I'd like to thank the organizers for giving me the opportunity today to tell you about the genetic architecture of non-alcoholic fatty liver disease. I'm a professor of gastroenterology and hepatology, as well as computational medicine and bioinformatics at the University of Michigan. I also direct precision medicine for internal medicine and the MSTP program, and I'm an elected counselor and institutional representative for the American Society of Clinical Investigation. I'm interested in using human genetics to define subtypes of human disease and develop targeted precision therapies for them. I have no permanent disclosures for this presentation. As many of you know, NAFLD has become a global health epidemic. It's increased with the rise of obesity. Its global prevalence is around 24 percent, and its prevalence is higher in individuals of Hispanic or Asian ancestry than European ancestry than African ancestry. It is going to become the number 1 cause of liver disease worldwide, and unfortunately, we have a very poor understanding of its etiology and therefore no effective treatment, making it a very large unmet medical need. Work that we and others have done to develop non-invasive ways to measure hepatic steatosis using CT scans in the population has been useful for genetic studies, and we use CT, but others have also applied MRI and FibroScan. As you know, these are now routinely being used to diagnose NAFLD without biopsy so that we can study this in larger populations. We've also developed methods to genotype large populations, and combining that with the phenotyping, we showed that NAFLD is about 25-34 percent heritable or genetic in the population. We've developed ways to integrate data across groups using imputation and meta-analyze this to increase our power, and using that, we've now identified many specific variants that associate with non-alcoholic fatty liver disease. As you can see here, we use this to identify multiple variants near genes that associate with CT-measured hepatic steatosis in 2011, and then verify these in histological samples, and confirm that these variants affect fatty liver disease, as well as confirmed another variant, PMPLA3, that was reported earlier to associate with LFTs and MRI-measured hepatic steatosis. Since then, others have also identified and confirmed other variants that associate with fatty liver disease, either from liver function tests, imaging, and histology. What you can see is that some of these variants associate with not only non-alcoholic fatty liver disease, but have also been associated now with alcoholic liver disease, and some of them also predisposed to hepatocellular carcinoma. But you can also see that there's multiple variants suggesting that this is indeed a polygenic disease. Even from this slide without any details, what you can see is that we can reproducibly identify human genetic susceptibility to common NAFLD, and that genetic ideology is polygenic, heterogeneous, but definable. We also looked at whether our variants explain heritability and or variants in different populations. In this study in 2013, we showed that our five variants that associate with non-alcoholic fatty liver disease explain more of the variants in individuals of Hispanic and European and African ancestry. In fact, most of this difference ends up being genetically mediated and mostly driven by PNPLA3 as Romeo et al had also reported in 2009. But furthermore, another interesting finding, as many of you know, is that liver steatosis can lead to advanced liver disease or can help promote metabolic disease or cardiovascular disease. But really, this correlate is imperfect. About 90 percent of individuals as they get obese will deposit fat in their liver. We think about 20 percent of these will go on to get inflammation, 20 percent of these will go on to get fibrosis and cirrhosis, and about 20 percent of these will get cancer in work that we have shown in the Framingham Heart Study, individuals that have more than a third of their liver composed of fat, 66 percent of them meet criteria for having the metabolic syndrome. Some of these also have cardiovascular disease. But really, we don't have a very good understanding of why certain people are going to deposit fat in their liver and which ones are going to go on to get advanced liver disease versus metabolic disease. Again, we're very interested in whether or not genetics can help us with that. One of the things that we observed back in 2011 is that our variants that increase steatosis, many of them can also increase NASH and fibrosis, but they have very different effects on other traits. Some of them really do not associate very much with serum lipid traits or glucose traits, which are metabolic syndrome related traits. Some of them actually are directly opposite of what you would expect epidemiologically in that, for example, TM6SF2 reduces your LDL and triglycerides, which is opposite of what's expected epidemiologically, and GCKR reduces your fasting glucose levels in home IR, which is opposite of what you'd expect epidemiologically. This is extremely exciting because what this tells you is that genetics can dissociate epidemiologically correlated traits. This is really the basis for precision medicine. Now you have genetic markers that could help you put these individuals that carry these markers into different subgroups as you can see on the right. Not only could you put them into different subgroups, but now you can link them back to a molecular cause, which is on the left, which is the genes that they implicate. This now allows you to subsegregate people into disease groups and then link them to particular treatments so that you can make precision therapies for them. Another very convenient thing that these kinds of analysis allow you to do, is that if a pharma company was interested in targeting a gene, they could see not only the on-target effects that they would have on steatosis, but they could look across at these pleiotropies and really see some of the side effects potentially that that medication would have. For example, if you wanted to target TM6SF2 to reduce steatosis, you're going to increase LDL and triglycerides in the serum and potentially increase your risk of a heart attack. That was back over 10 years ago and more recently, we've done very large association analysis for serum ALT, AST, all across and we've crossed these hits with CT-measured steatosis. Now we have up to 23 variants that implicate genes in the genome that associate with fatty liver disease. If you do a similar pleiotropy analysis, you can see again that now we have even more subgroups that associate with non-alcoholic fatty liver disease. On the right, you see that all of these increase serum ALT here, so that I've anchored it on the ALT increasing allele. You can see that they have various different effects on other metabolic traits. The group at the top increases serum ALT and also increases, for example, serum LDL and triglycerides. The next group increases serum ALT and also increases serum LDL but a portion of them actually decrease serum triglycerides. The next group increases ALT but decreases serum LDL but increases serum triglycerides. The last group increases serum ALT and decreases both serum LDL and triglycerides. If you look really carefully, you can actually split these into more than four groups. For example, APOE, which is at the bottom of this, actually isn't quite the same pattern as the other three that it's clustered with here. If you really wanted to, you could subsegregate these into finer subgroups. When you look at some of the genes that have been implicated from GWAS, they do give us new insights into the pathophysiology of this condition. For example, GCKR and TRIB1 seem to play a role in promoting glucose to triglyceride synthesis and de novo lipogenesis. Genes like PMPLA3 help promote triglyceride breakdown from lipid droplets and HSD17B13 may play a role in this process also, but it may also play a role in steroid or retinoic acid metabolism. Genes like ERLIN1 and TM6SF2 play a role in packaging triglycerides and other lipids to get them out of the liver in the form of VLDL. Genes such as Mk1, GPM1, and MBOD7 play a role in mitochondrial biology and maybe phosphatidylacetal metabolism. You can already see that before genomics, we really didn't know that this was the pathophysiology of NAFLD, but now we're getting a lot more insights into the various metabolic processes that can go awry to cause this disease. But as you've just noted, a lot of the causes from genetics seem to be pointing to NAFLD being a metabolic disease. As you know, we before tried to classify this as non-alcoholic versus alcoholic, and that really brings up the issue of whether that's the right terminology. Many people based on this work have now proposed that perhaps we should be thinking about calling this metabolic-associated fatty liver disease, and I certainly support that. Certainly what they have proposed is to have hepatic steatosis in adults, again, based on the fact that you can now easily get this non-invasively. Then also see just whether or not they're overweight or have diabetes, in which case you can call them NAFLD. But then if they don't have that, if they're lean or normal, you can also look for other metabolic syndrome-related traits, as I mentioned before, like elevated waist circumference, blood pressure, serum triglycerides, or HDL abnormalities, prediabetes. Or HOMA-IR, and then also classify them into a NAFLD category. Certainly I favor moving in this direction, but as you can see from the previous slide, I would probably say we're not far enough with this classification. I personally would favor more of molecular mechanism classification. For example, eventually, I think we probably will need to go to something like PMPLA-3, promoting fatty liver disease and other more molecular causes. Why do I think that? Well, I think very much we're moving in a direction of now getting more molecular-based treatments for this disease. Here I show you a study where they've used ASO against PMPLA, the mutant version of PMPLA-3, and actually ameliorated liver disease. This is now going into human trials. Very soon, we're going to be able to have genomic and very molecular-specific therapies for non-alcoholic fatty liver disease. In fact, there is precedent for this now happening outside of liver disease. Already, we have oral MC4R agonists for the treatment of rare genetic obesity. As I mentioned, there's this ASO trial for PMPLA-3. People are using CRISPR now to correct genetic defects for sickle cell disease using PCSK9 and base editors for that to try and treat heart disease. Really, we are going into a new phase of medicine where we're going to get very direct and molecular treatments for these conditions, and that's why I favor a more molecular diagnostic grouping. But as I mentioned, naphrodes, a polygenic disorder, in fact, in this paper that's based on BMI, we showed that variants not only genome-wide contribute to variation, but really, you can go genome-wide and contribute to variation. You may ask yourself, well, how are we going to use genomic therapies for a common disease? Certainly, we can use these genetic variants to identify people in the population that are increased risk of having hepatic steatosis or even cirrhosis. Shown here is a polygenic risk score that we developed for ALT that can help you identify people at high risk of getting hepatic steatosis or cirrhosis. Certainly, you can then take this to the next level and actually develop molecular therapies against multiple targets to target the particular genes that are causing the susceptibility in individuals. But in addition to the genetic, and while we wait for those genetic therapies to come about, we're also interested in identifying environmental variables that can mitigate genetic susceptibility. As most of the disease, I would say more than 70 percent is environmentally driven. Here in this study, what we have done is we have looked for environmental variables that multiplicatively interact with genetic susceptibility to cause fatty liver disease. What that means is if you have the genetic variant and the environment, 1 plus 1 doesn't equal 2, 1 plus 1 equals 8. What we found in looking at one of the largest meta-analyses for gene-environment interactions for this trait, multiple traits like serum insulin, glucose triglycerides, and BMI interact with this gene to cause fatty liver disease. But when we did conditional analysis, it's really the insulin resistance that's most causal for promoting fatty liver disease. This was surprising to some people because we know that diabetes promotes fatty liver disease and most people would have thought that it's the glucose part of it. But really what we showed is that's really more the insulin resistance part of it that maybe changes utilization of fatty acids or other parameters to cause and exacerbate this disease. This is important to know because we did this actually in pre-diabetics or people that are not really getting treated for diabetes. Really just this variant and the environmental risk that account for about eight percent of the population variants in steatosis, and about five percent of the general European population is at risk. As I told you earlier, a greater proportion of Hispanics and Asians will be at risk and a smaller proportion of African-Americans are going to be at risk. Even knowing how to change your environment will make a huge difference in terms of public health for these individuals. Towards that end, what I mean by that is that what our study shows is that if you have insulin resistance and the fatty liver promoting allele of GG, if you are able to get yourself from a high insulin resistance state to a low insulin resistance state, you are actually going to have twice the benefit of that in terms of improving your fatty liver compared to somebody that's not genetically at risk. In some ways, the people that are most at risk by changing their environment will get the most benefit from that. To follow that up, what we're doing right now is we're actually recruiting individuals based on genotype from our biobank here, we're putting them on a low carbohydrate versus controlled diet, and we're measuring liver fat and fibrosis at five months to see whether or not that hypothesis is true. This is essentially moving towards precision therapeutics. In addition to that, we can also look for specific dietary interactions with these genes. Shown here is work that's being presented at this meeting by Dr. Chen, poster 564 in genomics and precision medicine, where we've looked to see whether or not some of this genetic risk interacts with quartiles of the Mediterranean diet. You can see that the answer is yes. Basically, individuals that are high polygenic risk for NAFLD, if they're on the Mediterranean diet, they have about a 15 percent reduction in their risk of having hepatic steatosis compared to individuals that are in the worst category of following the modified Mediterranean diet, which is in red. This is a larger difference than if you weren't really genetically predisposed to having NAFLD in that if you improved your diet and you weren't genetically predisposed, you'd have about an eight percent reduction in your hepatic steatosis. Whereas if you were at genetic risk and you improved your diet, you would improve by over 15 percent. In fact, you can do this not just with a Mediterranean diet overall, but you can start to look at what components in the diet are helpful, and you can look at his poster for more details. But shown on the right are two things, for example, that improve your risk, which is fruit intake and vegetable intake. Again, you could start to do precision diet recommendations based on genetic risk in the future as we know more. Really, I hope that I've shown you that human genetics can help guide precision medicine for NAFLD-B by identifying disease subtypes, underlying causal biology, targets for therapeutic intervention, and that these plus environmental information can help inform precision medicine. I'd like to thank my collaborators in the Gold Consortium for their work, as well as our laboratory and our funding sources, and you for attending this talk. Thank you. Hello. My name is Alina Allen, and I'd like to thank the program chairs for inviting me to speak at this symposium. I will speak about subtyping of NAFLD by non-invasive testing. As you know, the biomarker field has continued to develop and expand at fast pace with novel non-invasive tests or revisiting old biomarkers in different contexts of use and combinations. Today's short talk is not an extensive review of this field, but a concise presentation of key points that should help understand the role of non-invasive tests in subtyping NAFLD. The objectives are to define the clinically relevant subtypes of NAFLD and in terms of non-invasive testing to understand the context of use, what are we estimating, diagnosis or prognosis, use of single or sequential or combination tests, and lastly, cost considerations. It is important to keep in mind that the goal achieved by using a non-invasive test is different at various stages of NAFLD in the natural history, which are typically managed in different settings. For example, in early NAFLD, managed in primary care mostly, the goal is to exclude severe disease. The ideal biomarker has a high negative predictive value to rule out advanced fibrosis. In the indeterminate or intermediate subtypes of NAFLD with clinically significant fibrosis, noted here in the middle. The goals are to monitor for disease progression, to identify candidates for randomized controlled trials and to initiate aggressive management. In this scenario, the biomarkers performance should weigh more towards a better positive predictive value. Lastly, at the other extreme of disease severity, which is advanced NAFLD, largely defined by stage 3 or 4 fibrosis, the goals are to identify those patients who need screening for hepatocellular carcinoma or esophageal varices, to identify candidates for randomized controlled trials and to institute aggressive management. In this specific population, the stakes are high. The ideal biomarker should have the highest positive predictive value. Secondly, we need to understand what outcomes does the biomarker measure? Is it diagnosis or initial evaluation? This is the main setting where combination or sequential biomarkers have been studied. Or is it prognosis or prediction of outcomes? Or lastly, is it longitudinal monitoring or response to treatment where the field is evolving, needs more data, and I will not spend time talking about this in this short talk. Let's dig in. Most of the non-invasive tests are methods of fibrosis estimation. As you know, these are serum biomarkers or imaging biomarkers or elastography, which can be used either independently or in combination depending on availability. Let's start by talking about single biomarker performance in a quick review. You are all aware about the abundance of serum biomarkers. I've listed here the ones that were mostly studied, with the top ones being the ones that use routinely available data from clinical practice. The bottom ones highlighted in gray, the patented tests, which include other proprietary parameters. These biomarkers can be used, especially the top ones, as first line in patients who are suspected to have fatty liver disease and further risk stratify. The performance of these serum biomarkers to identify advanced fibrosis, such as stage 3-4 compared to early fibrosis is noted here. As you can see from the routine biomarkers, FYP4 and NAFLD fibrosis score are the ones who are noted in most of the papers to have the highest performance. The patented tests noted at the bottom have higher performance at the expense of availability and cost. I think you are all aware of the recent approval by the FDA of ELF, the bottom one listed here in the United States earlier this year. When it comes to imaging biomarkers, this snapshot lists the most common ones, which are transient allostrography or fibroscan, MRE, or shear wave allostrography. Their performance is compared on the top line here. This is again to identify advanced fibrosis of stage 3 or 4. MRE has been consistently the highest performer, but fibroscan has the advantages of being a point-of-care test at a lower cost, but with higher failure rate and a lower precision of measurement. Important to know that there are confounders that interfere with all these tests. I'd like to mention the most important one because NAFLD is a disease that occurs in mostly an obese population, so it is important to mention that the performance differs based on BMI. In this study by Kossi et al, MRE is not impacted by BMI and central adiposity, but there's discordance between MRE and fibroscan in detecting clinically significant fibrosis of F2 or higher, as the BMI increases, especially at BMI of 35 or higher, where there's discrepancy between these tests in over half of the population. In head-to-head comparison studies between fibroscan and MRE using liver biopsy as reference, MRE maintains high accuracy of about 0.9 to differentiate fibrosis across all stages. As you can see, the performance difference increases as the disease stage increases, so differentiating stage 2 or lower from stage 3 or stage 3 from cirrhosis. This concludes the presentation of the individual biomarker testing, so let's talk about sequential testing. Blood biomarkers, as you know, are mostly used with two cutoffs, which leaves a large gray zone. For example, for NAFLD fibrosis score and Fib4, any values below these cutoffs give a very high negative predictive value to differentiate stage 3-4 fibrosis from earlier stages, especially for Fib4, which has a negative predictive value of 93 percent. At the other extreme, values above the listed cutoffs give a high positive predictive value for advanced fibrosis. Although in Fib4, which was on the contender on the other extreme, the PPV is only 66 percent, so suboptimal. The major issue though, is that the intermediate zones here include almost half of the patients where the score is uninterpretable. This is where sequential tests can help, because they decrease this indeterminate zone from 50 percent to 20 percent if Fib4 is followed by Fibroscan, for example, or 24 percent if Fib4 is followed by ELF. Why combine non-invasive tests? One is to decrease that gray zone mentioned before. Second is to reduce false positives, which are more dangerous than false negatives, especially for drug therapy. A high positive predictive value is needed for safe pharmacological intervention. The concept of sequential testing strategy was explored in this paper where confirmatory tests followed a blood-based biomarker. Confirmatory tests included MRE, ELF, or other, including Fibroscan or other blood biomarkers. This is the likelihood ratio to detect advanced fibrosis noted here. MRE had the highest likelihood to rule in disease in this combination approach. What about combination testing using two biomarkers at the same time? There are several of them that have been studied. The first one was Fibroscan AST or FAST score, which was designed to identify patients who would be candidate for randomized controlled trials, which currently includes mostly patients with NASH and stage 2 fibrosis or higher, so that intermediate subtype of NAFLD. This biomarker includes liver stiffness measurement, continued attenuation parameter, and AST, which are included in an algorithm. The area under the curve was 0.8 in the derivation cohort and between 0.76 and 0.85 in a validation cohort. As mentioned before, using FAST score for identification of patients who need clinical, or actually drug trials, we need to focus on the positive predictive value. As you can see, this varied from 0.83 in the derivation cohort to 0.69 in a pooled external cohort for validation. But more importantly, this gray zone, which is between those two cutoffs of rule out and rule in was quite high, varied between 39 percent of patients in the derivation cohort and 30 percent of patients in the validation cohort. A good biomarker combination, but still suboptimal. Another score is MIFIB index, which is a combination of MRE and FIB4. Again, for identification of NAFLD stage 2 fibrosis for inclusion in randomized controlled trials. The purpose of this combination is to increase the positive predictive value by MRE, using two cutoffs or higher than these in a combination of MRE plus FIB4. In two independent cohorts, the positive predictive value was increased from 87 to 97 and from 85 to 91 respectively. The limitation with this approach is, although we do not have a gray zone by different cutoffs, we do have a gray zone of patients in whom these two biomarkers are discordant. For example, a high MRE with a low FIB4 or vice versa. Lastly, in one of the most recent papers in this combination testing, the combination of ELF and FIB4 was studied, this time for identification of patients with F3 and 4. In this study, Seattle shows that the positive predictive value of ELF above this cutoff increased from a range of 43 to 62, to 81 to 95 by adding a FIB4 of higher than 2.9. Again, this is when both biomarkers agree, and there's a gray zone which remains to be clarified in the situation where the two biomarkers are discordant. Let's talk a bit about cost considerations. There's a multitude of biomarkers. How do we make the best of cost effectiveness? This issue was studied in this paper by Villar-Gomez published in 2020, where the cost-effectiveness of several combination approaches to detect cirrhosis this time was studied in this paper. This graph plots the accuracy of the test versus the cost. Note that a accurate and low-cost test should be in this corner here. The top two contenders were the combination of FIB4 and Fibroscan, which had the lowest cost. Followed by FIB4 plus MRE, which had a higher accuracy with only marginally more expensive costs than the other combination. Note the cost of liver biopsy alone at this extreme of cost, which is not uncommonly used in clinical practice. We talked about diagnosis of disease so far. In the last part of this talk, I'd like to discuss the use of a non-invasive test to predict future liver outcomes. In this field, blood biomarkers have been studied recently in a paper by Hagstrom published in 2020, using a large Swedish cohort of over 100,000 people. The most commonly used biomarkers noted here is what was studied. This population is a unselected population in the general community. You can see that the performance was modest in this unselected population, with area under the curves ranging from 0.5 to 0.7. The performance was slightly better in those who had a high ALT, so those who were at higher risk to progress to cirrhosis, but this was in the short-term. We studied imaging biomarkers using MRE to identify the role of liver stiffness measurement at baseline to predict future events. This was a large cohort from Mayo Clinic of over 800 patients with NAFLD and available MRE. We showed that the age-adjusted probability of cirrhosis in those without cirrhosis increases with each increase in kilopascal. Each increase in one kilopascal is associated with a threefold increase in future development of cirrhosis. For example, the three-year risk of cirrhosis development in a patient who has a kilopascal of three is 1.7 percent, whereas the one with kilopascal of five is 13 percent. In those with cirrhosis, each one kilopascal increase in stiffness is associated with a 32 percent higher increase of future decompensation. Liver stiffness measurement can be used not only to stage liver disease, but also to prevent future risk in an individualized fashion because the outcomes are not dichotomized in high versus low risk, but there are used on a continuous scale. Although patients are cirrhotic by all conventional criteria, their risk of future liver events is different. Somebody with a liver stiffness measurement of five has a 22 percent risk of decompensation in three years, whereas a cirrhotic with a liver stiffness of eight has a 43 percent risk of decompensation. In summary, the setting and the goal are the first critical issues to consider when assessing non-invasive tests. In primary care setting, we need to rule out advanced fibrosis. We need a biomarker with high negative predictive value, so fib4 is fairly good in that situation. However, in specialty care setting where we need to rule in advanced disease, we need high positive predictive value. In my opinion, imaging is essential in that stage. In clinical trial enrollment, the data seems to show that combination therapy is better than imaging alone, which is better than single serum-based non-invasive tests. A big caveat to remember is that these combinations of tests were examined in tertiary centers. In primary care centers where the prevalence of advanced liver disease is much lower, their performance remains unclear. Lastly, there is emerging data on use of non-invasive tests to predict future liver events in an individualized fashion. With this, I thank you for your attention. Well, it's a pleasure to be speaking to you all today. I want to thank the organizers for inviting me to talk about lifestyle interventions for non-alcoholic fatty liver disease. These are my disclosures. Today, we're going to review weight loss goals for adults with NAFLDs who have overweight or obesity, and we're going to identify optimal dietary and physical activity recommendations for those patients. First, let's start with the case. This is a typical patient that I see in our fatty liver clinic at Mass General. It's a 59-year-old gentleman with obesity, a BMI of 41, presents with abnormal liver enzymes. His routine blood work had shown an ALT of 105, an AST of 66, an alkaline phosphatase of 68, a total bilirubin that's normal at 0.4, with a normal INR, albumin, and platelets. His additional workup, and you can see suggestions for this, recently printed in the Journal of Hepatology from 2020, was negative. He underwent a video-controlled transient elastography or FibroScan, which showed a cap of 400 decibels per minute, suggesting fat within the liver, and a normal liver stiffness measurement of 4 kilopascals. What would we recommend to this patient about lifestyle intervention? We have a patient who has NAFLD, because we've excluded other causes, and we've confirmed on CAP that he does have fat within the liver, but he has normal liver stiffness measurements. Well, weight loss is the foundation of treatment for NAFLD, for NASH, and for fibrosis. So when we think of any form of NAFLD, we really want to be thinking about weight loss. What's the magnitude of weight loss that's needed, though, to impact liver histology, and it's not an insignificant amount. Data from Vilar Gomez that was published in Gastroenterology in 2016 illustrates this. This was a prospective cohort study that looked at lifestyle changes defined by a hypocaloric diet, so decreasing the diet by 500 to 1,000 kilocalories per day from the baseline, and aerobic exercise up to 150 minutes per week, and looked at 293 adults with biopsy-proven NASH, and they had paired biopsies available for 261 of those adults, so adults who underwent biopsy at baseline and then one year later, and I'll show you this in more detail in our next slide. So patients underwent a routine liver biopsy at zero weeks. Then they were advised to undergo a dietary intervention. Now, this was not a randomized controlled trial. This was a cohort study where all patients were advised to undergo dietary intervention and physical activity, and then followed for one year. The dietary intervention was a low-fat, hypocaloric diet, so on average decreasing by 750 kilocalories per day, 64% carbs, 22% fat, and 14% protein, focusing on dietary fiber of at least 20 grams per day with 10% or less saturated fat, and they were urged to keep a food diary. They were also recommended to start physical activity, generally walking at moderate intensity starting at 90 minutes per week and increasing to 200 minutes per week. They did have a number of touch points with the medical system. They underwent two-hour individual visits with a registered dietitian at weeks 8, 16, and 24, where they also completed a three-day dietary questionnaire and physical exam questionnaire, and then they had two-hour group visits and completed their questionnaires at 32, 40, and 48 weeks. So even with these number of touch points and this counseling, individual and group, weight loss was difficult, and it varied among the subjects. 70% of subjects failed to achieve even a 5% total body weight loss during this one-year study, but 30%, as we can see here, did achieve at least a 5% total body weight loss, and 10% achieved a 10% or more total body weight loss, which allowed the investigators to look at the varying impact of weight loss on histology. You can see here on the y-axis some changes in histology, and you can see on the x-axis change by weight, so weight loss less than 5% total body weight loss going up to weight loss greater than or equal 10% of total body weight loss, and impressively, when patients were able to achieve at least a 10% total body weight loss, 90% of them had NASH resolution on biopsy at 12 months, and even those that were able to achieve at least a 7% total body weight loss, the majority of them, 64%, were able to achieve NASH resolution, which we think is a driver of ongoing fibrosis. I think even more impressively was that for those patients who were able to achieve a 7% total body weight loss, no patient over that year had a worsening of fibrosis, so 0% compared to the overall total of 16% and 21% in those who achieved less than a 5% total body weight loss, and impressively, for those patients, again, who were able to achieve at least a 10% total body weight loss, 45% of them had an improvement in fibrosis of at least one stage, and the remaining 55% had stability of fibrosis, suggesting that really this 10% mark is really our sweet spot, although we can have a lot of benefit even when patients lose at least 7% of their total body weight. Other limitations to this study, it only assessed the impact of weight loss on histology after 12 months, and so we don't have longer-term follow-up, and the impact of weight regain, which we know is very common, and that impact on histology, and of course, we know that in this study that was close to real world, but did have more counseling than often were able to achieve in the clinic, only 30% were able to achieve at least a 5% total body weight loss, but based on this study and others, we do recommend that patients try to achieve at least 7% total body weight loss with the goal really of 10% to improve NASH histology and potentially prevent the progression of fibrosis. Something else to consider, although we won't touch on this further today, is that if patients have failed lifestyle interventions such as this before and have other indications for weight loss surgery, they should be considered for referral for weight loss surgery, as we know that that is the most sustainable method to lose weight. Now, what about the diets that we should use for our patients? Well, there's limited data about the best diet for patients with NAFLD, but this interesting study was published in the Journal of Hepatology Reports in 2021 and looked at intermittent fasting and low-carb, high-fat diets in NAFLD to see the impact on liver fat. So this was a 12-week study of a diet that was So this was a 12-week study of adults with NAFLD who were randomized either to intermittent fasting in the 5-to-2 method, so eating normal number of calories in five days and about 500 kilocalories on two days, a low-carb, high-fat diet, or a standard of care for 12 weeks, and the primary outcome was change in liver fat fraction by MRI. So you can see illustrated here that patients were randomized in a 1-to-1-to-1 ratio to standard of care, the intermittent fasting in the 5-to-2 form, or the low-carb, high-fat diet, similar to the keto diet that were often asked about by our patients. Those who were in the low-carb, high-fat diet had the largest change in diet composition with a significant increase in fat intake and a significant decrease in carb intake. And what we can see here is that both the intermittent fasting and the low-carb, high-fat diet had a significant decrease over 12 weeks in the liver fat fraction. Those in the standard of care, who we presume were a highly motivated group, did have a decrease, a relative decrease of 16.8%, but those who were in the intermittent fasting and the low-calorie, high-fat had significantly greater decreases, again, in just 12 weeks in relative liver fat fraction of 50.9 and 53.1%, although there was no significant difference between these two groups. So this at first suggests that at least in terms of liver fat, and for a pilot study, we can potentially recommend intermittent fasting or low-carb, high-fat diet for our patients who are interested. But I think there's an important caveat to that. The low-carb, high-fat diet was associated with a trend toward an increased LDL level, and we know that NAFLD is an independent risk factor for cardiovascular disease in our patient, and LDL is one of the strongest risk factors. And so it does give us some pause about whether this is a cardiovascularly healthy diet for our patients with NAFLD. This is distinct from intermittent fasting with the 5-to-2 method, which was the best tolerated of the diets, reduced LDL, and also on fiber skin, reduced liver stiffness, which was not seen in the low-carb, high-fat diet. So overall, intermittent fasting seemed to be very well tolerated in this short study, reduced LDL, reduced liver fat, and reduced non-invasive markers of liver stiffness. Now, again, this is a short study. How long patients can maintain this remains to be seen, and the impact on histology remains to be seen. But I think this is really interesting and encouraging preliminary data that warrants further evaluation and would give me a little bit more enthusiasm about recommending intermittent fasting to highly motivated and interested patients. So recently, Zobair Younosi, Joseph Kim, and I published the AGA clinical practice update on lifestyle modifications using diet and exercise for NAFLD, and we followed the data based on the studies that I just recommended. For our patients, we would recommend a hypocaloric diet, reducing calories by 500 to 1,000 kilocalories per day. We'd recommend a diet with fresh vegetables, fruits, legumes, minimally processed whole grains, and fish, as well as other lean proteins with a focus on omega-3 fatty acids as primary fat sources, minimizing saturated fatty acid intake, including dairy, red, and processed meat. And it has been shown that this diet, even in the absence of weight loss, can reduce liver fat. Other things that it's important to guide our patients on are to limit commercially produced fructose, especially things that contain high fructose corn syrup, to focus on a diet that's high in fiber, which may be beneficial in NAFLD, to restrict or eliminate alcohol use. But right now, there really is insufficient data on other types of diets, specifically meal replacement or vitamin supplementation for the benefit in NAFLD. Now, what about exercise? A recent meta-analysis of 20 studies with 1,073 NAFLD patients showed that exercise of varying types improved ALT, AST, and intrahepatic triglycerides, regardless of weight loss. And this really is an important point that I make to my patients, that exercise, even in the absence of weight loss, can help the liver and certainly help ALT, AST, and liver fat. Exercise with diet changes was also shown in meta-analysis to improve the NAFLD activity score. And we can see to our left here, the benefit of exercise alone favoring exercise and the benefit of exercise plus diet on the NAFLD activity score. And interestingly, there was no difference between aerobic exercise and resistance training, although some studies have suggested some benefit with aerobic exercise using resistance training as supplemental. So for our recommendations, we generally recommend that patients engage in moderate intensity activity of 150 to 300 minutes per week, which is certainly a lot of exercise, or vigorous intensity activity of 75 to 150 minutes per week. Resistance training may also be beneficial, and the data suggests that 120 to 140 minutes per week may be enough and can complement aerobic exercise. And it may also, for our patients who are deconditioned, be less intense, use less energy consumption, but therefore may be feasible, especially for those with limited cardiorespiratory fitness, especially at the beginning of their exercise journeys. One problem, though, with exercise is that exercise is difficult to start, and exercise rates are very low among adults with NAFLD. So this was a very interesting study published by Jonathan Stein out of Penn State that looked at the barriers for exercise so that we could better understand them. They surveyed patients with NAFLD, and 91 percent of those surveyed agreed and understood that exercise was important in improving NAFLD, and 88 percent desired to be more active. However, 75 percent did not achieve that at least 150 minutes per week of exercise, and 41 percent performed zero minutes per week of physical activity. They found that fitness tools and resources are underutilized compared to other populations, with only a third of patients using a fitness tracking device or belonging to a gym, and only 30 percent involved in an exercise program. Interestingly, some of the top barriers that were found were a lack of exercise resources or education from providers, found in 47 percent, suggesting that we could all do a better job at talking to our patients about the exercise that we need to do and helping them identify resources, physical discomfort from exercise, including fatigue and pain, which is something I know I don't ask enough about for my patients who aren't exercising, and time constraints. Cost was also an understandable barrier for 21 percent of patients, and lack of resources or access to exercise resources was a barrier at 24 percent. But reading this really gave me pause about the counseling that I do for my patients with NAFLD, and trying to not just advise them to exercise, but when they aren't exercising, trying to understand what their greatest barriers are and work with them, such as if fatigue is a barrier, thinking more about resistance training, or working with them to identify resources or educate them about exercise in their community. We talk a lot about how we need 150 to 300 minutes per week of exercise, and that certainly can be intimidating for our patients, but one thing from the EASL guidelines that I really think we need to take to heart, and these from the clinical practice guidelines of 2016, was the quote, any engagement in physical activity or increase over previous levels is better than continuing in activity. And so I think it's very important for our patients that we emphasize that even if they get on that treadmill for five minutes or walk around the block, any increased exercise is wonderful. It's something that they can start with and something that they can build upon. And focusing too much on the high numbers can sometimes be intimidating and be a barrier to patients. So for our patient, we would recommend a hypocaloric Mediterranean diet, decreasing kilocalories by 500 to 1000 kilocalories per day. We'd recommend whatever the patient was most motivated to start, aerobic exercise or resistance training, and make sure that we talk to the patient about available resources for exercise, perceived barriers to exercise, and any discomfort associated with prior exercise. And for patients who are interested and motivated, consider intermittent fasting, either in the five to two method or as data accrues, perhaps in the 16 to eight method for 16 hours of fasting and eight hours of consumption. This may be a viable method for our patients that can also lower LDL and improve liver stiffness. And with that, I'm happy to take any questions. Thank you. Hello, I'd like to thank the ASLD Governing Board and the program organizers for the opportunity to speak on emerging medical therapy for NAFLD in obesity. My name is Manal Abdul-Malik, I'm professor of medicine and director of the non-alcoholic fatty liver disease clinical research program at Duke University. Here are my disclosures. So, my learning objectives in the next 20 minutes are to review the emerging therapy, which has potential to treat NASH with particular attention for those patients with obesity. I'll be reviewing data on Orlistat, GLP-1 receptor agonists, SGLT2 inhibitors, and highlighting a little bit on other gut-derived hormones such as FGF19 and FGF21. Pharmacotherapy for NAFLD currently is reserved for patients with biopsy-proven NASH and fibrosis in accordance with the ASLD guidelines. The ESL guidelines also are aligned to the ASLD guidelines and pharmacotherapy being reserved for patients with NASH and fibrosis, but does suggest that those patients with less severe disease who are at high risk for fibrosis and NASH, less severe disease, who are at high risk for fibrosis progression, particularly those patients with diabetes and metabolic syndrome or persistently elevated liver aminotransferases or high necro-inflammatory scores on liver biopsies, may also be candidates for pharmacotherapy. The Asian Pacific guidelines state that patients who don't have steatohepatitis or fibrosis should not receive any pharmacotherapy for the primary treatment of NASH and should receive counseling for a healthy diet and physical activity. So clearly our patients want to lose weight. This gives them such joy and any therapies we offer them that also help facilitate weight loss are typically embraced. Now we do have medical therapy for obesity, which induces weight loss. And there are with the primary intention that therapeutic endpoints for NASH are NASH resolution without worsening of fibrosis or an improvement in fibrosis without worsening of NASH. There are several FDA approved anti-obesity drugs. And as you could see, many have been approved since the late 1950s with the most recent being some apatite. Now there are common side effects for these drugs such as dry mouth, insomnia, nervousness, increased heart rate or blood pressure, fatigue or nausea or GI symptoms. There's also the risk of polypharmacy when utilizing such drugs in the context of obesity and diabetes. And that the placebo subtracted weight loss varies from three to 10%, but there is limited data on these anti-obesity drugs on the histologic outcomes of NASH. I'm going to highlight only those drugs for which we do have data and that have been studied in randomized controlled trials of NAFLD and NASH. Let's start off with Orlistat, which has been demonstrated in a parallel group open label 24 week randomized trial of 170 patients with obesity and NAFLD to improve hepatic steatosis compared to placebo. And as you can see here, there has been a downgrading in the severity of hepatic steatosis in the Orlistat group and a representative image of MR-PDFF baseline 29% down to 8% after 24 weeks. And so Orlistat compared to placebo improved also weight and correlated with a reduction in hepatic steatosis. Another study evaluated weight loss due to Orlistat versus diet. And in 50 overweight subjects with biopsy proven NASH who received either diet and vitamin E 800 international units or Orlistat 120 milligrams three times a day for 36 weeks. The Orlistat group lost on average about 8.3% and the diet and vitamin E group 6%. However, the change in the NASH score on follow-up liver biopsy, in fact didn't correlate with the treatment arms, but the proportion of weight loss. And so those patients, irrespective of the treatment who lost more than 9% body weight were the patients who had a significant reduction in macro inflammation, steatosis, and even inflammation compared to those that lost less than 9% body weight. What about GLP-1 receptor agonists? Now this is a class of drugs that has gained a lot of attention in the therapeutic landscape of NASH. GLP-1 receptor agonists improve insulin sensitivity, they decrease lipolysis, they have significant cardioprotective effects and improve insulin secretion in the pancreas and can decrease hepatic glucose production. There are several GLP-1 receptor agonists that have been studied in phase two trials with somaglutide currently in evaluation in phase three global study. GLP-1 receptor agonists decrease weight. In the Sustane-3, Sustane-7 and Sustane-10 trials, somaglutide was compared to exenatide versus dilaglutide versus loraclitide. But what you can see here is the somaglutide in this class had the most potent weight loss effects compared to the other GLP-1 receptor agonists. And in a recently published New England Journal paper in March of 2021, a double-blind randomized placebo-controlled trial, phase three study of patients with obesity or overweight states with comorbidities who did not have diabetes. All subjects received standard weight loss intervention, 17 counseling sessions or a hypocaloric diet and increased physical activity versus somaglutide at a dose that increased to 2.4 milligrams subcutaneously once weekly. And as you could see here, somaglutide had a significant improvement in weight loss approaching 15% reduction in body weight. And in this 68 week study, 32% of patients lost more than 20% body weight with 50% of patients enrolled losing more than 15% of their body weight. Now, how does the GLP-1 as a class apply to the treatment of NASH? Well, loraclitide was evaluated in the 48 week phase two study of 52 patients with biopsy proven NASH. Loraclitide at a dose of 1.8 milligrams versus placebo was evaluated with a primary outcome being resolution of definite NASH with no worsening of fibrosis. And as you can see here, loraclitide improved NASH in 39% of patients compared to 9% on placebo. And this was associated with a significant reduction in weight and a significant improvement in serum ALT compared to placebo. And when we drill down into the individual histologic features of loraclitide, not only are we seeing a resolution of NASH, but you could see that loraclitide improved steatosis, improved ballooned hepatocytes, and while it did not improve hepatic fibrosis, fewer patients who received loraclitide had fibrosis worsening compared to patients who received placebo. Somatotide has also been studied in a 72 week phase two study of 320 patients with FOGSI proven NASH fibrosis stage one to three. The intervention was placebo versus somatotide ranging in a dose of 0.1 to as high as 0.4 milligrams subcutaneously once daily with resolution of NASH and no worsening of liver fibrosis being the primary outcome. And somatotide met the primary outcome in all the doses studied. Not only in those patients who specifically had stage two and three liver fibrosis at baseline, but in all randomized patients. However, it did not result in improvement in fibrosis without worsening of steatohepatitis. Somatotide demonstrated a dose dependent reduction in body weight as did we observe also in those patients with diabetes, a dose dependent improvement in glycosylated hemoglobin. Now, somatotide did not improve hepatic fibrosis at any of the doses studied. However, when we look at the data a little differently and evaluate those patients, the proportion of patients that have fibrosis worsening, you could see that those patients that were treated with somatotide and particularly at the highest dose 0.4 milligrams compared to placebo, fewer patients 4.9% on somatotide 0.4 milligrams versus 18.8% on placebo had worsening of fibrosis. Suggesting as we did see in the laryngotide study, that there may be a protective effect with use of GLP-1 on fibrosis progression. Although in these studies as of yet, we have not seen fibrosis improvement. Now, there is another GLP-1 receptor agonist. This is a triple agonist that has combined GLP-1, gastric inhibitory peptide and a glucagon agonist and a triple agonist from Harmony Pharmaceuticals in a 12 week phase 2a study, multiple dose ascending trial of obese non-diabetic patients with NAFLD as defined by MR-PDFF of greater than 10%. Patients received HM15211 in increasing doses from 0.1 to as high as 0.8 milligrams per kilogram, subcutaneously once weekly versus placebo with the primary outcome being more than a 30% relative reduction in MR-PDFF. This was very well tolerated with only mild treatment, mild to moderate treatment, emergent adverse effects that are known and in keeping with the gastrointestinal symptoms of GLP-1. And as you can see here, the majority of patients that received 0.2 milligrams or higher had a greater than 30% reduction in MR-PDFF with 100% resolution of liver fat at the highest dose in as little as eight and 12 weeks. This is a representative MR-PDFF here on the right showing no change in quantitative liver fat and placebo, but a complete defatting of the liver in eight to 12 weeks with HM15211. And there will be more to be seen with this compound. GLT2 inhibitors have been evaluated in NAFLD, both on the effect of liver fat and ALT. And then the ELIFT study, which is a randomized open-label study of impegloflosin versus standard diabetes treatment in 42 patients with diabetes and NAFLD. Impegloflosin improved liver fat compared to placebo as well as demonstrated an improvement in serum ALT compared to placebo at the end of 20 weeks. And in a separate double-blind placebo-controlled trial of 37 patients with diabetes and NAFLD, impegloflosin 300 milligrams per day was associated with lower hepatic triglycerides, which correlated with weight loss. So SGLT2 inhibitors in NAFLD has been studied in an umbrella review of seven systematic reviews of SGLT2 inhibitors, which included 67 to as high as 498 patients and these seven systematic reviews, four of them evaluated the effect on liver enzymes, four on reported changes in liver fat and two reported changes in fibrosis biomarkers. Overall results, as a summary, noted that none of these studies really were rated of high quality and only one of the studies had moderate quality data. Five of the systematic reviews indicated that SGLT2 inhibitors could decrease liver fat and liver enzymes. And one single-armed histologic study showed an improvement in hepatic steatosis. There was no evidence of liver fibrosis improvement in this umbrella systematic review. Just a comment on FGF19 and FGF21 for NASH. These VAT-derived hormones work via FGFR4 or FGFR1C beta-clothyl and in the liver reduce hepatic steatosis by improving insulin sensitivity, decreasing de novo lipogenesis and improving fatty acid oxidation and also reducing lipotoxicity. And in doing so can reduce hepatic steatosis, inflammation, hepatocyte ballooning and potentially even fibrosis. Aldofermin, which is an engineered FGF19 was initially demonstrated to improve fibrosis in as little as 12 weeks in an open-label study of aldofermin evaluating one milligram in three milligrams once daily for 12 weeks. There was a suggested improvement in fibrosis at both the one milligram and the three milligram treatment arm with a mean change in baseline of 0.5 fibrosis change with three subjects having an improvement in fibrosis of more than two stages with a median reduction also in the NASH score. In a follow-up study of 78 patients with biopsy proven NASH fibrosis stage two and three who had an MR-PDFF of greater than 8%, greater than or equal to 8% who were randomized in a one to two ratio to placebo versus aldofermin one milligram. You could see that aldofermin decreased liver fat content. However, there was no improvement in fibrosis without worsening of NASH and no improvement in NASH without worsening of fibrosis. However, when the dual surrogate endpoint was combined with both fibrosis improvement and NASH resolution, aldofermin had a statistically significant improvement over placebo. Now, unfortunately, in recent top line results of the Alpine 2-3 study, which evaluated biopsy proven patients with biopsy proven NASH who had fibrosis stage two and three for which the primary endpoint was greater than or equal to a one stage reduction in fibrosis without worsening of NASH, aldofermin failed to achieve this primary endpoint compared to placebo and will not be developed any further specifically for the treatment of NASH. However, when you take a look at the data, all surrogate markers of disease activity, whether they're a change in liver fat content or an improvement in liver aminotransferases, both AST and ALT, or an improvement in surrogate markers such as Pro-C3 in fibrosis, there was a suggestion that aldofermin did improve the biomarkers of disease activity, but failed to improve the histologic endpoint. Aldofermin is still being evaluated for patients with compensated cirrhosis, and that study is currently enrolling. Just a quick comment on FGF21. There has been evidence that FGF21 through Bio 89-100 at a dose of 27 milligrams once weekly did result in a significant reduction, a greater than 30% relative reduction in liver fat in 86% of patients, and greater than or equal to 50% relative reduction in liver fat in up to 71% of patients. Up to 43% of subjects normalized liver fat completely, and greater than or equal to 30% relative reduction in liver fat has been shown to correlate to a resolution of NASH and an improvement of fibrosis, and so there will be more to come in our evaluation of FGF21. Afroxifermin is a long-acting FGF compound that has been evaluated in patients with biopsy-proven NASH fibrosis stage one to three, who had liver fat content of greater than 10%. 80 patients were randomized in a one-to-one-to-one ratio of 25, 50, or 70 milligrams of Afroxifermin over 16 weeks, and these subjects who achieved 30% reduction in MRI-PDFF were the only subjects that had a liver biopsy, and so only two patients achieved this in the placebo treatment arm and underwent liver biopsy versus a significant proportion of patients who achieved this endpoint and underwent liver biopsy in the Afroxifermin arm, and as you can see, Afroxifermin improved hepatic fibrosis in all patients, including those patients who had fibrosis stage two and three at baseline. So the key takeaway is that medical weight loss therapies can hold promise in the management of NAFLD and NASH and obesity. It's unclear if medical therapy for weight loss is superior to effective lifestyle modification that it can achieve the same thresholds of weight loss. GLP-1 receptor agonists improved NASH without worsening of fibrosis, and the effect of anti-obesity medications on NASH-related fibrosis at this time is unclear, but we are going to see emerging therapies for NASH and obesity across the classes of GLP-1 receptor agonists, SGLT2 inhibitors, as well as FGF19 and 21 emerge in the coming years. With that, I thank you for your attention.

genetic architecture

non-alcoholic fatty liver disease

precision medicine

targeted therapies

prevalence

obesity

hepatic steatosis

genotyping

heritability

variants

lifestyle interventions