Good morning, everyone. Just a reminder, please silence your cell phones during the presentation. I hope everyone's enjoying the conference so far. My name is Blisma Sarsic, and I'm here from the Hospital University of Pennsylvania in Philadelphia. I'm going to talk about the new paradigm antimicrobial stewardship in the ICU. I don't have any disclosures to report. First off, my objectives, I'm going to talk about a myriad of things, just touch on some things superficially, just some interesting things I found in research. Not going to go deep into anything in particular too much, but I'm going to talk about the importance of antimicrobial stewardship and list and compare some educational opportunities that are available, discuss some studies that provide some new information on antimicrobial stewardship practices, also talk about our government involvement in antimicrobial stewardship practices, and describe some new methods and theories to impact a stewardship program. So our goal overall is to provide tools and resources to educate and raise awareness about the importance of antimicrobial stewardship. So our introduction to antimicrobial stewardship in the ICU, we all know the standard antimicrobial stewardship practices. We care about duplication of therapy, double coverage, the correct dose, optimization for CNS infections for the extent of fusion beta-lactams. We care a lot about like renal adjustments, and obviously IV to PO. We care a lot about de-escalating antibiotics once the sensitivities and cultures come back. There's a lot of things that we're currently looking at, and they've been working to a certain effect so far. But why is antimicrobial stewardship so important? Historically, it started out as a financial focus. Basically, it was pharmacists printing out reports of the expensive antibiotics on formulary that we were using and trying to figure out how to get the patient on something cheaper, more cost-effective. Multidrug-resistant pathogens have been around since antibiotics have been around, penicillin in the 1940s. So there's been a growing trend that as every antibiotic comes out, a resistance pattern has been quickly gaining traction and emergence. So antimicrobial stewardship is so important because we are trying to decrease these multidrug-resistant pathogens, which is when you have resistance to at least one drug in three classes. So it also helps with patients that are having treatment failure, and it helps with prolonged duration and mechanical ventilation. So overall, studies have shown that these stewardship programs reduce the rate of the resistance, our duration of ventilation, days of antibiotic use, and healthcare costs over on the critically ill patients. So it really is a firm belief that this is a public health threat at this time because we're facing so many emerging drug-resistant pathogens. And as with the decrease in new antimicrobial agents, antimicrobial stewardship should be practiced in all critical care units. So in 2014, there was a presidential executive order that was called CARB, Combating Antimicrobial Resistant Bacteria. So the president was telling us that we need to pay attention to this. Now most recently, in September 2019, CMS has taken action. They're basically saying, if you're getting reimbursement from us, you need to have an antimicrobial stewardship program in place. They want to help curb the inappropriate use, reduce the risk to patients, and obviously they want to advert the increasing healthcare costs by these excessive antimicrobials. There are a lot of educational opportunities that you can get involved in. I only listed a few, but there are many. You can find which one that maybe suits best for you. Some of them are self-studies, such as the IDSA. Some of them are geared more towards pharmacists, SIDP. There's some that help with, you know, revolutionizing your antimicrobial current program, how to start a new program. There's so much resources out there. I really encourage and even challenge you to see what you can do at your own institution. So rethinking the paradigms, first I'm going to talk about the gut microbiome. So we have a lot of ongoing concerns for antimicrobial resistance, especially gram-negative bacteria. In our GI tract without antibiotics, we have an amplifier of antimicrobial-resistant bacteria occurring, and broad-spectrum antibiotics cannot target these specific microbial species. So this results in decreased bacterial diversity, and it has the potential to cause overgrowth of resistant bacteria. The gut microbiome is something that was pretty new to me, but it has actually been studied and known about since the 1980s. We've known that the antimicrobials obviously affect the normal gut flora, and current ASP practices favor oral administration because it is better for the patient because it reduces their risk of infection, the complications from the lines, helps them get discharged and home sooner. You know, it's cheaper, decreased length of stay, everything like that. But there is a study by Zhang et al. that has challenged oral administration. They used a mouse model testing oral amoxicillin and tetracycline, and they found that PO versus IV administration of these actually increased development of antimicrobial-resistant genes. So IV to PO remains the standard, but it's worthy of further consideration. And in some cases, oral administration may reduce mortality because it decreases your gut bacterial overgrowth, and you're clinically unwell patients because they just have a lot of different bacteria in their gut. But for some that aren't typically sick all the time and like frequent flyers in the ICU, it may cause more harmful effects. So these antibiotics need to be studied, and they're binary, and like their fecal excretion rates, how they're affecting, because every antibiotic is different on how it affects the gut microbia. And we really need to get a more informed and sophisticated approach to how we're going to prescribe these antimicrobials. We're going to need PCR and Sanger sequencing for sure to further our understanding. This is something that even with a lot of research, as I said, and microgenomics, there's a lot of stuff out there, but it's still not fully understood because it's more like an individual type of treatment effect that you have to approach it as, and that's going to take some time to get to. So delay of appropriate treatment and mortality. We know that we need to, when a patient is sick in the ICU especially, we need to, that time to first antibiotic needs to be very quick. And there's been tons of studies about it and how it helps with mortality, but administration of subsequent doses receives less attention. So if they start all their stat antimicrobials, and then it takes more than the timeframe of, you know, 6, 8, 12 hours, whichever, until the next dose, or even 24, this isn't studied as much. So Anderson did a study to investigate the effect of early treatment of 28-day mortality. This was a study done in Sweden. It focused on appropriate administration of first and second doses, and for patients with community onset sepsis and severe sepsis. So the culture confirmed focus of infection occurred in 73% of patients. It was UTIs followed by pneumonia, skin and soft tissue infection, and then the others was unknown. The mortality was 37% among the 46 high-risk patients. The high-risk patients were noted by their age, their maximum lactate level, if they had multiple organ failures, and if they were on mechanical ventilation. So in all, 23% of patients receiving early antibiotic treatment included first and second doses, 34% in the high-risk group, and this correlated to a lower 28-day mortality, a p-value of .022. So with this study, we found the combination of inappropriate first and second doses mortality was increased, especially in the high-risk group, by threefold. So this study kind of shows that there are shortcomings in the management of sepsis. It was, however, a single study, and it also had some limitations because the inclusion criteria was outpatient to a specific ICU setup for septic patients at this hospital. So they were very specific in their study, and above all, you still need to look out for multidrug-resistant factors in your patients, so they were doing a good job at that. So it's kind of interesting to know, like, what would happen in a normal setting, how often first and second doses get done appropriately on time, especially with prior authorization through some hospitals have restricted antimicrobials that require prior authorization, and so there may be a delay in therapy until they can be seen by ID or the clinical specialist or whoever gives the approval. So another paradigm rethink is broad-spectrum antibiotics. So our most common infection treated with antibiotics in the ICU is pneumonia, and many practitioners defer to broad-spectrum antibiotics for CAP, even if there's guidelines and clinical pathways that say we just need maybe ceftriaxone azithro for your ICU patient, they will cover for MRSA, pseudomonas, and they'll go the vancosin broad-spectrum route. So there is data that says that there's worse outcomes in mortality for these patients that are put on broad spectrum unnecessarily. So clinical pathways are in effect at most hospitals, I would imagine, but maybe not ICU-specific, and the clinical pathway needs to be more than just a generic algorithm. It should hopefully address a lot of different specific traits about the patient themselves, and there's also a computerized program that has been studied for three years in a tertiary hospital that is actually working on taking about 15 different factors of a patient and determining to guide whether or not the provider should give antimicrobials at all, what dose, what duration, what drug, everything. And it looks at white blood cells temperature, past microbiology data, it looks at your chest X-rays, it looks at surgical history. There's a lot of different factors that go into it, so it's much more patient-specific and telling. So the next is something huge that we could talk about all day, but I'm going to briefly touch on some that I found. So the providers are trying very hard to make the right decision, and it's very difficult, and especially if they have to wait a couple days for sensitivities to come back on cultures. And so now the lab is part of this. It's no longer acceptable to wait that long. We really need to find out what are bacterias that we're treating and what it is resistant to in much less than 72 hours, even hopefully less than 24. So this Maldatov is something that is revolutionary. It's supposed to help actually do this, where you can isolate the bacteria and the sensitivities within about 24 hours. There's also the polymerase chain reaction, which you can use to help with determining if there is a need for MRSA coverage, so you can de-escalate a lot of empiric vancues and even lenazolid. It has a 99.8% success rate when it's done via bronc or BAL. And then there's the seroprocalcitonin, which we used to use for determining whether or not to use antimicrobials. Now it's being used in some studies to determine by trending the PCT to determine if they need to continue antibiotics or not. So the CDC is also getting very involved in this containment strategy. It's a really interesting document. So it came out in 2017. And they're listing their, this is a public, antimicrobial resistant is a public threat. They have 18 bacteria they're focused on of urgent and serious and concerning. Of the urgent are like our CREs and C. diff, as well as resistant Neisseria gonorrhea. So basically what they want to do is get these diagnosis tests everywhere possible. So the rapid detection of these targeted pathogens and their resistant mechanisms. They also want on-site infection control assessments by trained experts to identify gaps in infection prevention. For additionally screening of the exposed contacts to identify if there is asymptomatic colonization occurring. And this should be a coordinated response among facilities, especially if your institution frequently has bounce backs from say a long-term care clinic or somewhere close by. And you need to continue these interventions until transmission is controlled and detection and control efforts can extend to other facilities that share the same patients within your community, city or state. Okay. So I wanted to mention this just because I thought it was interesting. So it's another aspect of our broad spectrum antimicrobials in a way becoming more specific. So there has been a recent trial that came out, the Aspect NP. And it was randomized study phase three multicenter perspective with Zerbaxa, which we know is used to treat complicated UTIs, phylo-nephritis and complicated intra-abdominal infections with metronidazole versus meropenem in the treatment of adult patients with ventilatory associated bacteria pneumonia or ventilated hospital-acquired bacteria pneumonia. And the study found that Zerbaxa was non-inferior to meropenem, the primary endpoint of 28-day all-cause mortality and the intention to treat a non-inferior margin was 10%. The secondary endpoint was test of cure and the intention to treat population was 12.5. So now Merck is approaching the FDA to get a new indication for their drug. And this is increasingly helpful because as I said before, pneumonia is one of the higher causes of mortality in the ICU. So and the good thing about this study actually was that it only looked at ventilated patients in the ICU setting. So in summary, at our institutions, we all need to ensure that we're part of the leadership of our antimicrobial stewardship. We need to be collaborating with everyone. Everyone can be involved in infection prevention, the micro lab. It's not just provider and pharmacist. The nurses should be involved. Everyone should be looking at this because it's a very growing concern. We know that we have the capability to do better than what we're doing. What we need is more motivation and we have the opportunities. We just hopefully can get more of these. Here are some of the barriers to our successful antimicrobial stewardship programs. So firstly, in the ICU especially, there's diagnosis of uncertainty and then there's fear of not adequately covering a certain pathogen, especially with septic shock. So we all know when the patient is really, really sick, we are on more and more broad spectrums for longer periods of time and there's less chance of de-escalation. Overall there's about a 30 to 50% chance of de-escalation in the ICU, which isn't really huge but at this time, we're a little bit scared to do it because we don't have the proper tools to make us feel good about what we're doing. So these two, number one and two, basically act synergistically to increase the use of multiple and broad spectrum antibiotics. We have a lack of ICU specific clinical pathways and other resources that can help us treat a patient individually and look at their antimicrobial plan with their disease state specifically and past micro data. And we need more availability of these diagnostic testings so we can get quicker results on our, quicker results on the resistance and things. So overall, the giving antibiotics is thought of as like, oh, it's not a big deal. It's better than not giving them, but there are toxic side effects, adverse drug reactions and a lot of things that we need to be thinking about in addition to causing more antimicrobial resistant in our gut microbiota. So there are a lot of things that need to be considered and research is still pending. So hopefully there'll be some exciting new things coming our way soon in our guidelines that we can start instituting in our own practice. Does anyone have any questions or should I save it to the end? Save it to the end? Okay. Cool. Thank you. Good morning. Good morning. My name is Mike Klompas. I'm an infectious disease doctor at the Brigham and Women's Hospital in Boston and I'd like to thank the organizers for the opportunity to be able to talk today about antibiotic de-escalation. It's a very big topic, so I've just chosen a few points of interest, but don't make any claim to being comprehensive. These are my disclosures. And here's some information, actually, how often is de-escalation done in practice in a contemporary ICU? So this is a Diana study. It's only been presented in abstract so far. It's not been published, but this was a prospective survey of de-escalation practices amongst about 1,500 critically ill patients on empiric antibiotics for suspected or confirmed infection in 152 ICUs, mostly in Europe. Two interesting sort of factoids from this analysis. The median duration of antibiotics was 10 days. Okay. That might be good. That might be bad, depending, of course, on what syndromes they were treating, but I'll tell you my instinct is it sounds long. Here's a more striking figure over here. De-escalation on antibiotic day three only occurred in 16% of patients. Now they did do an analysis on the patients who had de-escalation compared to those who did not on day three. They did find that de-escalation was associated with higher rates of clinical cure, shorter ICU stay and hospital length of stay, no difference in mortality. We'll have to wait for the full publication to see how good a job they did adjusting for differences in people who got de-escalation versus not. The worry always when you see these figures showing that better outcomes of de-escalation is, well, presumably you only de-escalate in patients who are doing okay, right? You don't de-escalate in the guy who's continuing to crash. So it's not clear what to make of that, but thus far, at least it's saying that it's not clear we're doing harm from de-escalation. So here are some points that I'd like to share from just my clinical experience and my read literature on strategies that you can take as you try to think through a de-escalation approach in your own practice. My first point I'd like to make is we need to be cognizant of the fact that we're often wrong about the diagnosis of infection. And that, in turn, leads to a great deal of unnecessary antibiotic utilization. Here's one study published last year, a prospective two-day audit of antibiotic prescribing in 17 hospitals in France, 453 patients with antibiotics, 37% of those antibiotic courses were deemed to be unnecessary. So a full third, more than a third of those antibiotics simply not needed at all. The factors that were associated with unnecessary antibiotics was a lack of a diagnosis and a lack of blood cultures. So if you don't know what you're treating, both in terms of the diagnosis and you don't know what the pathogen is, you're going to have difficulty making an appropriate antibiotic plan for that patient. Ipso facto, work as hard as you can to find your diagnosis and to establish a microbiological pathogen. Here's another look at this common phenomenon, overdiagnosis. A series from John Hopkins University, where they look back over a one-year period, look at all the VAP diagnoses in their intensive care units. So in this one-year period, there were about 220 patients who were clinically suspected of VAP. By day one, day three, an expert panel of infectious disease doctors and critical intensivists and pharmacists looked at these patients and said, well, as more information emerges, do we actually think this patient has VAP or not? By day three, they had ruled out VAP in almost three-quarters of the patients. Nonetheless, on day three, almost all of them were still receiving antibiotics, despite the fact that there was no longer clear evidence of persistent diagnosis of VAP. And indeed, we've seen the same phenomenon again and again. So here's a series out of the Netherlands, where they retrospectively looked at 2,500 patients who'd admitted to ICU with a diagnosis of sepsis, were unantibiotic for sepsis. They did a retrospective analysis when the dust settled, when all information was back about the patient, what was the actual likelihood that those patients did have an infection or not? And you can see that a good 43% of the time, infection was deemed to be absent or unlikely. Now, that's not to say that the folks made a mistake up front in giving these patients antibiotics. We all know the initial fog of presentation. You don't know. So some rate of overdiagnosis is probably appropriate. But I think it's important context for us as we think about antibiotic de-escalation to realize that a third or more of the time, the patient probably does not have an infection. The next thing I think you want to keep in mind as you go about an approach to de-escalation is to be mindful of the risks associated with excess antibiotics. So here's a study that I did with my colleague, Dr. Chandra Rhee. And what we did is we did a retrospective analysis of about 17,500 patients admitted to 136 US hospitals. And we looked to see, amongst patients with culture-positive sepsis, using sepsis-3 criteria, what fraction of those got treated for different kinds of potentially resistant pathogens? What fraction of those actually had confirmation of the presence of one of those pathogens? So you can see over here is the fraction of people who got treatment for MRSA. And here's the number who actually had confirmed presence of MRSA. Likewise, seftra axon-resistant organisms, VRE and ESBL. In essence, only one in six patients treated for MRSA had MRSA. One in six treated for a seftra axon-resistant organism actually had that. One in 16 treated for VRE had VRE. One in 70 treated for an ESBL had an ESBL. And I'll say that I think our results are conservative because we limited it to culture-positive sepsis. So all those additional patients with culture-negative sepsis who are getting treated, no reason to think that they're more likely to have resistant organisms. And all the people who have viral sepsis, which is common because we heard that pneumonia is the most common cause of antibiotic utilization in the ICU. And one-third to one-half of patients with critical pneumonia have a viral etiology. The other part of the study, though, is that we found that we know that undertreatment, inappropriate treatment or delayed treatment is associated with worse outcomes. But it turns out that over-treating is also associated with worse outcomes in this observational analysis. So here's the risk of hospital death. If your antibiotic was too narrow, you failed to cover for an organism that was present. But likewise, there was also increased risk of death if you over-covered. If you gave that patient Vanco, they did not need it. If you gave that patient Cefepime, Peptazole, et cetera, where they didn't actually have a pseudomonas present. Also increased rates of acute akinity injury and increased rates of C. difficile with over-treatment. So our classic paradigm has always been to say, listen, if you undertreat the patient, you're going to put them at risk. Therefore, be aggressive. Give a broad-spectrum treatment. We need to over-treat immediately. I think that we have to balance by saying, listen, there's also harm when you over-treat. We need to give a bit more balance to those two sides of the equation. And so how do we face this challenge of trying to know who do we need to be immediate and aggressive and broad? And who can we take more time to actually acquire diagnostics to both confirm the diagnosis and to try to establish a microbial cause of the patient's illness so we can target the antibiotics? Well, we know that in septic shock, time really matters. Our colleague, Dr. Kumar, is right here. This is one of his classic papers showing that each hour of delay in time to antibiotics in patients with septic shock leads to increased mortality. So in septic shock, we cannot play any games. But what about sepsis without shock? So we have a great deal of emerging data from this five in sepsis campaign and from sepsis bundles giving us some insight into what happens to patients without shock. So this is the data from New York State, about 50,000 patients analyzed, showing that the time to antibiotic was associated with increased mortality in those who required vasopressors but not in those who did not require vasopressors. Here's the same information essentially coming out of Northern California. They did an analysis of 35,000 patients being treated for sepsis, those who had septic shock, a high increase in the absolute rate of mortality per hour of delay, but much less so with those severe sepsis alone or sepsis or sepsis. So the thinking right now that's coming out of the Infectious Disease Society of America is that we ought to tailor our approach to both your certainty of infection and your severity of illness. If you're certain the patient has an infection, you're in this zone over here, it doesn't matter how sick they are. If they definitely have an infection, treat it, right? Why not? Likewise, if the patient is critically ill, they're in septic shock, even if we're going to be wrong, we can't afford to make a mistake and withhold antibiotics from that subset who does have infection. Give antibiotics. Immediately be aggressive. But if you're in this gray zone over here where you're uncertain if the patient has infection or not, you're more sort of on the edge, and the patient is not critically ill, let's say we'd say, hold your antibiotics, get more data. You have a little bit of time to play with. Exactly how much time you have to play with is a topic we can get into if you'd like to. It depends on sort of many factors, but the point is that you don't have this one hour window with these patients in the low zone over here with low degrees of illness and low certainty of infection. And we actually do have some studies showing that this more conservative approach to antibiotic prescribing works, it's safe, might even actually have mortality benefits. So here's one example I'll show you. This was a prospective cohort study in the University of Virginia Surgical Intensive Care Unit. It was a quasi-experimental study. In year one of the study, they were practicing the way most of us do, which was an aggressive strategy if your patient sort of looks at you the wrong way and sniffed like infection, you gave him antibiotics immediately. But in year two, they said, let's put the brakes on. Let's wait and get definitive evidence of infection before we give antibiotics. What was definitive evidence of infection? Here we go. Basically, you had to have a positive culture, varying of course, but the patient's site. Now surgical site infection, you can start immediately because often there is no diagnostic microbiology for those, and if the patient was on pressers, of course, it was fine to start immediately. What happened in terms of results? Well, not surprisingly, in year two, fewer patients got antibiotics. And of those who did get antibiotics with confirmed infections, the antibiotics were more likely to be appropriate. Here's the fascinating finding over here. Amongst the patients with confirmed infections, in that second year, that conservative year, even after adjustment for all the differences between patients they could find, there was a lower mortality rate associated with this more conservative watch and wait approach to antibiotics. Here's another example showing basically the same phenomenon. This is an old study over here looking at diagnostic strategies to treat suspected ventral associated pneumonia. They took 413 patients in France with suspected VAP, and they randomized to diagnostic endotracheal aspirate versus bronchoalveolar lavage. And empirical antibiotics were only given if the patient had severe sepsis or a positive Gram stain on endotracheal aspirate or a positive BAL. So you have to have a positive Gram stain or you had to have severe sepsis. And directed antibiotics were then only given per culture of susceptibilities. If the cultures were negative, they gave no antibiotics at all. What this led to was this, was that amongst the patients with endotracheal aspirates, basically all of them had a positive Gram stain. And so just about all of them got immediate antibiotics. Whereas in those that they used BAL, only about 50% of them had a positive Gram stain or had severe sepsis. Only about 50% of patients got immediate antibiotics. And everybody else, they waited for confirmation of infection. That was associated with a lower mortality rate by day 14. Now that signal disappeared by day 28. But again, showing us that the strategy of watch and wait for the patient who is not severely ill is really safe and might be beneficial. So those are, I think, some contexts as we try to think about who needs antibiotics and who does not and how to start them. In terms of setting our approach now to stopping those antibiotics, I think the key thing, you've got to know the randomized control trials and duration of treatment. And there's an industry of these trials coming out. So intra-abdominal abscess over here, beautiful study, 518 patients with complicated intra-abdominal infection and source control who were randomized to antibiotics until resolution of fever, leukocytosis, and ileus, plus two days versus a fixed course of four days. Primary outcome was a competent of surgical site infection, recurrent abscess or death in 30 days. And it turned out in practice that led to a median duration of eight days in people getting usual care versus four days in people randomized to intervention. Outcomes, no difference in surgical site infection rates, recurrent abscess, death, or the compound outcome, showing that this short course therapy in patients who had source control, that's the key point over here, and were clinically showing signs of improvement, four days was perfectly fine. And that now has become standard of care. Gram-negative bacteremia. Here's another randomized controlled trial, 604 patients with gram-negative bacteremia randomized to seven versus 14 days of antibiotics. They were randomized so long as they were clinically stable on day seven. Again, you can't apply these short versus long course treatments blindly. Your patient has to be on the pathway to clinical improvement. The outcome over here was a composite of 90-day mortality, relapse, suppurative complications, readmission, or prolonged hospitalization. Here are the outcomes. You guessed it, no difference. No difference in mortality, readmissions, new infections, severs, seal, or multidrug resistance. And therefore, I think the emerging standard is for gram-negative bacteremia, particularly for interbacteriaceae, is seven days of treatment. Pyelonephritis. Five days versus 10 days. You guessed it, no difference. Ventilator-associated pneumonia. Our old favorite, our old classic over here. The study, I think, that started off the trend of short versus long course therapy over here. 400 patients with VAP randomized to eight versus 15 days of treatment over here. No difference in mortality. Ventilator-free days, organ failure-free days, or ICU length of stay. Community-acquired pneumonia. Five versus 10 days. Here's amongst all patients. Clinical success rates identical with five days versus 10 days. And even those who had severe disease, a port score of four or five, likewise. No difference in success rates with five days. So the standard is for VAP and HAP, seven days. For community-acquired pneumonia, five days. Now, could we get even lower? Here's another study. This one's only been published in abstract form. It's not yet come out in manuscript form, so don't take it with a grain of salt. But a randomized double-blind multi-center trial, 310 patients with severe pneumonia admitted to a hospital in 20 hospitals in France, randomized to three days versus eight days. And again, no difference. And there is actually a prior study from 2006 with the same result, three days versus eight days, showing no difference. Next point. Re-evaluate your patient and re-evaluate your plan daily. Just because you committed to antibiotics three days ago doesn't necessarily mean you have to continue them today. That old treatment, that old story that our mothers told us, our grandmothers told us, the doctor told us, which is that if I gave you 10 days of treatment, you must take 10 days of treatment. If you don't take 10 days of treatment, you're going to get antibiotic resistance. It's completely wrong. The more antibiotic you get, the more likely you get resistance. You want to try to find that minimum duration that's going to be effectively treat your infection without exposing you to additional antibiosis. In terms of a strategy to try to decrease your duration of therapy, which I think is the most potent thing you can do to minimize your risk of adverse outcomes from antibiotics, you can use procedural pro-carcinogens to try to tell you when you might be safe to stop antibiotics. The strategy goes as such. You check your pro-carcinogen value daily or every other day. And if your patient gets down to a value that's in the normal range or decreases by 80% or more compared to the peak, that's a signal that it's probably okay to stop antibiotics, all else being equal, and go ahead and do so. So two randomized trials looking at sepsis patients in the ICU showing that this strategy is safe. Here's the first one. 621 patients from seven ICUs in France over here. The daily pro-calcitonin strategy was associated with fewer days of antibiotics but no difference in ICU length of stay and mortality. So you decrease antibiotic days without putting patients at harm. And then here's the kicker study over here, which was a similar design but expanded now to 1,500 patients, 15 ICUs in the Netherlands over here. Again, showed the strategy was associated with fewer days of antibiotics. No difference in ICU length of stay but lower mortality rates. Again, this notion that excess antibiotics can be associated with harm. Less antibiotics can be associated with lower mortality rates. Don't underestimate potential adverse effects of antibiotics. Finally, just a word of caution over here. There are some risks associated with trying to aggressively promote de-escalation. I'll give you one example of that. This is actually, I think, the only randomized controlled trial we have of de-escalation versus not. It was a small study, only 116 patients who were randomized to de-escalation versus continuation of antibiotic courses. Again, done in France over here. So check this out. The patients who were randomized to de-escalation actually had longer antibiotic courses. And there was no difference over here in ICU days and, likewise, no difference in mortality. What's going on over here? So I'll say this, this finding over here that longer antibiotic courses in people who are trying to get de-escalation has been observed by others. This is not a unique phenomenon. I think that the speculation over here is that there are two risks associated with encouraging de-escalation. Number one is that it normalizes, it makes it okay to give very broad spectrum antibiotics up front. I'm going to de-escalate, so don't worry. I can pound them as aggressively as I want day one, two, because I'm going to be de-escalating day three. And as we saw, in most cases, most patients with sepsis in the U.S. actually do not have resistant pathogens. So that's risk number one, is you're encouraging the use of up-front broad spectrum antibiotics. It's appropriate for some patients, but it's not appropriate for all patients. The second factor is that clinicians sometimes have a tendency to reset the clock when they change antibiotics. So you've gone down from your broad spectrum, your vancopeptasia, your vancocephepine, levofloxacin, what have you. You switch down to just ceftriaxone alone or penicillin alone. You say, okay, now I'm going to give them their five days or seven days, what have you. And we've seen that phenomenon as far as well with discharge antibiotics as well, that docs tend to restart the clock when they discharge antibiotic patients. We think that's probably not necessary. There might be some pharmacodynamic, pharmacokinetic issues that might maybe could argue that might have some rationale to that. It's, I think, a bit of a gray zone over here. But you have to recognize that even that paradox of encouraging de-escalation can sometimes lead to more antibiotic utilization. You want to be on the lookout for that and try to protect yourself against that happening. Okay. Summary. Number one, be humble. We're often wrong about the diagnosis of infection. We're often wrong about the presence or absence of bacteria. Many of our septic patients actually have viral infections. Overtreatment may be as harmful as undertreatments. We have to do a better job of finding the balance. That might be to tailor the urgency of treatment, the breadth of treatment, to the probability of infection and the patient's severity of illness. Pursue an accurate diagnosis, an accurate antibiotic biology. If you know what you got over there, it's easier to feel comfortable making a tailored course. Know the randomized control trials and duration of treatment. We said seven days for HAP and VAP, five days for pneumococcal pneumonia, four days for intra-abdominal abscess, five days for pyelonephritis, seven days for gram-negative bacteremia. Re-evaluate your antibiotic plan daily. If the patient no longer looks like the patient has infection, you can stop the antibiotics, if you don't have to be committed to an upfront course. Consider using serial prokaryotonin monitoring to inform the stopping of antibiotics. And finally, recognize that encouraging de-escalation may paradoxically increase the use of borospectrum agents and extend duration of therapy. Be on the lookout for that. Protect yourself against that. Thank you very much. Thank you. Well, I apologize for being so informal today, but I talked my wife into coming down here and we're going to head off to the Keys right after the lecture, so you'll have to forgive me. Well, that's, I'm getting in the mood, just lift my hat up there, just look, I'm getting in the mood, thank you, definitely getting in the mood, thank you, thank you. I actually got my four children, I got the last of my four children out of the house like six months ago, so my wife and I are going, we're going to travel now. Anyway, I get to speak to you on procalcitonin, well, it was supposed to be biomarker-directed antibiotic therapy, but I might not be the best person to be talking on this because procalcitonin is pretty much the only biomarker commercially available, it only became available in Canada where I lived like six months ago, so I haven't used it. On the plus side, I've had to actually go through the literature and kind of develop my own take on it, so there might be a little bit more balance there. So you know why we want to talk about procalcitonin and biomarkers and all the other things we've done today, it's the global antibiotic resistance problem. There is antibiotic resistance pretty much popping up all over the world, and every country, like we tend to think this is, you know, an Eastern Asia kind of problem, but the fact of the matter is we've had resistance pop up, unusual initial unique resistance in North America, Ephesium, Staph aureus, Vibrio in South America, as well as all the other things that have popped up around the world. So this is not by any means, you know, an Eastern Asian problem. This is a worldwide problem. You know, of course, that essentially every time a new antibiotic has become available, within often within a couple of years, and if not within a couple of years, a couple of decades, there's invariably resistance developing, and that's a, you know, universally, that's a universal truth, basically. You put out an antibiotic, there's going to be resistance, and I know some of the companies say, well, you know, our product didn't show any development of resistance. My answer is always, just wait. We know the reason for it. My wife, who I just spoke about, is actually an allergist, and she says, you know, every patient who comes in, they've got a cough, they're hacking, they all want antibiotics. So I say, please don't give it to them, but nonetheless, it does happen. A lot of that happens outside of the ICU, of course, but we have our own issues, as Mike pointed out. Up to 50 percent of inpatient antimicrobial use has been shown to be inappropriate and or unnecessary, particularly, as Mike also mentioned, antibiotics for viral infections. Very common reason that antibiotics are given unnecessarily, and then, of course, again, as he mentioned, antibiotic course duration has been recognized to be unnecessarily long for a variety of infections, including VAP, abdominal sepsis, and Mike showed you also for CAP, and I would argue for pretty much everything. Absolutely, we've had this idea in the past, and I remember the training that, you know, if you don't give them a 14-day course, you're not going to eliminate all the organisms and that's going to drive resistance, right? That was the standard teaching back then. But I'm in absolute agreement. That's simply not true. There are certain things that are in medicine that we take as truths. They're so old and so fixed that nobody argues with it, but those are actually the things that we should go back and examine. Now, a friend of mine sent me this quotation. It actually came from a friend of his who I think said it to him directly in respect to biomarkers. I think the one thing that we all need to know about biomarkers is that they are never standalone tests, but rather adjuncts to what the clinicians bring to the table with respect to their clinical acumen, basically. There are always caveats and learning curves when using biomarkers. Biomarkers will make a bad doctor worse and a good doctor better, and I think that's absolutely true. If you use biomarkers in the context of your clinical judgment, you will make better clinical judgments. If you use biomarkers because you don't trust your clinical judgment, you will be making worse decisions. All right, so what are the characteristics of an ideal biomarker? Well, you want it to be routinely accessible, right? So that's been one of the problems with procalcitonin, at least in Canada, it wasn't accessible to us. You want it to be rapidly available, so it's got to have short turnaround time because if you get your answer kind of late, it's not of much use. It should be reproducible, accurate, and cost-effective, of course. You want good sensitivity and specificity, high predictive values. You'd like to have a good correlation to severity of disease, and you want to see a response if you apply appropriate therapy, whether that be antibiotics or source control. You want it to be a predictor of mortality along with being a predictor or a correlate of severity. And then you want it to be easily interpreted, objective, dynamic, rapid kinetics are useful for the reason I mentioned. That way you can follow the condition through the course of the illness. You want, ideally, kinetics that are essentially independent of organ dysfunction or other concomitant steroids. So for example, if you're immunocompromised, you don't want your biomarker to disappear. You want it to be minimally invasive, and a continuous variable would be useful. Excuse me. So people don't completely pick that up, but in fact, we use biomarkers all the time. Virtually all the tests we do in the ICU or in the rest of the hospital are, in fact, biomarkers of some sort, right? So the common one that we use is the white cell count, and it's a pretty reasonable biomarker. Certainly up until recent decades, that was probably the best we had for infection. We tend to use serum lactate, particularly in the ICU, as we start to move into the more severe end of things. But in fact, all of these things, serum creatinine, anion gap, the presence of toxic granulation or dolybodies, platelets, INR, those are all significantly altered in septic conditions, and in a sense, they're biomarkers. And then there's a whole bunch of sepsis-related experimental biomarkers, and some or any of these could potentially be of use in the future, TREM, we've heard about in the past. It's never been developed far enough that you could use it commercially. There are all these other ones as well. Precepsin is something that people are looking at these days. All of them have some relationship to progression of disease and sepsis and mortality risk. None of them are commercially available. And in fact, if you go further, I spent the first decade of my career as a laboratory scientist-slash-clinician, and I can tell you, I used to work with a lot of these things. And there are literally, I mean literally, 250 to 1,000 biomarkers that people have looked at or proposed on an experimental basis. So there are lots and lots of potential biomarkers out there. Procalcitonin, though, is the one that really has come to people's attention, and that's in great part because it's been developed as a commercial product. And so if you exam the literature, what you'll find is there's simply a lot of data available on it. That doesn't necessarily mean that there are not other biomarkers that may be in fact superior, but this is the one that we have data on and that we can really assess. So it's a 116-amino acid precursor of calcitonin, normally produced by thyroid C cells, with a serum concentration that's usually less than 0.05 nanograms per ml. Now, when you are septic or with systemic inflammation, it'll be produced in large quantities by a variety of cells, particularly adipose cells, and it's particularly responsive to bacterial infections. And we'll get back to that in a minute. It is pretty good in that it's detectable within two to four hours if it spikes, and then peaks within six to 24 hours, and that's actually earlier than the other commonly available biomarker, which is CRP. So we've used CRP in the past. There's not as much data on C-reactive protein, but we do know that it tends to rise after 12 to 24 hours and peaks at about 48 hours. So it's a little bit slower responding. Now, the nice thing about PCT is that its production isn't impaired by neutropenia or other immunosuppressive states. So if you're in those conditions, you'll still see something. And it's got decent utility as a prognostic indicator with higher serum levels related to the risk of death and, of course, severity of illness. Now, I said it's relatively specific for bacterial infections versus inflammation in general. The one interesting thing about PCT that I came up, as I did my reading, is that there actually is potentially a difference between gram-negative infections, where the spike is the highest, and gram-positive or even fungal infections, where it still goes up. So there's some, you know, soft utility there, although I have to say, on my review of the data, it's not such that I would look at it and say, well, I'm going to just cover gram-positives or just going to, you know, cover fungals or whatever. It's not that useful, but it's a little tool that may be able to give you a little information one way or the other. As I said, it has a relatively rapid rise after an insult, so it makes it kind of quite good that way. Good correlation with severity of illness and, of course, association with mortality. And I mentioned this last one. So this is basically how procalcitonin works. Can I ask, how many of you use procalcitonin in your institution? Can you put up your hand? So maybe about a third of you. OK. We're still trying to decide whether we're going to do it or not. I don't know if it adds to us in the ICU, depending on how much of you think of yourself as a clinician. And although I tend to think it wouldn't help me that much, I am sufficiently humble that I said to myself, you know, we should just trial it and see if, in fact, it helps us. My guess is no, but, you know. And I suspect a lot of you would feel that way. As I said, healthy, you're usually less than 0.05. If you have a local infection, it hasn't disseminated, maybe you've got a cellulitis or a urinary tract infection, something like that. If you have local, then it's 0.05 to 0.5. Start getting systemic infections. This is where we talk about sepsis, where you're having systemic signs and symptoms, getting up to about 2. Severe sepsis, which we don't use the term anymore, but we're basically talking about sepsis with organ failure, then you're getting up to 10. And septic shock, you can go up towards 100 or even past 100, depending on the severity. Now, just to look at the relationship to other biomarkers, as I said, some of these, in fact, none of them are commercially available with the exception of CRP. But you can see that TNF, for example, and IL-10, IL-6, will tend to spike relatively quickly with the onset of an infection. But then they'll come down. Procalcitonin will go up within a couple of hours, will peak around 12, and then back down. And then, of course, CRP, as I said, kind of tends to lag behind all the others. Unfortunately, these are not available. They're also, in a lot of ways, not as specific as procalcitonin. So high stress states, shock due to cardiogenic shock or hemorrhagic shock, you'll spike a lot of these things anyway. But PCT, you tend not to. If you look at the kind of kinetics over time, it can be used to help you determine whether a patient is responding to therapy or not. So a patient whose infection persists, they have inadequate source control or infection management, they'll basically stay up. But it's been shown that if you have a decline of greater than 80% from the peak PCT level, you'll have a much lower mortality rate. In fact, in one study, Schuetz here basically showed 20% mortality here, 10% mortality here. So it can be useful along those lines. And of course, that potentially can be helpful for de-escalation as well. You want the... we said that you wanted the test to be sensitive and specific. And again, here's another study looking at different markers that we typically use, lactate, IL-6, CRP, procalcitonin. And you can see that the area under the curve for this is particularly good. So procalcitonin, certainly in some assessments, would seem to be one of the best, if not the best, available biomarker for septic conditions. And one of the questions, this is what we pointed out in terms of being humble, my guess would be that if you add a procalcitonin to my unit, that it wouldn't give me personally any improvement. But on the other hand, I think virtually every clinician would probably say, I don't think so, but I'm open to the possibility. So in this particular study, they looked at a clinical judgment without PCT, clinical judgment with PCT. And as you can see, these clinicians might not have been as good as us in this room, but it did improve their outcomes. So we should be a little bit humble. It is definitely related to disease sepsis severity. So you can see here, this is PCT, SERS, sepsis, severe sepsis, septic shock, the old description. This is PCT, again, SOFA score, 1 to 6, 6 to 12, et cetera, going up. You can see a strong relationship here. This is CRP for the same thing. You don't see that kind of relationship. And this is IL-6. And again, there is a bit of a relationship there, but not nearly as strong as what we see there. So what can you potentially use PCT for? Well, I think it has some potential utility for bacterial versus viral respiratory tract infections, diagnosis and risk stratification, and monitoring of sepsis and septic shock. That is to say, particularly, are you getting better or not? Now, I'm going to be honest with you. If I'm dealing with septic shock, I can pretty much tell you're getting better because you're coming off vasopressors. So that's not an area where I feel that we need a lot of help. But if you're talking about sepsis short of shock, there may be some utility there. Monitoring the response to antibacterial therapy, the diagnosis of systemic secondary infection after surgery, post-organ transplant, severe burns, multiple organ failure, severe trauma. That's an area where I think it could potentially be very useful. And there is data on that. And if you're interested in that data, you should go and check it out. But certainly, as I've reviewed the literature, there's a particular, to me, a particular utility for that kind of thing because that situation's always very confusing in the ICU. The diagnosis of bacteremia and sepsis in adults and children, I think it has some potential utility there. Again, my own judgment is that we tend to be fast enough on the trigger for treatment of those kinds of conditions that it may not add much there. In fact, the literature is kind of both ways for that kind of thing. Differentiating between bacterial versus viral meningitis seems to be very good for that kind of thing. Diagnosis of renal involvement in pediatric urinary surgery, bacterial infection in neutropenics, and diagnosis of septic arthritis. But the one that I really want to talk to you about is the determination of the antibiotic treatment length because that's particularly where stewardship kind of comes into play. Now, all of these, you know, it potentially adds something to stewardship in that it may give you support for not starting antibiotics. But given the subject of this particular session, I think determination of antibiotic treatment length is of particular importance. So I'm going to show you some studies that have looked at this. Rather than looking at individual studies, I'm going to focus on meta-analyses because I have to tell you, given that this product was commercially supported in its development, there are literally dozens and dozens of clinical studies with this. And in fact, there are at least, you know, correct me if you think I'm wrong, but there are at least two dozen, if not three dozen, meta-analyses in this over the last 15 years. So I'm just picking a couple out that particularly show what I think is important to show. So this is PCT-guided antibiotic de-escalation with respect to pneumonia and lower respiratory tract infections. This is an older study, 2013. I think we've been studying procalcitonin clinically since about 2000, so a little more than halfway through. But what you can see is if you had PCT-guided antibiotic de-escalation done as a study, you could see that you had decreased antibiotic duration. You had decreased antibiotic prescription rates. You had decreased ICU admission, or at least a trend in that direction, not quite reaching significance. You had no impact on short-term mortality. So basically, you're getting less antimicrobial therapy, no impact on survival. Win-win. This is by the same group. They looked, in this case, at adult critically ill patients. Those are patients that are, of course, more relevant to us. And again, what you see here is antibiotic duration decreased, ICU length of stay unchanged, and short-term mortality, at least in-hospital or 28-day, also no change. So again, what's not to like? This is a more recent study, PCT-guided antibiotic de-escalation, antibiotic duration. And again, you'll see here, this is in, I'm sorry, in ICU patients. And you can see here that there is evidence of decrease in the antibiotic duration, both in studies that examined it specifically, and there were other mixed studies where they examined both initiation and cessation. So they both trended in that direction. This is where it gets fairly interesting to me. In fact, if you look at those studies, particularly those that dealt just with cessation, there was actually some evidence. Where is the? There it is. There was some evidence in the direction of favoring survival. That is to say, if you used PCT-guided antimicrobial de-escalation, you could actually had some evidence of an improved survival signal, which I think goes exactly with what Mike was saying. And I think that's incredibly important. Now, there are some issues with these studies, as far as I'm concerned. In that, we are in a dynamic changing environment. It used to be, as I said, 14 days was the standard. So if you used PCT and you dropped your duration of therapy from 14 days or 12 days or 10 days down to 5 days, well, that could be a big difference. These days, we're getting smarter about things, and we're going to shorter and shorter treatment courses. And the question comes up that, if your standard course duration is 7 days, for example, in the ICU, and you use pro-calcitonin, is it going to come down, and is it going to have an impact on survival? And the answer at this stage, in my assessment, is that we don't know for sure. But the preliminary evidence, I think I've seen one study, and I couldn't find it. But in fact, it seemed to suggest that it would even be helpful in that case. That is to say, if you went from 8 days to 6 days, I think they had in the study. In fact, they saw decreased antibiotic utilization, of course, but they also saw a signal towards increased survival. So something to keep in mind. And you're going to ask me, naturally, well, yeah, you say pro-calcitonin is good, but you need an algorithm. And the interesting thing is that there's actually a lot of algorithms. A lot of these studies used specific algorithms. And I don't have any specific algorithm to tell you is better than another. There are actually several lots of algorithms. They're all a little bit different, but they all get it roughly the same thing. And this is just one good example of it. So here's the PCT values that you're getting as the days go on. Here they're suggesting, are antibiotics encouraged to continue or discouraged? And basically what they do is they get a PCT level every one to two days. This is for lower respiratory tract infections. And if they're unstable, high risk, or strong suspicion of pneumonia, then you continue to initiate antibiotics regardless of PCT. So remember what I said. Your clinical judgment overrides anything the test tells you. So if the test is telling you something that you just disagree with, I would say go with your gut. If you depend strictly on the test, you're going to be in trouble. But continuation of antibiotics beyond the standard duration of therapy is not recommended regardless of PCT level, and if it remains high, consider it a treatment failure. So basically, at these levels, you should discontinue. At these levels, you probably don't want to discontinue your drugs, antibiotics. Now if you come down here, sepsis, the thing with this one is the numbers are just a little bit different, but the paradigm is pretty much the same. So there are differences between different types of infections. This is another protocol. Basically, in this case, obtain baseline PCT, but don't delay antibiotics. Repeat daily now. The other one was every one or two days, and then stop antibiotics when the PCT is less than 0.5, or it decreases by 80% from the peak, and the patient is clinically improving. So as I said, there are a lot of different protocols you can use. I'm not sure that anybody has shown that one, in fact, I'm sure that nobody has shown that one is particularly superior to another one. By the way, it's probably worth knowing that there are several companies that make these products. Last thing I want to mention is that as much as there's a lot of data for PCT, in fact, to the extent that it's been studied directly, CRP actually seems to be probably as good. There was a nice study done by Oliveira back in 2013, and there have been two or three other studies along these same lines where they compared them head to head, and just used, of course, different cutoff values. And as you can see here, this is what they tended to find, in fact, that there's no difference in the de-escalation effect, and there's no rate, and there's no difference in mortality. When do you want to use this with caution? There are situations where you want to be a little bit careful. If you're under massive physiologic stress, it could pop your procalcitonin levels. If you have non-bacterial cytokine activation, like vasculitis or graft-versus-host disease, some fungal infections, chronic renal disease, you'll get a mild increase in procalcitonin. And then if you have dysregulated PCT production associated with certain agents or certain perineoplastic syndromes. My conclusion, kind of looking at all the data, and remember, I don't use it routinely, I think PCT may or may not be useful for the initial diagnosis of sepsis. We didn't go over that, but the studies seem to break in either direction, basically. The more severely ill you are, probably the less useful it is. PCT may be helpful for differentiation of bacterial pneumonia from various mimics. I think the data for that is reasonable. But PCT is probably useful for direction in de-escalation in critically ill septic patients. Thank you. Oops, that's it. Thanks, Anand. That was great. Good morning, everybody. Thanks for hanging in there with us. I'm Phil Berry. I'm Professor of Surgery and Professor of Public Health and Medicine at Weill Cornell in New York City, and a trauma surgeon. I've had a lifelong interest in host defenses and the management of nosocomial infections in surgical patients. Before I start, I would like to underscore one very important point that Dr. Kumar made in his last slide, and that is that you need to be careful interpreting procalcitonin concentrations in patients who are recent to post-ops, especially if they've had major injury or major tissue injury from an esophagectomy or a whip or something like that. The literature, the medical literature, looks at a cutoff point of 0.5 nanograms per mil for the most part for procalcitonin. But in our evaluation of fresh post-ops, it's actually almost 10 times that high. So the contribution of tissue injury to the early procalcitonin response is quite substantial, in fact. Here are my disclosures. I'm only going to be talking about specific antimicrobial agents in the context of a couple of studies where they were used, so I'm not going to have any therapeutic recommendations for you. I think we're all familiar with all of this. Antibiotic resistance has been a problem since the 1950s, really. Most notably, the introduction of methicillin in 1958 led to the identification of MRSA in 1961. So this is not a new problem. But I want to point your attention down to the last two bullet points, and that is that how we prescribe antibiotics, both in terms of actual tonnage and in terms of appropriateness, is a major factor in driving resistance. And my talk today will have a surgical perspective, wherein, as we will show, lots of patients can have a pro-inflammatory response as a result of their injury or the therapeutic intervention for it, and not have infection at all. When one looks at the standard risk factors for patients to either present with or acquire MDR organisms, antibiotic exposure, or even outpatient oral antibiotic exposure within the previous 90 days, is an independent risk factor, as is recent major surgery. Well, why would that be? Well, surgery breaches a normal barrier against invasion of the host by pathogens, that being the skin. So too do percutaneous catheters and other interventions. But remember also that trauma and burns upregulate a profoundly immunosuppressive phenotype almost immediately after injury that persists for several days. And moreover, blood transfusion, which is also an essential adjunct in many, many cases that we face, itself confers an immunosuppressive phenotype, to the point where, in trauma, the administration of 15 units of red cell concentrates confers essentially a 100% risk of subsequent nosocomial infection. This is one of the earlier studies from Spain, which at the time had a serious MRSA outbreak, looking at the total days of antibiotic administration during the course of therapy and the odds ratio for acquiring MRSA. And you can see that signal is detectable even for a single day of antibiotic exposure, as might be characteristic of the administration of perioperative antibiotic prophylaxis, but increases to more than a tenfold increased risk of acquisition of MRSA if 15 or more days of antibiotics are administered. This does not account for dosing. This does not account for the particular type of antibiotics. For example, subsequent to this, we have recognized that the administration of fluoroquinolones to patients poses particular risk for the emergence of resistance, as increasingly does the exposure of patients to carbapenems. Back in 2013, and beginning the process of revision now, was a Quad Society guideline, including the Surgical Infection Society, on antimicrobial prophylaxis in surgery. It's a big document. I don't want you to belabor with it. We're not here to talk about antibiotic prophylaxis per se. But I just want to remind you that for long cases or complex cases with blood loss, intraoperative redosing of antibiotic prophylaxis is more important than postoperative administration. And why is that? Antibiotic prophylaxis protects only the incision. And the incision is vulnerable only while it is open. And remember that in the postoperative period, particularly if the electrocautery has been used, you've got a zone of coagulative necrosis about a millimeter wide, because the tip of that bovie is 2,200 degrees Fahrenheit. So you've got a zone of coagulative necrosis on both sides of that incision. And antibiotics, literally, because of the disruption of the microvasculature, literally cannot penetrate all the way to the wound edges. So remember, very important, that intraoperative redosing. We redose cefazolin and cefoxidin when used intraoperatively, Q4 hours, and became much more effective at it when our anesthesia colleagues agreed to accept responsibility for the timing of the dosing. We have had evidence since the 1990s that single dose antibiotic prophylaxis is equivalent to multiple dose, even within 24 hours. This systematic review was published in the Australian Surgical Literature in 1998. And you can see that it reflects a wide spectrum of surgical disciplines. And what you see highlighted in the arrow is the only one of these 28 studies that actually showed benefit of multiple dose surgical prophylaxis, as opposed to single dose. And over on the far right, you can see the fixed effects and random effects over here. Very narrow confidence intervals showing no benefit, even to a more than a single dose. Of course, if your case is of sufficiently short duration, you don't need to redose intraoperatively. There's more data than this now. Many of you may be familiar with something called SCIP, the Surgical Care Improvement Project, which recommended 24 hours of postoperative antibiotics for almost everybody except cardiac surgeons who complained loudly enough to get 48 hours. But all of that, in my opinion, was political and not supported by the data such as you see it here. Let's look at some of the consequences of prolonged antibiotic prophylaxis. This is a retrospective study from Nick Nimias, also a past president of the Surgical Infection Society, looking at the situation at Jackson Memorial Hospital, which is a large trauma center and also one of the largest solid organ transplant programs in the country. And they looked at the duration of practice with respect to subsequent bacteremia and the incidence of catheter-related bloodstream infections. Now this is retrospective. And since this study was reported, we've seen a dramatic decrease in the incidence of catheter-related bloodstream infections. But nonetheless, you can see that there was a substantial increase in the risk of subsequent infection after prolonged antibiotic prophylaxis. Here's another study of similar type that comes from the very busy trauma center at Los Angeles County USC Medical Center. And they were, again, looking at a multivariable analysis of drug-resistant infection. In this particular study, most of the patients who were receiving antibiotic prophylaxis for a prolonged duration of time were going back to the operating room repeatedly for multiple operations for fracture repair. But nonetheless, an independent predictor of subsequent drug-resistant infection was prophylaxis that exceeded 24 hours in duration. There are some certain areas where it's really not sensible to administer antibiotics as prophylaxis. The open wound that is not infected does not need antibiotic coverage. If you have an open abdomen and you're managing it with a VAC dressing, that does not need antibiotic prophylaxis. The presence of central lines, the presence of chest tubes, I can go on and on. They don't require antibiotic prophylaxis despite what your consultant surgeons may be trying to sell you. One area in which we hear often from our maxillofacial and neurosurgical colleagues is that patients with basilar skull fracture, patients with facial fractures need antibiotic prophylaxis sometimes for a week and sometimes triple antibiotics. Believe it or not, we get recommendations for vancomycin, ceftriaxone, and metronidazole prophylaxis of skull fractures for a week. There's no data whatsoever. This is a meta-analysis published now about 20 years ago. Antibiotics don't prevent meningitis in basilar skull fracture and they don't prevent antibiotics in basilar skull fracture even when there is a cerebrospinal fluid leak. In this particular circumstance, although the 95% confidence interval is a bit wide, the same is true for pediatric patients. An interest of ours for many years has been severe acute pancreatitis. We're seeing a bit less of this, probably because in the laparoscopic cholecystectomy era a lot of patients are getting their gallbladders out before they become too symptomatic. In our patient population, the severe pancreatitis is divided almost equally between alcohol and gallstone-related pancreatitis. But there's a large and very low-quality literature that goes back 25 years that suggested that antibiotic prophylaxis for pancreatitis should be given. Turns out the answer is no. This was our study, 100 patients randomized. We confirmed that they had severe pancreatitis, as defined by CT evidence of necrosis of at least 30% of the gland. And we randomized them to muropenem versus placebo with the primary endpoints that you see. No differences whatsoever. And we looked at meta-analysis of antibiotic prophylaxis for severe acute pancreatitis before and after our paper. Before the paper was published, the odds ratio was 0.78, with the confidence interval being from 0.44 to 1.39, reasonably robust inference that antibiotic prophylaxis is not necessary for severe pancreatitis. With our paper added in, the odds ratio went up even, went up, and the 95% confidence interval narrowed, strengthening the conclusion that antibiotic prophylaxis is unwarranted. Now fortunately, I see fewer and fewer patients with severe pancreatitis recommended for antibiotic prophylaxis. But really, in this day and age, the number probably should be 0. If there's an important message from me in all of this today, is that we tend to conflate fever and infection. Remember that fever is the host response. Remember also that fever is actually an adaptive host response. So the febrile patient, unless you have reason to believe that there is an infection, probably should not be started on antibiotics. And antibiotics are not antipyretics. They do have important sedative effects, however, but we're not going to talk about that today. It's a joke, but how often do you hear or even think in your own mind, phew, patients on antibiotics now, everything's going to be OK? Well, as we've seen and are about to see more, that may not be the case. In revision now are the SCCM fever guidelines, also a multi-society undertaking. It's been a long time. I can't tell you exactly when the third iteration of the guidelines will see the light of day. So we're still using the 2008 document, which was robust. And I want to point out that basically fever in the first 72 hours after an operation is either due to non-infectious causes or the resolution of the pro-inflammatory response that was the nidus of the reason that the patient needed operation in the first place. So we don't do chest X-rays in the first 72 hours post-operatively if fever is the only indication. And you don't need to do a urinalysis in culture. After 72 hours, or if the patient has been in the hospital for a period of time and is getting repetitive operations, this rule no longer applies. But basically, post-operative fever, particularly if there was no infection present at the time of the index operation, does not need the same kind of workup. Now I want to tell you about a great study that was published in Science in the 1950s. An investigator took Gila monsters, who are, of course, poikilothermic, and needed to find a good rock and a couple hours of sunshine to maintain their core body temperature, and infected half of them with Eremonis hydrophilia. And then randomized them to whether or not they could get to a place where they could sum themselves or not. And the mortality was lower, significantly lower, in the Gila monsters that were able to sum themselves. So the first implication that fever is actually a salutary host defense response. We shouldn't be reflexively suppressing it unless the patient has cardiovascular toxicity therefrom. We know that neutrophil function is enhanced, at least up to about 39 degrees. And we know that bacterial growth and culture is suppressed. So put that acetaminophen, please, back in your pocket, because it doesn't help patients. And in fact, a study that I'm not going to show you in surgical patients suggested that mortality is higher when fever is suppressed in the post-operative period. This classic paper now from a quarter of a century ago in JAMA looked at about 3,000 patients in the ICU to attempt to determine whether or not their fever was related to a source of infection. And in fact, the majority of patients with febrile in an intensive care unit, after appropriate workup, were found to not have a source of infection. So here's another take-home point. More than 50% of fever episodes in ICUs are non-infectious in origin. And thus, empiric treatment of the febrile patient may lead to overtreatment. And of course, we've all heard of the deleterious consequences of overtreatment with antibiotics. The list of afflictions that can result in fever that is non-infectious in origin is very long and highly apropos of the critical care patient population. Basically, anything that causes the death of tissue or a hematoma can be associated with fever. Please bear this in mind. We could create a more exhaustive list, but nonetheless, this makes for a pretty busy slide covering only the highlights. There has been interest in Staphylococcus aureus bacteremia in particular for a long time. Staphylococcus aureus, because of cell surface adhesions, has a particular propensity to adhere to soft tissues and cause metastatic infections. Some estimates of patients with Staphylococcus aureus bacteremia suggest that a third to a half of patients will develop a metastatic infection. The heart valve is especially susceptible. And because of that, some infectious disease experts recommend that all Staphylococcus aureus bacteremias can be treated with four to six weeks of antibiotics. But as the data are beginning to emerge, it looks as though you can consider a shorter duration of therapy. We have not gone to seven days for Staph aureus. Staph aureus bacteremia remains dangerous. But in some patients, you may be able to go to 14 days of therapy if they're not diabetic, not neutropenic. If the source, such as an infected central line, is removed, there's no endocarditis and fever and bacteremia resolved less than 72 hours after initiation of therapy. There's some evidence in the literature to suggest now that if your follow-up blood culture after initiation of antibiotics is negative, you may not even have to do that echocardiogram. I'm not up here to recommend that to you, but that literature is available. So we looked at 800 consecutive fever episodes in 600 patients and found that, indeed, 50% of our fever episodes were non-infectious in origin. We used the same standardized six-test evaluation protocol, and we were able to ascribe a source of fever to 94% of the patients. But having the courage of your convictions to not start antibiotics immediately until you have your cultures come back in patients who are hemodynamically stable may allow you to make a substantial dent in the antibiotic tonnage that you prescribe in your intensive care units. Indeed, even 15 years ago in our unit, when we were treating bona fide infections, our mean duration of therapy was just under six days. You're much better off hitting them hard using big doses of antibiotics initially, provided organ dysfunction and PK permit that. The first 24 hours of antibiotics, in my opinion, is far more important than days 9 through 14 of a 14-day course. Dr. Klompas showed you the STOP-IT trial. This is a randomized prospective trial of intra-abdominal infection with adequate source control. Essentially, what was important about this is intra-abdominal infection studies tend to be larded heavily with patients with appendicitis because it's commonplace, and the patients aren't too sick, and it's easier to get them to consent and some of that. Some of the antibiotic registration trials for intra-abdominal infection contain 50% or more of cases of appendicitis, which is not really informative for making inferences about giving antibiotics to sick patients. So this did not have a lot of antibiotics in it, and as was mentioned, four days versus eight. The composite endpoint of death and surgical site infection or recurrent intra-abdominal infection was met by 22%, 22% versus 22% in the two groups. Not much wiggle room for interpretation there, with an absolute difference of 0.5%. When looking at subgroups as to whether the protocol was followed, they were sick based on Apache score, they had a hospital-acquired intra-abdominal infection as opposed to a community-acquired intra-abdominal infection. They were treated with percutaneous drainage, surgical drainage, had appendicitis or not. No differences amongst the subgroups. And taking that one step farther, several post hoc studies were performed on this data set because people were asking the questions, well, okay, what if they had sepsis? What if they were obese? What if they were older? What if enterococcus was administered? Did the administration of vancomycin do any good? Now, all of these are retrospective and post hoc, I grant you. But none of these substats were shown to benefit from longer courses of therapy in complicated intra-abdominal infection. So can you wait until you know what you're treating, particularly if the patient has stable hemodynamics and organ dysfunction? Dr. Klompas showed you this paper as well, too, from the same group in Charlottesville. Mervyn Singer, a very prominent intensivist from London, doing a year in review of critical care in 2013, called this the single most important paper published in critical care in the year 2012. It was a quasi-experimental before and after study design. And you can see that the results were that in surgical patients, at least, it is safe if they have stable hemodynamics to wait and figure out what you're treating. Remember, half of those fever episodes are going to be non-infectious in origin. And whether stratified by the degree of hypotension, the time from blood culture to initiation of treatment, and all of that, no difference. And the death rate, again, lower with the withholding of antimicrobials. Here's the multivariable analysis of mortality. Actually, an aggressive approach in this study to treating with antibiotics immediately was an independent risk factor for mortality. I'll thank you very much for your attention, and our last speaker, Dr. Dagger, will be coming up. Thank you. Thank you. Well, it's a pleasure to be here, and thank you for hanging out this last bits of our Congress. And can I get my slides loaded up? So I'm going to talk to you from not only the clinician's perspective, and I have over 30 years of experience of treating infections in the ICU, but I'm a sepsis survivor times three times. So I'm going to share a little bit of the perspective of being the patient working with sepsis, as well as some of my experiences in monitoring and managing patients requiring antibiotic therapy. So in my case, being 56 at the time, I developed some nausea and chills while I was at the hospital. I thought I had the stomach flu, so I kind of got in my car and did the 40-mile commute home because I didn't want to be in the hospital where those hospital things invited to go in, but mostly I just thought I had the stomach flu. About halfway home, I'm driving, and here's what I'm doing. I am shaking so hard that I could barely get myself to drive, and I got home and immediately had my family take me to the hospital where I threw up and didn't have a great experience, was admitted into the ED, and I got virtually no physical exam. And in that process, my antibiotics were started about three hours later from being in the ED. My procalcitonin was 6, greater than 6, and my temperature was 104, and I had a pretty elevated white count. I don't remember the number at the time, but they started me on vancomycin and did my blood cultures. It turned out to be streptagylactate, but it was kind of an interesting experience because they ordered procalcitonins every eight hours. So in the course of one day in the hospital, I had about 16 different pokes in my elbows. On day two of the hospitalization, they no longer could draw from my sights because there was just too much trauma to my elbows or whatever that they lost access. So then nurses doing all the ice things, ultrasounds, anywhere to try to draw blood in me, because it's problematic because there's so many tests ordered, and they were not coordinated. And also in my vancomycin levels, interesting, the dose before they drew my level, they stopped it halfway because they sent me to my MR. And technically, with all the scans I've had now, I should be glowing up here. But it's like they stopped it halfway through, and then they finished it, and then they drew my trough level after they had a four-hour break between the infusion of the vancomycin. And so then luckily, the levels were low, and they actually increased my dose, but it was quite an experience. And so I ended up on a vanco dose of 2 grams Q8, and then when I was transferred to a higher level facility, the pharmacist promptly said, no, we'll start him back at 1 gram Q12 because they didn't have any of my level results. So it was quite an experience going through this. And even when they wanted to put prophylaxis on my leg, well, my mechanical prophylaxis and I had cellulitis, they wanted to put the squeezers on my legs and everything. I was arguing, no, don't do it. I said, well, the doc ordered to do it this way. We have to do it. And I had virtually had no physical exam, and they had never seen this raging cellulitis on my leg, which I'll give you a shot of what it looks like shortly. But this is all the things I was going through in the process. And there's even more. I got neutropenic. I had a wipe count of less than 1, and no one had caught it. So I ended up back up in the ED with rigors and fevers again. They measured it, my wipe count less than 1. And nobody did anything. They just sent me home. So I was also in ADE as well as in ADR because no one actually looked at my results. So I had quite an interesting ride in this experience. And so you can see, this is what my leg looked like on the second day. And my chest CT, all these different things done for me. And actually, because they put me on an IV, my IV was running at 150 cc's an hour for almost 20 hours. And I had trouble breathing. I was struggling. And no one would do anything until the doctor could come and sit me. And believe me, I don't know if any of you have experienced pulmonary edema, but it is not fun. What was even more interesting is that I'd been put in a step-down unit at the ICU at this time. And during that process, the family, there was like 10 family members in the room. I mean, I'm looking at them from my bed, and it's like, I got to pee, and I can't do it. So I just had to get right up in front of everybody. And you know those canisters? Well, I filled like five of them in two hours. And so it was just like an interesting experience. And also, since my roommate was on viral precautions, but he's sweating. So what does a nurse do? And nothing against the nurses. I appreciate it. My mom was an overseas flight nurse during World War II, so I have the utmost respect for nurses. But the nurse takes a fan and puts it on the patient to cool him down. And he says, don't worry, sir. You're a little bit more than three feet away, yet I can feel the fan. So up goes the mask and such. So these are all the things I experienced in my stay. But things can happen, and there's ways we can avoid it. But I think one of the most important things that frustrated me is my antibiotics were delayed, and they were timid. I'm not small. And I just go, gee, and I try to argue the importance of this. So this is what I'm going to share with you today, is how you look at antibiotic dosing in the ICU and how it's structured. Because the one survey showed, looking at ICU clinicians, that most didn't understand the pharmacokinetics. So this is where I thought I would focus my time with you today, is kind of go over some of the innuendos and intricate things and thoughts with pharmacokinetics. Because just as Dr. Baer mentioned, the importance of a good dose right up front. So I'm going to show a little bit of how you might want to approach that. So it's always important when you look at pharmacodynamics, first of all, is you're looking at the MICs. You're looking at the Cmax to MIC ratio for certain drugs. You're going to look at the duration over the MIC. And then also you're going to look at the area under the curve to the MIC. And it depends on which one of these values you're going to put in the perspective of how you dose these drugs. So these are some of the things you want to keep in mind when you're making decisions on the kinetics, is also how the drug works as far as the pharmacodynamic effects. And also the post-antibiotic effect, which can change. So one of the issues, if you're dealing with patients on hemodialysis, is for genomycin or aminoglycides, we like this big old peak and get a big high concentration. But we can't do that in the dialysis populations. And it really is an issue where you actually now have to get a post-antibiotic effect that's shorter. So you have to end up with peaks of maybe 8 to 10. But you have to keep your trough levels higher. The old adage of getting the trough less than 2 in these patients' dialysis is out the door. And now we keep it around 3 to 4 because the dynamics have changed. We have to change our kinetics. So when you start antibiotics in the ICU, it's important to look at what's been discussed. You're looking at the indications. What are you treating? And the initial dose is really, to me, the most important thing. In all my years in the ICU, I learned early that patients were dying. Back in the day, we treated pneumonias with dual therapies with a drug like pipericillin. We didn't have PIP taste, though. We had pipericillin. We added an aminoglycoside, and that's how we treated these patients. But we grossly underdosed patients because of fluid status and such, and we had to catch up. And so I learned early in my career the importance of really hammering them because I rarely see toxicities in the first couple days in most cases, and don't be afraid of that. We use intermittent, extended, and continuous infusion strategies, and you have to look at the organ function. Where is it going, and which direction? Getting better, getting worse? And what's the fluid status of the patient? Because when they come in and they're 70 kilograms, and today they're 85 kilograms, that's not fat. That's fluid. So account for that fluid in your weight assessment, how their hemodynamics are. Keep in mind that that's a measure to see how bad the infection is. The immune status of your patient, because the antibiotics work with your immune system, and the location of infection, how aggressive you have to be. And then I already think when I'm in the first day, what is my de-escalation plan going to be? Now, I also want to show a little bit when you do kinetic monitoring. This is one of the first studies I did. It's old. I only presented an abstract, but this is HPLC data of vancomycin in patients in renal failure, whether you're a 10-day-old, a 13-year-old, or a 70-year-old. And there's studies that show this in graphs. But in reality, I never draw a vancomycin level in the first 12 to 20 hours in a patient after a load in dialysis, because they're still distributing the drug, and you cannot do an elimination rate. You might use it in this time period to see if you gave enough drug, but not for calculating out elimination. And that is a helpful tool when you don't know whether they're fully loaded. Let's check a level. Oh, the level's 6, and you're septic. You know, then you do need to get a load, and how much you need to give is kind of drive on the fluid status of the patient. And we frequently make that oversight. And my colleagues will frequently give everybody a gram of vanco for the first dose. In some patients over 100 kilos with a lot of fluid, they should be getting more like 2 grams. It's usually 20 to 25 milligrams per kilo for these patients with their initial dose, especially if you're not sure where you're going in the next 24 hours. And that is aggressive initiation of vancomycin. So when you start these antibiotics, also keep in mind that there's different states. When we walk through some of these, massive trauma, big fluid resuscitation. You may have to give higher loading doses to these patients. And if you're afraid of giving more than 2 grams at a time, you can also give the next dose earlier to get an adequate load in these patients. Open surgical wounds, what was discussed, should you or should you not, but you have to keep in mind. Is this a patient that even needs antibiotic therapy in the first place? I found that burn patients were unique to be burned, hyperdynamic and such, probably for the first 10 days. And after 10 days, they become more like your normal ICU patient. But there's going to be a lot of fluid issues. And if they have a lot of still oozing, there's going to be antibiotic ooze through the eschar wounds. And also hardware. We always kind of worry about is there hardware in place, because the last thing we want to do is have infected hardware. And so that sometimes drives how we're going to do our antibiotic therapy. Organ system failure we're going to discuss in a moment, and also some of the mechanical support that we provide. But one of the things when you look at organ dysfunction is actually there's very little data on heart failure. But think about it. Heart drives blood to the kidney or the liver. So if your heart's failing, your flow of blood with your drug to be eliminated is also going down. So cardiac dynamics has a lot to do with the ability to eliminate these drugs. And also you have separately injury to those perspective organs that the body may be eliminating the drug via. So you have to think of heart failure also as something that affects drug elimination. And you have to look at how bad the dysfunction is and what direction is it going. Is it getting better or is it getting worse? Is it stable? And then you can kind of, OK, we're kind of cruising at whatever level we're at, but there might be a change somewhere in the future. And then what's the impact of mechanical devices? Whether they're on CRT, whether they're on ECMO, or they have an L-band or such circulatory support that also can help drive our decisions. But I think the most important thing is don't delay and give a loading dose. I think from my panel, unless they argue, I think what they're saying is don't wait if you think you have infection. Get something on, but get it on adequately. Get your patient to an adequate level, especially if their immune system's shot, and you really have to make sure that the antimicrobial therapy is bridging that gap. So when you're monitoring, obviously, this has kind of been discussed, so I'm not going to spend a lot of time on it, but temperature. But temperature, as was mentioned, is not necessarily the best indicator. And also, if you have a patient on ECMO, the temperature is set by the circuit. Same with CRT. So these circuits are warming or cooling the blood, so it really becomes useless in those situations. Procalcitonins, well, I do have, and this was kind of mentioned, I do find in practice that when I can't get my team to stop the antibiotics, I use a procalcitonin. And to me, if it's less than 1, in other places, if it's less than 0.5, it's a good argument to say, you're OK, you can stop it. They just need something to give them that last assurance. But otherwise, I'm not a big proponent of constant procalcitonins, especially Q8. And I don't even think there's a value with doing it every 24 hours, unless you're trying to make decisions on these patients. But every so-and-so till you can decide on, really, can I not only de-escalate, or can I stop the antibiotic therapy altogether? Now, when I'm estimating drug dosing and renal insufficiency, I first keep in mind that the same creatinine is always yesterday's renal function. And I look at urine outputs, or also all these other parts, to assess how their organ function is, because urine output today tells me more. I'll talk to the nurse and say, well, what's the urine output in the last four hours? Well, it's actually gone down to 0. So I'm actually being time sensitive in making my decisions today, instead of waiting for the creatinine to rise tomorrow. Also keep in mind the importance of dosing these medications based on the evidence. And in 2000 or so, we changed the way we calibrate the creatinine. So the creatinine assays are now down about 0.2 to 0.3. So if you're looking at literature based on a creatinine of 0.8 in 1995, and you look at one today, the creatinines are going to be lower today than they were at that time. So you have to think that you're going to get different creatinine clearance estimates based on today's data, which could give you a more robust, higher creatinine clearance that may have been experienced in the trials or the studies that looked at dosing these drugs in renal insufficiency. And also, if it's moving, you can't use that hover over if you have these systems hover over and gives you the creatinine clearance. If that creatinine clearance is changing, that value is not going to be helpful to you. It might tell you where you're going or what the limit is or bookend, but it's not telling you truly what it is. In dialysis, I'd laugh. People don't go to the bed, and they don't see. And my colleagues would go, oh, oh, look, the creatinine is 1.5, and we can be super aggressive on the antibiotics. But the patient was getting dialyzed. And so dialysis was putting off the creatinine, again, creating a mismeasure. And when you're dealing with dialysis, there are all sorts of ways we dialyze these patients. And you have to deal with the effluent rates as well as the blood flow rates and the durations of these dialysis to get a better handle on how drugs are eliminated. So when you look at elimination, here's the thing. So I did a study in aminoglycosides and intermittent hemodialysis, and I'll tell you that creatinine clearance or drug elimination in hemodialysis patients with AKI is very different than CKD. Aminoglycoside measurements are a better depictor of renal elimination of drug than the same creatinine is. And what I found in those patients with AKI is that the elimination is about 30% to 35% faster in AKI than it is in CKD. Yet we pull the CKD data to dose them, and we're actually going to get a dose that's going to be more conservative. In reality, with AKI, they need to be more aggressive, because they are removing these drugs faster. And we have no data on drug elimination, for the most part, in a setting of AKI. So be cautious of that, especially when someone looks at a dosing chart and says, well, it says this. If they're in AKI, you have a good argument to be more aggressive with the dosing, especially initially so you don't underdose these patients. So when I look at patients who have changing organ function, I kind of look at what is going on with the patient dynamically to assess how I'm going to adjust my therapy. If they go into renal failure, I'm looking at that acute phase where they're actually dropping. So I have to think about my dosing up front while they're changing their system function and either adjust or make a change in the dose. And then they have this phase of a plateau phase, and then maybe you have the recovery phase. And when they start recovering, you've got to be on top of it. This is where looking at urine output and such helps me determine, hey, they're getting better. Their infection is getting better. They're recovering. I can now start increasing the dose, and I'm not chasing my tail. Again, I think good antibiotic management up front and for shorter durations of time is probably the best way to go. But you have to be on top of it during the phase in which you want the antibiotics to work for you. So you have to think of this dynamic change and not just put them on a dose. We're set, and we'll go for the duration. You have to adjust accordingly. So in renal replacement therapy, you have to think about how the circuit removes it as well as the patient, noting that I just mentioned that if they're an AKI, they're going to remove it faster. So this is a paper we did. We looked at different dosing strategies. This is actually what feeds into LexiComp as far as how you do your dosing in CRT. And depending on the various modes of CRT, you have different dosings because the dialyzer circuits will change that effect. Keep in mind, biofilters are on the horizon. And when the biofilters come out, this is all going to change because we don't know the nature of the filter and how it eliminates. So you have to track the technology with the doses. So you might start with the table, but you have to adjust for what the patient needs. So it's important to keep that in mind. So I like to say you start with the table. And I usually don't like to go to the table until day two or three. I dose most drugs at their normal dose, except for vancomycin and aminoglycide, for the first 24 to 48 hours. And then I go to the table and adjust when I know more about what I'm doing with the patient. Going to the table initially, well, it's fraught with the risk of underdosing your patient. So here's also the issue with all the data you look at. And you look at all the literature and such. Almost every pharmacokinetic study in renal failure is severely flawed. And here's the issue. The FDA requires a single dose. With a single dose, you actually measure these parameters. And so you look at the elimination. But it's all about elimination out of the plasma. But the problem is that distribution of tissue is depicted as elimination of plasma. But it's not. It's distributing. So hence, elimination rates are actually reported as much faster than they truly are in patients because they do not do it after subsequent doses. And studies that have shown this for some of the anticoagulants, like dabigatran, apixubam, and noxapiram, have showed that actually the elimination rates, or the AUCs, are two to threefold higher. When you look at it two or three days later. And so this is one of the things you have to think about when you're looking at this data, is that data is kind of crappy in the first place. So you have to consider that it's not all that great. And I encourage you to keep looking at what your patient's doing and responding. And as you get to be more aggressive with your renal replacement therapy, also looking at when they stop it, because it's important with your antibiotics, if you're doing aggressive management during the CRT and then they stop it, you have to then say, OK, do I need to adjust my antibiotics? They stop it. OK, we adjust. Oh, no, we're going to start it again. Because sometimes the nephrologists are driving when we start it or when we stop it. You have to stay on top of this to know if they restart it, you have to get your antibiotics back on board to a higher dose. So we have to watch the renal replacement therapy and coordinate it with your antimicrobial dosing. And so when you consider, sometimes you have to consider also, how do we sample it? Are we doing it pre or post circuit? When was it on? When was it off? Were they dialyzing for the day? Did they change the blood flow or the ultrafiltration rate during the process? What's the residual renal function? Because when that's starting to pick up, especially in AKI, you know that their kidneys are starting to recover. And then also, you have to look at, actually, the interesting thing is this kidney, in full failure, sends a signal to the liver and decreases cytochrome P450 metabolism. So for some of the antibiotics, some of the cephalosporins, this has been shown to be a 30% or so reduction in the clearance of these drugs. For certain ones with certain enzyme systems. And also, what kind of filter are you using in the circuit? And are you adapting for that? If it's a high-flux filter, larger molecules can transition across. So how... what size does that filter do is one of the other things. For ECLS, I'm going to be very brief. These circuits, they're constantly changing. The technology is changing. There's sequestering in the tubing. So for drugs that are highly lipophilic, highly protein-bound types of drugs, they can sequester onto the circuit. So you get drug picked up by the circuit. But then also, when you stop your antibiotics, there still could be some drug leaching from the circuit in that process. But you have to look at the volume that you do with the circuit. You have to look at, also, organ dysfunction. And also, sometimes you can ultrafiltrate off the circuit. So you have to keep in mind that there are other parameters that could be driving things in a setting of ECLS. And then continuous infusions. So this is one little trick. So a lot... You know, some people use, like, vancomycin continuous infusion. And I do it. I'm comfortable doing it in the setting of renal failure or organ failure because I actually do a process where I iterate the clearance. And that way, I know the direction that the clearance is going. And I can be a day ahead of everybody by seeing where I'm going and adjusting the rate accordingly. So those of you that do continuous infusion vancomycins, by iterating the clearance, look at the change in the dynamic change in clearance. You can be ahead of the game instead of reacting with each level that you get. Again, a tool to be aggressive up front with your antimicrobial therapy. And hopefully getting to your goals and stopping it earlier. So for conclusion, some of the things that I found. So obviously, I mentioned the blood draws, the levels, my white count being low on ceftriaxone, which actually landed me back in the hospital. I found for the... I chuckled with the PANAS guideline folks and everything. They said, the most important thing to me starting was earplugs because I got sleep. The face mask. I could sleep. I found without those, for three days, I was starting to get loopy. Things weren't making a lot of sense, and I was starting to get that sleepless effect. So don't forget these measures to me. I just can't over tell you how important that was as a patient. The pulmonary edema, I mentioned that. And trying to be a good neighbor, but also it's important that... Why did I have 12 people in my room 24-7, and I couldn't sleep? I couldn't even get to the bathroom. These are things that we have to think about with good care and stuff for our patients. And I could go on about other things, but I just wanted to say that there's also the small stuff you have to pay attention to as well. And I also want to say importance. I can't over tell you the importance of thinking about the kinetics of your patient and getting adequate dosing up front, and hopefully have a plan to de-escalate and get them off the antibiotics sooner than later. So thank you for your time. I guess we'll call off the panel, and for any questions, there's microphones here. So if anybody would like to ask questions to any of us, please do so. Start with the gentleman over here. Thank you for outstanding talks. My question is, you know, in the era of the CYP1 that mandates starting early antibiotics, how do you deal with that? That you have to, patient suspected CYP says, you have to give antibiotics within three hours. Well, you know, I think that I don't worry about the metabolism, the CYP impact initially. The importance is loading them and getting them started. And then after you get that initial load, you have to follow up. And one of my admissions, they gave me my first dose of antibiotics and ED, and then he discontinued them. And 12 hours later, it's back in rigors. And the thing is, there was no oversight to me. I kind of got lost in the packet, and then I had to start over. So really, importance, again, is starting the therapy and then sorting out the other components. But don't be timid. I think metabolism issues, you're not going to worry about for the first 24 to 48 hours. I don't know if Melissa agrees with that, but I wouldn't, where then down the line, you have to look, okay, what's going on? What do I need to adjust according, if I worry about toxicities? And a good example is cefepime. We worry about neurotoxicities, but also stop the compazine and things that lower the seizure threshold. So there's a lot of different things to think about. Gentleman over here. Question about duration of antibiotics for pseudomonas, seven days or, you know, there's prior studies have shown benefit for longer courses of treatment. Sure. Thank you. I think the best evidence we have for this comes from VAP. I think what you're referring to is that that study I showed comparing eight versus 15 days showed them the subset of patients with pseudomonas was actually the patients who had non-fermenting gram-negative rods, mostly pseudomonas. There was a higher rate of recurrent pulmonary infections in that group. However, if you dig below the surface of that study, what you'll find is that their definition of recurrent pneumonia was basically a microbiological definition. You had to have a repeat positive BAL, and the study protocol mandated repeating the BAL if there was any kind of change in the clinical condition of that patient, even things that might not be pneumonia-related. And so we know it's very difficult to sterilize a patient from pseudomonas, particularly if they have an endotracheal tube. So microbiological persistence alone is not necessarily true invasive pneumonia. What you have to do is you have to look at the outcomes in the patients who had non-fermenting gram-negative rods also reported inside that paper. And the subset of individuals who had non-fermenting gram-negative rods, their outcomes were identical where they got eight days or 15 days in terms of ICU clinical stay, mortality, time to organ failure recovery, et cetera. So both the European and the American HAPVAP guidelines now say seven days for all organisms, again, presuming it's an uncomplicated infection that's improving. No difference for pseudomonas. Any other questions? Hi. Yes. Yeah. I work in a big burn unit. And, you know, we take all these patients with extensive burns. And they're usually febrile, you know, from the serous reaction, their procalcitonin is up there. And we're always in the dilemma of if we start the antibiotic, we don't really know when to stop it. Because, I mean, they have lost their skin barrier. And they're going in and out of theater for deprivement and grafting and stuff. Do you have any advice about that? Yeah. I would say culture aggressively and trust your cultures. If your cultures come back negative, you have to have the courage of your convictions to stop. You know, burn patients, particularly major burns, are up there. I mean, it's a pro-inflammatory swamp. So clinical decision making comes to the fore. We have a big burn unit at Cornell, too, 850 burn admissions a year. And I think that we ought to treat burn patients the same. You can't rely on procalcitonin. And you can't necessarily rely on white count. And all of this literature, to me, about using biomarkers to determine duration of therapy essentially becomes moot. If you take an approach that says, I'm going to set a date certain for stopping antibiotics at the moment I start and stick to it. You know, I'm going to treat this patient with peritonitis from perforated diverticulitis for five days and stop no matter how the patient looks. If the patient doesn't look good, you're going to get a lot more information from reculturing and regrouping and making a decision to restart antibiotics than you are to start piling more and more antibiotics on into a 14 or 21 day course of therapy. That's how we get resistance. Can I just mention one thing on that issue? I think one of the things that I found that people don't do enough of is when they have uncertainty and they've been using antibiotics, just stop. Just stop and see what happens. And what I end up doing, if I'm not certain I should continue at all, I just stop. And if the patient needs to be on antibiotics, they'll tell me and we're watching them like a hawk so they're not going to have a significant deterioration. And the fact of the matter is, in a lot of those cases, at least in a reasonable subset of them, when you stop and if the infection does come back, most times it doesn't or there's no infection, but if it does, at least at that point you can culture it and you know what you're dealing with. So you shouldn't have a hesitation about doing that. And another way to look at this is the way that I manage my practice when I'm on ICU runs. I am trying to stop every antibiotic every day in every patient and somebody needs to talk me into continuing. I like it. Question here. I'm a pharmacist in a mixed ICU and kind of on the flip side of this coin, at the 48-hour mark where I'm trying to de-escalate and narrow our therapies, I get a, well, they're improving on vancomycin and zosyn, why would we narrow now? And what's some ammunition to combat that and keep providing good patient care? Do you want to take this one? If they're on vancomycin and piperacillin and tazobactam, that's a bad combo for renal dysfunctions. That's one good reason to stop it right away, but I think you need to recourse to the foundational principles I think we've all discussed, which is to establish what actually is the diagnosis, what's the causative pathogen, what's susceptibility, and there's no data that broadening the therapy or double-treat covering the therapy in an improving patient leads to any kind of better outcome. You know, the last set of the surviving sepsis guidelines, we try to address that to an extent. As you know, about a third of the time you don't get an organism isolated, and at least according to our data, at least for septic shock, you're still responding to antibiotics. What we've said is that it's not just based on isolating a pathogen that you could narrow and de-escalate. Even if you don't have a pathogen, if your patient is clinically improving, you can de-escalate at that point. The probability of running into it, like, I'll tell you what I typically do. If I've got a patient in septic shock, for example, and I don't get a pathogen isolated, if they're coming off pressors in 48 hours, I narrow it back. If I'm on pip-tazo, I go to cefetaxime. If a couple of days later they're coming off the ventilator, I'll, you know, if it's five days, I'll go to cefuroxime or something, or I'll just stop the antibiotic at that point. So in those one-third of cases, it's perfectly legitimate to stop just on the basis of clinical response. And I look at de-escalation in the perspective of two things. One is removing antibiotics that aren't necessary. So do I need everything here? What are the MRSA nerves or other things driving my decisions on the vancomycin and such? But then also for my organ-impaired patients, I may de-escalate because of concerns for potential toxicities down the line. So those are other drivers for me to think about when I'm doing the concept of de-escalation. I'd like to ask Dr. Klomp his question, not to put him on the spot, but I think it's something that we need to talk about. And that is double coverage of pseudomonas. Everyone wants to make sure that the waterfront is completely covered, yet certainly when source control is adequate, it's not clear at all that you need to cover every isolated pathogen. But setting surgical patients aside, the meta-analyses, at least to my interpretation, suggest you get no mortality benefit, but you get more acute kidney injury. How do you formulate all of that? So I agree with your conclusion. There were some early meta-analyses that said that maybe there was a survival benefit in pseudomonal bacteremia with double coverage compared to monotherapy. Subsequent analyses with more modern data, more additional data, more modern regimens have not borne that out at all. The one point of – that would be for the scenario where she knows what the pathogen knows susceptibilities is, there is still a rationale for the critically ill patient in septic shock where you don't know what the diagnosis is. You don't know what the pathogen is or the susceptibilities where double coverage can be indicated to make sure you have at least one active agent on board. But as soon as you know what the pathogen is, and this is true for pseudomonas as well, you know the susceptibility profile, there is no benefit, in my opinion, to continuing double coverage. And in fact, there's increased risk of harm in doing so. Do you believe that there is any value to a double coverage regimen that does not include an aminoglycoside? And can I add to that question is a lot of times what we see is that first day they give a single dose. I don't know what you see, Bliss, but we give a single dose of aminoglycoside, so we're really worried about septic shock, and then try to sort it out, but they're hesitant, at least in practice, to just completely go to single coverage for at least the first day. I don't know whether you're answering that. There is actually data on this. There's a quasi-experimental study out of the Netherlands comparing two hospitals that had two different regimens for empiric treatment of sepsis. In one hospital, they did recommend empiric aminoglycoside. The other hospital, they did not. And the outcome that was reported that even though in the aminoglycoside hospital, the median duration of therapy was one to two days, that was associated with a higher rate of acute kidney injury and a higher rate of the combined outcome of acute kidney injury and mortality. So on a routine basis, it would not be the preferred approach. The complicating factor is that if you practice in a setting with a very high rate of antimicrobial resistance, and it might be that once you have a beta-lactam on board, that quinolone is not going to add much if most of the isolates in your hospital are resistant to quinolone. So you might be forced to go to something like an aminoglycoside just to assure you have adequate coverage. That's a very exceptional circumstance, particularly in the United States, where we have to have a hospital with such a high degree of antibiotic resistance. The other factor, of course, is taking into account what you know about the patient. If you know the patient's colonized with a resistant bug that's resistant to everything except for aminoglycoside, you might be forced, your hand might be forced. People should recall that the patient in septic shock can have renal hyperperfusion from multiple mechanisms. The first is blood pressure. But the second is that the pro-inflammatory milieu, particularly if prostanoid metabolism is upregulated, can cause intraparenchymal vasoconstriction. There's even some data to suggest that aminoglycosides cause efferent arteriolar vasoconstriction. So even a single dose of aminoglycoside, as pointed out by Dr. Klompas, is potentially hazardous to the kidney in a patient in shock. Any other questions? Hi. I'm also a pharmacist at a mixed ICU. So for pro-calcitonin, understanding that it can be affected by tissue damage, and also a lot of the data lies within or looking at the duration of therapy or de-escalation, have you started seeing practices where clinicians want to use it to initiate therapy, and what is your response to that? Do you want to take it? You know, frankly, I defer more to people who actually have used it. But what I can tell you from my analysis of the literature is that, basically, my sense is the less severe the presentation, the more utility for initiation there is. The thing is, for example, in septic shock, or at least if I clinically think it's septic shock, I don't really care what the pro-calcitonin says right off the bat. And that's what the guidelines from Surviving Sepsis say. Basically, it's probably not that useful in critically ill patients where you've got to make a clinical judgment. On the other hand, in less ill patients where you're uncertain, it can be the deciding factor, and probably legitimately so. So can I answer that question? What about the concept of having that initial baseline pro-calcitonin to track trends? Would you think that's valuable or not, if you want to see which way it's going in your patient? Again, I'll just say that this is based on my analysis of the literature, is that if you are inclined to use it, then I think getting a baseline level, because if you want to see an 80% reduction, you need to get a baseline level. But again, that's going to be in situations where there's a certain degree of illness, basically, that's driving the process. So certainly, if I was using it in the context of septic shock, I'd get a baseline number. Or in somebody who's in the ICU, I'd get a baseline number. How useful that is in an outpatient, I don't think that's going to be that useful. And in somebody who's a hospitalized patient, again, I'm not sure, to be honest. I don't know if others have a different opinion. My only thought on that is if you have a policy in place that looks at that 80% reduction or level 0.5 for de-escalation, you have to have something to start from. But if that's part of your policies in your institution, that's where I think it's not to make a decision what to do today. It's really to make a decision on stopping the therapy. Other questions? Well, thank you for sticking out all these days in Congress and being in our session. We appreciate it. And we'll be here for any additional questions you have.

Procalcitonin

biomarker

antibiotic therapy

septic patients

bacterial infections

inflammatory markers

infection control

antibiotic duration

mortality

clinical judgment

cautious approach

individualized approach

elevated procalcitonin levels

interpreting results

antimicrobial stewardship