SCCM Resource Library-Pro/Con Debate: WhatSUP With Stress Ulcer Prophylaxis in the ICU?

Video Transcription

Well, hello there. I thank you all for coming in. I thank you all for coming in because this truly promises to be an interesting session.  And I use the word interesting because it is about something that all of us do and practice nearly every day of our critical care practices wherever we are.  So we walk around our ICUs every day and stress ulcer prophylaxis, in one way, shape, or the other, is on at least 50% of our patients' charts, I believe.  I think the number is probably higher, but 50% is just a conservative estimate. Of those 50% patient charts that stress ulcer prophylaxis is on,  maybe half of them don't really need it or don't really qualify for it. But it's one of those things, right? It's just Pepcid. Who cares? Just leave it on.  Next thing you know, they go home with that Pepcid. It's on their home medication list. They come back six months later. And then you're like, well, if this is a home medication,  we should keep them on it. So the bottom line is we think that stress ulcer prophylaxis is benign. We can just give everyone the Pepcid or their Protonix.  It'll probably help them feel better. But the answer is probably not that. The answer is more complicated than just that.  And I have for you all today two of the best in the business, really, so to speak, especially with all of the major stress ulcer prophylaxis  trials that have been published in the last few years, including very recently. We thought it would be a good time to really take this up in a debate, which  would involve all of you and help answer your questions. So when you leave this room today, you're able to really practice more evidence-based medicine  when it comes to stress ulcer prophylaxis in the critically ill patient. Without further ado, I'm going to introduce my speakers today.  So we have all the way from Copenhagen, Denmark, Dr. Anders Berner. Dr. Berner is going to take the con side of the debate and then from the University of Colorado,  we have Dr. Rob McLaren, who's going to take the pro side of this debate. They have told me before starting that though this is a pro-con, they will still  be friends afterwards, and they will actually converge their ideas a little bit. So they're not going to go at each other with knives on this.  And this is all about you all, so there will be enough time at the end for you all to bring up your questions and engage in discussions.  Dr. McLaren is going to go first. I'll quickly tell you guys about him. Rob McLaren is a professor in the Department of Clinical Pharmacy at the University of Colorado.  He's a clinical pharmacist in the medical ICU at the University of Colorado Hospital. And he did his undergraduate degree in pharmacy at the University of British Columbia,  and then the University of Utah was where he did his PharmD, then completed a critical care specialty at the University of Tennessee.  He has nearly about 100 peer-reviewed publications related to the pharmacological and nutritional therapies in critically ill patients, a fellow  of the American College of Clinical Pharmacy and Society of Critical Care Medicine, and most importantly, he serves as a co-chair of the Stress-Ulcer Profile Axis Guidelines Task  Force for the SCCM. Before I introduce him, get him to the stage, I have been instructed to go through some of these housekeeping items for you guys.  It's a long list that includes things like know where your exits are, et cetera, et cetera. What I'm going to say is please silence your cell phones,  leave your questions for the end. If there are any slides that says copyrighted material or do not publish, then don't broadcast them on social media. Tweet a lot.  We all love tweeting. If you tweet, use the hashtag CCC49. All right, without much further ado, Dr. Rob McClaren. Thank you.  All right, good morning. Thank you, SCCM, for inviting me. Let me apologize a little bit in advance. I'm getting over a little bit of a cold, so if I lose my voice,  please endure and bear with me. OK, so I have the pro side. And I noticed in the handbook that the program has the acronym WHAT'S UP?  And I challenge anyone who is going to do a future study in critical care medicine related to stress ulcer prophylaxis to figure out a way to use this acronym in your title.  I think it's a good one. OK, so I'm the pro side. I will say right now that you'll see a little bit of my presentation being schizophrenic. I'll be a little bit con and pro.  And my goal for today is not to make things simple for you, but rather to muddy the water, because I think most people in this room have some understanding  and some thought process that they already think that the people that need stress ulcer prophylaxis are mechanically ventilated, coagulopathic patients,  and they're going to use either an HRA or PPI. And my goal is to bring some lack of clarity for you. I don't think I have any conflicts.  I am co-chair of the Stress Ulcer Prophylaxis Guideline. But everything I'll be talking about today are my thoughts, my ideas, and I don't represent that task force whatsoever.  Obviously, off-label content. OK, so why provide stress ulcer prophylaxis to patients? From a pathophysiologic standpoint, it makes sense. These patients have ischemia.  They have not necessarily increased production of acid, gastric acid, but increased exposure. So it makes sense to use HRAs or PPIs.  There is reperfusion injury going on in the gut mucosa. And so at least HRAs we know inhibit that to some extent. If somebody does have a bleed, obviously as a clinician  we don't want to see that. But there are clinical and economic outcomes of those bleeds, including prolonged ICU stay. The three that I will distinguish today are risk factors.  And again, I think everybody in this room probably has a preconceived notion what the risk factors are. My goal is to make that not so clear.  There's lots of studies to support prophylaxis. And then I'll say that there's probably minimal risk, especially as we know it today with HRAs,  although I will admittedly say PPIs may be dangerous. Just by a show of hands, here you have a patient. Patient Jack, 67-year-old male, comes in with diabetes, hypertension.  He's admitted to the ICU with septic shock. He's on a ventilator. He's on pressers, elevated lactate, AKI. By a show of hands, would you provide stress ulcer prophylaxis?  All right, so almost everybody. Pretty easy one. What about Jill, 67-year-old female now with community-acquired pneumonia, COPD exacerbation? She has AFib at home, takes Warfarin.  INR is 2.7. She is confused, but she's put on non-invasive ventilation admitted to the ICU. By a show of hands, would you provide stress ulcer prophylaxis? A few hands go up.  What if I told you that non-invasive ventilation is just a bridging method and that maybe in 6 or 12 hours from now she'd be intubated? Would that change your opinion?  What if she was on a PPI at home? So automatically, there's a lot of hands going up. So what are the risk factors for stress ulcer prophylaxis?  Well, up until recently, the only two guidelines we had were now 12 and 20 years old. And what you see is mechanical ventilation, coagulopathy stands out in both guidelines.  The Eastern Association of Surgery and Trauma goes a little further and distinguishes level 2 and level 3. But again, we don't really know what  the incidence of bleeding is based upon these two guidelines. If you follow the literature, you'll realize in BMJ last month, there was a consensus statement  guideline for who should receive stress ulcer prophylaxis. And it broke it down to high-risk patients, high-risk being those who are on a ventilator for at least 48  hours and did not receive enteral nutrition or had chronic liver disease. Notice coagulopathy now drops down to slightly lower risk of 4% to 8%.  Now, there's some things I like about this. They do have a differentiation of very, very high-risk, high-risk, moderate-risk, and low-risk.  They do a nice job of grouping the risk factors into those delineations. And then they have this in the middle, an area of what they call a cut point, where maybe  the value preference of the patient or provider should be considered. So that if your rate of bleeding is presumably less than 4%, now you start thinking about whether or not  it's actually worth giving these patients stress ulcer prophylaxis. I will point out that I think the 8% to 10% is substantially higher than what we see in the ICU.  I would argue that 1 in 10 patients do not show clinically significant bleed. And probably we're more in the 2% to 4%, which means we're always considering  the value preference of the patient or the risk benefit profile. Now, I hate to harp on this study. It's now a quarter of a century old.  It's kind of hard to believe that this study was published in 1994. But it is really the premier study that everybody talks about for stress ulcer bleeds.  This study was done in Canada, a multi-center study where the investigators basically asked the attending physicians to withhold stress ulcer prophylaxis unless, and I repeat,  unless patients had head injury, burns, transplant, or a recent GI bleed. Why? Because those are already known risk factors or presumed risk factors.  The result was that about a quarter of the patients received stress ulcer prophylaxis. Three quarters did not. And so as far as I'm aware, this is the only study  out there still in which patients did not receive stress ulcer prophylaxis. And we're looking at risk factor analysis. On univariate analysis, you see there's  lots of things that come out as predictors of bleeds. Multivariate, coagulopathy, and respiratory failure. This is where everybody quotes coagulopathy  and respiratory failure as the two risk factors. If you look at the rate of bleeding, it's still less than 4%. So it's less than that value preference of the previous slide.  I will point out that even in multivariate regression analysis, you see that hypotension, there was a trend towards bleeding with hypotension.  Now, I do want to point out the type of patients in this study, because we forget that. We think it's multi-ICU. And really, this is a cardiovascular study.  So these are not my patients in the medical ICU. They're not your patient in the trauma ICU or neurosurgical ICU. Rather, they're cardiovascular patients.  At a time in the early 1990s, when we treated patients with these cardiovascular diseases, totally different than we do today. I will say that if you look at actual bleeding rates,  those who had prophylaxis, about 16%. Those who did not get prophylaxis, excuse me, those who had prophylaxis, about 16%. Those who did not, about 1.5%.  And so you might say, well, patients who got prophylaxis were at higher risk of bleeding. Really, this tells me the bedside clinician,  those who provided prophylaxis, knew their patients were at risk and therefore provided prophylaxis. So when I look at this study, it's not just mechanical ventilation, coagulopathy  that are the risk factors. It's the reasons I said not to consider giving prophylaxis. Hypotension was a trend in trauma. More recently, a couple of pharmaco-epidemiological  studies, these studies were not designed to look at no prophylaxis, but rather PPIs versus H2RAs. These are large database analyses, one we did. You can see almost everything was  a risk factor for clinically significant bleeding in our study. The Lilly study, not as much. Now, the Lilly study, the point estimate of bleeding  was about 0.5%, whereas ours was about 4%. So maybe they just didn't show prediction based upon the low rate of bleeding. But again, it sends a message that we don't really  know what the true risk factors are. When you put this all together, this was a pseudo-meta-analysis looking at mostly observational studies.  So this included the three studies I just looked at, plus one more study in terms of low-risk studies. We put it all together. Those things that are associated with bleeding,  with clinically significant bleeding, were coagulopathy, shock, chronic liver disease, and true nutrition was protective. So notice ventilation is not part of this.  So when I look at it, I'm a little confused. And I sit on the task force to figure this. So I would say that patients who deserve stress ulcer  prophylaxis, at least initially, are critically ill patients with anything here in the blue. I would say hepatic and renal failure is still up in the air,  but looks like it's trending towards maybe giving. And then you have some other things that we just don't know about. Now, I would ask you, how frequently  do you see these in your ICU? Basically, everybody in the ICU would get stress ulcer prophylaxis. Now, this comes from actually Andrew's group.  And I like this, because what they did is a large cohort study. And they said, when are patients at the highest risk of bleeding? Roughly 1,000 patients, 26 clinically significant  bleeding, rate of 2.5%. Half of those bleeds occurred within the first two days of ICU admission. And the majority occurred within the first five days.  So I'm not saying everybody needs stress ulcer prophylaxis indefinitely. My personal belief is somewhere in the two to three to four day range is all you really need,  even if patients are still critically ill. Good. Let's talk about which agent. Let's go back to this case of Jack. Remember, he's a patient admitted with septic shock.  He's on pressers. Let me ask you, which of these agents would you use? PPI, hands up. HGRA, hand up. Sucrophate, hand up. Enteral nutrition, non-needed. No, that's Anders.  I'm being nice to him. Just kidding. All right, if you look at sort of cross-sectional analyses, you look at surveys, prescribers will say they're wanting to give HGRAs.  When you actually look at what they give, they give PPIs. Classic example, you all said about 50-50 PPIs, HGRAs. So you'd fit in this box.  If I was to go to your ICU, I bet you most of the patients are getting PPIs. If you ask the guidelines, you see this mix. Way back 20 years ago, sucrophate HGRAs,  the surviving sepsis campaign, change your minds a couple of times. Easter Association for Trauma, HGRAs, PPIs. The Danish Society of Intensive Care Medicine  actually says don't give it. But if you are going to give it, we use, say, PPIs over HGRAs. And the most recent BMJ article suggests PPIs over HGRAs.  Now let me just start with, yes, we should give stress ulcer prophylaxis. This is all placebo studies. Looking at either PPIs or HGRAs, and you see that overall there's a reduction  in clinically significant bleeding. No difference in pneumonia, no difference in CDI. You'll see different infections. So it looks like acid suppressants reduce bleeds.  If you look at the subgroup analysis in this, it seems to be some benefit mostly in neurosurge patients. It might make some pathophys evidence of that,  because these patients have hypersecretion of acid. Enteral nutrition didn't seem to matter for whether we provided enteral nutrition or not.  Now the most recent BMJ article, the guidelines would look at sucrophate versus placebo, no benefit. Both PPIs and HGRAs versus placebo,  there is a benefit in clinically significant bleeding. And then they also categorize it by low risk, moderate risk, and then interpret it for you such  that PPIs and HGRAs are most likely to be beneficial in those patients that are at least 4% risk of bleeding. So I know Anders is going to talk about this study  a little more, but this is the SUP ICU study. Roughly 3,000 patients. It was a randomized double blind study, pantoprazole versus placebo. Primary outcome is mortality.  You see no difference in mortality. However, there was a reduction in clinically significant bleeding with PPIs. So if you look at this study, you  would say PPIs, it makes sense to give. I will point out the average duration here was only four days of therapy. So again, patients don't need stress ulcer prophylaxis  entire ICU stay, just four days or so. What's interesting in this study is that while PPIs reduce clinically significant bleeding, those at the highest risk  or those who are the sickest actually showed an increase in mortality with acid suppression, so with PPIs. And I'll talk a little bit about that as we go through,  but just remember that as we get to another study. Does enteral nutrition affect therapy? This is an older meta-analysis now. Those who did not get enteral nutrition  look at H2RAs versus placebo. Seems like the H2RAs are beneficial versus placebo when enteral nutrition is not provided. However, no real difference in clinically significant bleed  when enteral nutrition is provided, but there is an increase in complications when you give H2RAs and enteral nutrition. That kind of makes sense,  because you're doing acid suppression, you're putting volume into the stomach, you're at high risk of aspiration, maybe it increases pneumonia.  Pharmacological profile timeline, I'm gonna move into PPIs versus H2RAs. You know this, when we were back in the early 90s, all of us were running around with litmus paper,  dose-suggesting H2RAs and antacids based upon what color a litmus paper was.  Somewhere along the way, Deb Cook and colleagues in Canada did the randomized double-blind study of sucrophate versus ranitidine,  found ranitidine reduced clinically significant bleeding, so we all changed to H2RAs. PPIs made headway because of numerous meta-analysis showing PPIs compared to H2RAs  reduced clinically significant bleeding, and so there was a trend to move towards PPIs. If you look at PPIs versus H2RAs, yes, all comers, clinically significant bleeding  is reduced with PPIs, but I think it's important to look at what studies drive this meta-analysis, and there's numerous meta-analyses like this. The Conrad study and Levy study  are two of the bigger studies that drive why PPIs are beneficial compared to H2RAs.  The Conrad study, the primary outcome, although they say it's clinically significant bleeding, was really did the gastric aspirate clear relatively quickly with lavage,  and if you look at that, sometidine was no different than omeprazole. The Levy study was simplified omeprazole suspension versus a drip of ranitidine,  and they did use clinically significant bleeding as their outcome, however, patients in the ranitidine group were much sicker, and they had a bleeding rate  of 31% in the ranitidine group, so that's pretty high, that's one in three patients in the ICU showing clinically significant bleeding. So I just want to point these two studies out  because they're somewhat flawed, yet they drive a lot of these meta-analyses.  If you look at the BMJ recent recommendations, both H2RAs and PPIs seems to be beneficial versus sulcrophate, however, they both increase pneumonia rates, all right?  Now, the BMJ recent recommendations suggest using PPIs over H2RAs,  primarily because you see this point estimate of .58 of PPIs being beneficial. Now, all of these, depending on the risk, you see that it gets better and better  as patients get sicker and sicker, PPIs are more beneficial, however, they all overlap zero, meaning there is a chance that H2RAs  are actually better producing clinically significant bleeding.  I'll point this out, a couple of pharmaco-epidemiological studies, one being ours that showed H2RAs are actually beneficial compared to PPIs  at reducing clinically significant bleeding. So now we're dichotomizing here, we have some data that says PPIs are better, some data that say H2RAs are better.  All right, so this is the most recent study, again, I know Anders is going to go into this with more detail, the PEPTIC study, it's a pragmatic study,  it's a cluster-randomized study of 50 ICUs,  roughly 27,000 patients. If you look at all-cause mortality, there was a trend towards lower mortality with H2RAs versus PPIs,  although there was a statistically significant difference in terms of clinically important bleeding favoring PPIs.  Again, if you go to the sicker patient populations, the APACHE-2 score greater than 18, mortality was increased with PPIs. So we're seeing signals,  yeah, PPIs might reduce clinically significant bleeding, but it might increase mortality.  All right, so I've shown you some of the data, I'm trying to focus you a little bit on H2RAs, let me go a little bit into the side effect profile  for these drugs and maybe try to explain why PPIs increase mortality rate. And again, this is all speculation.  We go to the guidelines, do these agents cause adverse events? Even over the last 20 years, we really haven't made any headway. We went from we don't really know, it's unresolved,  to maybe they increase pneumonia, but probably do not have any real effect on mortality.  All right, so why do I care about this? I'm a big believer that in the old days, we used to run around and get gastric pH above four, we get litmus paper, right?  And I actually think that that's not therapeutically beneficial, but actually harmful. Why? Because as we increase gastric pH and we keep gastric pH above,  we change the microbiome of the gut. And it's been well documented. And if you use intermittent H2RAs, you see that the pH comes back down, allowing the microbiome to be reset.  If you look at outpatient data, or hospital data that's not ICU, acid suppression is associated with pneumonia. All right, and in our pharmacoepidemiological study,  the stronger acid suppression with PPIs compared to H2RAs was associated with pneumonia.  You look at C. diff, you see the same thing. What's surprising with C. diff, though, is the rates of C. diff go up earlier on in treatment. And it's directly correlated  with how strong is the acid suppression, such that H2RAs are actually associated with less C. diff than PPIs.  Again, if you look at meta-analysis, or ICU patients in pharmacoepidemiological studies,  PPIs seem to be associated with more C. diff. So again, acid suppression, the stronger it is, the more it upsets the microbiome,  the more potential there is for C. diff and pneumonia.  Slide a little bit off the top. But something I want to sort of speculate here is that how do we treat patients with hepatic encephalopathy? Sure, we all give antibiotics, right?  Rifaximin, metronidazole, whatever it might be. Why? Because it changes the microbiome of the gut,  and in theory, reduces any harmful exposure that causes encephalopathy. Well, guess what? Patients with critical illness have altered microbiome of the gut.  PPIs also change the microbiome of the gut, such that you actually see reduction in some of the beneficial bifidobacterium,  bacterium, and an increase in some of the harmful bacteria. All right, if you look at what these bacteria produce, for instance, lactobacillus and bifidobacterium  produce GABA, streptococci produce dopamine. E. coli, saccharomyces produces norepinephrine and NMDA. So these are all neurotransmitters.  Whether or not they get into the brain, I don't know. But there are potential changes.  I'll show you what we're working on at the University of Colorado. We have a mouse model of stress. We've stressed mice. We give them a PPI,  and only in the scenario of stress plus PPI, so not either one individually, but only in the scenario of both, we see substantial changes in microbiome.  If we go to the hippocampus and splice it up and centrifuge it and look at RNA sequencing, what we see is a lot of changes in gene expression,  again, only in the scenario of stress and PPI. And those gene expressions, once you map it, start affecting things like dopamine, locomotive behavior, associative learning.  So now we're trying to tease out whether or not these actually contribute to altered cognition of our mice. It's all speculation, but I do believe there's something there.  The other issue with acid suppression, and again, this is mostly PPIs, in fact, versus H2RAs, PPIs are associated with increased kidney dysfunction, or AKI, and increased  cardiac dysfunction. So why? Because PPIs have been shown to reduce nitric oxide production, they increase free radicals.  And while I don't know for sure, and there's not great data in the ICU to say PPIs increase  MIs or increase AKI, I think as we move forward into various studies, we're probably going to have to put these into our studies.  So my take-home message, and hopefully I've confused you a little bit, that was my goal. I think risk factors are variable. I think they're not consistent with current practice.  There's more than just mechanical ventilation. There's more than just coagulopathy. Studies assessing acid suppression, our agents are conflicting.  I would say that it looks like PPIs might be harmful, at least in terms of mortality, even though they reduce bleeding rates.  The most recent DMJ guidelines suggest PPIs over H2RAs, although they did not take into account the PEPTIC study.  I think no prophylaxis warrants further study, especially versus H2RAs. I don't know about enteral nutrition. I can't say whether it's actually beneficial or not.  I don't have great data for that. I will say that maybe using enteral nutrition with stress ulcer prophylaxis is harmful.  And maybe just patients tolerating enteral nutrition is an indicator that the gut functions back to relatively normal, and we don't need to give stress ulcer prophylaxis.  I do worry about the adverse event profile of our agents, especially PPIs. And I would especially encourage stopping stress ulcer prophylaxis when patients are extubated.  And I think the duration of therapy and studies are shortening. You're seeing three or four days.  And so if I was to do a pseudopeco, these would be my risk factors as to where I would start stress ulcer prophylaxis. I would use H2RAs, and I probably would say three days.  What I'm starting to push for now is just like we do with our antibiotics, I think we probably need to put stop dates on our stress ulcer prophylaxis.  So there's no reason why we can't say automatically discontinue stress ulcer prophylaxis after three days.  If patients still need it or we think they need it, you can keep going, just like we do with antibiotics. These are just the questions that the task force is trying to answer.  I put this up here just to show you that there's more than just what the BMJ article was looking at.  So we are looking at things like frequency, combination therapy, what happens if patients come in on these medications, what do we do.  So just sort of a hopefully we'll see something out of the task force in the next year or so, but we're doing a little more.  With that, I'll let Anders contradict everything I just said. Thank you, Rob. Clearly set us up really, really nicely, and especially thoughtful how you brought out  the mechanisms of how some of these adverse events may be actually happening at the more molecular level.  So without further ado, we will now talk about why we should not use stress ulcer prophylaxis. You saw all of your hands went up, but one hand did not go up consistently, and this  man is sitting right here. Dr. Anders Berner, I can tell you that I don't have a lot of lines for him, but he is way beyond most lines that I can speak for him.  He is a professor of critical care at the University of Copenhagen, and we appreciate him coming all the way.  He leads the Center for Research in Intensive Care and also been the chair of the Scandinavian Critical Care Trials Group and section editor for intensive care medicine since 2013.  He's the face behind a number of sepsis, septic shock trials that have come from Europe and have influenced us.  He's also the face behind the sub-ICU trial and a number of other trials looking at stress ulcer prophylaxis in the ICU.  So, get ready for seeing the other side of the coin, Dr. Anders Berner.  Thank you, Ashish. You're way too kind. Thank you very much for the invite. I've spent a few days in the beautiful state of Florida and together with my family, we had a great time.  As I said, this will not be a traditional pro and con debate. It will be more about uncertainty, and as I've already alluded to, I'm uncertain about this issue. I'll show you why.  A few conflicts of interest. Most importantly, I'm a sub-ICU investigator, I was part of the management committee. I'll discuss some issues with you. Is preventive care special?  I think it is. Is intensive care special? I definitely think it is. We'll look at the burden of GI bleeding in ICU patients and the safety efficacy of pantoprasol  in particular, the choice between PPIs versus H2 blockers, and as I said, uncertainty will remain, so this will be my agenda.  Two distinguished colleagues expressed concerns about preventive care. So here's a quote. If a patient asks a medical practitioner for help, the doctor does the best he can.  Sorry, ladies, this was some time ago. He's not responsible for the defects in medical knowledge.  If, however, the practitioner initiates preventive procedures, he's in a very different situation.  He should have conclusive evidence that the prevention can alter the natural history of disease in a significant proportion of patients.  In addition, criteria to consider before embarking on preventive care has been suggested here in a textbook of epidemiology.  We should look at the burden of suffering from the disease that we try to prevent and here particularly the frequency and the severity, and obviously also the effectiveness, safety,  and the cost of the intervention. This intensive care special, for us in Copenhagen, we like to claim that intensive care was invented  in Copenhagen during the polio epidemic in 1952, positive pressure ventilation out of a number of, in particular, kids. Very beautiful photos.  There are many like this quite simple personalized care, whereas today it's much more complex.  I would state that with this complex setting, we probably need better evidence, particularly in preventive care, because we have so many difficulties in seeing side effects and the  natural history of any complications, because we interact a lot. So it's very complex, therefore we need better data than most other settings, I think.  So returning to the burden of clinically important bleeding in the ICU, the classical papers, you saw one of these referenced already.  So 1.5, 3.5% incidences of clinically important bleeding. More recent estimates, you saw these already, 2.6% in a cohort study where many patients  did receive stress ulcer prophylaxis, and then in the placebo group of the sub-ICU trial, we'll probably get a more natural estimate, 4.2% of the patients in the placebo group  experienced clinically important bleeding in the ICU. So not a lot of patients. More importantly, is this associated with mortality?  Again, in the classic papers, yes, in an unadjusted analysis. The second classic paper, both in unadjusted and in a matched cohort, there were associations  between clinically important bleeding and mortality. And in the latest study, yes, in unadjusted analysis in the cohort, but not when we adjusted for markers of severity.  And in the randomized control trial, as I'll show you in a minute, there was no association between the occurrence of clinically important bleeding and mortality.  So maybe the burden is less. So this is the most important question to answer. What's the safety and efficacy of stress ulcer prophylaxis?  I'll spend a few slides on the sub-ICU trial. As said, it was a fairly large trial on pancreas versus placebo in patients at risk of GI bleeding  who were acutely admitted to the ICU. So the European multicenter parallel group, Rheindel RCT, more than 3,000 patients who  were acutely admitted, adults, had at least one risk factor for GI bleeding, randomized to panthroposol, 40 mg once daily, versus saline placebo.  Primary outcome, 90-day mortality, the Kaplan-Meier curve, so no apparent difference in mortality. This is quite sick patient groups, so 30% mortality at 90 days.  If anything, numerically fewer deaths in the placebo group. So I think to answer the safety question in general, there was no apparent differences  in the PPI versus placebo group in this trial on mortality, days alive or their life support,  what we term clinically important in the event, being pneumonia, C. diff, and myocardial ischemia. So overall, it appeared safe.  Rob already alluded to this subgroup analysis, so it was pre-planned that we would look at six subgroups.  One of them were the sickest versus lesser sick patients, dichotomized with a SAHPS-2 score of 53 above or below.  And as you see, there was statistical significant hereditary indicating increased mortality at 90 days with the use of Pantoprazole in the sickest patients.  It's hypothesis generating, it could be a chance finding, but still a worry. So we decided to describe this a bit better.  We did sort of the full analysis of all outcomes in the sickest subgroups, and again obviously saw the same result than in the sickest patients.  So now it's only those with SAHPS-3 above 53, increased risk of death with the use of Pantoprazole versus placebo.  This is a post-hoc analysis, data-driven you may say, still a potential concern. We've done another more detailed publication looking with different types of analysis on  this finding. And one way of looking at it is to look at SAHPS-2 score on a continuous scale.  So that's there in the bottom, mortality on the y-axis and the blue curve is the Pantoprazole group and the placebo group is green.  And you see the sicker the patients, the potentially higher the increase in mortality with the use of Pantoprazole.  The same goes for clinically important events, so that's GI bleeding, myocardial ischemia and infectious complications.  So this composite endpoints the differences between placebo and PBI increases the sicker the patients are, potentially explaining the increased mortality.  So here is sort of a potential explanation, which I guess increases the worry. The same goes for infectious adverse events alone, the sicker the patients, the higher  the risk of a PBI-induced infectious adverse events versus placebo. Whereas with the clinically important bleeding, there was a clear separation between the groups  with more bleeds, it appeared in placebo group across all SAHPS-2 strata. So it does appear to prevent bleeding in all patients. So how effective is this drug if we look overall?  Number needed to treat to prevent one GI bleed, Rob showed this already, 59. If you want to prevent one blood transfusion with the use of Pantoprazole, you cannot do  that based on the data from the sub-ICU trial, because there was no difference and numerically  there were more use of red cells in the Pantoprazole group, so you cannot spare blood with the use of this drug. To prevent one endoscopy, you need to treat 140.  One coiling or laboratomy, 500 cases have to receive prophylactic Pantoprazole to prevent one of these events.  If we change this around, and as I said, there were numeric differences between the two groups in mortality, not a lot, but it's number needed to harm for death at 90 days is 143.  So around the same number as for endoscopies. This has been touched upon already. If you want to use stress-alt-prophylaxis, would you use PPI or H2 blocker?  As I said, there is a recent publication, only a month old, the PEPTIC trial, an amazing undertaking.  It's not that easy to interpret because it's a cluster crossover trial, so the full unit used either PPI or H2 blockers for all patients in whom the clinician found it indicated to  use these drugs, and then the unit crossed over. So this is the primary outcome, so hospital mortality up to day 90.  Again, signal of increased mortality with PPI, now compared to H2 blockers. These are the numeric results in the main outcomes, mortality, clinical board bleeding,  ventilatory associated event, and see not big differences. You see the absolute reductions there. And here are the number of patients needed to treat or needed to harm.  And it does appear that if you were to choose one of these drugs based on this trial result, you would choose a H2 blocker because you need to treat 200 patients with PPI over H2  blockers to prevent one bleed, but you only need to treat 125 patients to cause a death. So that's, I guess, concerning in general. Are these numbers true?  So there was a recent network meta-analysis, I was part of the investigator team, looking at this, and this was just before the PEPTIC results.  So these are the point estimates of PPI versus H2 blockers in the PEPTIC on the right side for you, and in the network meta-analysis without PEPTIC. And they look sort of the same.  So the direction of the point estimates are the same, and the size of the estimates are somewhat the same.  So I think it's probably right that compared to H2 blockers, PPIs may increase mortality in these patients.  As said already by Rob, there was also a subgroup analysis of the sickest patients in the PEPTIC trial.  This was done post hoc as requested upon by reviewers, but the signal is the same as in the super ICU trial.  The interaction here with disease severity and the intervention effects on mortality of PPI versus H2 blockers again indicates that in the sickest patients, PPI increases mortality.  This time it's against H2 blockers. So to wrap this up, I think it's beyond reasonable doubt that prophylaxis use of PPIs do reduce GI bleeding.  They do not go away, but the rate is reduced. But I think there is a big question mark whether these drugs are safe to use in the sickest of ICU patients.  And intuitively and also in the guideline you saw that these are the patients for whom we want to give prophylaxis. So I personally do not use stress ulcer prophylaxis anymore.  If you still want to use it, I would say that you should consider using a H2 blocker. And reserving PPIs for the treatment of those who bleed, because there it's an excellent drug.  The number needed to treat for those who bleed is a lot less. It's 10 to 15. So there it still works. Reserve it for those who bleed. Uncertainty remains.  We do not know the safety of H2 blockers, so we need a placebo-controlled trial in this patient. And I think this is highly relevant. Hopefully someone will be doing this.  If clinicians continue to use PPI for prophylaxis, I think those clinicians should participate  in a trial randomizing patients to this strategy, because I think they are potentially harming their patients. We know very little about patients' preferences. What do they prefer?  How do they look at these number needed to treat, number needed to harm figures? We need to involve them in the decisions. Thank you.  Thank you, Anders. We have about 10 minutes to answer all your questions.  queries, concerns, so come on up. There's two mics over here if you have any questions at all. Please introduce yourself. Hi, I'm Jessie. I'm a pediatric intensivist from Nashville.  I have two questions, one for Pro and one for both of you. The first question for Pro is, did I miss about steroids risk of GI bleeding?  And we have a tendency to put all of our kids who are on steroids on gastric prophylaxis just because they're on steroids. And so if I missed it, I apologize.  But if I didn't, if you could speak to that, that would be great. And then for both of you, I am a pediatric intensivist. So have you come across anything in the PEAS literature  about how this might change for critically ill kids? Thanks. So in terms of steroids, it's more of a theoretical potential that it increases mucosal damage.  There's not any data that I'm aware of that says that having patients on steroids, regardless of the dose, actually increases bleeding risk. It's more theoretical.  Can I answer that? So we asked this question ourselves  and found all steroid trials versus placebo in ICU patients to look at reported GI bleeds in these trials. So there's obviously a lot of trials on steroids.  So potentially, we would get a good estimate. But it appears that the rates of GI bleeds reported in the trials are fairly low. So I think reporting of side effects, even in trials,  is pretty poor. But from the estimates, there may be a slight increase risk of GI bleed in general, but not in critical bleeding.  But the uncertainty around those estimates is very high. So it's not a gigantic risk. There may be a slight risk with some uncertainty. And just to follow that, so if you only  have a kid on steroids, I don't think that's a reason to use PPIs or H2Os. Now, if they're critically ill, that's probably a different scenario.  And going to your second question, there are studies in the pediatric ICU world, but not nearly to the same number of patients, the population.  So most of them are looking at serograde outcomes like pH control, PPI versus placebo, or PPI versus HRA. If you're lucky, they might have 100 kids in them  and not any more than that. So I wouldn't say there's any great outcome data. Yeah, we will submit a systematic review on this within a few months.  And I think there are seven trials, very low quantity and quality of the evidence. So we really need a trial in kids. I'm not sure we know the incidence of GI bleeds  in pediatric ICU patients either. So they may be at lower risk. Yes, sir. Bruce Friedman, JM Stillburn Center. Georgia. A couple of questions.  For H2 blockers, if that would be the recommendation, how do you deal with the problem of well-known of the tachyphylaxis that occurs within 24 to 48 hours  in terms of their responsiveness to protecting the gut? Second question is, based on your evidence, how do you explain the mortality since there is no statistical difference?  You manipulate the numbers, but the statistical significance is not there with regards to PPIs and mortality. And third, if you would ever recommend carafate,  how do you explain giving a phosphate binder to mechanically ventilated patients that would then deplete their phosphorus already, which is always a problem in these mechanically  ventilated critically ill patients so it will worsen the hypophosphatemia that they already have? So those are the three questions.  With respect to H2RAs, I mean, the tachyphylaxis usually is more of a 24, 48 hour thing. And from my perspective, I'm looking really to give it while the patients are the sickest.  So usually it's in the first 24 or 48 hours. And then I recommend usually trying to stop it as quickly as possible.  And I think H2RAs have some other effect beyond just acid suppression. I think it limits reperfusion injury, et cetera, et cetera. So there's other reasons to give an H2RA that  still need to be explored. With respect to a phosphate binder, we don't give cercophate that often. So are you referring to using like a cercophate  or an antacid or something that binds the phosphate?  No. Yeah. So I'm not suggesting we use Sucrofate. There's not great data for Sucrofate. And again, I'm using an H2 blocker short term, 48, 72 hours, as patients get better.  So I wouldn't at all suggest moving to Sucrofate, and if anything, the data would suggest that Sucrofate's inferior for preventing clinically significant bleeds. All right.  We'll take the next one. My name is Zach. I'm a pharmacist from North Idaho. And just two quick questions.  Is there any association with these acid suppressive agents in contributing to delirium and the poor outcomes associated with delirium, such as older patients and anticholinergic  effects and with H2s? And then second, how you might handle a STEMI PCI patient on dual antiplatelet therapy, and there's cardiology data to suggest they should be on the PPI.  I only heard the first part because there was an echo going on, but the delirium portion of it.  There is outpatient data that suggests that PPIs increase the incidence of delirium or dementia in the outpatient world.  Now, obviously, it's a little hard because it's a chicken and egg situation. It's all cohort data.  My belief is that change in the microbiome might actually contribute to the change in delirium. Now, I don't have any data to support that in the ICU, but that's what I'm trying to  work on. And then the second part was something about STEMIs or something. The dual antiplatelet therapy that STEMI patients are on, if they should be on a...  I mean, I guess it's a theoretical drug interaction from that standpoint. But again, I'm promoting H2As, not PPIs, so I would put those patients on an H2A.  And again, the clinicians who continue to use PPI in these patients should probably do a trial to show that it's beneficial. Moving quickly.  I want to tie in with that dual antiplatelet. Glenn Whitman from Johns Hopkins.  I'd like to tie into that dual antiplatelet therapy question. Patient has acute coronary event, ends up getting cardiac surgery, and now you're immediately  putting them on aspirin and a P2Y12 inhibitor, but they've had cardiac surgery and frequently their hearts are not normal.  Could you please discuss H2 blockers versus PPIs just in terms of their toxicity to myocardial function and then also address the issue of the guy's doing well, but he's just been  given 60,000 units of heparin, he's on the ventilator for however long, is there a population that you would be treating for ulcer prophylaxis and would you use an H2I versus a PPI?  So again, I would use an H2A no matter what, up front even. So I wouldn't use a PPI even a dual antiplatelet.  I think there's, I mean, the drug interactions, it's real, but it may not be as clinically significant as what we think.  And then actually in those patient populations, the bleeding risk is extremely low. If you look at the cardiology literature, it's pretty low.  But in the long term, I would put that patient on H2A, not a PPI. And myocardial toxicity? I don't know of any great direct myocardial toxicity with H2A that would be significant.  Versus PPI? Versus PPI, yeah. So myocardial infarction was, or myocardial ischemia was an endpoint in the SUP-I2 trial  and numerically there were more myocardial ischemic events in the PPI group. So there is still a concern, I think.  And pathophysiologically, it probably makes a little more sense that a PPI would be associated with it than an H2A from a nitric oxide release, et cetera, et cetera. Thank you.  Last two questions. Good morning. My name is Jessie. I'm a pharmacist here in Florida. I have a quick question for you about our population.  We have a lot of elderly patients, either self-prescribed or already taking H2RAs or PPIs prior to hitting our ICU. We don't always know why they're on them.  But either way, how does that play a role in whether you titrate off after three days or don't prescribe at all in your settings? Yeah, that's a great question.  It's one of the times where I still get a bit wobbly. So I tend to take them off. But we probably need better data on that.  And that would be another great research question for a trial. Those who are on it, for better or for worse, what should we do with those patients?  I think the PEPTIC study did not discern. So if you came in on a PPI and it was the cohort that was being randomized to H2RA, you got H2RA.  I shouldn't say the cohort, but the ICU was randomized to H2RA. You got an H2RA even if it came out on a PPI.  So I'm comfortable changing to an H2RA provided the PPI wasn't for a true erosive esophagitis, recent bleed, something of that nature. I'm comfortable changing to an H2RA.  And we did that too in the super ICU trial. So we took patients off if there was no good reason. At least in some centers. I cannot account for all centers.  So there is still uncertainty about that. If you go to nothing, I think there is some uncertainty. But again, if you use a H2 blocker, I think it's fair to switch them to that.  That's actually really interesting. And we'll hold this for afterwards. But the chronicity of being on the acid suppressing agent, some patients walk into your ICU, they've  been on it for 20 years versus someone who's been on it for like two years. So it certainly is some difference in the two patients, right? One way or the other.  We don't know the answers to that part yet. We don't. Yeah. Okay. Last question. Hi, I'm Sarah. I'm a pharmacist. I'm from Pittsburgh. And my question is for the pro side.  You allude to short durations of prophylaxis, like two to three days, and then evaluating for discontinuation or continuing if risk factors are still present.  So I'm wondering specifically about mechanically ventilated patients, do you continue prophylaxis for the duration of them being on the ventilator?  Or do you discontinue at some point, for example, when they start enteral nutrition? Yeah. So that's a good question. So I don't know the answer with enteral nutrition.  I think enteral nutrition is a surrogate marker to say that patients, their gut ischemia that's causing the ulceration is starting to reverse.  So to me, it's a surrogate marker to say the patients are physiologically stable enough that you can probably stop it.  I don't know if the enteral nutrition itself is doing anything to prevent stress ulcer or bleeds. But I think it's probably a surrogate marker.  But I think not all ventilator patients are the same, right? Somebody coming in, they're on 100%, 20 a peep, maybe they're prone.  That's a different patient than somebody who's on 50% and 7 a peep ready to be moving towards extubation. So I think you really need to look at the patient.  Even if somebody's on 50% and 10 a peep, I think it's probably OK to go ahead and stop. They're looking better. They're not truly critically ill anymore.  They're transitioning to that recovery phase, ready to be extubated. So I'm comfortable stopping those patients. And the vasopressor part, does that play into all of this?  Again, it depends on what the rest is going on. If they're on a low-dose, you know, 5 mics a minute, 10 mics a minute of norepinephrine, that doesn't concern me.  Now, if it looks like they're getting sicker or they're on 50 mics a minute of norepinephrine, that's a different scenario. Andrew, do you have a different?  I mean, we basically stopped using stress-off prophylaxis in all patients. There are some primary, some of our, I mean, there are some of the referring teams that  are really keen on renal, the renal nephrologist, for instance, and some of the surgeons. And they all will agree. But if we were to decide all of it, we would stop all.  Still a discussion point, but I'm happy to stop in shock. We should get Dr. Perner to come work on this side here with us, save us a lot of money in our ICUs across the country.  I thank you all. This has been great. Thank you for the speakers. Thank you.

Video Summary

The debate discussed the use of stress ulcer prophylaxis in critically ill patients. The pro side argued that stress ulcer prophylaxis with medications like Pepcid or Protonix is commonly prescribed in ICU patients, but that it may not always be necessary. They emphasized that about 50% of ICU patients receive stress ulcer prophylaxis, but only about half of those patients actually need it. The speakers discussed the potential risks of unnecessary stress ulcer prophylaxis, including drug interactions, increased risk of pneumonia, and altered gut microbiome which could contribute to delirium and other complications. The con side argued that stress ulcer prophylaxis should not be used in critically ill patients, especially those who are considered to be the sickest. They presented evidence that stress ulcer prophylaxis with proton pump inhibitors (PPIs) may increase mortality in these patients. They suggested that H2 blockers should be used as an alternative if stress ulcer prophylaxis is deemed necessary. Overall, both sides agreed that more research is needed to clarify the risk and benefits of stress ulcer prophylaxis in different patient populations.

Asset Subtitle

GI and Nutrition, Hematology, 2020

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 2020 Critical Care Congress held February 16-19, 2020, in Orlando, Florida, USA.

Learning Objectives

Apply the results of the SUP-ICU trial and understand its implications for management of ulcer prophylaxis in the ICU
Discuss the negative effects of unnecessary stress ulcer prophylaxis in the critically ill population
Determine when and when not to use stress ulcer prophylaxis in the ICU

Pro: Stress Ulcer Prophylaxis: What Are We Really Trying to Achieve?

Robert MacLaren

Con: Life After the SUP-ICU Trial Results: Prophylaxis for All?

Anders Perner

Meta Tag

Content Type Presentation

Knowledge Area Hematology

Knowledge Area GI and Nutrition

Knowledge Level Intermediate

Membership Level Select

Membership Level Professional

Membership Level Associate

Tag Peptic Ulcer Disease

Tag Hemorrhage

Year 2020

Keywords

stress ulcer prophylaxis

critically ill patients

ICU patients

medications

risks

complications

mortality

PPIs

H2 blockers

	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.