All right, so my task is to discuss the epidemiology of obesity, diabetes, and HCC. I will start by one of those maps showing the global incidence and mortality of hepatocellular carcinoma. It's a very common cancer, third or fourth largest killer, 700,000 cases. Many places shown in the very dark red have very high incidence rates. And then as the color fades, these are the lower incidence rates. Keep that in mind, especially for the very red areas, because changes are happening. And these changes are shown in this slide. So the top three lines where the incidence is declining, these are the high incidence areas shown in the previous slide. So this would be Korea, China, and the bottom one is Italy. While you have areas where the hepatocellular carcinoma incidence is going up, and these belong to the low or the light colors. So as the incidence in areas that are traditionally high density is declining, it's going up in traditionally low incidence areas. This is a slide about the US, and I know this is a global talk, but I want to make a couple of points here. The incidence in the US is going up, but as part of that increasing incidence, the major racial and ethnic groups that are affected by HCC are changing. So traditionally, Asian Americans or Asians who recently immigrated from their home countries had the highest age-adjusted incidence rates of HCC. Only in 2012, this group has been surpassed by Latinos or Hispanic in the United States. And so try to hold on to these pieces of information. They'll become important as this story unfolds. So quickly, you don't need to know that obesity, or you don't need for me to tell you that there is a global epidemic of obesity. Many countries have 20-30% proportions of obese people. Some people term this as globesity. Then diabetes seems to parallel the incidence and prevalence of obesity, estimated 371 million around the world who have diabetes. So you have two very common conditions that are increasing, a cancer that is not that common that is also increasing. So if you're a traditional alarmist epidemiologist, you would look at tables like this. It shows you the prevalence of the risk factor in the general population. And in this case, I want to turn to you your attention to metabolic syndrome, which includes obesity, diabetes, and a few other things. And the prevalence, based on the maps that I've shown you, is 30-40% in the world population. The risk estimate, which I will get to, which means how many folds people with these entities have an increase in HCC, is really low. So it's 50% to 100% increase. But if you take those numbers, then one would expect that 30-40% of the burden of HCC would be attributed to one or more component of the metabolic syndrome. This is really becoming more prominent. What used to be a hypothetical risk in the days of hepatitis C and hepatitis B that is active, now it's going to become even more. Because what happened to hepatitis C and hepatitis B, their prevalence is dropping. Their estimate that the virulence, their ability to cause liver cancer, has dropped significantly after ACE-VR and sustained suppression with HBV. So their population attributable fraction is dropping progressively all over the world. And therefore, this estimate of 30-40% attributed to metabolic syndrome is probably going to be larger in the future. More than a hypothetical risk, this abstract was presented in this meeting. This reflects the distribution of the risk factors in patients with cirrhosis in Texas, U.S. Multiple cities, multiple patients with cirrhosis. These estimates are 2018, 2017, 2019. To me, it's stunning. To look at cirrhosis, in general, hepatology practices where hepatitis C that is active present is only 16%. Hepatitis B is only 2%. Alcoholic liver disease, really a big chunk. But the rest of this wasteland of cirrhosis is NAFLD, obesity, and diabetes. This is what is happening now in practices all over the world, I would argue. So what is the link between this entity, the metabolic syndrome, and the risk of HCCs? Many epidemiological studies have defined metabolic syndrome in different ways. And a meta-analysis done of those studies, and it shows that the presence of metabolic syndrome, however you define it, is associated with a 76% increase in the risk of HCC. Great introduction, but really not very useful information. Two very common or multiple common conditions increase the risk of a rare cancer by a little bit. Nice introduction, but I will get deeper. So my talk is going to be less useless as it progresses. So by the end, you are really going to get something useful. So now I'm going to go less useless. Let's dissect this metabolic syndrome and talk just about obesity. Medical studies from all over the world that shown in this table, this is the variable here is obesity, defined by BMI greater than 30, or overweight, BMI 25 to 30. And for those who are obese, the risk is about 2. So obesity confers a twofold increase in the risk of hepatocellular carcinoma, overweight a little bit and probably not significant, 20% or so. What about diabetes? Again, a good introduction to the discussion. If you look at cohort studies, which is probably the best way to examine this association between diabetes and HCC, they show something like this original study from our group years ago that shows basically type 2 diabetics are at approximately twofold increased risk of developing hepatocellular carcinoma over a follow-up of 10 to 15 years. Since that study, 24 other studies with a similar design from different countries, and they remarkably show the same point estimate. It's around two. So do these entities mean the same thing? And the answer is no. While obesity and diabetes overlap, they seem to confer an increased risk of HCC independent of each other. Having the two together is really bad, but each one of them increases the risk independently. And this is but one example from a meta-analysis that examined the effect of obesity with or without diabetes. And this last box I'm highlighting in red shows that whether people have diabetes or they don't have diabetes, two studies in the former, three studies for the latter, show anywhere between 70% to 1.8 fold increase in the risk of hepatocellular carcinoma. So now try to put it together. You have two common entities that are overlapping. You have a rare cancer that is rising. It's a problem. It's a problem because just telling you that is not useful. You can't really act on it. It's a condition that affects half of the globe. How are you going to even move towards preventing HCC? And the other concept is distal versus proximal association. So basically obesity and diabetes defined as someone saying I'm diabetic, someone who looks like obese, leading to HCC. There's a whole vast land between being here and going there. By just telling someone because your BMI is 30, you're at a higher risk of HCC twofold doesn't do anything to you. You cannot use that information. So one needs to look at proximal associations, meaning things around here and around here and around here, away from BMI and diabetes and more into things that are meaningful to understanding the pathogenesis and biomarkers for diagnosis, prevention, and treatment. So what am I talking about? I will discuss a few things. Moving away from the big picture, metabolic syndrome, useful but not too much, diabetes, useful but not too much, obesity, almost not useful, into abdominal obesity, humoral mechanisms, medications, presence of NASH or NAFLD, and genetic factors. This is where the field is moving to make the epidemiology more meaningful. So I told you about the obesity thing, and I said the twofold, but for overweight it was almost not significant. If you move proximal and discuss abdominal obesity, multiple studies have shown that those with waist circumference, waist-hip ratio, in the highest tertile of the population have three times higher risk of HCC than those in the lowest tertile. Some people have seen this picture, but for those of you who haven't seen it, I'll keep showing it. This is me before going on my last round of diet and exercise, and that's how I look right now. To tie it into hepatocellular carcinoma in the United States, remember that part when I said Latinos are the group that is now surpassing Asian Americans? Well, it turns out to be that as a population, intra-abdominal as well as hepatic adiposity varies among ethnic groups, with African Americans notably the least of the groups, and Asian Americans, Native Americans, and Latinos have high that. So if you want to sort of put the pieces together, this piece of abdominal obesity fits into the overall epidemiological patterns. Now I'm going to talk about biomarkers. This is still an area of evolving findings. The studies have not really found anything consistent, but this is what I mean by humoral mechanisms. Rather than saying it's BMI that is associated with an HCC, this is an example from a population-based multi-country European study that examined biomarkers, C-reactive protein, IL-6, C-peptide, adiponectin, leptin, et cetera, and it shows the association with things that are associated with obesity, but not all the time. So it's a little bit more useful to look at things like IL-6 and C-peptide, et cetera. Admittedly, most of the studies have not found the same biomarker elevated in most studies, so just stay tuned. So now I'm moving to the third proximal thing. So I said proximal things, abdominal obesity, humoral mechanisms, treatment. Someone's just messing up with my slides. Is that a divine intervention? Maybe. This is time for the music. Here we go. So now the treatment. So among diabetics, one could narrow things down instead of just seeing diabetes. So it turns out to be in all the studies that examined the association between type 2 diabetes and the risk of hepatocellular carcinomas. The subgroup of patients who were treated with insulin had the highest risk of HCC. The subgroup that received metformin had the lowest risk. Now you could say this may reflect the severity or the chronicity of the diabetes, but there is reasonable translational science that indicates some inherent properties for metformin to reduce oncogenesis, etc. That's slightly more useful, again, than saying type 2 diabetes increases the risk of HCC. Diabetics who are treated with statin have a lower risk of developing HCC than those without statin. Another thing that I mentioned in the intermediate pathways is genetic predisposition. This magical PNPLA3, and this is the number of the polymorphism discovered from GWAS studies associated with the risk of NAFLD, but also cirrhosis and also hepatocellular carcinoma. The beauty about it, it's not associated with obesity or insulin resistance or plasma triglycerides or cholesterol. This idea of, oh, it just happens in the same groups that are anyway predisposing to HCC is not there. And it is associated with the risk of hepatocellular carcinoma, particularly among Caucasians. And if one looks at the population distribution of this SNP, it corresponds to the high groups that are affected, such as Mexico, Guatemala, Latin America. So now I'm going to move to my last and biggest intermediate or proximal association to explain why metabolic syndrome, diabetes, obesity associated with hepatocellular carcinoma. One has to look at NAFLD nowadays. So NAFLD clearly overlaps, as well as predisposed and possibly caused by several features of the metabolic risk factors, obesity, type 2 diabetes, hyperlipidemia, hypertension, metabolic syndrome. These prevalence rates are at least two to three times higher than the general population. Maintaining the international or the global theme, you see that the global prevalence of NAFLD among type 2 diabetics is greater than 50%. So clearly this is a condition that lies somewhere in the pathway between diabetes and hepatocellular carcinoma. So what evidence do we have that NAFLD increases the risk of hepatocellular carcinoma? I'm slightly biased, so I'm going to present the best study out there, which is ours. So this is a large study from the VA, National VA, that examines approximately half a million patients with NAFLD based on biochemical criteria. In other words, elevated ALT and the absence of other causes. And looking at all comers, comparing them to another half a million without NAFLD, there is a seven-fold increase in the risk of hepatocellular carcinoma. The absolute risk is really, really small, but the relative risk is quite elevated. So can we make this slightly more useful? Yes. So what about the metabolic risk factors that were associated with NAFLD in this large cohort in the VA? So this shows the distribution of diabetes, hyperlipidemia, hypertension in different combinations in the beginning of following that cohort in the VA. And the next few clicks will hopefully demonstrate the point that over time, the proportions of patients with NAFLD who have not just diabetes, but diabetes, dyslipidemia, hypertension enlarges over time as the proportion with cirrhosis, HCC, or death increases over time. So this is an evolving process in which accumulating more metabolic risk factors on top of NAFLD is associated with increased risk. If you look further into the risk of HCC among NAFLD, something jumps at you, and I don't want this to be lost in the shuffle of too many slides and too many messages. Cirrhosis or advanced fibrosis in the setting of a NAFLD is clearly the major risk factor that predisposed people to develop hepatocellular carcinoma. Second is diabetes. Among all the metabolic risk factors, diabetes seems to confer the strongest risk of HCC on top of an existing NAFLD. Almost reaching the end, but I'm going to leave you with a strong visual message. This is the VA cohort. All of those have NAFLD. The Y-axis refers to the risk of developing hepatocellular carcinoma. This shows the power of accumulating multiple metabolic risk factors. So, dyslipidemia alone, dyslipidemia and hypertension, dyslipidemia and diabetes, all three together. And if you layer that on a BMI with obesity, you really get more. So what is that group that has eightfold of HCC? It's the group that has high BMI, diabetes, cholesterol, and hypertension. So this is now the closest that one can get from saying, oh, you have obesity, you get hepatocellular carcinoma. Now we're talking about you have obesity, you also have NAFLD, and you may have cirrhosis, and then you have those things on top of each other. It gets closer to where epidemiology might translate into predicting risks. I'm going to skip this and just go for my last slide here. Someone is going to say, because I know the rest of the talks, it's not all cirrhosis. NAFLD is associated with this HCC in the absence of cirrhosis. And the answer is correct. In this large study from the VA, approximately 13% of all comers with HCC did not have underlying cirrhosis. But what did they have? They almost invariably, all of them, either had a known NAFLD or a metabolic syndrome. So it's not this mysterious entity, and it still funnels through some of the pathways that I delineated through the talk. So in short, there's really a change in the underlying risk factor for hepatocellular carcinoma, and that change is happening globally as the incidence of active hepatitis C and hepatitis B is declining. At least in the U.S., the face of a contemporary hepatology practice is almost no active viral hepatitis, all metabolic risk factors. This creates a real dilemma. You have lots of people who are at an elevated risk of hepatocellular carcinoma, none of whom have a dramatic increase in the risk. So the good old days when you saw someone cirrhotic drinking and having hep C, and you tell them, I know you're going to develop cancer in three or four years, they're gone. You don't see those people. You have more dilemmas. The relative risk related to metabolic syndrome is modestly elevated. The absolute risk is small, and we need more knowledge related to risk stratifications, mechanisms, and the effect of treating the different components of the metabolic syndrome. Thank you very much. We have time for a couple of questions from the audience while we are waiting for these questions. Can you estimate when HCC related to NASH is going to take over HCC related to hep C in the U.S., for instance? Excellent. So the question was, have we started seeing the effects of NAFLD related HCC in explaining the current trends? And the short answer is, implied from the smart question, is it hasn't yet. Most of what we see is still related to hepatitis C, whether it's still active or recently treated and cured. One has to look at the story of hepatitis C to be able to answer your question, which is it takes three to four decades of a cohort being inflicted with the metabolic risk factors in order to see some of that effect. So my guess is, in the next 10 years, we're going to see 20% increase related to NAFLD-HCC. And unless some dramatic change in the natural history, then I'm talking 15 to 20 years. And that's kind of concurrent with some of the decision analytic modeling studies. Dr. Traubein. This is a very nice talk. Thank you. Hypertension makes more risk. Patient hypertension, of course, take much more drugs. So can I differentiate the risk between hypertension and taking more drugs? Because the drugs, of course, will also contribute to higher risk of HCC. Completely agree. The question, to repeat, you have one of those other big useless conditions, as I say. You have hypertension. How can you peel it further to say, is this an effect of hypertension? Is it an effect of comorbid obesity and diabetes? Or is it related to the types of the medications that they're using? The short answer is, I don't know. Good question. As you're sort of implying, there are categories of medications that have been recently implicated with the risk of cancer, like angiotensin-related blockers, et cetera. Thank you. Thank you. Very short questions. Very short answers, please. So typically, there are sex differences seen in HCC incidents. So is the NAFLD epidemic actually collapsing them, or still the women show lower risk than males? The sex difference. Great question. I removed that slide. So it's about sex difference in HCC, which may or may not reflect the underlying risk factor. The short thing is, in the days of active hep C and active hep B, the man-to-woman ratio was astronomical, 4 to 5 to 1. Almost all the studies that are coming out show a much lesser male-to-female discrepancy related to NAFLD. It's still higher in men, particularly postmenopausal. We just have a study accepted, looking at this question. As of today, the male-to-female difference in the incidence of HCC in the U.S. has not dropped or changed. Going back to Dr. Lovett's question, maybe NAFLD hasn't kicked in with full gear. Thank you. Very fast, please. Hi. Great. Brad Winston here, Washington, D.C. Great, great presentation. Thank you. Thank you. Two very quick questions. One, I'm struck by the 13%- Better one. Beg your pardon? Better one. We don't have much time, please. Go ahead. We'll incorporate it into one question. The 13% number that you quote for non-serotic HCC seems a little bit high. I know there was a Kaiser poster that had a much lower percentage. One, can you explain the discrepancy? And two, what percentage would you think would justify HCC screening in the non-serotic population? Excellent, two-part question. Under the 13%, there are two categories. One of them that says absence of cirrhosis with the highest possible certainty. And that proportion is actually only 3%. 10% is absence of cirrhosis with reasonable certainty. So I wouldn't put it beyond anything that half of this number of 13% is truly not absence of cirrhosis. And that would explain the discrepancy with other studies. What proportion would justify surveillance? A big one. 50%, 40%. I think given how large the denominator is, it would be at this point in time insane to advocate screening in non-serotic people with the metabolic syndrome. So I don't even have a percentage, but it has to be huge. And as of this point, the numbers don't justify it. Okay, thank you very much. Thank you.

global epidemiology

obesity

diabetes

hepatocellular carcinoma

metabolic risk factors

NAFLD