Okay, here we go. Good morning, everyone. Welcome to the Hepatotoxicity SIG December seminar. It's the final seminar of the year. We are very delighted to have Dr. Dean Karvelas join us. To anybody who studies acetaminophen and does translational research and transplant, Dean is very well known. He is at the University of Alberta, and I just realized that he is a professor in the Department of Critical Care Medicine, but he has a joint appointment with Gastroenterology Hepatology. He is a hepatologist and is involved in the liver transplant program as well. But as he told me, he's primarily an intensivist, and that's the majority of his practice. And of course, he's a very well-known researcher. And he has been very active in the Society of Critical Medicine, and he's a co-chair on the liver, recently published liver management guidelines and the ASLD. He's the chair of the ACLF SIG and a co-chair of the recently published ACLF guidelines. He is an associate editor for the Journal of Hepatology. Dean is the only Canadian co-investigator on the recently completed NIH-funded ALFSG group. I met him through Dr. Will Lee, and a lot of you know him through his work with the ALFSG, which a lot of us hepatologists read and enjoy. He has over 150 peer-reviewed publications, and his career has really been focused on ALF, acute and chronic liver failure, and transplant in the setting of liver injury. And today, since we're in the hepatotoxicity SIG, he's going to tell us about his work and focus on acetaminophen and idiosyncratic delay, compare and contrast what we know regarding a transplant. Welcome, Dean. Thank you, Lily. And it's a real pleasure to speak to this SIG through this webinar. And so I thought for this topic, what I would do is kind of bridge you through the story, and I know that this is a very erudite, sophisticated audience, but kind of bridging through the story of kind of some of the differences between ALF from APAP and DILI, idiosyncratic DILI, I should say, and then particularly look at some of the differences in management and then kind of also relating it to that impact on transplant. And this is kind of an outline of what I will be kind of presenting today. So to kind of start with, I know that we've got kind of a range of people in the audience today. So we know that ALF comparing to patients with cirrhosis and ACLF, while there are similarities in some of the different complications, really the thing that we tend to talk about, particularly with hyperacute liver failure, in particular acetaminophen-induced ALF, is the intracranial complications, partly because of the rapidity of onset. And this obviously relates to the fact that ALF is a necroinflammatory process with a large amount of hepatocyte dropout, where cirrhosis is a chronic fibrotic condition, inflammatory condition in some cases, where it can take months to years to develop. So this is probably one of the most important slides that we're all familiar with, is that acute liver failure, or by definition, you know, synthetic liver dysfunction with the development of hepatic encephalopathy, can be kind of subdivided into different categories. And really on one hand, we've got hyperacute liver failure, where classically this is acetaminophen toxicity, where the timeline from the development of a liver dysfunction to the development of encephalopathy is very short. These are patients that have a high burden of multi-organ failure, often are admitted to the ICU for management, have a significant potential for the development of intracranial hypertension, but paradoxically also have the best chance for recovery. And these are the ones where, particularly with better neurocritical care and ICU supportive care, more and more of these people are walking out of the hospital. And it's a proportion of these patients that end up going for transplant. And I will say that that proportion of patients that you end up making this decision about transplant can be incredibly challenging. At the other end of the spectrum, as mentioned, is really what you're all very familiar with, is idiosyncratic drug-induced liver injury, which in some cases patients initially present to their, it might even be to their primary care provider, where it can in some cases follow an indolent process over several weeks, where these patients tend to get very cholestatic. And it's often several weeks later that they develop encephalopathy. And while they don't necessarily have the same burden of multi-organ failure, and some of these patients even initially are admitted to the ward setting, the potential, once they meet poor prognostic criteria, of recovery without a liver transplant is significantly lower. Meaning in some cases it's more of a, the decision to go ahead with transplant is not so uncertain. So the reason why I became interested in this condition is I always felt that ALF is a multi-system condition, and I call this the classic intensivist disease. And in many cases there's potential for reversibility. So patients can develop neurologic complications, and the one that we talk about the most is cerebral edema and intracranial hypertension. They develop a high-output cardiac output state. Patients can develop acute lung injury, and often this can be a challenge, particularly in somebody that's on the transplant list that is very unstable. We are all well aware of the hepatic complications of ALF. I just highlight them there. We'll talk a little bit more about hyperammonemia and why it's important. They can develop pancreatitis, adrenal insufficiency, AKI, bone marrow suppression, and septic shock. And this is another thing that I'll go through in a little bit more detail, where particularly in your subacute DILI patients, they can develop a compensatory anti-inflammatory response where in fact they're immunocompromised and they're at high risk of sepsis and why bacteremia is more of an important predictor of outcome in idiosyncratic DILI ALF. So I can't give this presentation without giving credit to Will Lee. I met Will back in 2011, and he has been a wonderful mentor to me. And for any junior faculty that are on this call today, Will was kind of the trifecta to me of mentorship, data, and funding, and it's very rare to get that in today's world. But Will was a true pioneer in this, given the fact that ALF is an orphan condition. And one of the few ways to study orphan diseases is through networks. So he started the U.S. Acute Liver Failure Study Group back in 1998, and we completed enrollment in 2019. However, we have been continuing to have investigator calls even to the present day. So in terms of looking at this registry that's now more than 3,000 patients, you can see that almost half of the patients in this registry, and this is very reflective of North America and Europe, is due to acetaminophen toxicity developing ALF. And probably the second most common cause of ALF in the registry is drug-induced liver injury from idiosyncratic causes. And once again, this is likely reflective of North America in general and furthermore Europe. I will say too that there is a group of patients that we call other or indeterminate. Because of recent causality work that's been done by the group to try to do further kind of biomarker testing to elucidate if there are patients here that truly have a cause of ALF, we used to quote the rate of this at about 5%, but it's actually probably less. But what we do know is that these seronegative or indeterminate patients tend to follow a similar course to DILI patients, where they follow a subacute process. And once they develop kind of poor prognostic criteria, most of them need to undergo a liver transplant. So I know that probably within the liver toxic, that you probably talk about cytochromes all the time, so I won't discuss this in too much detail today. But we know that the cytochrome P450 pathways lead to unstable compounds. And with acetaminophen toxicity, NAPQI is really what we talk about that we're concerned about is the highly reactive oxygen species, that if it is not dealt with with the phase II reactions in the liver or with glutathione, that this is what leads to hepatotoxicity. For example, we know that particularly with staggered toxicity with acetaminophen, acetaminophen levels might be negative, but more work has been done on things like acetaminophen adduct, particularly with Laura James's group at the University of Arkansas, potentially looking at a point of care test where we can test for acetaminophen adducts. And this is really why we often make decisions about using things like N-acetylcysteine based on the phenotype and rather than the acetaminophen level, because almost 50% of patients with APAP-ALF will have a negative acetaminophen level. So I think this is, I mentioned kind of the dichotomy between hyperacute and subacute liver failure. This is also probably highlights the difference between APAP and idiosyncratic DILI in terms of when you see this hyperacute phenotype with APAP-ALF, you know, high burden of multi-organ failure when they come in, you know, more than 50% of patients in the registry had high-grade hepatic coma, in most cases meaning patients needed to be intubated for airway protection, massive amount of hepatocyte necrosis with, you know, with an ALT of greater than 3,500. And I'll mention that in terms of an implication with biomarkers down the road. In contrast to this, you tend to see that the idiosyncratic DILIs, which are more slow, you know, progressive, more cholestatic, tend to have a higher median bilirubin. But probably the most important statistic is that it's a minority of patients with APAP-ALF that get transplanted. It's almost less than 10%. And furthermore, if you look at their transplant-free survival, it's almost 70% with APAP, where in contrast with DILI, more than half of these patients are listed for transplant. And in patients with poor prognostic criteria, the transplant-free survival is only 30%. It's significantly lower. So this was one of the first reasons why people have now discussed the concept of favorable and unfavorable etiologies in ALF. So traditionally when we talk about favorable etiologies, it's acetaminophen, hepatitis A, ischemic hepatitis, where really it's related, the liver is more the bystander, and then fatty liver of pregnancy. In contrast to that, the unfavorable etiologies are DILI, indeterminate, hepatitis B, and autoimmune. And this factor is one of the factors that's in the ALFSG prognostic index that was published back in 2016. Furthermore, though, to take it a step further, hepatic encephalopathy coma grade is also an important factor, where it's easier or kind of best to stratify patients between 1 and 2 and 3 and 4 because this is more of a more objective, meaning that somebody who's grade 3 or 4 coma is generally somebody that is barely rousable or completely a GCS of 3, meaning these are patients that need airway protection, where sometimes the difference between grade 1 and grade 2 can be a little bit more subtle. So here you can see that irrespective of whether you had a hyperacute or a subacute etiology of ALF, having low-grade coma generally portended a better outcome. So the example being with acetaminophen toxicity, if a patient is grade 1 encephalopathy and talking to you, there's a high chance that they're going to walk out of the hospital without a transplant. So we know that, you know, there's a large amount of literature that goes back to the 90s even for the benefits of N-acetylcysteine, where we are priming the glutathione pathway in APAP ALF, and we know that there's data showing decreased rates of intracranial hypertension, improved outcomes, et cetera. Really where one of the other questions came up is, you know, the example being what about in idiosyncratic DILI or other non-APAP causes of ALF? And this is a little bit trickier. So through the US ALFSG, we all led a randomized study of N-acetylcysteine in non-acetaminophen causes of ALF. This was back in 2009. And what they demonstrated was that in patients with a more favorable outcome, and this was essentially patients with a lower coma grade, that those patients that went on N-acetylcysteine actually had a better outcome. If you do look at the data, one of the criticisms of this paper was that the difference between a significant and non-significant outcome was only a few patients. But the challenges that we have to highlight is that ALF is an orphan condition, and it's very difficult to generate statistical power with randomized trials. So really, pragmatically speaking, if we have a DILI patient with ALF and encephalopathy, then we will put them on NAC. Probably the bigger question is how long do we run it for? There used to be kind of biochemical metrics like an INR less than two, for example, where you could stop it. In reality, there isn't really a lot of good data to support this. So in most cases now, we're running it for about five days at the most. So really the real concern in ALF is the development of intracranial hypertension and cerebral edema. But because of improvements in medical care and identification of these patients, one of the positive factors is the rates of ICH, and particularly death from intracranial hypertension, has actually been declining. We reviewed this with a whole ALFSG database back in 2020, and what we found is that the prevalence of ICH is dropping, which is when you look at the two decades, and also mortality from ICH. So really to remind us again that there are kind of two important pathophysiological pathways. There is astrocyte swelling leading to cytotoxic edema, which we talk about the rapid accumulation of ammonia, and in turn the accumulation of glutamine and astrocytes. You get disturbances in cellular metabolism and ionic gradients, and the brain itself is not able, or astrocytes in particular, to compensate by expelling other organic osmolites, where in contrast to patients with cirrhosis, where this might happen at a molecular level, given the duration and time course of this, that's why we don't tend to see intracranial hypertension that's clinically significant, because of the compensatory mechanisms. The other factor is the massive hepatocyte necrosis leading to production of damage-associated molecular proteins and patterns, or DAMPs. This leads to vasogenic edema and loss of autoregulation. You get breakdown of the blood-brain barrier and extravasation of macromolecules into the plasma. We mentioned that ammonia and ammonia levels are one of the few biomarkers that gives us useful information with regards to discriminating patients at high risk of developing intracranial hypertension that's clinically significant versus those that don't. I highlight this is an old study now, it's back to 2007, but what Will Brunel and colleagues at King's were able to demonstrate was that the independent factors that were associated with the development of ICH were an ammonia of greater than 150. We use this as a cutoff now with regards to potentially either instituting neuromonitoring techniques, and furthermore, blood purification techniques like continuous renal replacement therapy. Not surprisingly, the other factors that were independently associated really reflect this massive hepatocyte necrosis that you get with hyperacute liver failure, so APAP patients more likely, and other reflections of this overwhelming release of DAMPs, vasoplegia, high-grade hepatic encephalopathy, and the need for renal replacement therapy. Up till recently, a lot of our neuroprotective strategies in the ICU were borrowed from the general critical care literature, but there's kind of more and more direct data, particularly with regards to targeted temperature management that we talk about now, as opposed to therapeutic hypothermia, hypernatremia, and furthermore, the role of blood purification strategies, and in particular, continuous renal replacement therapy. So one of the things that has changed has been kind of the shift to doing less direct intracranial pressure monitors and using more noninvasive techniques like transcranial Doppler, optic nerve sheath diameter, and this kind of highlights also the more widespread use of point-of-care ultrasound in the ICU setting. So I highlight this study that we published now almost 10 years ago now, where this was a cohort study looking at direct intracranial pressure monitoring, and we wanted to really highlight what is the role of this in high-risk patients. And what we found in terms of, you know, there were about 630 patients with high-grade coma in the registry. About 143 of these patients had some form of ICP monitor. Most of these are bolts or kind of subdural catheters or interparenchymal monitors. We're not really talking about putting in extraventricular drains. About 50% of patients that had a monitor in place had clinically significant ICH. So on the good side, it means that, you know, more than half of patients that got a monitor had high ICH. And really what we found was that often, you know, these monitors are being put in where you're trying to make a decision about whether or not somebody needs a transplant or somebody's on the list. And the example being is the intracranial pressure manageable at 20 or 25, but it's high, or is it 60 where really it's a stop sign? And that's really what most centers were doing. And also those patients that had monitors in place received more, you know, therapies to mitigate intracranial hypertension. But the monitor itself did not change outcome. And it kind of highlights to a lot of things that we do that there's, first of all, we're talking about a monitoring technique where there are multiple management strategies that have already been implemented. So this is the incremental effect of a monitor. But probably more importantly, it means that this is personalized medicine, that it's not really for everyone. And primarily we use this now in patients that have abnormal noninvasive findings. So if they have a TCD that shows abnormal, you know, a calculated ICP that's elevated, particularly in a patient that is potentially needing a transplant, so it will alter your management decision. So, you know, generally also, you know, the people that we were putting these bolts or Codmans in, generally they're patients high-grade coma. Most of them are APAP patients. They meet poor prognostic criteria. They're already receiving kind of ammonia lowering therapy with CRT. And generally we tend to see this more in younger patients under 50 because they have less cortical atrophy. It's actually generally rare in patients over the age of 60 to develop intercranial hypertension just because we tend to develop cortical atrophy with time. So as mentioned, you know, I'll highlight that some of these noninvasive techniques, particularly using transcranial Doppler and looking at the pulsatility index of the middle cerebral artery have really added information. Although I do admit that we still need, you know, larger validating studies. So the example with this is that really what we're looking at is, is markers of impaired auto-regulation that you really have impaired regulation of the cerebral vasculature in response to changes in profusion and cerebral blood flow. And the idea is that you, when the, when the brain gets, when you get swelling of the brain and, and things get tight in the skull, that you lose this elastic reservoir, and we call this the Winn-Kissel effect, meaning dampening the disparity between the systolic and diastolic part of your waveform. So the example being on this on the left, generally what we measure is something called the pulsatility index, which really graphically is that this patient on the left, which there is this kind of short, fat waveform, implies that you have fairly good elasticity in the brain, and the ICP is normal, whereas the brain gets tighter and you lose this elastic reservoir, you get a narrow waveform, you lose the dichrotic notch of the systolic waveform, and furthermore, you get a narrowing of this. So the other way also analytically that we measure this is actually with an area under the curve, which when we do that, you can see very clearly on the right that the area under the curve is significantly less. So really, in terms of evidence for this, this was one study that came from France, and really what it highlighted was that those patients with an abnormal pulsatility index of greater than 1.2, essentially virtually all of those patients either died or ended up needing transplant. Another technique that's also used, and just highlighting again that these patients are generally high-grade coma intubated, is doing an optic nerve sheath diameter where you take the linear ultrasound probe and you put it basically across the eyeball with gel and with the eyelid closed. Generally, what we're measuring here is the width of the optic nerve sheath, about three millimeters behind the retina. As you can imagine, there is some inter-observer variability with this. And for that reason, generally we take four measurements, two perpendicular measurements in each eye. And there is some data to show that this correlated with survival from this study from colleagues of mine in Portugal. And generally, what has been published from normative data, what is the upper limit of normal is generally five millimeters. But I can kind of state that in terms of doing this at the bedside, differentiating six versus five millimeters can be very challenging. And that's why we tend to rely more on TCD than optic nerve sheath. So in looking at these together, I know that you all know Bob Fontana very well. The Neurocritical Care Group looked at this at the University of Michigan and kind of highlighted a few things that optic nerve sheath diameter, if you look at the area under the curve alone in terms of detecting an ICP of greater than 20, the area under the curve was only about 0.59, which is not great. And also if you only use the pulsatility index by itself, the area under the curve is not great. However, when you actually look at flow velocities, and this is where you would measure that area under the curve and what we call the velocity time integral, that when you incorporate more of this information you're getting from the TCD, your area under the curve goes up dramatically to almost 0.9. So nowadays, most of the time, this is our initial screen in a patient before we would make the decision of doing something invasive. So one of the things that we'll highlight is that so far to this point, there are neuro biomarkers besides ammonia that have been measured in the blood. And we know that probably how we measure these things, what we're measuring in blood is probably different than what's in the cerebral spinal fluid. People have looked at neuron-specific enolase at S100B. Our group looked at, with Chris Rose in Montreal, looked at brain type fatty acid binding protein. And to date, none of these markers really discriminated those patients that went on to develop intracranial hypertension from those that didn't. So we don't, beyond ammonia at this point, we really don't have another good biomarker that we can measure in the blood. So the other thing that we see, there was a case study that came out from Rajiv Jalan's group in 2003 that said moderate hypothermia, cooling people down to 34 degrees, saw improvement in cerebral blood flow, significant reduction in intracranial pressure. And this was also a story that we saw in a parallel population in critical care in the cardiac arrest literature. And one of the challenges is that when you, particularly when you cool somebody that's number one, hemodynamically unstable and number two, potentially coagulopathic, it makes their management incredibly difficult and there are other complications. So, you know, there have been, you know, I highlight two studies of moderate hypothermia. One was a cohort study that we did, which has a lot of limitations because patients weren't randomized and management otherwise was at the discretion of the investigator. And we were not able to demonstrate a benefit with moderate prophylactic or preventative hypothermia if you did it for all comers to prevent the development of ICH. However, you know, there's a lot of biases potentially in that type of a study. However, Will Burnell and colleagues at King's with colleagues in Denmark as well, did a prospective randomized study, numbers are small because orphan disease. And even after stratifying patients between 34 and 36, there was no difference in clinically significant outcomes. But what it probably highlights is that, you know, generally actively treating fever and keeping somebody's temperature around 36 is probably important, which means that we have to actively treat this and we don't ignore the temperature. And I always kind of highlight this to trainees. And one of the things that we do with more widespread use of blood purification strategies, like continuous renal replacement therapy, is if I put somebody on a circuit and I turn off the rewarming circuit with the replacement fluid, I can actually passively cool patients to 36. And that is probably reasonable. And that's something that we do kind of in the preventative setting. And that's how we achieve kind of targeted temperature management. I will say in terms of the personalized medicine approach, if somebody blows a people in front of you, we will in some cases, you know, actively pursue therapeutic hypothermia in the salvage or rescue setting if you don't have any other options. So one of the other things that I mentioned before, how like point of care ultrasound is also helping us because one of the biggest challenges in a lot of these ALF patients is they come in with a lactic acidosis and you're trying to figure out is it the liver or is it perfusion? And using bedside echocardiography really helps us with this. One thing to mention too, sorry for the slide issue, is that norepinephrine is kind of our first line treatment. Generally, if there are concerns particularly with the hyperacutes that they might have intracranial hypertension and you don't have a way to measure this, you can target a higher immune arterial pressure of 75. So one of the things that we will say too is that one of the other kind of organ failures that tends to really impact outcome is lung injury. And this is why it's also important that if you are going to secure the airway in a patient with ALF, that you probably have somebody with significant expertise because if a patient aspirates and then develops severe lung injury, you get two problems that can make it very difficult to proceed with transplant on a high FiO2. And furthermore, the other factor is that we tend in acute respiratory distress syndrome, we tend to use lung protective strategies with where we let the CO2 climb. And in ALF, we can't do that because of the brain. So you end up with competing organ failures. And as you can see from this study that Victor Dong published with us a couple of years ago, that those patients that develop lung injury, and we've seen anecdotes of this where we have patients where they, we weren't able to proceed with transplant, where essentially your survival was less and you were less likely to get a transplant. So probably one of the biggest thing that's kind of changed over the last kind of 10 or 15 years is blood purification. And the one thing in particular is the use of continuous renal replacement therapy, not in acute kidney injury settings, but actually in the setting of as a blood purification tool for things like hyperammonemia, lactic acidosis, and volume overload. And the reason why we like continuous renal replacement therapy or low and slow is that you minimize solute shifts, you minimize hemodynamic changes, or you don't dump their pressure, which can affect the brain. And you kind of mitigate ICP changes that way. I will mention that intermittent hemodialysis with higher blood flow rates, you will get higher ammonia clearance, and it might be appropriate in a urea cycle disorder in the setting of somebody without acute liver failure, but in ALF really it's been demonstrated to exacerbate intracranial hypertension and brain injury, and that's why we avoid it. And what I wanted to highlight is that, while this is a practice that's probably been going on for 10 years, it was this nice study that Phil Cardoso published with our group in 2018 that showed that, as we know, that generally patients that are sicker, we will do more interventions. So in a cohort setting, you need to adjust for severity of illness. This was a cohort of almost 1,200 ALF patients with and without acute kidney injury. And after you adjusted for things like severity of illness based on organ failure, the year of enrollment, if patients received intermittent hemodialysis, which can dump your blood pressure, your odds of mortality actually went up by 70%, where patients that went on continuous renal replacement therapy, which were in the more recent era you had a reduction in mortality by almost 50%. So now this, particularly in a patient with an ammonia greater than 150 is our standard of care. There's been more data that's come from Ronaldo Balomo's group in Australia that also shows that if you are gonna do it, we do it early, that you get more bang for buck. And also while in, just for reference, in the general critical care population we're talking about potentially a dose or an effluent dose we talk about of about 25 mils per kilo per hour. We tend to run this at significantly higher levels in ALF patients. And what they showed in Australia is that we're doing it at a effluent dose of almost 40 mils per kilo per hour. And just to mention furthermore that this is a systematic review. Obviously this was dominated by that one study, but there are also other studies that demonstrate the benefit of CRRT over not. Another thing, this is more for us as intensivists that prescribe renal replacement therapy. You know, there's two ways you can remove solutes. That's either through diffusion which is traditional hemodialysis, or you can do hemofiltration with convective clearance. And probably in the setting of ALF, it really doesn't matter. And there are other factors that will impact our decision in terms of what kind of a circuit to use. And really the one example of a simpler circuit with just continuous hemofiltration or CVVH is that simpler circuits don't tend to clot as much. So we can often get away without running any anticoagulation. So just for your reference, you know, the use of CRRT has made its way into multiple guidelines including the Society of Critical Care guideline, the American College of Gastro guideline, and also the ESL clinical practice guideline. Now, I'd mentioned that the other factor in ALF is that we also build up a lot of these damps and a lot of these toxins and damps are not water soluble. And in fact, they're protein bound. So things like CRRT will not clear these. So not surprisingly, there has been interest for many years in terms of albumin dialysis and other, you know, liver support or blood purification circuits to try to clear these to potentially mitigate the development of other end organ complications like cerebral edema and circulatory dysfunction and vasocollision. So, you know, for a long time, different parts of the world, we were trialing the MARS albumin dialysis device or the molecular absorbent recirculating system. And this was the largest study that was actually done in France. It was done by Fauzi, Saliba and colleagues. They did the study in a year. So to enroll 102 patients in 16 centers in an orphan disease is very commendable. One of the differences that you can imagine certainly compared to the US and Canada as well is that the organ donor rates are significantly higher in Europe. So the median time delay from listing to transplant was only 16 hours. So that was a significant confounding factor in the patients randomized to MARS where about 14 of the patients never completed their first run. As I say that this turned out to be a negative study, although potentially there was an underpowered effect in APAP patients that I haven't shown you here. But really what's changed recently, and I will mention this, that when we looked at MARS in the North American setting, the only way we were able to do this, this was done with my former grad student, who's now Andrew McDonald, who's a transplant fellow at Mayo, was that we did a propensity match study of the three centers that were running MARS, which was Edmonton, us, Atlanta, Emory, and the University of Kansas. And then we propensity matched them to patients in the registry, in the LFSG registry that didn't get MARS. We actually demonstrated that there was a benefit with the MARS circuit, even after adjusting for CRRT. And probably the reason for this was that the confounding factor of transplant was significantly lower, where almost three quarters of the patients in the full MARS study from France got transplanted, where it was only 22% in this North American propensity match study. So what has actually changed this, and why MARS is actually being sunsetted by Baxter, is that plasma exchange has also demonstrated to remove these DAMPs and PAMPs and contributed to outcome. And I kind of previously mentioned this kind of dichotomy of differences between DILI in terms of bridging somebody to transplant versus hyperacute or APAP patients, where there's a higher potential to bridge to spontaneous recovery. So when this study was done by, it was done by Finn Larson and colleagues in Denmark, along with Finland and the UK, was that they showed in patients that underwent transplant, and the vast majority of these patients had unfavorable etiology, and in particular DILI, plasma exchange likely did not impact outcome, and it wasn't statistically significant. And part of the rationale for this is also, you don't get the massive necrosis ongoing like you do in an APAP patient, where in patients that didn't get transplanted, where the vast majority of these patients were acetaminophen patients with a high kind of DAMP load because of massive necrosis, but also a potential recovery potential that is high, the use of plasma exchange was associated with a mortality benefit. So for example, in our center, for the treatment of significant vasoplegia, particularly in APAP or hyperacute ALF, this is a strategy that we use, we'll use plasma exchange. Now, if you, and in terms of a mechanistic benefit of this, the late Harry Antonides, who unfortunately passed away five years ago, had the pleasure of working with, they did a nested cohort study looking at particular DAMPs in ALF patients, and they were actually able to demonstrate a significant drop in many of these DAMPs, including histone-associated DNA, TNF-alpha and IL-6 in patients that got plasma exchange compared to those that didn't. One of the criticisms of that study, and there is a paper that just came out in hepatology, and I just do that to make you aware, also has to do with the volume of plasma. They were suggesting, you know, two and a half blood volumes, which if you go to your blood bank and ask them for, you know, nine, 10 liters of plasma, you might run into problems. And most of us think that there's probably a rate of diminishing returns, where probably even with high normal volumes of plasma exchange, there's benefit. And this study actually came from Raki Mayawal's group in New Delhi, where they demonstrated even at normal blood volumes, or kind of 1.5 blood volumes of FFP replacement, that there was a mortality benefit with PLEX. So generally we're using plasma exchange predominantly in hyperacute liver failure patients, predominantly APAP, and we generally do it for the component of vasoplegia. So I had mentioned before, you know, the other challenge of Delhi patients is an immune component that these patients, you know, they tend to be supportable, but the challenge is that because they've got this cholestatic liver, they can develop this compensatory anti-inflammatory response dominated by IL-4 and IL-10, and can develop a form of immunoparesis. And I mentioned before that generally, if you want to kind of stratify these two patient groups, again, APAP patients tend to die of complications of their liver, whether that's intracranial hypertension or multi-organ failure, where often the subacute patients, which they're predominantly listed for transplant, and we have cases where these were you know, in the setting of somebody that's difficult to find a match, you might wait a while on the transplant list that really the most common cause of mortality in a idiosyncratic Delhi patient is infection and sepsis, and bacteremia increases the odds of death as we showed in this study by almost two times. So, you know, one of the good things is we acknowledge that while many centers are still using the King's College criteria, it is dated the initial criteria were based on patients from a cohort from 1977 to 1983, and there's been, you know, studies down the road that have kind of reevaluated them, but still, you know, they tend to lack particularly sensitivity. I will, you know, compliment, you know, this group that I've worked with, in that they were one of the first groups to kind of identify this dichotomy of APAP and hyperacute versus non-APAP or your subacutes and your idiosyncratic Dhillies. That the decision to list somebody for transplant is different that, you know, an INR of 6.5 in a subacute patient will get them on the transplant list with HE, or it won't on its own in a patient with hyperacute liver failure. So I don't say that there are any, you know, that one particular scoring system is advantageous. I will say at least that the ALFSG prognostic index and I quantify here that while the area or the curve is better, it was, this is only in the population it was derived from, but really it's more of the, kind of the general philosophy of things. So high coma grain, we talked about, etiology, whether it's favorable or unfavorable, INR and bilirubin. This looked at, for example, if patients were vasoplegic and on pressors, there are probably other factors, things like lactate that has been shown in different studies and the challenge with the reason why it doesn't, lactate probably doesn't show up in a lot of these criteria is we just don't collect it well. So there would be a lot of, there's a lot of missing data and hence you would lose a lot of patients from your model. So this is how it kind of ties into transplant is that there was this misconception probably about 20 years ago that transplanting ALF patients did significantly worse than other causes of, or other indications for transplant. And while that might've been true in the 2000s and 1990s, it's not really the case anymore. And I kind of show this older study from Europe as an example. We recently collaborated with UNOS, with Jack Lake, where we merged data from patients that underwent liver transplant in the ALF-SG registry, where we had a lot of granular data on their physiology when they were admitted in their first seven days of study, along with the long-term and graph data that we got from UNOS. So what you can see here is that within the ALF-SG, one-year survival is actually quite good. It's over 90%, which is similar to cirrhosis. But there are a few caveats with that. And I think this is really the take-home points, is that the vast majority of patients that were transplanting with ALF are predominantly subacute patients. And you can kind of see here that only a minority of patients that are transplanted for ALF are acetaminophen. And within the overall kind of this registry, it was about 16% or about 65 patients. And I was actually surprised it was that high. But the other thing you see though, is that there is a high wait list dropout with acetaminophen. And this is why acetaminophen for a lot of us is the bane of our existence, because there's always this high potential for recovery. So you don't want to commit somebody to a life of immunosuppression. We have more blood purification therapy. So maybe if I push them a few days longer, but we know also that they have a high wait list mortality. And partly it's because of this higher burden of multi-organ failure. So once again, that generally your DILI subacute patients from an anesthesia point of view and a transplant point of view, in many cases are more like an ACLF patient or a cirrhosis patient. And they don't tend to run in as much trouble when they get them through the operating room and through transplant. So I know that there are a lot of people on this webinar today that have an interest in biomarkers and liver injury. And there are a variety of different approaches that have been taken in terms of looking at biomarkers. Some of them are related to necrosis and regeneration. We've talked about this immunoparesis in subacute liver failure, that potentially looking at immunology and macrophage function, there are a lot of different techniques. And I wanted to highlight again, why the biomarkers in APAP versus idiosyncratic DILI in particular are likely different. And once again, it's because the hyperacute patient, you get massive necrosis, high circulating levels of a variety of things. Also, furthermore, APAP is a much cleaner population to study. Often there is, in an intentional overdose, you've got a defined insult, even in stagger toxicity, there's a well-described pathophysiology. Perhaps not so much with DILI, it can be more unpredictable, but furthermore, other subacute etiologies that can be a more indolent process. You might not see as much massive necrosis by the time we're measuring blood samples and also this component of immunoparesis. So one example, and there are many other examples out there, I just use a marker that we studied, that we published the work in collaboration with Chris Rose's lab at the showman Montreal, where we looked at liver type or FABP1 fatty acid binding protein, which is a small cytoplasmic protein abundantly expressed in hepatocytes. Primary function is intracellular transport of long chain fatty acids. And you get cellular expression upregulated in response to inflammation. And we found that this marker in APAP patient, both at early and kind of on admission, and then kind of between day three and day five, discriminated very nicely between survivors and non-survivors. And furthermore, in using this on top of existing scores, like the King's criteria or the ALFSG prognostic index, it actually, when we do this net reclassification using the DeLong method, that it added information, meaning it's telling us something that those scores are not. But as I kind of mentioned this cautionary tale, what about DILI? Where we looked at non-APAP etiologies were about a third of these patients were DILI, and you can see that it did not discriminate very well between these patients, certainly at an early time point, and in particular, once we adjusted for other variables. Mentioning as well that we've also collaborated with people at King's with Varuna Alvahari's group, they've been looking at different micro RNA signatures. And really this is kind of the progress that's coming forward is really, are we able to identify those patients where the liver is more likely to regenerate? And this once again is only in acetaminophen. Or furthermore, at a later time point, are there people where the liver is burnt out? Prior to this, we sometimes looked at factor five levels as a crude way of doing this, but we think that probably there are more sophisticated ways to do this. There were a lot of other markers that have been studied and will continue to study, but I just kind of highlight these as examples of the challenge or the dichotomy between APAP and idiosyncratic DILI. So, kind of one question to conclude is, are we getting better at this? And certainly in acetaminophen induced ALF, we are seeing improvements in overall survival with decreased rates of ICH and cerebral edema corresponding, particularly when we looked at this study over kind of the last 20 years with increased use of CRT, particularly in improved neuroprotective strategies. So, in summary, transplant-free survival has significantly improved over the last 20 years in ALF, particularly in APAP ALF. Rates of ICH and death from cerebral edema have decreased likely due to improved neuroprotective strategies. We're implementing more non-invasive strategies, but these need to be further validated. We consider the use of CRT now in patients with hyperaminemia and will up-regulate to plasma exchange, particularly in those patients with vasoplegia. One of the dichotomies between hyperacute and subacute or idiosyncratic DILI is the DILI patients tend to be more at risk of sepsis when they're on the transplant list. So that's why potentially getting an early donor in some parts of the world like India and Korea and Japan, they're transplanting these patients if the overall burden of multi-organ failure that's extra hepatic is low, even with a living donor. And generally liver transplant outcomes in appropriately selected ALF patients are similar to other populations with the caveat that the vast majority of transplants we do certainly in North America are in subacute liver failure. And I just want to thank my collaborators, particularly with ALFSG and the funders. And I really appreciate this opportunity to speak to your group, and I'm happy to answer any questions you might have. Fantastic talk, Dane. I'm already getting requests for the slides. So, yeah, and you know, what we're going to do is, you know, if you're a member of our SIG and you're getting our emails, every time Julie sends out the registration link, the recording to the previous talk is always at the bottom, and you can go back if you join late and watch it from the beginning. So I'm just going to open it up and I see Skip, you've turned off your mic, I think you have a question. Yeah, thanks. That was really a great update on literature, and thanks, I appreciate that. I wondered about the point-of-care ultrasound technique. You know, it's great, and I think that's a great avenue to change care. How do you see that rolling out? Mainly, how teachable is it? Because I see this as being maybe part of our training, you know, not to have a technician come up, or how are you seeing this roll out? How teachable is it, and should a fellow be taught this, and we as hepatologists should do this, or what? So I think because it's, like, you know, the whole thing with a lot of POCUS, point-of-care ultrasound, and one of my colleagues, Brian Buchanan has, you know, he's in Canada, he's kind of one of the big kind of opinion leaders on point-of-care ultrasound in the ICU, and I think one of the things that he piloted, just to give you that example, is the repetition, is kind of the number of scans you do. So I was part of a study of old dog new tricks, is what I called it, where mid-career, I'll still call myself that, intensivists did a curriculum of point-of-care ultrasound. And I would just say that I think it has to be something that you're doing, particularly the neural one is much more tricky than doing an echo, you know, looking at the heart. We're looking at big ticket items, we're looking at LV function, at, you know, the inferior vena cava, we're looking at filling, you know, a pericardial effusion. You know, definitely doing, like trying to identify like the circle of Willis with when you're looking at, to try to do your transtranial Doppler is more tricky. So certainly I think if this is something that you do regularly, and I think that's why for an intensivist, it's probably valuable. Unfortunately, if you're, I would say on the hepatology side, the problem is that if you're seeing this only a couple of times a year, I think it's important to probably, you know, collaborate with, in most critical care groups, now you have somebody that's going to be a point-of-care ultrasound lead. So the thing you mentioned too, is that we do actually have a TCD nurse that works in our neuro ICU, and occasionally she will come up, but most of the time it's a couple of us that end up doing this. Can you not teach it on normals though, and get people, it has, because I mean, we can, can we just do it on each other? And is that- The example is that essentially the idea is that, you know, you're, you're, you're kind of in front of the ear, you are, you know, you need to get, the finicky part is you're looking at a small vessel, you want to make sure that you're in the right plane, and then you're placing a cursor over it to get the waveform, but you're right. Like it is something that is certainly learnable, and obviously in a healthy person, it's going to be, you know, it's going to be a little bit easier to kind of acquire that. Right, thanks. I have a question too. Just a practical one. As, you know, a lot of times when patients come to the ER, regardless of coma grade, regardless of anything, the ER just starts knack. What's your practice there? Yeah, so I think like for, you know, the, that's a good question. The, you know, APAP for any reason, and often we, it's pattern recognition. You see the, you know, the high enzyme pattern. The high R before the billy, right? Exactly, exactly. So most of those patients, there's no downside to knack. In the non-APAP patients, generally if they're encephalopathic, I'll use knack, but I think probably what you're talking about too, Lily, is that everybody that comes in with a, with an ALT of 900, it's like, I want to do something. And we actually had a, I don't know if it was the same thing in the US, but we actually had a shortage of knack a couple of years ago, where they were kind of clamping down on us to make sure we had, you know, good indications that we should be kind of, you know, publicizing these. You know, the other example is that every time we would get a, you know, a patient after transplant with a deceased cardiac donor organ, and there was a high perfusion ischemia, ischemia reperfusion injury, they were getting thrown on knack. So I totally agree with you. The discussion about when to stop it is more difficult. And I think this is something that certainly needs to be teased out. Well, we think it is indolent and safe. There are some questions that potentially, are you contributing to this overall immunoparesis, particularly in the subacutes? So particularly in those patients, you know, after certainly after a couple of days, I don't run it ad infinitum. Certainly in those patients where you're likely heading down the pathway of transplant. Right. And then I guess it's kind of like to tag onto that, we usually do it for 48 hours. I have to tell you, I don't remember the last time anybody used oral knack at USC. Everyone's at IV knack. And the data is pretty similar. If you're not chiropractic, right. And if you have an OGNG tube in, how about in Canada, do you use IV or PO? We've been mostly using IV just because in some cases, it's been difficult just to get, like you said, if they have an NG, it's one thing. A lot more expensive, you know, that thing, the IV. So much more expensive. And that is dear to my heart from a, you know, to we certainly want to look for things that are cost-effective. And the outcomes aren't that different, right? It's just as good. Yeah, the hard part for us was just, if you look at the registry, like there was only a small proportion of patients in the LFSG registry that got PO and knack. We do have some of that data. Okay. Yeah. Any other questions from others? If no questions, I have another one. So the cutoff that you use, I just want a clarification on this. So if a patient has an ammonia cutoff more than 150, and I'm guessing this is arterial ammonia collected on ice, right? This is like a very clean ammonia, right? And for the trainees, you know, the ammonia is drawn, it's incorrect when it's drawn incorrectly often. And so over 150, you mentioned that only for knack, then you cool them? No, no, no. What I mentioned was if the ammonia is more than 150, we start CRRT. Oh, I'm so sorry. We do that as a blood purification treatment. And we know that those are going to be the people that also probably those people with an ammonia of 150, particularly the APAPs, you're, if they're, you know, pretty well anybody who is high grade coma, we're going to do what we were talking with Skip about, about TCD and we started doing our whole kind of package of non-invasive techniques. But certainly, you know, generally the way we step up therapy then is also, let's say they're on two pressors, they're hemodynamically unstable, then we'll get PLEX and we'll do that as a one to 1.5 blood volume. We'll do it at generally with our, the other thing I didn't mention is that, is that just so you're aware, there is kind of like centrifugal plasma exchange where, which are usually done by your phoresis, your hematology people, but there also are inserts on some of the Baxter Prismaflex machines where you do it through a membrane and that's called membrane tandem plasma exchange, but it's not nearly as efficient. If I do a centrifugal run, I can do the PLEX run in almost in less than three hours, where if you do the tandem run on the dialysis machine, it can take several hours. So we find that most of the time, you know, like in a 24 hour day and somebody that is hyperaminemic and vasoplegic that might still recover, we'll, you know, we'll run their PLEX for three hours and then 21 hours of the day, we put them back on the CRT machine. And Dean, how often would you get a patient that has a high ammonia, but doesn't actually meet a renal replacement therapy criteria for CRT? Does that even happen? Right? Like they already asked a few minutes ago, they already called. All the time. I'd say probably 70% of our patients we're doing, they don't have a renal indication. The one luxury we have here, and I know it's different at different hospitals is as an intensivist, I run my own, I prescribe my own renal support. We run CRT and hemo. I'm not calling nephrology. And I've had a few webinars like this, where I've been trying to convince nephrology colleagues of why we do this. And I know that that just adds a layer to this. Luckily, I know that Mitra Nadeem is at USC. Yeah, Mitra is a big CRT proponent. We can get CRT pretty quickly here. But, but so you're saying that when the, how I don't know the correlation of like, when the ammonia hits 150, a lot of times they're not acidemic. They're right. They're not yet. So. So we're using this purely as a blood purification strategy. Somehow make it into guidelines, Dean. You need to like update the ALF guidelines. So, so that, that obviously has been like, we've done a, like we've done a few reviews, but also we've also mentioned through the SIGs that this is one that hasn't been updated since. So the interesting thing is like, as I kind of mentioned, it's in the, this is in the, the, we just finished the society of critical care medicine, liver management guideline. Also, you know, it's in the ACG. In the US you got to get the payers to pay for it, right? So it needs to really like make it to the ASLB guidelines. Oh, that's really fascinating. I have two questions. Zonya, I think you were first, go ahead. And we can go over nine. I can stay as long as we need to. Thank you. Thank you. Great talk. I have a very basic question, actually. I was wondering, you mentioned that the goal for Dilly is to bridge to transplant and the goal for APAP is to bridge to spontaneous recovery. And I was wondering why that is. Is that mainly because of the risk of sepsis or are there other reasons? Generally, it's not really the goal. Like obviously the, you know, I have, you know, I'm on service this week and I've got two Dilly patients that I'm really, neither one of them are transplant candidates. So we're really hoping that this kind of turns the corner. So really it's just that the, because of this idiosyncratic phenomenon that, you know, that we think might be immune mediated, it's just this cascade that has started in Dilly patients quite often, you know, once they get, you know, a Billy of 20, 30, 40, we know that the liver is not going to recover. So it's just that we know based on, we are better at prognosticating Dilly patients in terms of spontaneous recovery than APAP patients. And it's just more difficult because if I use a traditional, like an organ failure score, like the ACLF score, the reason why a lot of them don't tend to work is because in APAP patients, they can come in with four or five organ failures and walk out of the hospital without transplant. And that's just because of this, the fact that you still have potential for regeneration. And some people talk about hypophosphatemia as one test to say that they're regenerating their phospholipid bilayer. It's just that we don't, you know, by the time these patients get cholestatic with Dilly, the liver's not, the liver's burnt out. It's not coming back and it's been happening over, you know, 10, 8, 10, 12 weeks. We use alpha-fetoprotein and FOS. I know we track people like that. And I thought it was because of ATP generation. I thought that if your FOS is low, you're like. Yeah. Very interesting. Anand, do you want to ask your question? Dr. Kulkarni, you have a lot of like great comments in there, Dean. There was a vet attending about dogs. That was such a sweet thing to hear. I'm excited that they saw our post and joined the. Yeah, I was going to say that for Anand here, it's a good question. So we don't do ICP bolts routinely. So really the step up there again is, you know, ammonia greater than 150, despite being on CRT and an abnormal TCD or optic nerve sheath. And then really the question then becomes, are you, is it going to affect management? Because I can look at a TCD and empirically treat somebody with Manitol. We've, you know, one case we had, we were using, you know, like phenobarb coma, we were throwing everything at them. But if you're not, you know, going to transplant, generally it likely won't change your management. So we generally don't tend to bolt those patients or certainly those people where they have a concerning TCD and they're listed for transplant. And as I said before, really, I want to know, is the ICP 25, but I can manage it medically and get them through an operation? Or is it 50 and we're, and it's a stop sign. And really it means that we need to pull out. And that's really the setting where we use ICP monitors, partly because also, you know, we, it's different in different centers. We get our neurosurgeons to do them. We were actually purchasing the Codman boxes, but as I can say, really that's kind of our systematic approach. And to your veterinarian colleague, yeah, I don't know, I have a, I have my dog, but yeah, I do the same thing. No, dogs get a lot of liver failure, actually. Especially in LA, there's these like, you know, palms that are very like toxic and they just eat them. So Lily, that leads to my other question, if you don't mind, I just have one last question. You didn't dice out the idiosyncratic dilly between herbals and non-herbals. My understanding is you've had literature in the past that said that they did worse. Is that still holding true? Is that your general impression with the herbals that they do generally do worse? So that's a good question. Like I don't have that kind of data on the tip of my tongue. I know that like certainly through the, and I know that Will and Bob Fontana are both with ALFSG and the Dillon Network. So I think, you know, they've probably got more updated literature on this. And also I think sort of like you said too, it's like when people actually admit to doing like, and we're kind of curious if they start as an indeterminate and they say, oh yeah, by the way, I was taking the herbal preparation. Yeah, we think sometimes they come in later. Yeah. They're just presenting later and we don't have a chance to get care to them. Thanks. Perfect. Okay, no more questions. Then I'm going to thank Dean again. If I ever have acute liver failure, I'm going to ask to be flown to Ontario. I want you to be my doctor. And really it's amazing, amazing data. You're really the trifecta. Amazing clinician, amazing researcher. And I know that you mentor a lot of people. So great, thank you for joining us. And we're going to say bye. Thank you. Thank you very much.

Hepatotoxicity

acute liver failure

acetaminophen toxicity

drug-induced liver injury

liver transplant

biomarkers

neuroprotective strategies

continuous renal replacement therapy

plasma exchange

prognostic models