SCCM Resource Library-Early Antibiotic Administration in Suspected Sepsis: Slam Dunk or Shrouded in Controversy, With Case Discussion

Video Transcription

the Antibiotic Administration and Suspected Sepsis, Slam Dunk, or Shrouded in Controversy with Case Discussions. I am Rahul Manjal. I am an intensivist at the Medical College  of Wisconsin and the moderator for this session. We have two talks scheduled for this session. The first talk is going to be given by Dr. Anand Kumar, who I will introduce shortly,  and in the second talk are some case-based discussions that I will be giving myself.  After the two talks, we hope that we can have a robust discussion with some questions from the audience. So without further ado, let me introduce Dr. Kumar. Dr. Anand Kumar is  Professor of Medicine, Critical Care Medicine, Infectious Diseases, Medical Microbiology,  and Pharmacology Therapeutics at the University of Manitoba, and he is an attending physician  at the Health Science Center and St. Boniface Hospital in Manitoba, Canada. His talk today is entitled, Antibiotics Absolutely Every Time Lives Are At Stake. So welcome, Dr. Kumar,  and let's get started. Hi. This is Anand Kumar. It's virtually a pleasure to be here with you. I wish we were  all in the southern U.S. somewhere enjoying the sun and the margaritas, but I guess that's not going to be possible this year. So I'm happy that you can all attend virtually and  we can do this talk over the internet. So my subject is going to be early antibiotic administration and suspected sepsis. Is it a slam dunk or is it shrouded in controversy?  A couple of things that I have to address, of course, right off the bat, and that is disclosures. I have had grants and study support spoken for. I've been on advisory boards for  virtually every antimicrobial and rapid diagnostics company in existence. The good news is that  although I might be bought and sold, I am equally bought and sold by all of them. Therefore,  I don't feel that I have any specific allegiance to any of them. I'm going to speak about antibiotics  in general, antimicrobials in general. You won't hear me speak about any specific drug,  but I will speak about general concepts around it. Now, I'm going to try to help you understand  how I think about sepsis antimicrobial therapy, and that will help you understand the results  of some of the studies that I'm going to review with you. Now, many of you have probably seen this diagram. This was developed back in the 1990s by Roger Bone, who was a very famous  intensivist back then, and he developed this idea of SERS, the systemic inflammatory response  syndrome. Now, you've probably heard of that in the past. It's been superseded by more recent  sepsis definitions, but back at that time and up until fairly recently, the idea of SERS was that  it underpinned sepsis. SERS was basically a clinical syndrome defined by two of the following,  an increased heart rate, increased respiratory rate, an altered white count, up, down, or left  shifted, and altered temperature, up or down. Now, if you had two of those four in the context  of an appropriate clinical insult, then by definition, you had SERS. Now, if that appropriate  clinical insult was infection, then the overlap of the two defined sepsis. Severe sepsis was a  subgroup of sepsis wherein you had organ failure with sepsis, and septic shock was a subset of  that where you had basically cardiovascular failure. All fine and good. The problem with  this paradigm, I think, the one thing that I really don't like is that you don't get the idea  that infection is directly related to SERS. They seem like two separate phenomena, and if you think  about it, that's actually the way that historically intensivists have tended to think about sepsis and  septic shock. We've tended to think it's an immunologic process. We have an immunologic view. The idea is, yes, we have infection, but we give these great antimicrobials right away,  and the infection is gone. Nonetheless, we've triggered what we call the cascade. We always hear about cascades in the context of sepsis, inflammatory cascades, cytokine cascades,  eicosanoid cascades, coagulation cascades. It's all about cascades. The idea is that infection  triggers this cascade, and once you've started this cascade, even if you give antimicrobials,  you get rid of the infection right away. Nonetheless, you have SERS, you have a compensatory anti-inflammatory response syndrome, and the interaction of the two over time gives you  organ injury. Now, you know, the ID guys among you, and I'm sure there are some as I am,  would look at this and immediately say, well, that's not realistic. You can't give antimicrobials  in a situation of infection and then expect the infection to be gone just like that. It takes time.  And so, basically, microbiologists and infectious disease docs have a different way of looking at  the septic processes. They tend to think of sepsis beginning with a small nidus of infection within  which there is a microbial load, and the microbial load, if untreated, increases over time. So we're  talking about peritonitis, pyelonephritis, pneumonia, whatever. Microbial load is increasing  over time. Now, there's a toxic burden overlying that. So that's your endotoxins, exotoxins, even  bacterial nucleic acid has antigenic activity. Overlying that is your body's endogenous inflammatory response. Now we're talking about your cytokines, eicosanoids, and the whole  inflammatory cascade as it relates there. And then finally, there's what we clinicians see at the  bedside. We actually see evidence of cellular dysfunction and tissue injury, which basically  is organ injury. In its most ultimate manifestation, we're talking about septic shock. That's the most  dramatic manifestation of cellular dysfunction and tissue injury. Now, there's an issue with that,  too. And that is something that physiologists, I think, will tend to recognize. This is Carl  J. Wiggers. He is arguably the most famous cardiovascular physiologist of the 20th century, famous for a lot of things, not least of which is defining the cardiac cycle.  But I think he ought to be most famous for this idea. He actually was one of the first,  maybe the first, to have a really solid definition of shock as we understand it today. He said,  shock is a syndrome resulting from depression of many functions, but in which the reduction of the  effect of circulating volume and blood pressure are of basic importance and in which impairment  of circulation steadily progresses until it eventuates in a state of irreversible circulatory  failure. What's he talking about? Well, he's talking about the idea of irreversible shock, basically. Now, you know that irreversible shock has been well-defined in the context,  particularly of hemorrhage. In fact, what he did was he took dogs and he bled out their blood. In  fact, he was the one who came up with those four grades of hemorrhage, you know, the ATLS  four grades. Grade four hemorrhage is 40% of the volume. So what he did was he wanted to ask two  questions. He asked, how much blood do I need to take out of an animal to make it hypotensive?  And B, how long will the animal tolerate that hypotension before whatever, before it dies?  So what he found was that essentially 40% of the blood volume would give you a shock. And he found  something very interesting. If you resuscitated the animal, if you gave the blood back within an  hour, all of the animals lived. If you gave it back after about two hours, pretty much all of  them died. Now, the interesting thing was they didn't die right away. They, if you continue to  support them, they would die, you know, in 12 hours, 24 hours, two, three days. They'd all die,  but it took time. They had been at that point irreversibly shocked and it didn't matter what you did. Now you can actually broaden that concept to other things. Now, for example,  in traumatic hemorrhagic shock, which we've discussed, that definitely occurs. The interesting  thing about traumatic hemorrhagic shock, it's not enough just to give fluids and resuscitation.  You actually have to stop the site of the bleeding. That's the critical determinant of survival. You have to get to the underlying source of instability. For cardiogenic shock,  say due to MI, you need to thrombolyze or angioplasty the source of ischemia or you're going to die. It doesn't matter how fast you give heparin. It doesn't matter how fast you  give nitroglycerin or pressers or anything. The primary determinant of outcome is how fast do you  thrombolyze or angioplasty, that is how fast you address the underlying source of ischemia. Similarly with obstructive shock, primary determinant of outcome is simply how fast  do you thrombolyze or embolactomize the source of obstruction. In each case, you have to get  to the underlying source of the problem, that which is driving instability. The question comes,  well, what's driving the source of instability in septic shock? I would suggest that in fact, it's the microbial load. You can create a new paradigm of sepsis and septic shock  that actually combines all the elements of what we looked at before. It's very similar to the microbiologic model that ID people like. Again, you start with  anitis of infection. The microbial load is increasing over time if it's untreated. Overlying there is the toxic burden, the endogenous inflammatory response,  cellular dysfunction and tissue injury, basically organ dysfunction.  This is different in that I've added this line. This line is the shock threshold, basically.  Shock threshold is simply the point at which any individual is under so much inflammatory stress  that they cardiovascularly, hemodynamically decompensate. Essentially, you are septic, you progress, and when you can no longer tolerate the inflammatory stress,  your blood pressure drops. That for septic shock is when your golden hour begins, in my opinion.  Now, essentially, at some point thereafter, some point after you've exceeded the shock threshold, you'll be irreversibly shocked. You'll be irreversibly committed to death.  Now, it's not the same for every given individual. We all have different phenotypes and genotypes,  but the basic idea holds. Once you reach shock, you have a limited amount of time to intervene.  Now, this kind of model is useful for different things, the way I think about it, but one of the  most useful things about any kind of model is that it has implications, it has ideas.  There are two interesting implications of this model. First is it suggests that sepsis and septic shock are fundamentally two different diseases. That is to say, if you don't have  hypotension, then you have sepsis and that's one disease. If you have septic shock, that's another  disease where time becomes highly dependent in terms of your intervention. The other thing this  model suggests is that delays in antimicrobial therapy are just an indirect way of saying  that the microbial load has increased too much. Now, do we have some evidence to support that?  In fact, we do. If in fact, infection without sepsis, sepsis and septic shock, and these are  the criteria here, sepsis in the old terminologies, this would have been considered sepsis. This would  have been considered infection without sepsis. This would be called sepsis now, and of course,  septic shock would be the same. If in fact, they were different diseases, you would expect  fundamentally different mortality rates, and what you see is exactly that. Essentially, if you have  infection without sepsis or sepsis, you have a fairly low mortality rate of about 15% over 14  days. It's only when you get to septic shock that the mortality jumps way up, really suggesting  we're talking about two distinct diseases because otherwise you'd expect similar mortalities.  Now, this is another. Whoops. Hang on a second. This happens every time. It will stop, I hope.  There we go. Here's another study looking at the issue of the microbial load. This is basically  people with pneumococcal pneumonia looking at the risk of septic shock and the risk of death.  Jordi Rello here looked at the log PCR copies. If you have less than three logs,  then in fact, your probability of getting septic shock is less than 10%. If you are three to four  logs, you're getting up 15%. It really starts to jump up when you get four to five. Then of course,  when you get greater than five, you have 100% chance of pretty much of having septic shock.  People with this kind of microbial load almost always have septic shock with pneumococcal sepsis.  Similarly, if you look at mortality, you see a significant difference in outcomes. Again,  suggesting that delays of initiation, delays in terms of intervening in the process just means  you allowed the microbial load to increase too much. What do you do when you give antimicrobial  therapy? Well, this is basically what happens. You hit the organism, the microbial load decreases  over time, the toxic burden follows, the inflammatory response, cellular dysfunction,  tissue injury, all coming down, not immediately, that is to say they're all delayed to each other,  but if you're fortunate and you intervene quickly enough, the person inflammatory response and  cellular dysfunction settled down early enough that essentially they don't pass the golden hour.  If we consider there's a golden hour or golden hours of septic shock, if you bring them down  fast enough, you'll be lucky, but you might want to do better than this. How can you do better?  Well, the simplest thing would be simply give the antimicrobial faster. If you give early antimicrobial therapy, this is what should happen. You truncate the microbial load earlier,  all of the following elements settle down. If you're lucky, you don't even go into shock or  you go transiently into shock. You've seen this in the emergency room, you get called downstairs, you know, they tell you, oh, I got a patient with pneumonia, you look at them,  they're looking, you know, toxic but not hemodynamically unstable and you go, okay, yeah, just give them cefetaxime, they're just probably a community-acquired pneumonia,  give them your antibiotic, you think you've done something really good, you head out and then you  find they call you in an hour and say, hey, your patient went hypertensive, does that mean you did  anything wrong? No, they're just having, you know, basically what you'd expect, you're having a bit  of a delayed inflammatory response, usually just giving them fluids will settle them down.  Now, conceptually, when is it that giving faster antimicrobials can actually make a difference?  Well, the thing you have to ask yourself is, is this a condition where faster pathogen clearance matters? And really it matters if there's a delay-dependent risk of  irreversible or irreplaceable organ failure. And it's a limited number of situations where that occurs. I mean, if you think about it, meningitis is one of those cases. If you give  antimicrobials late meningitis, what's going to happen is you're going to irreversibly damage  your brain, you can't fix it and you can't replace it, you're going to die. Encephalitis is similar, rapidly necrotizing soft tissue infections. Again, you have that necrosis,  it's very time dependent. There are certain types of pneumonias, necrotizing pneumonia is also very  important. Generally speaking, you can't get a lung transplant. And then of course, septic shock  is in that category for a different reason, and that is over time with septic shock, you get more  and more organ failures. Now you can replace one organ failure, you maybe replace two, but once  you get three, four, five, you're going to die no matter what you do. And so you got to ask yourself, is this a delay-dependent risk of irreversible or irreplaceable organ failure?  And the answer for septic shock is certainly yes. And one could make that argument for sepsis as  well, at least as it relates to organ failure as opposed to shock. So if you've got to eliminate  the pathogen more quickly, there are a couple of ways you can do it. Hit faster, hit the organism  faster, increase the speed of attainment of therapeutic levels, or you could hit harder, potentially, at least in theory. We're not going to talk about this on this occasion,  but if you talk about hitting faster, then you're thinking about antimicrobial appropriateness,  loading doses, and speed of initiation. All of those will give you accelerated pathogen clearance.  Now, I'm not going to talk at all about loading doses. You guys know that you need to use loading doses with certain groups of antibiotics, colistin, for example. There's some data,  actually, that if you're using intermittent infusions, you want to give a loading dose.  But like I said, that's not what this talk is about. You guys are all very familiar with the  data. And again, I'm not going to belabor the point. But basically, you know that if you have  inappropriate antimicrobial therapy for patients with a variety of severe infections, you will increase the mortality risk. This is data from COLLIF. It's about 20 years ago,  looked at patients who had bacteremia, actually, but they were going into the ICU, so a lot of them had septic shock as well. But you can see that if you give appropriate  antimicrobial therapy, you have about 27% mortality. If you get inappropriate, it's about 55% mortality. Basically, you double the risk of dying if you give inappropriate  antimicrobial therapy initially. So that's something very much to bear in mind.  Now, if you look at it, by the way, in the context of septic shock, the difference is even more stark. What you see here is that if you give  inappropriate antimicrobial therapy, you have 10% survival. If you give appropriate antimicrobial  therapy for all cases of septic shock, you have 50% survival. So you're no longer talking about a  50% increase in mortality or doubling in risk with inappropriate antimicrobial therapy. You're talking about basically a five-fold difference in survival if you give inappropriate  antimicrobials right off the bat. And that's even if you correct it. Now, people have looked at this  issue of early antimicrobials, but they looked at it very loosely. This is data that a friend  of mine, Charles Natanson, provided many years ago. I don't think he's published this, but he  looked at a whole group of studies that look at speed of antimicrobials prior to 2006 when I did  the study that I'll show you. And basically what he found was that if you did a meta-analysis of  sorts, you found just about every study at least trended in the direction of better outcomes with  faster antimicrobials. Now, this is data that we generated. I think it's become quite well-known  at this point. Back in 2004 or so, I was doing some animal research and found a very startling  increase in mortality of mice that were made septic. In fact, basically, if you made them septic and they went hypotensive, basically, if you started your antimicrobials within,  I believe it was a few hours of hypotension, then basically all the animals survived. And if you were  more than a few hours, they all died. It was about a three-hour window where they went from 100%  survival to 100% death. So I thought we should look at this in humans. I asked my friends,  how long do you think it takes to give antimicrobials once you've documented hypotension  in patients who have septic shock? And they said, well, what do you think they said? Well, they said,  eh, an hour, maybe an hour and a half. So I naturally went and started measuring in a couple  of hundred patients. And I was quite flabbergasted as I went along. The reason the study became so  big is because every time I looked, we were just startled because the first time we looked in our  place, it was six hours. We were astonished. And so I went to the other side of the river. We're  in a city with two rivers and went to the other side. And they're the French side of the river.  And they said, of course, we are much better than you, Dr. Kumar. We're the French side and  we're clearly better. But to their credit, they checked. And in fact, they were the same as us.  And so I went to my friends in Toronto who were absolutely brilliant. If there are people from  Toronto here, you can ask them. They'll tell you they're brilliant. They're very good guys. But  they actually also thought they were somewhere around an hour or so. So when they checked,  they were also six hours. Then we did it down in the States because we were still flabbergasted  what we were seeing. It was still around six hours, anywhere from five to nine hours basically.  So we were astonished. Yet for every hour delay, what we're seeing was a decrease in survival.  And so if you got your antimicrobials within 30 minutes, then you had about a 90% survival. If  you got it within the next half hour, then you were bound down about 83% survival. And thereafter,  for every hour delay, you had about a seven and a half percent decrease in survival out to six hours. So it was just quite amazing. If we looked at the running average of survival,  so if you took say 250 cases, we had I think 2,000. So if you took 250 fastest and then just  shifted down one each time and did a new average. So this is a running average of survival. You can  see that you start in the mid 80s and then you just drop like a stone so that once you get to  about 12 hours, you're down about 25% survival. Once you get past about 24 hours, you're looking at under 5% survival. Stays pretty stable thereafter. I actually suspect these are  misdiagnoses as much as anything else, or maybe they're places where surgical intervention has  been sufficient. When we looked at the adjusted odds ratio of survival with increasing delays,  what we actually see is that if you got your antimicrobial in the second hour compared to the  first, your chances of surviving were already significantly decreased. They just skyrocketed  from there. We've rerun this, by the way, with much larger numbers. We can show a break at  half an hour, maybe even 20 minutes. Basically for every hour delay, your chance of dying  increases. It's about what? 5% per hour kind of thing out to probably about 12 hours.  Other people have looked at this kind of issue. We actually looked at this. This is one of my  colleagues, Dr. Ketheretti, who is currently at, I don't know, he was in Louisiana. He's moved  recently. But we looked at culture negative and culture positive. We actually saw the same response there. The in-hospital mortality just skyrockets with increasing delays,  whether you're culture positive or culture negative. We're actually looking at it now in  terms of gram positive, gram negative, same kind of thing. Dr. Wiles, who works with Steve Simpson  in New Mexico, I believe, looked at a really interesting question, basically asking,  were increasing delays of people with sepsis associated with increased progression of septic  shock? The answer is, yep, definitely. This isn't the best graph. I actually have a somewhat better  graph, but you can see this overall trending in that same direction. I think he's recently  shown it in a new paper that's coming out in ER patients specifically. One of the questions that  we thought was interesting to ask was whether if you correct your antimicrobials, so say you  get it wrong initially, so give it inappropriate antimicrobials, but then you give the right one,  does that have an impact still? Does giving fast correction essentially improve outcome? Again,  interestingly enough, the answer is yes. We had 55 patients who got it within 6 to 12 hours.  They had about a 25% survival. Not great, but okay. 115 got 12 to 24 hours. Their survival  was down to 25%. As you can see, though, as you get out to 36 hours, you're getting under 5% survival. Even in the context of giving inappropriate initial antimicrobials, it's good  to get them corrected quickly. Does speed of antimicrobial impact acute lung injury? This is a study by Isamen et al. You can see here that delayed antimicrobials was, in fact,  an independent risk factor for acute lung injury and septic shock. Delayed antimicrobial therapy  is associated with worsening of inflammatory response and also organ failure. This is a study  by Garnaccio-Montero in 2006 showing, again, delta SOFA and delay of initiation of adequate  antimicrobial therapy. Actually, it is associated with worse inflammatory markers and organ failure.  Now, accelerated antimicrobial therapy is usually given in the context of bundles. People often ask,  well, maybe it's a different part of the bundle that's having an effect. People have looked at  that. In fact, this is a nice study by Ferrer et al., AJRCCM 2009. What you can see is that, again,  more rapid antimicrobials is associated with better outcomes. Fluid challenge didn't seem to have an effect. Low-dose steroids have an effect. In this study, activated protein C  seemed to be beneficial. There was a very interesting study by Brock and colleagues. Actually, Charles Natanson was the senior author on this. Basically, he asked the question,  similarly, in severe sepsis bundles, what are you actually changing? Severe sepsis bundles clearly  improve outcomes, but do they decrease the time to antimicrobials? What they found was clearly the  bundles seemed to help outcome. Okay, so the bundle was favored. What changed though? If you looked  at crystalloid use, vasopressor use, inotropes, RBCs, you can see that inotropes were used a bit more, but the I squared value is fairly high, if I recall correctly. No, I squared  is good, and so inotropes were significantly different, but not significant. Crystalloid use, I squared very high, vasopressor use, and RBCs all very high, overall basically  not reportable, might be some benefit from the inotropes. However, what they found was  that antibiotics, the one thing that was significant consistently was that antibiotics were given  faster. Let me go back. There was no consistent change in the use of fluid suppressors, Paxil  steroids, or activated protein C. There was a small increase in inotropes. So, as I said, what seems to be important there is the speed of antimicrobials. This is, oh, this is the  study by Wiles. Again, I'm sorry, this is the one I should have showed you before. So, again, what you see here is that with increasing time until initial antimicrobial therapy in  sepsis, you have an increasing risk of progressing to septic shock up to about 55% with increasing  delays up to 18 to 24 hours. What about bacteremia, by the way? It's interesting that delayed antimicrobials don't seem to have much of an impact with straight up bacteremia. Now,  there's actually a bunch of studies, probably 20, I think, and I should probably do a meta-analysis  on that one of these days, but in fact, about 80% of the studies don't show a benefit to more rapid antimicrobials, at least within the timeframes that we're talking about, which  is hours. Certainly, if you get out to days, you can start to see some effect, but if you're talking about hours, you don't really see an effect with delayed antimicrobials. What  about sepsis without shock, in particular? Is there a mortality effect there? This is another study by Ferret, colleagues using a large database. I think this was the surviving  sepsis database, and you can see here that with increasing delays out to six hours, the potential for mortality goes significantly up over time, 95% confidence interval. Again,  this study suggests that in sepsis, time does matter. Now, this is an interesting study by Seymour et al. This is probably the highest quality  study that's been done on this subject. Seymour looked at the adjusted rate of completion  of a three-hour bundle among a bunch of different hospitals. There's quite a bit of variation  in initial time to antimicrobials among these being different hospitals. As you can see, the best were significantly better than those that were at the far end, at the low ranking.  If you looked at the completion, the cumulative probability, time after protocol initiation, you can see, again, that the majority of patients got stuff done within about three  hours. You're talking about intravenous bolus completion, fluid bolus, broad-spectrum antimicrobials.  This is in light blue. You can see getting up to about 60% at just two hours, and then, of course, a three-hour bundle. The real thing I wanted to focus on, though,  with Seymour's study is that they did look at the initial bolus, the three-hour bundle, the administration of antibiotics. I'm not going to get into details here. Suffice to  say, though, that administration of antimicrobials was significantly associated with mortality risk, much more so than the initial bolus of intravenous fluids. Here's the thing I  want to show you, though. If you look at vasopressor use, so essentially, if you're talking about  septic shock here, no septic shock here, you can see that those who are on vasopressors  had a very significant effect in terms of the impact of time to antimicrobials, whereas those that did not have septic shock, i.e. they had sepsis without shock, seemed to have  no significant effect overall. Now, it's close. I wouldn't argue with you, but you can see  the effect is much more significant there. Again, if you look at it straight from vasopressors,  I think that last one was hospital ratio for in-hospital mortality. I'm sorry. Let me just go back for a minute. In-hospital mortality, death. Oh, I'm sorry. I must have duplicated  it. I won't get too far here. Let's just see. Oh, septic shock, you can see significant differences in the three-hour bundle completion, et cetera, here. Yeah. Now, the question  of the impact of speed of antimicrobials has been looked at in meta-analysis in many cases. Data can be somewhat contradictory depending on whether you look at just septic shock,  sepsis without shock, or a combination of the two. This is basically one study that looked at this question. I believe these were septic shock primarily these cases. Obviously,  they found a significant beneficial effect here. On the other hand, this is another analysis by  Sterling looking at sepsis-septic shock, and what they found was no significant effect. You can see  those studies going either way, but a lot of it has to do with the heterogeneity of the cases in  there. Finally, I'm going to show you this one study that I think is pretty explanatory. This  is a nice study by Lewedal. It came out in AJRC-CM in 2017. Again, they looked at a large group of patients who had sepsis, and they actually categorized them. I'm going to show  you the results. First off, they found the adjusted odds ratio of death went up significantly for  every hour delay, particularly going up around here at five to six hours, kind of more stable  in between there, but overall, clearly going up. Now, what they found when they were adjusting the  models is that if you had septic shock only, you had a very solid relationship with time to  antimicrobials. If you had severe sepsis, what we call sepsis these days, that is sepsis, basically  infection with organ failure, then you had a significant effect, but not as robust. Finally,  if you had sepsis only, what we would call infection these days, severe infection these days, then in fact, the effect was quite marginal in terms of the statistical significance.  I'd like to show you what you can potentially do here. This is data from our sites in Winnipeg.  We've never published this data, but this gives you an idea of kind of what's possible. This is  the antimicrobial delay versus the year that we stratified it into. As you can see, initially, when we started collecting data, we were up around eight hours median delay.  Now, these are six different hospitals. We had a lot larger numbers at HSC. That's our major center.  The other one is St. Boniface here. You can see here, as you get out to around 9901,  you know, we're still around probably seven hours as a median overall, not really improved,  and not really making much effort on it. This is when we started studying this issue, and we're  starting to get a Hawthorne effect here. We didn't really intervene, but we were kind of getting down  to about five hours at this point. Then 05-07, we had the data in hand, and we were analyzing it.  People knew what it showed. Got down to four hours. Here, we were actually making specific  interventions. As of 2011, we had gotten down to about three hours. In fact, we're down to about  two hours now overall. Now, if you look at what the survival does during this period,  it's just what we'd expect. Basically, we go from, what, about 45% survival to about 65%  survival during the period that we decreased the time to antimicrobial for about three hours.  Basically, we would estimate we'd see about a 21% absolute improvement in survival. We actually  see about 20%. It has actually continued to go up. What lessons can we take from this?  Well, accelerated antimicrobial therapy is strongly indicated for septic shock. I think there's pretty good consensus on that. In fact, the consensus guidelines will say that.  Accelerated antimicrobial therapy is probably important for sepsis. Without shock, probably less so, but still important. Accelerated antimicrobial therapy may or may not be useful  for infections without sepsis. Depends on the severity. Depends on the kind of infection.  But certainly, if you look at bacteremia, for example, most of the studies will tell you that  the kind of delays that we're talking about, where we're talking about hours or six hours or even 12  hours, probably doesn't have too much of an impact. Then the utility of accelerated antimicrobial therapy appears to be dependent on the severity of the underlying process,  which I think is very understandable, given the pathophysiology that we discussed earlier.  I believe my colleague, Dr. Nanchal, is going to do some case series, and then we'll be available for questions. Thank you.  Thank you, Dr. Kumar, for a most wonderful presentation. That was fantastic.  Now for this part of the session, we will go through a few cases, and then hopefully, this will lead to a robust discussion, and then we can spend some time answering questions.  So I have no relevant disclosures for this part of my talk, except unlike Dr. Kumar, I don't have an opinion on antibiotics. I'm just always right.  All right. So this is the outline of my talk, and truthfully, I really cannot argue that one  should not administer antibiotics early when sepsis is highly suspected. But I think we can  all agree that administration of antibiotics should be performed judiciously, and there are some harms associated with antibiotic administration. So our decision-making  should address some questions such as appropriateness of administering antibiotics in certain patient populations, duration of antibiotic therapy, choice of antibiotics,  and perhaps whether we can use biomarkers to guide starting and stopping antibiotic therapy.  In the next few minutes, in a case-based discussion, we are going to go through some of these important aspects and some controversies.  So here is case one. Case one is a 47-year-old man who suffers an out-of-hospital ventricular  fibrillation cardiac arrest. The ACLS protocol is initiated, and he achieves return of spontaneous  circulation in about 12 minutes. He is comatose on arrival to the hospital, and target temperature  management is initiated. He is intubated, mechanically ventilated. The target temperature is set at 33 degrees centigrade. There is no evidence of aspiration. He is hemodynamically  stable, and the chest X-ray is clear. Should antibiotics be administered to this man?  Now, in a recent randomized controlled trial in 16 ICUs in France, these investigators asked this  very same question. The premise of the study was that targeted temperature management is associated  with increased risk of secondary infections and is an independent risk factor for the development  of ventilator-associated pneumonia. In the anthotic study, the population was patients who had out-of-hospital cardiac arrest with a shockable rhythm and return of spontaneous  circulation and treated with targeted temperature management, which was defined as a temperature between 32 and 34 degrees centigrade. The intervention and comparator groups were  antibiotics administered within six hours of return of spontaneous circulation. For two days, the antibiotic of choice was amoxicillin clavulanic acid, and the comparator was no  antibiotic therapy, and the outcome was the development of early ventilator-associated pneumonia. And early ventilator-associated pneumonia was defined as the development of  VAP during the first seven days of hospitalization. And here are some of the results. There were 198  patients that were randomized, and 194 that were included in the final analysis. Except ventilator-associated pneumonia, there was no difference in any other outcomes, including  mortality, ICU length of stay, hospital length of stay, and the development of late ventilator-associated  pneumonia. As you can see from the cumulative incidence function curves, which are shown  on the top of the slide, the control group, which was the group that did not receive antibiotics, had a higher incidence of early ventilator-associated pneumonia, as well as  cumulative ventilator-associated pneumonia. Now, there were no other differences in either  catheter-related bloodstream infections, as I've already mentioned, urinary tract infections,  or any other outcomes. Okay, so what can we take away from this trial? The main takeaways are  that early administration of antibiotics may prevent early VAP within seven days after cardiac  arrest with shockable rhythms and return of spontaneous circulation, but the benefit of  outcomes besides VAP is very unclear. And there may be some undetected harm in administration of  antibiotics to everyone with cardiac arrest, versus just administering antibiotics at the  point of suspicion of well-ventilated-associated pneumonia developers. All right, let's come to  case two. Case two is a man, a 53-year-old man, who has predatonitis from a diabeticular abscess,  which has ruptured. He is taken to the OR for emergency laparotomy, where the abscess is drained,  part of a sigmoid colon is resected, and primary anastomosis is performed. The abdomen is closed,  and the abdominal drain is left in place. He is admitted to the ICU post-operatively. Post-operatively, he has also started on subtraxone and metronidazole as antibiotics.  Now, four days later, his temperature is 101 degrees Fahrenheit, or 101.6 degrees Fahrenheit.  He is hemodynamically stable. He has minimal drain output, and he has slight abdominal tendons. His  white cell count is 20,000, and he still has an ileus. On talking to the surgeon, the surgeon  confirms that he has had adequate drainage of the abscess. There is absence of an asthmatic leak,  and the surgeon believes that there is adequate source control. So what should you do in this  case? What is the next best step in management? Should you stop antibiotics? Should you continue  antibiotics for two weeks? Should you change antibiotics to meropenem? Should you re-explore  the abdomen? Or should you do something else? In this trial, a short course of antimicrobial  therapy for intra-abdominal infections, now published in the New England Journal of 2005,  2015, about five years ago, the investigators tested whether a short course of antibiotic  therapy for intra-abdominal infections was as good as a longer course. The premise of the study  was that the optimal duration of antibiotic therapy in intra-abdominal infections is unknown.  Prolonged courses of antibiotics are associated with adverse effects. And perhaps, perhaps,  if adequate source control is achieved, shorter antibiotics course may just be sufficient.  Again, the population was patients with complicated intra-abdominal infections, with fever or pelotonitis, and those that had had an intervention to achieve  adequate source control. The intervention was four full days of antibiotic therapy. After source control, the comparator was antibiotic therapy until two days after  resolution of the systemic inflammatory response syndrome. And the outcome was a composite of  surgical site infections, recurrent intra-abdominal infection, or death within 30 days of randomization.  Okay. And here are some of the results. There were 518 patients that were randomized.  A 30-day follow-up was complete in 99.8% of the patients. Now, there were some protocol deviations,  and interestingly enough, the patients that were randomized to receive antibiotics  for a longer duration of time had more protocol deviations than the experimental group.  However, there were, as you can see from the table that is shown on the slide, there were no differences in the baseline characteristics of these patients. They were  very, very similar. And here are some of the outcomes. Importantly, there were absolutely  no differences in death, surgical site infections, or repeated intra-abdominal infections within the  two groups. In every subgroup that was studied, no matter how you slice or dice this data,  the proportion of patients with a composite outcome were similar. In fact, in the group that  received antibiotics for a prolonged period, there was a delay in diagnosis of surgical site or recurrent intra-abdominal infections by an average of about seven days, and this was  statistically significant. So what are the takeaways of this trial? Takeaways of this trial  are that four days of antibiotic therapy seem adequate in intra-abdominal infections if source  control is achieved, and it really decreases the duration of antibiotic use and exposure.  Delays in manifestation of infectious complications in the control group was likely an adverse event. The time to recognition was delayed, and therefore the time to resolution  was also prolonged. And it is very likely that after four days, the ongoing systemic inflammatory  response syndrome is a host immune response rather than the presence of viable microorganisms. And overall, again, if source control is achieved, then there is no need for a longer  course of antibiotics in intra-abdominal infections. Here is case three. Case three is a 69-year-old  woman who was admitted to the ICU from the community with sepsis from UTI. She is resuscitated  and stabilized and started on ceftriazole. Two days later, both blood cultures are growing E. coli,  which is sensitive to ceftriazole. She has been hemodynamically stable for the past 24  hours. Acute kidney injury is resolving, and blood cultures, which are repeated, are negative.  Now, what is the next best step in management? Would you continue antibiotics for seven days,  14 days? Would you measure serial procalcitonin and determine antibiotic therapy based on procalcitonin? Or would you measure serial CRPs and determine antibiotic therapy based on  serial CRP measurements? All right. So here is a trial that was recently published in JAMA that  compared a fixed seven-day course of antibiotics with a 14-day course of  antibiotics or a course of antibiotics that was guided by repeated measurements of CRP. The premise of the study was that prolonged antibiotic exposure drives resistance and may  have other adverse events. The duration of antibiotic therapy generally in these cases, which is sort of uncomplicated gram-negative bacteremia, is just based on expert opinion,  and there may be clinical non-inferiority between seven- and 14-day courses, and biomarkers could  perhaps be used to individualize duration of antibiotic therapy. Again, in this study, the population was adults with one blood culture that was positive for gram-negative bacteria,  and treatment that was started with efficacious antibiotics. The intervention was a fixed  seven-day course of efficacious antibiotics, and the comparators were either a fixed seven-day  duration of efficacious antibiotics or CRP-guided therapy. Now, the CRP-guided therapy was defined  as a decline in the antibiotic durations. Antibiotics were continued until the CRP declined  to 75% of peak value. That is, it was left at about 25% of its peak value, and the patient  had been afebrile for 48 hours. The outcome was defined as a clinical failure at day 30, which  was defined as a relapse, recurrent bacteremia, local separative complication, such as a kidney  abscess for pyelonephritis, distant complication, death, restarting antibiotics due to clinical  worsening for the same organism. There were some prominent exclusions of this trial, which should be taken into account and kept in mind. The prominent exclusions were fever or  hemodynamic instability in the past 24 hours, severe immunosuppression, non-fermenting bacilli, polymicrobial sepsis or polymicrobial bacteremia, gram-positive bacteremia,  and complicated infections, such as endocarditis.  Now, this was designed as a non-inferiority trial. For non-inferiority trials, to demonstrate  non-inferiority, the p-value has to be significant. You can't have a non-significant p-value. For  example, in this study, the interpretation of the p-value that you see up on the screen, which is less than 0.001, means that the probability that the non-inferiority margins  between the two groups would be crossed was less than 0.1%. And the non-inferiority margin  was set at 10%, which means that non-inferiority in the trial was demonstrated. It was demonstrated  between both the CRP-guided group versus the 14-day group and the 7-day group versus the 14-day  group. However, it is important to understand that in the CRP-guided group, protocol deviations  occurred in about 20% of patients. And the median duration in the CRP-guided group of antibiotics  was 7 days as well. And the inferiority margin that was set was about 10%. And usually in these  trials, inferiority margins are set at about 5%. So there was a large non-inferiority margin that  was set. So the takeaways from this trial are that our fixed 7-day therapy or CRP-guided therapy  is non-inferior to 14-day therapy. And this can help limit the exposure to antibiotics because  in uncomplicated bacteremia, you can get away with 7-day therapy. The caveats, though, are again,  like I mentioned, that there were protocol deviations in 21% of the CRP-guided group. So it is unclear whether CRP can actually be used to guide therapy or not. And there  was a low event rate. And there was a very, very large non-inferiority margin.  Okay. My last case is a 69-year-old woman who has sepsis from a UTI. She's been resuscitated and  admitted to the ICU. She has started on piperacil and tezobactam, more commonly known as zosyn.  Blood cultures are positive for E. coli. Now the blood cultures are resistant to ceftriaxone,  but are sensitive to both zosyn and meropenem. So what is the next best step in management? Should you continue the beta-lactam and beta-lactamase inhibitor? Should you switch  to meropenem? Should you add ceftriaxone to the beta-lactam and beta-lactamase inhibitor? Or should you discontinue antibiotics? Now, here is a study that was published  about two years ago in JAMA that compared the effect of zosyn and meropenem on 30-day mortality  for patients who had bloodstream infections with E. coli or klebsiella and had ceftriaxone resistance. Gram-negatives that have ceftriaxone resistance produce an enzyme called  extended-spectrum beta-lactamase. And actually, as antibiotics are more... broader and broader  spectrum antibiotics are used, these infections are becoming more and more common. Carbapenems  are the treatments of choice. However, using carbapenems may also select for carbapenem resistance. And usually, there is in vitro sensitivity to zosyn. And therefore, the zosyn  may be used as a carbapenem-sparing drug in cases where you have bacteremia with ESPN-producing  organisms. And that was the premise of this study. So their population was adults with one  blood culture positive for either E. coli or klebsiella, which was non-susceptible to ceftriaxone  or susceptible to zosyn and meropenem. The intervention was administration of meropenem,  which was dosed at one gram every eight hours. The comparator was the administration of zosyn,  which was dosed at four and a half grams for every six hours. And the outcome was all-cause  30-day mortality at 30 days post-randomization. And as you can see here, this was again designed  as a non-inferiority trial. And the non-inferiority trial, the p-value was non-significant, which  means that the margin for non-inferiority, which was set at 5% in this study, was crossed,  demonstrating that piviricillin-tazobactam was actually inferior to meropenem for the treatment  of ESBL bacteremia. So my takeaways for this are, please do not use a beta-lactam. Beta-lactam is inhibitor for ESBL organisms, especially if they are in the bloodstream.  So the conclusions are early antibiotics and sepsis are appropriate. I absolutely agree with  Dr. Kumar. But I also would like to state that the antibiotics should be used judiciously. Sometimes it may not be necessary to give antibiotics. Please make the correct choice  of antibiotics. Prolonged courses of antibiotics should not be used unnecessarily. And whether one  can guide antibiotic therapies based on biomarkers or not, and the duration of antibiotic therapies  based on biomarkers or not, still very unclear. Thank you for listening. And now we will open the session up for questions.  Hello everyone and welcome to the Q&A session. I am Rahul Dhanchal and I am the moderator for the session and let's get started. So Dr. Kumar, the first question is mine actually  and would you be able to share, do you think that the pendulum for antibiotics has swung too far in an effort to give early antibiotics and like I said in my presentation, I really  can't disagree with early antibiotics and septic shock but now people are using antibiotics  indiscriminately too broadly and for too long? Yeah, I think that is an issue. I think that people should be aware of the data and you know certainly I think that we do routinely  tend to use antibiotics too long. I think we routinely use antibiotics where the risk of a serious infection is too low basically. I don't think it's ever a mistake to start  antibiotics when you are faced with the possibility of an immediately life-threatening infection,  meningitis, septic shock, even severe sepsis, etc. But when you have time, you should take  advantage of it and with a lot of infections that present, you may not need the antibiotics and even more centrally I think, if you start antibiotics and you then discover in a day  or two, no it actually wasn't bacterial sepsis, a lot of people tend to think, well you know  what I'll just finish a course and that's an absolute mistake. Another thing that frequently people do that's a mistake is that they in culture negative sepsis, which is a third  of cases, they start the antibiotics and they start a very broad spectrum. The patient starts to get better and they say, well you know what, since they got better I'm not going  to stop anything. And again that's a mistake too. At that point you can judiciously de-escalate to a more reasonable regimen. And then if you have any reversal, which is very rare  to be honest, then you can go back up if you need to because now you've got the luxury of time. They're not imminently going to die if you get it wrong.  Just so that the audience knows, we might not be able to get through all of the questions.  There were a lot of them on the site, but we will be sure that we answer all your questions the Q&A session has ended. One question that we got from the audience, which I thought  was a brilliant question, is that we actually don't know, we can really never ever define  time zero of sepsis. We just arbitrarily measure mortality when people come to the ER and they  may have had an infection for one day or three days or four days or five days. By that logic, the first question was that every hour of untreated sepsis increases mortality. They  should have all died by the time to get. So I think the person is trying to make the point  that maybe what we are measuring is just clinical worsening and that septic shock is an early  ante. So what you were measuring is a surrogate of when people arrive and septic shock is an easily measurable surrogate of how long they've had sepsis.  Really, what we should focus on is actually recognizing sepsis early and not administering  antibiotics early. I'm not sure how to respond to that. Certainly, we kind of looked at this  question in the study that we did way back and we actually found that there are a subset of patients, for example, that are hypotensive before they come to the ER so that you don't  really know how long they've been hypotensive. And if you actually look at their mortality,  their mortality is significantly higher than those who come in normotensive and go hypotensive in the ER or in the hospital. Having said that, they're still in a window mostly where  you can help with immediate antibiotics because generally speaking, I think human beings are  relatively social creatures for the most part. And so if somebody is really sick or they're lying on the floor of the apartment building that you live in, usually they call for help  pretty fast. So they're usually not lying around for days in septic shock. So I think early antibiotics are still very important in 95% of cases. There may be somewhere they've  been found kind of down somewhere. They've been like that for a while and giving antibiotics isn't going to make a difference, but I think that's a minority of cases.  We had another question that pointed to a recent systematic review that I think if I remember the systematic review right, said that there was no difference between administration  of antibiotics within one hour or administration of antibiotics within three hours. What is your take on that? As I remember it, the systematic review included a lot of studies  that were at risk of bias and included sepsis, severe sepsis, or what we call sepsis these  days in septic shock. And when they had separated out septic shock, although the confidence  intervals sort of crossed one, there was still a strong signal towards benefit towards mortality.  Yeah, I think the thing that you got to be careful about the meta-analyses is that a lot of them mix basically ICU level infection, sepsis, what we used to call severe sepsis,  you know, infection with organ failure and septic shock. As the study by Lewitt had mentioned  the last one I reviewed, basically the more severe the septic state, the more likely that rapid antimicrobials will be helpful. The other thing that kind of screws up these things  is the, and this relates to another question you asked, is the times zero point. You know, I think my data shows a very robust effect because I used a physiologic parameter of  when the hypotension starts as my index point, my zero point. And I did that because of an  animal study where we showed that hypotension was, was, you know, what mortality was linked to in terms of duration of hypotension for antibiotics. Most studies though, they don't  have that data. They don't do a case review. They just use administrative data. And that administrative data is basically, when did you come into the OR? When did you have your  blood culture taken and stuff like that? So it introduces a lot of, you know, variability and timing error into that relative to when did they become hypotension, hypotensive.  So that's why I used hypotension as my index because our animal data strongly suggested it was the strongest correlate of a time zero that was appropriate.  All right. Thank you, Dr. Kumar. I think that's all we have time for. And just as a reminder  to all our, you know, to our audience, we will make sure that all your, all your questions are answered online.  Yeah. So check, yeah. So check the, check that in, in a couple hours and should have it all done for you. Thank you so much, Dr. Kumar for this wonderful session. My pleasure.  Bye-bye.

Video Summary

In this session, Dr. Anand Kumar discusses the importance of early administration of antibiotics in suspected sepsis. He presents the idea that infection triggers a cascade of events that can lead to organ injury even after the infection has been treated, highlighting the need for prompt antimicrobial therapy. Dr. Kumar also emphasizes the importance of appropriate antibiotic choice, loading doses, and speed of initiation in order to achieve accelerated pathogen clearance. He provides several studies that support the idea that early antimicrobial therapy improves outcomes in sepsis and septic shock. However, he cautions against using antibiotics indiscriminately and for prolonged durations, stating that antibiotics should be used judiciously and based on appropriate indications. He also discusses the use of biomarkers such as procalcitonin and CRP to guide antibiotic therapy, but acknowledges that more research is needed in this area. The Q&A session following the talk addresses questions regarding the timing of antibiotic administration in sepsis, the use of antibiotics in specific patient populations, and the role of biomarkers in guiding antibiotic therapy. Overall, the session highlights the importance of early antibiotic administration in sepsis while also emphasizing the need for appropriate use of antibiotics to prevent resistance and adverse effects.

Asset Subtitle

Sepsis, Pharmacology, 2021

Asset Caption

The Society of Critical Care Medicine's Critical Care Congress features internationally renowned faculty and content sessions highlighting the most up-to-date, evidence-based developments in critical care medicine. This is a presentation from the 2021 Critical Care Congress held virtually from January 31-February 12, 2021.

Learning Objectives

Emphasize the importance of early administration of antibiotics in suspected sepsis and enumerate the benefits of this strategy
Discuss the pitfalls of a one-size-fits-all strategy in terms of antibiotic administration in suspected sepsis
Enumerate the harms associated with broad use of antibiotic therapy
Discuss situations in which early antibiotics in suspected sepsis should be administered and when a watch-and-wait strategy is reasonable

Early Antibiotic Administration in Suspected Sepsis: Slam Dunk or Shrouded in Controversy
Anand Kumar

Case Discussion
Rahul S. Nanchal

Meta Tag

Content Type Presentation

Knowledge Area Sepsis

Knowledge Area Pharmacology

Knowledge Level Intermediate

Knowledge Level Advanced

Membership Level Select

Tag Systemic Inflammatory Response Syndrome SIRS

Tag Shock

Tag Antibiotics

Year 2021

Keywords

early administration of antibiotics

suspected sepsis

infection

organ injury

prompt antimicrobial therapy

appropriate antibiotic choice

accelerated pathogen clearance

sepsis

septic shock

biomarkers

	Society of Critical Care Medicine 500 Midway Drive Mount Prospect, IL 60056 USA Phone: +1 847 827-6888 Fax: +1 847 439-7226 Email: support@sccm.org	Contact Us About SCCM Newsroom Advertising & Sponsorship DONATE	MySCCM LearnICU Patients & Families Surviving Sepsis Campaign Critical Care Societies Collaborative	GET OUR NEWSLETTER
© Society of Critical Care Medicine. All rights reserved. \| Privacy Statement \| Terms & Conditions The Society of Critical Care Medicine, SCCM, and Critical Care Congress are registered trademarks of the Society of Critical Care Medicine.