Our last speaker will be on the Portal Hypertension Debrief, Dr. Brett Fortune, Associate Professor of Medicine and the Medical Director of Liver Transplantation at Montefiore Einstein Medical Center. Thank you to the audience for spending the early evening with us, and thank you to the ASLD for the kind invitation to present the 2024 Portal Hypertension Debrief. These are my disclosures. So hopefully in the next 30 minutes, we will be able to discuss kind of the themes of the journey of portal hypertension during this year's meeting, both including diagnostic and prediction models for portal hypertension, as well as associated outcomes, updates in variceal bleeding, beta blockers, and tips, cardiac function assessment, ascites, kidney failure, hepatic encephalopathy, and actually hepatopulmonary syndrome. For the portal hypertension, I focused mainly on clinical abstracts. About 200 were accepted for this meeting out of over 15 countries, and congratulations to all the authors, and similarly disclosed. I was not able to include everyone's abstract. I thank the ones that were able to share, and I was not able to include this afternoon's late breaker, so please refer to the recording for those details. So we'll start off with noninvasive diagnostics for portal hypertension and outcomes. We'll start with our Chinese colleagues who looked at the ongoing need for noninvasive methods to diagnose clinically significant portal hypertension, particularly severe portal hypertension. They looked at this multicenter prospective study in China looking at three-dimensional MRI to detect. They enrolled prospectively 172 patients, nearly 90 with clinically significant portal hypertension defined by HVPG. They looked at a multivariant regression model showing that actually they were able to show that three-dimensional MRI by itself was able to predict very accurately with high performance, but also combining it with serum indices increased that performance to well over an area under the curve of 0.9. And with that, as you can see here with these discriminatory graphs, they have stellar prediction, and there's possibly great potential to see how 3D MRI could be used in the future, obviously further development there. In addition, another outcomes model called the portal decompensation score was designed as another strategy to noninvasively detect clinically significant portal hypertension. This was actually a large retrospective multicenter cohort study in Western Australia. The patients were identified by high HEPA scores and liver stiffness measurements. The outcome was model accuracy for their first decompensation event. They were able to acquire well over 1,000 patients for both a training set as well as validation cohorts with compensated advanced disease. And as you can see here in the graphs, they were able to show quite significant divergence in outcomes for their first decompensation event using these model cutoffs and were able to validate with very high discriminatory function of over 0.8 on the AUC curve. This was validated in all chronic disease etiologies simply using well frequently checked lab values of total bilirubin, ALT, ALKFOS, albumin, and platelets. And with that, there could be potential of using this score to help further identify those that are higher risk for compensating events, and they actually provided in their presentation that they could possibly use this as a secondary way to stratify for possible mitigation of the risk with beta blocker. Turning the corner to look at metabolomics, it was interesting to kind of look to see if you could detect hospitalization within six months just by checking serum. This study looking at stable outpatient patients with cirrhosis, they were able to do an exploratory study through the Naxold cohort, 3.0, that prospectively enrolls outpatients from 10 American centers. This includes those with etiologies that are under control based on definitions within six months, and they wanted to see if metabolomics could predict next hospitalization within six months. Very large cohort of almost 800 patients, 600 having six-month follow-up data. It turns out nearly a third had hospitalization within six months, nearly all of them liver-related, and they were actually able to find with this analysis that metabolite-only models had an extremely appropriate area under the curve. Using just six metabolites, they studied over 800, and interestingly, they were focused particularly on primidine metabolism, gut microbiome, and lipid peroxidation, and by also adding just one clinical variable, serum albumin, some slightly increased changes in the AUC, but this is an interesting investigative study to look at that, and despite control of etiology, it is interesting in these high-risk patients that still nearly a third required hospitalization within six months, and that there could be possible opportunity here for this to noninvasively detect risk for hospitalization to allow you to try to intervene. Something that I wanted to include, not so much due to it being related to portal hypertension, but I found it quite fascinating. As you probably have heard, many sessions today working with artificial intelligence. This very interesting abstract from the Mayo Clinic was looking at the capability of CHAT-GBT4-OMNI, which is a generative AI using large language models to help actually take complex clinical scenarios via voice command, and to see how accurate it could predict the outcomes with these complex sessions, and what was interesting, these are two examples of the 10 cases, that 10 out of 10, it was accurate, and with over 90% agreement based on a panel of hepatologists, that it was able to meet multiple criteria, including accuracy in response time, clinical appropriateness, and actionable recommendations. So hopefully it doesn't take our jobs in the future, but it clearly is something fascinating that might be able to assist, particularly those that may not have easy access to specialists like hepatologists in the rural community, where perhaps if this gets validated in a larger scale, that maybe primary care could use this. Moving on to bleeding beta blockers and TIPS-OMI. We'll first start off with Curvetilol. It's obviously become very popular recently with its recent evidence of showing reduction in all-cause decomposition and survival in the setting of compensated and decompensated patients with portal hypertension. Interesting, I wanted to present this paper that comes from India, that showed a randomized open-label trial at two centers, where they were able to look, based on the criteria listed, to look at whether or not Curvetilol would actually reduce all-cause decompensation within a year, and interestingly, they found in using 270 patients, 135 receiving Curvetilol, the median dose was around 12 milligrams per day, and compared to placebo, they found no significant difference in decompensation over a year in these patients that were meeting those criteria that had been tested with a Bavino. However, moving on to a different study based in a large multi-center retrospective cohort study using a large health system database in America that was relying on administrative ICD-10 codes and low platelet counts, they wanted to test a similar question, would Curvetilol versus the other non-selective beta blockers, particularly Natalol and Propanolol, in this large cohort, and as you can see here, well over 2,000 patients were included. They did use a risk-adjusted analysis to try to control for confounding, and they actually found a significant reduction in the hazards for both hepatic and decompensation and mortality against other types of beta blockers. Admittedly, they did not have a non-beta blocker arm. Curvetilol also had a significant reduction in decompensation, including ascites, axillopathy, and variceal bleeding, which is more consistent with some of the other prospective data from Europe and Asia. However, obviously, we need to continue to look at these different populations throughout the world to kind of look at the different effects and efficacies, as well as the different patients, whether they're compensated or decompensated. Something that is definitely gaining steam to help us with our non-invasive risk stratification for clinically significant portal hypertension is the use of spleen stiffness. Similar to our liver stiffness, you can use transient elastography or other elastography imaging modalities to look at the spleen, and by getting a stiffness measurement, this can actually help assist and increase your probability, positive predictive value of having clinically significant prior hypertension. This group in India wanted to look at the actual, to see if splenic stiffness could help dictate whether or not on patients on beta blocker, based on response, would there be a delta in the portal pressure, which we have yet to prove with the elastography modalities at this point. So they basically studied in those that just had an acute barocelial bleeding episode, measuring their HVBGs at the time of the bleed, and then six weeks later, and then measuring splenic stiffness serially in these patients. With 220 patients in a prospective cohort, they were able to look and show that with those that were started on Cravatalol, there were almost 40% were responders, and for those that respond, that liver stiffness was highly discriminatory in predicting the change in portal pressure based on the repeated HVBG, and that it was also noted that with a delta of 8% from baseline, that that can actually increase in splenic stiffness, that can actually predict recurrent bleeding, which was quite fascinating. So this being one of the first studies to show a positive response, a correlation with elastography. Obviously, we'll need to further validate this in other cohorts, but this could be something that could help us noninvasively look to see if patients are actually responding to their beta blocker. Looking actually at the management of acute barocelial bleeding, in these states, we still use octreotide, and per our guidance, we still use a five-day protocol. However, this interesting randomized trial in Thailand, they wanted to see if maybe you could abbreviate that course of octreotide, and so they basically performed this trial looking at one-day versus five-day courses of octreotide infusion, and those that presented with bleeding getting standard endoscopy care with their abandoned ligation, and looking to prove a non-inferiority approach of within 5%, and with 220 patients presenting with bleeding and receiving ligation, 109 received the one-day octreotide course, 111 received the five-day course, and as you can see here, it was a near less than 1% difference in outcomes for re-bleeding, and also mortality was also not appreciably different. So obviously, once again, I believe we need to validate this in other cohorts around the globe, but it is interesting and possibly true with proper patient selection and risk stratification, we may not need to always treat all patients with full five-day course of octreotide, and I think this needs further validation. Moving on to preemptive tips. For those that may not be aware with the terminology, because it can be confusing, preemptive tips, formerly known early tips, are tips performed in high-risk patients after their first variceal bleeding that get a tips within 72 hours in order to prevent their recurrent bleeding episode because they're high-risk of doing so within five days. But it's interesting to look specifically at certain phenotypes in this protocol that was done. They wanted to look at those patients that present with bleeding in the setting of severe alcohol-associated hepatitis, and so this study, looking at preemptive tips, and wanted to see what happens compared to other patients in the same scenario that received just drug and endoscopic therapy, standard of care. This was an aggregate pooled data from four European prospective observational cohorts looking at the bovino-approved six-week survival. 142 patients were accrued. 47 received preemptive tips. The other 95 had the standard of care with combination therapy. And as you can see here, there's a substantial improvement with the tips group. Even in these very high patients, the median meld was actually in the 20s. And as you can see here, nearly 90% had six-week survival in the preemptive tips group, and 70% with the standard therapy group. Re-bleeding was also lower in the tips group substantially, only 3% versus 19%. Less societies, no difference in hepatic encephalopathy. And in the competing risk analysis, preemptive tips was associated with a minus 14% rate of decompensation, death, or transplant need. So based on this, I think it continues to advocate more data that even in those very ill patients presenting with high meld, high bilirubin, this validates again that one should consider preemptive tips in the setting of variceal bleeding if they meet the high-risk criteria, which again is defined by child Pugh, class B, with active endoscopic bleeding, or child C, score 10 through 13. Also, what happens when you do a tips in someone that continues to drink alcohol regularly? And so this was an interesting study in Austria looking at defining post-tips alcohol use and what happens if they should continue to drink. This was a two-center study based in Austria looking at those specifically with alcohol-associated liver disease and defined as ongoing positive alcohol metabolites with serum checks, and noted that out of 248 patients who received the tips with a median follow-up of almost three years, they showed that actually over a third continued to have alcohol use after their tips placement. 70% were for bleeding, 30% were for ascites, by the way, for an indication of the tips. There were no characteristic differences seen. Unfortunately, 60% actually of the whole cohort died within this three-month time period. 30% were readmitted at that time, and only 5% received transplantation, which kind of tells you that it's very important to find other bridging therapies to help these patients. In a multivariable competing risk model, it demonstrated that post-tips alcohol use was associated with an extraordinarily high, almost two-and-a-half hazards fold increase in ACLF, and with that associated mortality. They noted that the patients that continued to drink after their tips had more bacterial infections after the tips, had more decompensation events, had more ACLF, and overall, and liver-related mortality with that. I think it's really important that you help educate these patients after their tips, and get them to addiction specialists, and to mitigate or anti-craving therapies to help try to reduce the possibility of continuing alcohol use after the tips placement, because clearly, the tips does not rescue them from ongoing liver injury from alcohol. Next, we have a couple of abstracts talking about fundal varices. These are varices in the gastrointestinal fundus diagnosed on the serine classification as gastroesophageal varices class II, or isolated gastric varices class I, and it remains still unclear what modalities to help treat these patients with gastric variceal bleeding, and this project was a multi-center randomized trial in France, involved 20 centers, because it was very difficult to recruit, looking at one arm of preemptive tips versus glue obliteration with beta blocker after the bleeding episode resolution, and they wanted to look at an aggregate of re-bleeding free survival at one year, and as you can see here, interestingly, that with the tips group, 75% remained alive at the one year mark, as opposed to only 37%, so a significant risk reduction in these patients, having a greater than double survival benefit, and with that, it definitely is one of the higher levels of evidence showing that preemptive tips is superior to endoscopic sirenacolate plus beta blocker for re-bleeding free survival among these higher risk of bleeding patients with fundal varices. On the follow-up with that, we also wanted to see what is the natural history and predictors of outcomes in those that have an index variceal bleed after fundal varices. This study was a prospective observational cohort in India, looking at 350 patients with fundal variceal bleeding over an eight-year period from two centers, 275 received sirenal acolyte alone, an additional 15% were followed by BERTO, balloon retrograde transvenous obliteration, 6% then received TIPS after the sirenal acolyte, and interestingly, those that received secondary endovascular therapy, either BERTO or TIPS, had substantially reduced re-bleeding episodes compared to sirenal acolyte alone, and no difference in mortality in this case. And these are the Kaplan-Meier curves, showing kind of convergence, because it did show significant reduction in bleeding risk for endovascular therapy, but without improvement in survival or further decompensation. So further to go on these algorithms, but we need to continue to study what are the best approaches for both prevention of primary bleeding, prevention of secondary bleeding, and so forth. So one controversy that I always hear is do beta blockers cause portal vein thrombosis? One of the concerns when you have decreased portal inflow, when you use your beta blocker, are you promoting possibly PVT formation? And this was an interesting study looking at trying to correlate that risk, because most of the literature has been limited to a lot of retrospective small studies. This is another retrospective cohort study, but it used a very large database that's been commonly tapped in for multiple studies, called the Trinex-X, which has over 110 million patients in 63 American health systems. This group wanted to look at this database to see if there was any effect of being on a beta blocker and the incidence of PVT. They, using ICD coding, which is obviously a limitation of this methodology, but used propensity score matching to try to control for confounders. They were able to get 8,300 patients on both the beta blocker side and the non-beta blocker side. And interestingly, this analysis found that PVT actually did develop in 13% of the beta blocker group, as opposed to only 6% in the control group, showing a significant odds ratio adjusted of 2.3. They did not find any difference in mortality, however, as both groups had around 30% mortality. So I think, again, there are limitations with administrator databases and retrospective studies, and I think we need to continue to look at this as some other prospective studies have not shown differences with beta blockers and PVT, but this is something that clearly needs to be continued to look at. Moving on to the next section, heart, ascites, and kidney failure. Lean up here. POCUS is something, obviously, that is taking some prestige in having the POCUS course here, and the group in India has been definitely the leaders at looking at the use of POCUS to try to predict the presence of cirrhotic cardiomyopathy and the effects on acute kidney injury. This was a prospective study looking at 370 patients that presented into the hospital with acute kidney injury. As you can see here, 2 3rds actually responded to fluid, 14% were diagnosed with HRS AKI, and almost a third had the echo parameters for the diagnosis of cirrhotic cardiomyopathy. Interestingly, those that had cirrhotic cardiomyopathy and HRS, over 10% died in three months, and over a quarter died at one year, and it was, as expected in other studies, AKI was associated with both 90-day mortality and one-year mortality. Interestingly, that there were a couple of echo parameters that were consistent with cirrhotic cardiomyopathy. The septal E to E prime and E prime alone were both predicted of those that needed renal replacement therapy. And so we need to continue to look at echocardiogram and other POCUS parameters to see if maybe we can have more point-of-care and ultrasonovic findings to help us diagnose our patients and help pronosticate what their outcomes could be so that they can allow us to test what interventions might be appropriate, as opposed to albumin, albumin, albumin. Moving on, this is an interesting study looking at a randomized controlled trial, also in India, of the effects of curvata law combined with this I channel inhibitor called Braddeen, Ivabraddeen, mispronouncing it, but it's something that lowers the heart rate and prolongs diastolic time. We know that left ventricular diastolic dysfunction is a key part of cirrhotic cardiomyopathy and it's linked itself to adverse outcomes. Looking at curvata law alone and this combination therapy, what were the outcomes in a year, including decompensation survival? With 76 patients and 77 in both arms, they did not find any difference in rates of decompensation or mortality or adverse rates, but they did notice in the combination therapy there was a significant improvement in the CCM parameters. So obviously more data to come to see if combination therapies like these can help patients that are diagnosed with cirrhotic cardiomyopathy, but as of right now, it did not show any benefit outside of those that had a heart rate of greater than 85 beats per minute. Moving on, this abstract is looking at something called effective albumin concentration. This is calculated by assessing the liquid chromatography and mass spec of baseline redox state of CIS 34 of albumin and then related to the proportion of native albumin. Don't ask me to say that again. But basically they felt that this was a more corrected albumin level for patients with decompensated cirrhosis as opposed to the traditional serum albumin level. And they wanted to look at short-term prediction of outcomes in acute decompensated patients that were hospitalized. This was a prospective study in China. They had a comparator arms looking at those with both healthy controls as well as compensated patients and wanted to look at one-year outcomes among these patients that presented just with acute decompensation without ACLF. They had 445 patients, 50 controls and 52 compensated cirrhosis patients. Interestingly, they found significantly lower concentration levels of this effective albumin in acute decompensated patients compared to the comparator arms. And with competing risk modeling adjustments, it was strongly associated with one-year mortality. As it gets higher, it increases by 55%. And the authors also found by combining it with MELD and CLIF scores that this can actually improve the discriminatory function of predicting these outcomes at one year. So with that, I think more to come looking at this measurement and seeing if this can improve predictability of outcomes in these patients and possibly also help translate it to compensated patients. Moving on to a theme called a community acquired acute kidney injury. This was a large nationwide US veteran study looking at the prevalence of this diagnosis of patients with cirrhosis that are in the outpatient setting that meet the criteria for acute kidney injury. This looked at the VA database of almost 56,000 patients. 13% were found to meet the criteria for AKI in the ambulatory setting. They were followed for a median of four years and accounting for competing risks, they were able to find that there were 3.6 cases per 100 person years and actually with the majority being kind of low stage one A and B and mostly fluid responding. But it did show that there was a substantial risk of being hospitalized and actually a quarter died within a year. So remember, these are patients coming from CIRP. The predictors based on their very intense analysis showed that gender, the presence of decompensation, prior decompensation including SBP, HE, and ascites were extremely predictive. But also interestingly, diabetes, hypertension, and using an ACE inhibitor were also very predictive. So it was felt by the authors to say that it is important to monitor your patients on the ambulatory setting to make sure you don't notice any increases in the creatinine and to be very cautious for those that have comorbidities like hypertension and diabetes because these patients might need mitigation and closer monitoring as they are at higher risk of developing AKI which obviously translates to bad outcomes. This is the global AKI study looking at socioeconomic disparities on a global scale. We all know that the World Bank and Universal Health Coverage Index are very predictive of disparities of healthcare access and the ability to manage chronic diseases. This study was to look to see if those thresholds would also be predictive of the outcomes of hospitalized patients with AKI. This was a large prospective international cohort looking at over almost 4,000 patients throughout the country on five continents, 67 centers. Over a third developed AKI. 40% were from underserved nations that met low income cutoffs. And interesting, they had a higher cumulative incidence of 28-day mortality, including lower rates of transplantation. So clearly, high disease burden, low access for curative therapy. So very important to address these disparities and really relates to maybe nationwide policies to address with our health services organizations. Moving on, this is the CLEAR consortium. This is looking at variations in 90-day mortality in this also international prospective cohort. This showed with 2,500 patients from 25 countries. Also looking at the income kind of thresholds. Also saw 9% death, 30% readmitted, 6% received transplant. Interestingly, that predictors of 90-day mortality in this group age prior infection within six months, prior listing for transplant but not receiving it. PPI use, which is something that's been brought up multiple times in prior literature. White cell count, melt sodium, and then the nation's income level. Also 90-day readmission can be seen with these other parameters here. And for time purposes, I will move on. Looking at this interesting cohort, looking at mean arterial pressure. When we see a patient with acute kidney injury, we always wonder what is the goal of the arterial pressure when we use fluids or other vasoconstrictor therapies. This was looking at almost 1,800 patients that presented in the ICU with septic shock. They had over 150,000 mean arterial pressure readings. 72% had AKI. And interestingly, with this very impressive time-weighted analysis, they were able to show these ideal cutoffs of aiming for a MAP goal of over 65, and in fact, an ideal goal of 75 for these patients, which would be different from prior shock management in the past. And they found that with those that achieved a 75 millimeter mercury MAP goal, early fluid administration in antibiotics, they had basically a 2 3rds reduction in death for these patients. So definitely something that we should talk to our critical care intensivists about of trying to help our patients that present in the ICU with septic shock as that these targeted goals of higher blood pressure might be more important. Moving on to the Poseidon study, which is the long-term follow-up of the alpha pump in refractory ascites. It's a percutaneous device where one end is in the ascites in the peritoneum, the other is in the bladder, and this is to help pump out the ascites into the bladder. This has been published before on their one-year data. This is actually the two-year data showing sustainability and safety of beyond a year. It was first studied at six months and then showed at 12, and now this is the interim at 24 months, showing that it remains safe and with good control of ascites. Hepatic encephalopathy, this is an interesting study, a systematic review and a meta-analysis looking at lactulose alone versus lactulose combined with polyethylene glycol or MiraLAX. This study was able to look at three trials, 216 patients, 108 in each group, and showed that, interestingly, that combination therapy with polyethylene glycol was significantly improved. Hepatic encephalopathy scores at both 24 hours and led to shorter length of stay. I think this needs to be further looked at because obviously we want to avoid over-treatment, but something to further mount. Moving on, Dr. Bajang presented this study looking at overt impact encephalopathy using only rifaximin monotherapy and those that had had a prior episode of overt impact hepatic encephalopathy. A lot of patients are very restricted taking their lactulose, so this study was looking at a pooled analysis of two trials, one phase three, one open-label phase four trial, and, interestingly, looking at 270 patients. Again, prior history of overt impact encephalopathy within the last six months, 125 were on rifaximin, 145 were on lactulose, and they found in this analysis that significantly lower six-month mortality was seen in the rifaximin alone group, and a number needed a treat of 19, and that lactulose, as expected, had a higher discontinuation rate. So I think more to come here, but it is interesting to see that possibly with higher adherence to medication because of better tolerance, maybe this leads to better outcomes. Real fast, because it's so rare to see something therapeutic for hepatopulmonary syndrome, we all know that it has very limited therapeutic options and basically liver transplant is destination. This was a study, a randomized trial in India looking at the drug pentoxifiline, 400 milligrams three times a day versus placebo, in patients that were diagnosed with moderate to severe hepatopulmonary syndrome. And this study, a small study, only 20 in each arm, showing that among the patients that received pentoxifiline, that they did have significant increases in their oxygenation, PaO2 levels, as well as AA gradient reduction. And they also, on secondary analysis, showed several inflammatory cytokines improving as well. It still warrants further study because it's a small study, but it is interesting and shows an interesting signal that maybe we could possibly help treat some of these patients without necessarily jumping straight to transplant. Conversely, this is a study looking at simvastatin to see if that would help in the treatment of hepatopulmonary syndrome since it does its pleiotrophic effects, including anti-inflammatory properties. This study also performed in India, was a randomized two-to-one double-blind placebo-controlled trial over six months to see if there were any differences using simvastatin versus placebo. As you can see here, I'll just get to the gist, is that basically they did not find a difference. They did see a response of 13% versus 2%, but because they only had 45 patients with 30 in the simvastatin arm, they weren't able to find clinical significance. So there is a trend of maybe something, and maybe if they also had longer-term follow-up beyond six months, maybe they would have seen something. There was also almost a tripling of the side effects, although not statistically significant. That is something we always worry with, the statins and something. But as we continue to look at the use of statins in the treatment of cirrhosis, perhaps this could be another indication in the future. So I will conclude, and I appreciate your patience. Much to cover, and so the key takeaways, the themes that I got from the review of these abstracts are there is definitely ongoing advancement in technology, both artificial intelligence and metabolomics and other modalities to improve our detection of both clinically significant portal hypertension and predicting the associated outcomes. There's continuing need to determine the efficacy of non-selective beta blockers in specific populations so that we can continue to validate and make sure that we're doing the right things. There's definitely ongoing evidence that endovascular therapy with TIBS or Berto remains superior over endoscopic therapies for bleeding gastric varices, and I think this is something that we need to continue to advise in our guidance. But there are some caveats. Further diagnostics and therapies to improve cardiac function seem to help outcomes, and I think we need to continue to look at that and work with our cardiology colleagues. There remain global disparities in healthcare leading to inferior outcomes of patients hospitalized with cirrhosis, so this should prompt more initiatives and public health direction to try to get policies to address these disparities. Perhaps lactulose doesn't have to be the first-line therapy to prevent HE recurrence. I think more to come with that. And then epedopulmonary syndrome remains very understudied and still remains without effective therapy outside of liver transplantation. However, there might be promising data from the Pentoxifilin trial. And with that, I appreciate your attention. I really am grateful for all the research that was done during this meeting, and I look forward to next year's meeting, and thank you again. Thank you.

portal hypertension

variceal bleeding

three-dimensional MRI

beta blockers

splenic stiffness

TIPS

hepatic encephalopathy