In non-COVID patients during COVID times, there was more VAP than in pre-COVID times. What does that mean? I don't know. I think, for one thing, it means we're no longer pretending there's no VAP. But also, I think that the other point here and the other concern is, although COVID may have some unique features that predispose to VAP, non-COVID patients during COVID period had more VAP than pre-COVID. And that may be an indictment of the fact that we just don't have the nurse-patient ratios now that we used to have. So I think we've got to deal with VAP. And biomarkers may help us here. This was a study done in 66 patients, half of whom had secondary bacterial infection. And they did serial measurements of ProCal and C-reactive protein. And what they found, very briefly, is with both C-reactive protein and ProCal, if you looked at those with and without infection, there was a rise in the biomarkers early on. And that may be a clue to the presence of infection. So to summarize, biomarkers, I think, are useful in sepsis. And they may specifically have a role for us now during COVID-19. We've got to avoid overusing antibiotics. We use them much too much when patients are admitted with infiltrates at the time of admission. We have to use them for VAP. And we may have to use them for other infections. But many patients get antibiotics without documented infection. And maybe biomarkers are a strategy to address that. Maybe BAL could help us. But I don't know that all of us are going to be doing that regularly. And ProCal may help us guide antibiotic use by withholding antibiotics in COVID patients if it's low on admission. And serial measurements may help us in ventilated patients to identify when they have VAP. So I think that the potential of biomarkers in sepsis is strong. But I think we need to learn a lot. I think that the gold standard that new biomarkers are being compared to is ProCal only because there's so much data out there. And I think, ultimately, the answer is going to be multiple biomarkers combined with clinical assessment and not one single biomarker that will do everything for us. Thank you.

biomarkers

sepsis

diagnosis

antibiotic therapy

procalcitonin