liver transplantation for ACLF, the North American experience. Here are my disclosures. So in North America, we tend to use the definition that was described by NAXLD, or the North American Consortium for the Study of End-Stage Liver Disease, which was made off of the first cohort and focuses really on extrahepatic organ failures outside of the liver, where respiratory failure is defined as mechanical ventilation, renal failure is defined as the need for dialysis, circulatory failure is shock, or the need for pressors, and neurologic failure is grade three or four encephalopathy, where at least all of the continents agree. This criteria does lead to marked differences in outcome for hospitalized patients, depending on how many organ failures they have acquired, and the official definition of NAXLD-ACLF is having two or more organ failures. This was separately validated in a second prospectively collected cohort of over 2,600 patients. The first cohort had only infected patients, however, this second cohort did enroll patients with infections as well as no infections, and the validated score definitely did allow us to predict 30-day mortality well in both cohorts with a C-statistic of .85. Turning now to look at transplantation, in our first cohort of NAXLD, or the patients who were admitted with an infection, of the 413, we initially analyzed 136 patients who were listed for transplant, and after six months, about a third of them did go on to have a transplant, whereas 42% unfortunately died or were delisted. In multivariable modeling, the predictors for death or delisting before transplant in those listed for transplant were the development of respiratory failure with the highest odds ratio at over 4 1⁄2, brain failure being second at over 2 1⁄2, and because in this cohort all patients had an infection in order to get into the group, having a less severe infection, specifically a UTI, was protective from death. In the second larger cohort of about 2,800 admitted patients with and without infections, you can see that we had about 750 or so patients that were listed for transplant among the group. When we compared the differences between patients who were listed versus not listed in this group, there were marked differences in demographics. Specifically, those patients who were listed for transplant were younger. They were more likely to have SBP during their admission or at admission. They were much more likely to develop AKI during that initial index hospitalization with more than half of them developing AKI, and that is the new definition for AKI that's accepted of a greater increase in creatinine of 0.3 or more within 48 hours or doubling compared to before admission. And the MELD scores, as one would expect, were higher amongst the listed patients with an average of 23 versus 18 in the non-listed patients. The number of organ failures that occurred during the index hospitalization was similarly different between these two groups, and patients who were listed for transplant were more likely to have more organ failures during admission than non-listed patients. And when we parcel it out by the type of organ failure, all types of organ failure other than circulatory failure were significantly different between listed and non-listed patients, with the biggest difference being seen in the patients with renal failure. Specifically, three times as many patients that were listed had renal failure compared to the non-listed patients, which may have simply been a bias in offering dialysis to patients who were listed compared to potentially not offering dialysis to patients who were not listed. When we turn our attention to outcomes of the listed patients, there were three potential options over the following three months. They could be alive without transplant, they could be alive with transplant, or they could be dead or delisted. There were significant differences amongst a variety of different metrics, but one of the largest ones was the number of organ failures. The number of organ failures, as one would expect, was lowest amongst the listed patients who remained without transplant and alive three months later, intermediary in those patients who were alive with liver transplant, and highest in those who died or were delisted. Similarly, the percentage of patients with ACLF, by the Naxal definition, was lowest in those patients who did not have a transplant and remained alive at 7%. About a quarter of patients who were alive with liver transplant had experienced ACLF, and about a third of the patients who were dead or delisted had ACLF. One of the other big differences between the three groups was the delta MELD, where we subtract the discharge MELD from the admission MELD, and found the greatest drop in MELD score amongst those who were alive with transplant versus an increase in MELD in those who were dead or delisted. When we turn our attentions to the post-transplant outcomes for the 250 patients that were transplanted amongst this group, the average MELD at transplant was 28 overall, and the average creatinine at transplant was relatively high at about 2.5. About 70% of these patients were transplanted from being admitted for another reason from the hospital, 28% of whom were actually in the intensive care unit at the time of transplant. And fortunately, despite this transplanting this very sick patient population, 84% were able to be discharged home after transplant as opposed to being discharged to a facility. When we specifically compared those patients who had had ACLF versus those without ACLF and their outcomes post-transplant, you can see that there was no overall difference in death between the two groups at three or six months. One of the biggest differences was the delta MELD, indicating that patients who had ACLF and went on to be successfully transplanted, their MELD score markedly improved before transplant, indicating that probably their clinical scenario had markedly improved. The MELD score on average, though, was higher amongst the patients transplanted with a history of ACLF than no ACLF by four points. There was a marked difference in the number of patients transplanted as an inpatient with almost all of the patients with a history of ACLF being transplanted as an inpatient versus two-thirds in the non-ACLF group. Three times as many patients with a history of ACLF needed perioperative dialysis, although fortunately the use of simultaneous liver-kidney transplants was similar amongst the groups. The creatinine transplant was markedly higher in those patients with ACLF than those without ACLF, but the three and six months creatinine was of borderline significance and fortunately had plateaued out at roughly the same. This concept of stabilization or improvement that needs to be seen after ACLF prior to transplanting a patient was also shown in other studies where they investigated transplanting patients with and without ACLF and showed an inferior survival in patients with ACLF than those without. However, in patients who had ACLF that went on to have significant improvement, there was no overall difference in survival than those patients without. Turning to other data that's available from North America, it's been shown there are several risk factors for death before transplant in ACLF patients, and this was using UNOS dataset where they attempted to use the easel-cliff definition of ACLF. However, because not all of the variables for the easel-cliff definition are actually in UNOS, they had to make some modifications. They used liver failure as a bilirubin greater than 12, renal failure as a creatinine greater than 2, coagulation failure as an INR greater than 2.5, neurologic failure as a Westhaven grade 3 or 4 encephalopathy, circulatory failure as vasopressors, and respiratory failure as mechanical ventilation. And they concluded that the risk for weight loss removal was high in ACLF3 patients even with lower MELD sodium scores. They also found also 21% of patients in the U.S. are now being transplanted with ACLF by these criteria, and risk factors in multivariable analysis for death after transplant within that first year in multivariable analysis were shown to be a donor risk index of greater than or equal to 1.7, the need for mechanical ventilation, but if you were able to get access to transplant within 30 days of listing, this was somewhat protective. When they parceled it out by the grade of ACLF, more severe grades of ACLF had worse prognosis post-transplant, but mechanical ventilation pre-transplant had the worst outcome post-transplant. The same group analyzed the UNOS dataset and looked at the outcomes of patients with ACLF grade 3 modified criteria again because of the UNOS dataset and compared it to patients who were status 1 and found that the mortality of ACLF3 patients was actually markedly higher than our status 1 patients who are supposed to have the highest mortality. And this morning we heard in the plenary session about another evaluation of the UNOS dataset more recent under the Share35 versus non-Share35 guidelines, specifically analyzing data two years before versus two years after Share35, reviewing approximately 4,700 recorded cases, and then, I'm sorry, that's a typo, 1,200 NAXL ACLF patients by that definition were used. And in era two where patients were given access through Share35 to organs more readily, they had a higher hazards ratio for being transplanted even after being adjusted for MELD. Another database has been used looking at ACLF in the U.S., specifically the National Inpatient Sample from 2006 to 2014. Two million admissions for cirrhosis were evaluated. The definition of ACLF that they used was greater than or equal to two organ failures, but this was only by ICD-9 code. And so 5.9% of these cirrhotic patients met the definition of ACLF and showed that NASH is actually the most rapidly growing etiology of ACLF. Turning to the VA dataset, the VA dataset over a 10-year period was analyzed looking at 72,000 hospitalizations for decompensated cirrhosis, and a quarter met the criteria for ACLF, the canonic definition, although the data in the VA computer will not allow you to make that strict criteria, so they did also do some imputation in order to determine that those patients met that criteria. You can see that if you use the canonic ACLF data, the way of diagnosing it, that you incorporate more patients with a diagnosis of ACLF versus the NAXLed ACLF cohort, if you use that definition, it applies to a smaller, more uniform patient group. And when evaluating the difference in mortality at 28 versus 90 days, the ACLF 1, 2, and 3 did increase mortality over the numbers. The take-home messages from the VA dataset analysis showed that approximately 26% of decompensated cirrhotics did meet admission criteria for ACLF by the criteria they used, and there was higher mortality associated with older age, white race, hepatocellular carcinoma diagnosis, a higher MELD sodium score, and being admitted to a non-transplant facility. Fortunately, over the 10-year period that was studied, ACLF prevalence did decrease, although unfortunately mortality in the ACLF 3 group did increase. So, because we would all like our patients to live to die another day, I'm going to turn our attention to looking at ways in which we may, some simple interventions that we may want to employ. Because these days, when we're trying to live to transplant, our patients, unlike in the hep C era, tend not to have a linear progression in their liver disease, but have a more punctuated course where they have acute events. And on the back end, where there is hopefully this improvement in MELD, and the patient may be getting priority for transplant because of a MELD that they were given while they were very sick, as they're recovering you hope to get them transplanted. In this single center prospective trial of admitted patients with decompensated cirrhosis, it was found that almost half had relative adrenal insufficiency, and that this was an independent risk factor for the development of ACLF. Certainly something that we could all look into more reliably in our patients, because it was also associated with an increased risk for death, even when controlling for ACLF. As previously mentioned, non-selective beta blockers may improve outcome in ACLF patients. However, we do want to make sure that we're only using them in patients who have reasonable systolic blood pressures. And the study that was previously cited did look at this in a randomized controlled trial of patients with a puzzle ACLF criteria looking at carvatalol versus placebo. And in the first 28 days there was mortality improvement. However, unfortunately by 90 days there was no improvement in overall mortality, although there was a decreased risk for AKI and SBP. Indicating that possibly non-selective beta blockers may be decreasing the dysbiotic bacteria that we have in our cirrhotic patient's lumens into the circulation and peritoneum. Other risk factors for recurrent infections, which is often the modus operandi for celestial discharge, as opposed to transplantation, are things that we can't control, such as age, whether or not the patient has SBP, and their MELD score. However, something we can control is whether or not we subject them to PPI therapy. And do remember that PPIs block the oxidative burst of the neutrophil and so we are further immunosuppressing our immunosuppressed cirrhotic patients by giving them PPIs, most often when they don't need them. And then I think, as our last speaker mentioned, there needs to be more investigation into extracorporeal liver support. This is a meta-analysis looking at the extracorporeal liver support of varying kinds in ALF and ACLF. Most studies are small in nature, and in general we don't have convincing data to incorporate this into care yet, but it is certainly an exciting area of research. So in conclusion, about 21% of patients are being transplanted in the U.S. with ACLF. ACLF3 has a worse prognosis on the wait list than status one. We certainly know that we need to stabilize patients, and you would see this by a significant negative delta MELD before transplanting them in order to improve survivals after transplant. But some important and easy things to do to improve the outcomes peritransplant in our patients with ACLF is to diagnose relative adrenal insufficiency earlier, minimize PPI use, and try to maximize non-selective beta blocker use in patients whose blood pressure will tolerate it. Thank you very much. I would like to invite our panelists to come up here for an opportunity for the audience to ask questions. If you have any questions, please don't hesitate to come up to the microphones. They will be open. I realize it's late in the day, but we have, we actually run at a reasonable pace, so we will not get you, we will not keep you here very long. Those of you who may be in the overflow areas, you can actually come up here and ask questions as well. Well, since I have you here, let me post the first question for the panelists. I mean, there clearly is a great deal of controversy around potential strategies to get patients transplanted. I mean, I think that we've all talked about the fact that the ultimate treatment for our patients with ACLF is a liver transplant, because right now our strategies for treatment are maybe very limited, although I do want to ask those questions relating to those patients with ACLF related to hepatitis B. So what would the APASL team members here let us know about how they approach transplantation in your setting? And maybe we can then go about discussing the projection of patients with ACLF into the transplant list in North America. Thank you for this question. For hepatitis B, Dr. Tawande has probably said I had to leave, but the guidelines would generally say anybody who has a MELD above 30 and has a donor who is non-B is safe. Generally living donor is very prevalent in the Asian region, so would be considered for a liver transplant. Also if there is a MELD between 20 and 30, one would try to stabilize this patient, and if it is less than 20, certainly we will not consider for transplant. For alcohol, it is very challenging, especially when there is a living donor available. Mostly these are women, and patient has high MELD, and we know that the patient will not survive for more than 30 days. So whether we should go, but after all the evaluations by psychiatrists and others, if the patient has more than two organ failures, which I think many definitions include, then we would certainly try not to rush through transplant because the outcomes may not be good. However, if the ARC score is improving or is around 11 or 12, patient may be taken up for transplant. And sometimes we be, now our surgeons generally would like a plasma exchange to be done to bring down the MELD score so that the outcomes are better, but this is not published. This is what we normally do. Dr. O'Leary, would you like to comment on the approach that we have in our centers? I mean, I know that there's positive data collected prospectively about what happens with ACLF, so we could just look back and see what people are doing, which is what we tried to do this morning, and what is your perspective on how you recognize somebody with ACLF, and rather than try to take them to the point where there's futility and they would be delisted to get them to the point where they can be listable and transplantable? Sure. So... Hello? Okay. I guess I'll just yell. No. Oh. Okay. So, you know, in general, pretty much consistently, I didn't present the European data, but it consistently patients on mechanical ventilation, if they have actual respiratory failure, those patients consistently do the worst. So if they happen to be ventilated for, you know, neurologic failure, and then quickly get off, that's very different, and that's frequently how patients are actually transplanted, is that they come in in the throes of ACLF, and the patients who get better, you know, where their MELD score sort of falls, especially after an infection, which is one of the biggest precipitants of ACLF in North America, would be infection-related. I think that a lot of times there's been a little bit of a delay between, you know, listing with that higher MELD score, because you don't want to transplant them within that side, that first 48 hours right after they've been so ill. But if you can get them, you know, remember when we give IV antibiotics, you sterilize the blood, especially in SVP, about 86% of the time within that first 24 hours. So especially when we have SVP, which is a big reason why we develop ACLF in listed patients, you can probably get them listed in about 72 hours after you've sort of sterilized the blood, you know, and then you still would have that high MELD score, so that in a few days you may be actually able to use that MELD score to get them actually transplanted. So I think that's where people have been the most successful. Okay. Very good. I have a quick response to this. One issue that we come up with here is the lack of uniform application of what people are considered comfortable with transplanting, and where is, how close to the edge they want to take it when this patient is close to its transplant. And every transplant committee has their own decision and their comfort level, because obviously these are not automatic yes-no decisions, and they have to be re-evaluated. So that is the challenge. Not only is it a challenge to actually, obviously we don't have enough organs available, also in patients who are being subjected to living donor transplant, our responsibility is even greater, because now we are harming a human being who is actually going to be a donor. Our responsibility for actually providing the appropriate organ in a person who would most likely survive, but is not already too far gone in this stage to actually not receive the organ in time, or the organ not being enough for them. So I think these are such important decisions that I think a lot more thought, and not just ethical thought, but a lot more objective measures have to be done to actually say that these things definitely add to the MELD sodium score, at least within the United States context, and that is consistent across centers. And it's not something that, you know, I eyeballed this patient today, he or she looks horrible, so therefore we cannot list him or her today. Those are things beyond the MELD sodium that we sometimes use in our transplant committees that we need to get away from if we have to actually make these patients higher up in the listing process. Well I think one of the things that's been a challenge, especially in ACLF, is the renal dysfunction, because that gives us the greatest priority. And so I think in general, at least in what we've published, the centers were doing a nice job at looking at that well, because what we know is that HRS, when you don't give a liver kidney and you transplant them early on in the course of HRS, they do very well, and that was borne out in the data. But if the patient has progressed on to develop ATN, that those patients have worse outcome after transplant with a 50% mortality fairly early on afterwards. So we also have to parcel out what type of renal injury they have. Okay. We have a question from the floor. Yeah. I'm Dr. Dushesh from India. Question for Dr. Bajaj. I mean, this is regarding the CDR, the cirrhosis dysbiosis ratio, and the relationship with the severity of disease. So the question is about the cause and effect and something like chicken and egg theory. So does the liver disease become severe first and then the dysbiosis comes in, or the dysbiosis comes in first and that is responsible for the severity? So that's the question. So the way to answer this is a very important question, because unlike in animals, these take decades to happen in these patients, and we are a sum total not just of our liver disease. People with liver disease also can be obese, can have a different diet, can have diabetes, can be a medicine that can affect their microbiota. But when we talk about patients in animal models where we can actually control this, and we put the stool from patients with cirrhosis into germ-free mice, only thing you develop is inflammation. You do not develop cirrhosis in those mice. However, if you take the stool from alcoholic humans and put it into mice who are not then fed alcohol, their amount of liver injury is way, way higher than the germ-free mice who are colonized with stool from healthy or non-alcoholic people. So microbiota propagate the liver injury, but they almost never cause the liver injury. So most likely, it is an effect of the liver disease rather than the cause. But when it starts, it actually worsens the liver disease progression. That's what the current understanding is. And of course, it might change as more research comes. Dr. Martin. Hi. I have a practical question for Jackie and Shiv. In non-alcoholic patients, you mentioned it briefly in your presentation, the use of steroids for relative adrenal insufficiency, I mean, these patients come in and they act septic in most cases. How comfortable are you giving steroids for presumptive adrenal insufficiency? Do you wait until you get all the negative cultures back or how soon do you give steroids? Well, so usually, we just give antibiotics and steroids at the same time if they actually have relative adrenal insufficiency. But once again, I think that that also would be dependent on how bad the sepsis was. Because if they were really hypotensive, I think it might have helped with the hypotension. But if they weren't hypotensive, I might hold off just because of the...until I was out of the infection, you know, with a more complete workup. So I think it might depend on the blood pressure and, you know, at the situation. What do you do in India? Yeah, for those who have come with sepsis and septic shock, they would receive 50 milligrams of hydrocortisone eight hourly. But that's, you know, very subjective with no actual studies, very few data to support that. And that's mainly for hypotension to be reversed. So I'm not sure we have logic behind that. But I have a question for both our other panelists. So you have a patient who has alcohol and has come, was drinking till, let's say, one week ago and now has come. And he needs a transplant. He has a high meld. So what would be your policy, who was drinking till one week ago and now is intubated, maybe whatever? Will you transplant such a patient or? So in the U.S., generally speaking, the criteria of six months is no longer applicable in most centers. However, in order to get past that rule, which this patient would need, you would have to actually talk to the patient. And then you would have to really have tremendous engagement from the family as well. And so an incredibly sub-select group of patients with acute alcoholic hepatitis, very limited. Have you done? No is the answer. So that's the thing. We have outsourced. There's dramatic. No, we have outsourced. There's a huge amount of discussion about it. And it is done in the U.S. But it feels like, because there's so much discussion about it, that it's happening a lot. But it is an extremely sub-select group. And if they were not speaking, they would not be transplanted. I can give you the perspective from what we do at the Mayo Clinic. We do have a policy, which is an exceptions policy. So like Jackie mentioned, we would not take a patient who is intubated, sedated, not really able to be assessed. Factors that we look into very frequently are, is this patient the first occasion for an alcoholic hepatitis presentation? A recalcitrant alcoholic that has been presented several times like this would never, ever be considered. And we have very strong guidelines relating to social support. And then we tend to try to get a contract with the patient to have alcohol rehabilitation. Now, do we succeed every time that we do this? The answer is, I don't think so. And I think their experience in the United States bears that out. Particularly younger people, despite social support, they tend to go back to drinking. That's been our experience in our center. So I think that we're probably going to see an uptick and then potentially kind of a downtrend towards being more cautious about who we choose to transplant in the setting of alcoholic liver disease. And then it behooves us to find out what we can do to keep those people alive and offering them something that's short of transplantation. Dr. Gallegos, you have a question? Yes. Yes, Dr. Bajaj. How do we control the data on the microbiota, the fecal microbiota specifically, in these patients that have very advanced liver disease? Some of them are on treatment with antibiotics, either for prophylaxis of spontaneous bacterial peritonitis or on rifaximin for their encephalopathy. So how does that come into play specifically for the fecal microbiota? I don't know if that has any bearing as well as the saliva or the CR microbiota. It does. It has bearing on saliva as well. Your question is very important. And, of course, when I was trying to study this very muddy field, it was a very problematic thing. So the best way me as a clinician can actually figure out is go to an outcome that is incontrovertible – death, transplant. No one can doubt that that happened. And then go back and backtrack and get enough samples, enough patients that you have the ability to either match or control for all of these things. So for example, the publication that we had in 2014, where we did a lot of these analyses using 16S sequencing, we wanted to find out whether infected patients had a different microbiome than uninfected patients. But uninfected patients often have a very high MELD score just because of their infection. So we matched them by the MELD score on the day of admission, infected versus uninfected patients. And uninfected patients, even before the first dose of antibiotics, had a very different microbiota in their feces compared to the others. PPI is an epidemic, so that worked in our favor because everyone was on PPI. Very few people were on SPP prophylaxis, so we were easily able to control for it. So the important thing that you correctly point out is just because you're studying microbiome, you want to still nest it into a clinical question, that autopathophysiological question that you need to answer. It is not enough just to study the microbiome. It sounds very cutting-edge, but it's not anymore. It's actually a thing. We also have to treat it as a tool to help us become better prognosticators and potentially better designers of therapy. Can I ask a question about hepatitis B? So with TAF, supposedly there's a better drop, there's data showing there's a better drop in ALT. You mean TAF? Yeah, TAF. Tenofovir and olefamidate, yeah. So for patients who get admitted with ACLF with hepatitis B, either a reactivation, is that your number one go-to, or as far as drug as opposed to the other two, or do you feel like it clinically doesn't make much difference? You know, right now, just Entecovir and Tenofovir are going to be, you know, for the study, you know, for recommendation TAF, theoretically, it's going to be better, but still need more study because the study for decompensated liver disease is still going on. But theoretically, that can be used as the choice for the treatment. But it's not, you know, bring down the virus is not just the only one that you can get the patient survive. I think most of the patients with acute hepatitis B flare, this could be very, you know, severe, and most of the organ involvement is going to be only liver. So extra hepatic is going to come very late. Okay. Any other questions from the floor? Well, if there are none, I want to thank the audience for staying this late. Thank you very much. On behalf of Dr. Payal and I, it was a pleasure having you here, and thank you to our panelists. Thank you. Thank you so much.

liver transplantation

acute-on-chronic liver failure

ACLF

NAXLD-ACLF criteria

organ failures

transplant outcomes

treatment strategies