Dr. Walter C. Willett - What Do We Know about Opportunities to Optimize our Diets

Welcome to anesthesia and surgery combined grand rounds. Um, this morning's lecture is part of our health and wellness lecture series, and we are fortunate to have a world-renowned expert today, Doctor Walter Willett, um, speaking to us on diet and health. Doctor Willett is a professor of epidemiology and Nutrition at Harvard School of Public Health. Uh, he's professor of medicine at Harvard Medical Medical School and previously until recently served as chair of the Department of Nutrition at The Harvard School of Public Health for 25 years. Doctor Willett studied food science at Michigan, Michigan State University and graduated from the University of Michigan Medical School before obtaining a master's and a doctorate in Public Health, um, at Harvard School of Public Health. Doctor Willett has focused much of his work over the last 40 years on the development and evaluation of methods using both questionnaire and biochemical process or biochemical approaches. To study the effects of diet on the occurrence of major diseases. And he has applied these methods starting in 1980 with the nurses' health studies one and two, and then also the health professionals follow-up study. He's published over 1700 original research papers and reviews, primarily on lifestyle risk factors for heart disease, cancer, and other conditions, and he has written the textbook, Nutritional Epidemiology. He's also written 4 books for the lay public, um, and Doctor Willett is the most cited nutritionist internationally. He is a member of the National Academy of Medicine, of the National Academy of Sciences, and the recipient of many national and international awards for his research. Please join me in welcoming Doctor Willett. Good. Well, thank you. And um, Pleasure to be here and see a lot of people showing up bright and early to talk about wellness. So, uh, there's a lot to talk about and I'm, in this short time gonna give just a bit of an overview of some of what I think are the most important things that we've learned over the last couple of decades. Uh, the picture of health in the United States is actually not a pretty picture. Uh, this is looking at life expectancy. I thank the economist for this nice little picture. Uh, basically showing about 1 year ago that life expectancy in America had declined for 2 years in a row. Probably most of you are aware that, uh, recently, uh, the data were out and we declined for a 3rd year in a row. Uh, that's despite the fact that we've made huge progress in And reducing smoking, we've eliminated trans fat. We have all the modern medical miracles that you know about that are created here in our medical area. And yet, uh, the picture is dismal. Um, life expectancy going down and very much going down relative to other, well, uh, uh, well-off countries. Uh, we're now living about 6 years less on average. Our life expectancy is about 6 years less than the Japanese. Uh, and the French have been doing pretty well. Uh, but, uh, actually all the country, major countries except, uh, Japan have had downturns in life expectancy. Uh, this has gotten a little boost from this downturn from the opioid epidemic which has been in the news, but this is exactly what was predicted by a paper that David Ludwig and others, uh, here, uh, published a few years ago in the New England Journal of Medicine because of the impact of the obesity epidemic on Uh, life expectancy. Yeah. Oops. I'm not getting this to go down. If somebody could help me here in AV, um, Uh, I'm not sure what happened. OK, thank you. Um, so, uh, the, a lot of this, a lot of these trends are driven by the obesity epidemic, and I'm sure most of you are familiar with this, but we thought a couple of years ago that maybe we were starting to flatten out at least, uh, and I should say that, uh, in the 1970s. The prevalence of obesity in adults is about 10%, but in the last few years, uh, the data are clearly on the upswing. That was just a temporary little plateau that we had or maybe had there or just a little statistical blip. We're now at about 40% obesity rate in adults in the United States. Uh, and, uh, children have also that we give, we thought we might have bent the curve, but, uh, no, the last couple of years have also shown that we're still trending upwards in the prevalence of obesity in children now at about 18%, that was 5% in the 9 in 1970s, early 1970s. Uh, also, you probably heard in the news a few weeks ago, a paper in The Lancet Public Health, that the rates of obesity-related cancers are now going up in, uh, adults under age 50. Uh, this is really expected because that generation that became obese. As children are now hitting middle age and the consequences are starting to be manifested. And even if we didn't increase obesity rates at all, we'd still see increases in these diseases for the next 30 or 40 years, but we're not, we haven't plateaued, so things are going to get worse and accelerate. Yeah, but the obesity epidemic is just the tip of the iceberg that's obvious visible. Uh, we can see it. The other part of, uh, nutrition and the concern about nutrition is diet quality. And so I'm, that is really equally as serious independent of obesity in terms of the health consequences. So, I'll focus most of my talk on that. It's helpful, I think, to go back a few decades to understand where we've come from in terms of our understanding. This is from the famous seven countries study by Ansel Keyes and others uh looking at coronary heart disease rates around the world. Uh, there were 14 different populations of men, about 10,000 each in 7 different countries. So it was called the 7 countries study. And for the first time, they used a standardized definition of incidence of coronary heart disease. Most striking. Finding was the, the fact that there was about a tenfold difference in rates from northern Finland down to Crete, which had the lowest rate of heart disease at that time. So the obvious question is why. Uh, genetics was raised, but that was quickly ruled out by epidemiologic findings showing that people who migrated from low-incidence areas like these Japanese villages to the US which was about there at this time. Uh, very quickly adopted the rates of coronary heart disease in, in the United States. So it wasn't, this is not due to genes, it's something about the diet or lifestyle or environment. There were some clues from the seven-country study that diet might be important because there was a strong correlation with saturated fat in the diet. But there were many other factors that could be confounders here that differed between these populations. So we clearly needed some more detailed studies. Other epidemiologists were looking at cancer rates around the world. And again, they found very large differences here. This is uh breast cancer uh rates around the world, US and the affluent countries up here, uh, Japan, traditional Asian societies and low-income countries down here with about an eightfold difference in uh breast cancer rates. Same applied to colon cancer and Uh, prostate cancer, most of the common cancers in Western populations, and other epidemiologists again looked at migrants who moved from these low incidence countries to the United States, and those rates of breast cancer do catch up to the, uh, in the Japanese Americans. It takes about two generations before they catch up. But again, it clearly shows that genetics do not explain these huge differences in cancer rates. The, there was a correlation with fat and rates of breast cancer like we saw for coronary heart disease. But again, lots of potential for confounding variables. Uh, lifestyle is very different in traditional Japan and low-income countries compared to the US and so that might be explained by other factors. Yet that correlation, even though it was really the, the lowest kind of epidemiologic evidence was so repeated so often that it became conventional wisdom that fat in the diet was the reason for these high rates of cancers and cardiovascular disease in Western countries. Um, and that was translated into food guidelines. Here, the US Food Guide pyramid that came out in 1992. Uh, so, uh, fat, the real emphasis was on getting fat out of diet. Fats and oils, all kinds, you sparingly. And if you're not going to eat fat, you've got to eat something. So it's inevitably gonna be large amounts of carbohydrate, large, large amounts of starch. And we were told we should load up on that, uh, 6 to 11 servings a day. And if that wasn't enough, they put potatoes there and the vegetable groups are up to, up to 13 servings a day of starch a day, and that was supposed to be good for you. There are some other things about the food guide pyramid that concerned us, uh, These foods, uh, uh, uh, were lumped together, uh, chicken, uh, red meat, nuts, legumes, fish, uh, called the meat group, uh, and those were said to be equal to each other even though we had suspicions that those might differ in terms of their health implications quite a bit. And you've heard that I'm from Michigan and I know that 2 to 3 servings a day is deficiency in dairy products. I was, I grew up, uh, Uh, being told I should have 10 or 5 servings a day of milk, uh, I'll come back to that. But, uh, it's really, there was already some, this, uh, uh, there were some red flags raising here that if we look around the world, it's actually the countries who don't drink milk, uh, at all as part of their culture that have, had the lowest rates of fractures and the reason for drinking all that milk was supposed to be that you were preventing hip fracture and other fractures. There are some early clues that this recommendation of loading up on lots of carbohydrate might not be good for us. A series of controlled feeding studies. These are a few dozen people, uh, given controlled diets for a few weeks and then you look at what happens to the risk factors. And this is a very simple study, but uh it's been repeated time and time again. Uh, starting off on a traditional 40% of calories from fat Western diet, 10% of calories were replaced with either olive oil, which made the same number of fat, uh, calories, or, uh, saturated fat replaced with complex carbohydrates. And this would be more like the recommendation to eat a lot of carbohydrate. This would be more like the traditional Mediterranean diet. And so what happens on these two diets is that Uh, total cholesterol goes down a little bit more, if anything, on the higher olive oil diet, but I think you all know that total cholesterol doesn't mean too much. You have to look at the fractions. Uh, HDL cholesterol on high carbohydrate diets goes down and triglycerides go up on high, uh, carb high carbohydrate diets. And again, that's been repeated time and time again. And we've known from other studies that high triglycerides and low HDL are part of the metabolic syndrome, exactly which is causal, but this syndrome does predict higher rates of heart disease, not lower rates of heart disease. So, this really raised some questions as to whether this high carbohydrate diet was gonna be good for us. We've also learned over the last couple of decades that the pathways leading from diet to coronary heart disease, and I should mention I'm focusing on coronary heart disease initially here because that was and still is the number one cause of death in the United States and now in most countries around the world. But that there was, uh, and there has been a major focus on blood lipids that's appropriate, but we've also realized that diet can influence. Coronary heart disease through many other pathways as well. Blood pressure, thrombotic tendency, influencing insulin resistance, having antioxidant or prooxidant effects through affecting homocysteine, uh, inflammation, endothelial dysfunction, and of course, importantly, a ventricular irritability and sudden cardiac death. So you can't just look at one pathway, for example, through blood lipids and predict how diet will influence heart disease. Ideally, you'd want to do large randomized trials to look at various aspects of diet and heart disease, but that's extremely difficult to do because that usually would mean tens of thousands of people randomized to diets for many years and keeping people on those diets. And we haven't been very successful in, in our attempts to do those kinds of studies, mainly because free-living people uh don't always do what rats do when you feed them. They, they don't stay on those diets necessarily. So, our group has focused a lot of our effort, our primary effort on conducting several large prospective studies. Uh, the first was the nurses' Health Study which started in 1976. Uh, Frank Speiser initiated that mainly to look at oral contraceptives and breast cancer and I joined the next year and was interested in nutrition and its consequences. So we started pilot testing dietary assessment methods and first, uh, distributed our standardized dietary questionnaire in 1980 to the nurses' Health Study. Participants and about 95,000 participants returned that questionnaire. And we've been updating uh diet as we go along over time. Every 2 years, we send out a questionnaire and we've gotten, uh, still about, uh, we have about 90% follow-up in the nurses' health study still after over 40 years now. And uh a lot of what we've learned really is, uh, We have, we have to acknowledge that the commitment of the nurses to the study has really been essential and we would not have been able to learn what we have learned without their very active participation. So, we're updating diet, uh, smoking, uh, physical activity, pretty much everything we could think about, uh, uh, uh, every 2 to 4 years, and then also tracking incidents of pretty much every major disease now. And so if we want to look at, say, some aspect of diet, saturated fat or trans fat and heart disease, we can adjust in great detail for smoking, physical activity, family history. We have the DNA from about 175,000 people. Now, we can adjust uh look at interactions with genetic polymorphisms. We have blood so we can look at pathways mediated through blood lipids and uh hormones and uh nutrient levels. Uh, that was only women, so we enrolled about 52,000 men in the health professionals follow-up study. Uh, we had already started the physicians' health study by that time. Uh, maybe some of you were in that, uh, so we didn't, uh, include physicians here. This is mostly, uh, dentists also, veterinarians, osteopathic physicians, pharmacists, uh, and a couple of other, uh, physician groups. And again, following them with the same, uh, collection of data. And then a lot of information was pointing that for breast cancer risk, which was definitely the primary focus of the nurses' health study in the beginning, uh, a lot of the action is occurring, uh, during childhood and, and, and adolescence. Uh, that's where a lot of risk factors for breast cancer seem to be operating. And so we enrolled another cohort starting younger, uh women who are 25 to 42 years of age, uh, in the nurses' health study to another 116,000 women. And so we're tracking all of these uh individuals over time. I should mention that this is really the work of many people and I've uh put here the names of some of our active colleagues, but, uh, literally several 100 postdocs and faculty members, um, and, uh, uh, doctoral students who have contributed in a major way to this work over the years. Uh, this is one example of fairly recent analysis looking at type of fat and total mortality. And the picture is exactly what we published 20 years ago, uh, in the New England Journal when we looked at type of fat in relation to coronary heart disease risk. And here, this combines our cohort. Uh, so this is about 133,000 individuals, about 33,000 deaths in the cohort. So Huge statistical power to look in detail at type of fat. And what's critical, and I'll come back to this, when you look at something in the diet that has calories, you really need to specify the comparison source of calories because the comparison, what you're comparing it to makes a huge difference. And so this is each type of fat compared isocalorically to carbohydrate in the diet. And uh this is a trans fat by far the worst type of fat in the diet. Uh, this is saturated fat weakly but significantly uh related to higher risk of coronary of total mortality as well as coronary heart disease. Uh, and this is A monounsaturated fat and even more importantly, polyunsaturated fat. And in these and everything, I'll show you this has been adjusted in detail for smoking activity and other potential confounding variables. And this fits quite well what we see with the impact of these. Different types of fat on blood lipids and controlled feeding studies that just go on for a couple of weeks with polyunsaturated fat being the most uh inversely related to risk of total mortality. So total fat is pretty much irrelevant here that that's a fact weakly inversely related to risk of, of total mortality now that we've gotten trans fat out of the diet. But the type of fat is uh very important. Uh, and I should mention you really don't want trans fat in the diet. We have eliminated that as of last year and, uh, and replace some of that saturated fat with unsaturated fats. Uh, there's one important controlled feeding study that has looked at this. It's not a controlled feeding study, randomized trial. That's uh the Predimed study done in Spain. Some of you are probably familiar with this. There about 6000 individuals were randomized to a controlled diet, which was a lower fat American Heart Association type diet, or a Mediterranean diet where they gave them nuts to consume on a daily basis, uh, or gave them extra virgin olive oil along with the Mediterranean diet. And on both of these versions of the Mediterranean diet. There was a lower risk of incident cardiovascular disease and lower risk of diabetes, quite a few other benefits as well, about a 25% reduction in, uh, in cardiovascular disease, uh, despite the fact that they just gave them more fat and told them to eat more fat, and they didn't weigh more in the end either, despite just giving them these very high fat sources of healthy fats. Now, the general, I think the idea that the, the concept that uh the composition of fat, the type of fat is more important than total fat was uh pretty well accepted. But then just a couple of years ago, This idea came along that we should really be loading up with butter and red meat. Uh, this is sort of the paleo kind of, uh, idea. And this is based on really, I think, uh, uh, a logical conclusion that if carbohydrates, especially refined starch and sugar are bad, then, uh, saturated fat must be good. And that's, that really has, uh, no, uh, really no evidence behind it. We've looked at this in detail. Uh, looking at dairy fat and butter, basically butter. In relation to cardiovascular disease risk, uh, total mortality looks the same, but again, this is now a larger number of women and men, about 220,000 followed for over, uh, for up to over 30 years. And the, the vertical line here is dairy fat, and this is looking at the relative risk adjusted for confounding variables, uh, gram per gram or calorie per calorie, uh. Vegetable fat as expected, uh, is related to lower risk of cardiovascular disease compared to the same number of calories from dairy fat. Uh, Animal fat from other sources was related to somewhat higher risk. This would be mainly, mainly from red meat. And, uh, polyunsaturated fat, more specifically, whether it's omega 6 or omega-3, uh, strongly related to lower risk of cardiovascular disease compared to butter. Uh, carbohydrate from whole grains, strongly inversely related compared to butter fat and carb but for carbohydrates from refined starches and sugar, just about equally bad, uh, really sort of a wash. So, again, the, uh, the high uh dairy fat, very high in saturated fat, but does not look at all as we would expect like an optimal type of fat in the diet. So, the point here again, uh, I've tried to show in this diagram is that if you're talking about reducing saturated fat, the, um, The comparison is really the difference. If you replace saturated fat with trans fat, then the risk of heart disease is going to go up. If you replace it with carbohydrates, it's about a wash. But if you then subdivide the carbohydrates into refined starch and sugar, that actually may slightly increase risk of cardiovascular disease, but replacing saturated fat with whole grains would be related to lower risk. Uh, Most importantly though, replacing saturated fat with unsaturated vegetable oils such as olive oil, canola oil, and other similar liquid plant oils would be strongly related to lower risk of cardiovascular disease. So, uh, very briefly, just to summarize, uh, the fat issue that coronary heart disease rates can be dramatically reduced by nutritional means, but this will not be achieved by replacing saturated fat with carbohydrate. And we should abandon recommendations about percentage of calories from fat and avoid pejorative references to fat or fatty food. Fat per se is not unhealthy. And advice about dietary fat should focus on replacing saturated and trans fats with vegetable oils, including sources of omega-3 fatty acids. And in practical terms, we should replace it, be replacing red meat with nuts, fish, uh, poultry, and legumes and uh replacing dairy fat with those foods as well where we, where we can. Now, back to the breast cancer issue where there had been really a very strong belief that high fat, high percentage of total fat in the diet was related to breast cancer was really driving dietary recommendations. The biggest randomized trial ever done with diet was the Women's Health Initiative, where about 48,000 women were randomized to their usual diet or a low-fat diet. And after an average of 7 years of follow-up, there was no significant impact on risk of breast cancer. But the sad truth is here that this was really an uninformative study because if you look at the biomarkers that respond to fat intake, They didn't change at all. There was no difference between these two groups, which really meant that people weren't following the low fat diet. We've looked at this and continued to look at this relationship over time because it, as it was, it was from the beginning, the primary hypothesis in the nurses' health study that high-fat diets increase breast cancer risk. And we just don't see uh any relationship here with over 20 years of up to 20 years of follow-up, uh, just flat relationship with percentage of calories and fat in the diet. And uh actually just a few months ago, we had a paper in Science, uh, uh, that was meant to be sort of a consensus paper uh among low-fat, high-fat enthusiasts, uh, David Ludwig here at Children's did organize that and uh Included a very low carbohydrate enthusiast, a very low fat enthusiast, uh, and after a lot of back and forth, we could agree that, uh, with a focus on nutrient quality, good health and low chronic diseases can be achieved for many people, but really most people on diets with a broad range of carbohydrate to fat ratios. So that, that, that, those percentages are just not what's important. to move on and talk about protein a little bit, that's in, um, in the public awareness for, uh, in recent years. Uh, we've looked in our courts at the sources of protein, uh, in detail. Now, we, we, we'd never eat protein in isolation. It always comes out of in packages with other, uh, with other nutrients, uh, uh, often with fat and again, the type of fat that comes along with the red meat will be really important. So we've looked at protein sources here. And here, the, the bottom line is that replacing red meat with pretty much anything will be uh better, with pretty much any other source of protein will be related to lower risk of coronary heart disease or total mortality. Uh, this is replacing, uh, using statistical models here, uh, red meat with poultry, uh, relative risk of one here is red meat and, uh, significantly lower risk of cardiovascular disease, uh, fish for red meat, lower risk and especially plant sources of protein, uh, replacing red meat are related to lower risk of coronary heart disease. And again, this is very consistent with the controlled feeding studies, uh, in short term looking at blood lipids and risk factors. Uh, similar kind of analysis here looking at total mortality. And again, uh, the, the strongest relationship is nuts compared to red meat, uh, legumes compared to red meat, low fat, dairy, whole grains, poultry, fish, uh, all these protein sources, uh, compared to red meat related to total mortality as well. Uh, we've looked at this in multiple ways. Uh, uh, one of the big, uh, more complex and interesting areas is dairy, which does have a nutritional value, uh, uh, but, uh, it also comes with a lot of saturated fat, about, uh, dairy fats about 65% saturated fat. So again, this is looking here as dairy as the comparison, and we can see compared to the same percentage of energy from protein. Uh, unprocessed red meat, eggs, and processed meat are all related to higher risk, uh, but, uh, plant sources of protein compared, uh, is, uh, isoprotein-wise compared to dairy are related to, uh, lower risk of, uh, uh, uh, overall mortality or cardiovascular disease. So again, the comparison really makes a difference. If you look at red meat or you look at dairy, it looks sort of neutral, but that's misleading, uh, in a sense, it's uh essentially these uh dairy sort of in the middle, not as good from a health, long-term health standpoint as plant sources of protein, but better than uh red meat. Uh, uh, given, uh, this is Children's Hospital, uh, uh, I thought this might be interesting. We've just now been, uh, are able to look at diet during high school years in relation to longer-term health outcomes and breast cancer in particular. Uh, this is from nurses' Health Study too, which is really designed to look at this kind of issue. And, uh, this is looking at breast cancer, again, looking at substitution of, uh, different sources of protein for red meat. And uh we do see that uh compared to red meat, consumption of legumes and nuts is related to lower risk, uh, poultry also related to lower risk, fish about equal, eggs about equal, or, uh, and of course nobody's gonna eat all legumes or all nuts or all poultry. If you look at all these alternative sources, which sort of describes a Mediterranean diet compared to red meat, uh, a significantly lower risk of breast cancer. Uh. So, uh, uh, you know, this is very interesting because if we look at these same relationships during midlife, we don't see much of any relationship with breast cancer at all. But it looks like, like, as we expected that diet during adolescence for breast cancer is particularly important. So what we feed our, our girls, uh, during their school years does turn out to be, uh, very important for cancer risk later in life. Again, we've looked at this issue from another lens, another perspective. Uh, this is looking at overall diet and plant-based diets versus uh more animal sourced diets. And what we, we've created a score called a plant-based dietary index. A very simple score, you get 1 point for a plant source food, a negative point for an animal source food, and uh then look at risk. You have type 2 diabetes, we see the same picture for uh coronary heart disease and total mortality. So this blue line is the relationship between this plant-based dietary score and type 2 diabetes. And importantly, this, there's no sharp cutoff here. This is really a continuum of risk shifting incrementally toward a more plant-based diet related to a lower risk of these health outcomes. But I think Uh, you're all aware that not all plants, plant-based foods are healthy. Uh, Coca-Cola and, uh, and donuts are plant-sourced foods. And so we created, uh, a healthy plant-based diet score, the green line here, which, uh, which included foods like, uh, whole grains, uh, nuts, legumes, uh, fruits, vegetables, and then this unhealthy plant-based dietary score which was Uh, refined starch, and sugar-based foods and not too surprisingly, uh, that was, that unhealthy plant-based diet score was related to higher risk of all of these health outcomes, uh, whereas the healthy plant-based diet score was related to lower risk. But the basic point here, again, this is really a continuum. There's not a sharp cutoff. So, uh, just to summarize this a little bit on protein sources, replacing red meat with a variety of other protein sources, especially soy, nuts, and other legumes, will have major health benefits. I think everyone here is aware, even though maybe not in Washington, that climate change is a really uh almost existential issue for the future of society as we know it. And what we eat also has major impacts on climate change. Uh, so, uh, uh, we need to keep an eye on that as we, uh, make our food choices, as we make our dietary recommendations. And there are huge differences in the effects of different foods on greenhouse gas emissions. This basically red meat, lamb, and beef have huge impacts on greenhouse gas emissions. Other animal source foods, sort of intermediate cheese, pork, farmed salmon, chicken. Uh, uh, quite substantial, but less than red meat. But most of the animal, uh, vegetable, plant source proteins actually have quite modest impacts on greenhouse gas emissions, lentils, uh, dry beans, tofu, for example. So, uh, the moving more toward plant-based diets has great benefits for planetary health as well as for individual health. A few words on carbohydrate and like the fat, it's really a type of carbohydrate that's more, much more important than the total uh percentage of calories from carbohydrates. Uh, thanks to General Mills. Uh, I have this nice little slide here that shows what we do with the grains in our food supply. Most, uh, we start off, of course, and we think, uh, I remember seeing something like this in 2nd grade. Uh, this is the whole grain, three main parts, the bran on the outside, and that's where most of the minerals and vitamins are. And the germ, which has the fat soluble vitamins. And the embryonic plant and then the endosperm, which is most of the grain, which is mostly starch, and we can mill all that, turn it into whole grain flour, or wheat, for example, but most of the grains in our food supply are refined, and that means we take out the germ, we rip off. The bran and then we uh mill that and that gives us sort of white Wonder bread and that, that's almo almost pure starch and that leaves the germ in the bran. What do we do with that? We mostly feed it to animals because that's where the good nutrition is and they grow big and strong and we eat this stuff over here. And what does that do to us? Uh, uh, harm, basically. Uh, this is looking at type 2 diabetes risk. We see the same relationship in all of our cohorts. Uh, and that these refined starches are basically giving us a lot of glycemic load. This is, glycemic load is just multiplying the glycemic index of a food which is really high in white flour by the total amount of carbohydrate. And as you can see going from low to high. Glycemic load, increasing risk of type 2 diabetes, and also going from low to higher cereal fiber is related to lower risk of type 2 diabetes. So, the worst possible carbohydrates are just like those white bagels and muffins and things outside. And that was related to a substantially higher risk of type 2 diabetes. Uh, the lowest risk was with lower glycemic carbohydrates and unrefined, uh, grains, which would be high in cereal fiber. So the type of fiber makes a very important difference, and there's no single number to Uh, that we can use for carbohydrate quality. It's a combination of the glycemic load, uh, the amount of cereal fiber in it, but also very importantly, of course, uh, sugar and water or sugar-sweetened beverages, and, uh, those we have seen are related to more weight gain and very importantly to risk of type 2 diabetes, even independently of weight gain, and this has been seen now in dozens of studies, uh, very robust statistically. Uh, so, Like, uh, dietary fat, carbohydrate quality rather than the the percentage of calories and carbohydrate, uh, is important for long-term health, and we should really be consuming our grains in the, in the form of high fiber, whole grains. And there's a huge amount of evidence now that will reduce the risk of type 2 diabetes, coronary heart disease, but refined grains is likely to, are likely to increase the risk. Uh, high intake of refined starch and sugar is particularly problematic with underlying insulin resistance. I didn't have time to go into that, but if you're really lean and active, we can, and have low insulin resistance, we can better tolerate carbohydrate in our diet. But if we're like most Americans who put on a little weight, uh, and are more, have even a modest degree of insulin, we have a much more adverse metabolic response to those high glycemic loads. And a reduction of soda and sugary beverages is a very high priority. A quick couple of words about uh dairy. Again, uh, very interesting topic and we still have a lot to learn because of the complexity. But the evidence is quite, is really quite strong now that drinking a lot of milk, uh, it does not reduce the risk of hip fractures. Of course, calcium is essential and we have to have some in our diet, but for reasons I won't go into the current recommendations for calcium intake are almost seriously overstated and they're much higher than the World Health Organization definition of adequate calcium intake is the basic reason is so. The DRIs, RDAs are based on studies lasting only 2 to 3 weeks and you're almost for sure gonna get a misleading answer looking at such uh short-term studies. So this is a meta-analysis looking at milk consumption over a very wide range up to 30 glasses a week. No reduction in fracture risk. And again, uh, because a lot of you, uh, deal with children. Uh, where we are, uh, with diet, milk consumption during childhood and adolescence has been said to be particularly important because that's when we're growing our bones and we need to build up that calcium bank to protect us from fractures later in life. It, uh, we, again, we definitely do need calcium. We need some more while we are growing. We, we have to be in positive calcium balance. But I, uh, a number of years ago, we Uh, saw very clearly that greater height is related to increased risk of fractures, and then we've come to see in more recent years, uh, when we learned that height is related to higher risk of many cancers, uh, trying to understand what determines height, it turns out the most important factor by far is milk in the diet because it has all these anabolic hormones in it and uh anabolic amino acids. That milk will accelerate growth rates and lead to taller height. And yeah, I'd probably be a few inches shorter if I didn't have those 5 glasses of milk growing up in Michigan. And uh because milk increases. Uh, height and also height is related to risk of fractures. It made me wonder whether it's possible that uh drinking large amounts of milk during adolescence might actually increase risk of fractures later in life. And that's what we found for boys in particular, uh, very substantially higher risk of hip fractures later in life with drinking large amounts of milk. During adolescence. And a lot of this was explained by uh the greater height uh with adolescent milk consumption. For women, there was no reduction in risk. We didn't see the same increase as we did with men. We're still trying to understand that. But, uh, that actually, yeah, I'm probably at increased risk of hip fractures because I drank all that milk during childhood. Uh, and there are some potential downsides of high dairy consumption too. Again, we do know that from controlled feeding studies that drinking a lot of milk will increase our blood levels of insulin-like growth factor one, which uh is also related to higher risk of prostate cancer and several other cancers. And, uh, in a number of studies, this is an older study, one of the first to look at this in Seventh-day Adventist men, an increase in risk of fatal prostate cancer with higher dairy consumption, and we've seen something similar in our cohorts as well. So, Uh, consuming 3 or more servings of dairy products per day will have little effect on fracture risk and is likely to increase the risk of fatal prostate cancer. Uh, so the exact amount of milk is optimal at different stages of life is not yet clear for adults. We really don't, it's pretty clear we don't need any. Uh, we need to have some calcium intake. Uh, the exact amount is still, I think, to be determined, but the World Health Organization of 500 mg a day being adequate probably is, is reasonable. Um. And uh I didn't have a chance to go into that, but vitamin D, uh, inadequate vitamin D is a serious issue in the United States, and the consequences do include increased risk of fractures and probably colorectal cancer and, uh, and, and other cancers as well. And for most of us who stay inside and live in northern climates, uh, some supplementation is, is, uh, I think, a reasonable thing to, to do. Uh, more recently, we've been looking at cognitive function and cognitive decline in our cohorts as our participants are aging. And uh what we see is that it looks like diet has a very important influence on our rate of cognitive decline. Uh, most of the research in this area, I think has been Uh, almost naively looking through a very little time window. And the reality is for, uh, uh, development of dementia, Alzheimer's disease, uh, this is really a very long-term process. Probably we are at our peak function in medical school or sometime around that and the sad story is that for the rest of our life, we're on a downward slope and So, what we really want to do is if, if we can just change that slope a little bit by the time we're age 70 or 80, uh, a little change in slope will have a big impact. But most of the studies looking at, uh, drugs, uh, which have all failed in terms of influencing, uh, cognitive decline or uh other lifestyle factors that they've been looking at studies lasting just 2 or 3 years. Uh, except for one study and nobody pays any attention to it. And that was a physician's health study where people, men were randomized to beta-carotene or placebo. And after 4 years, there was really no difference in cognitive function. But the study went longer than any other study and by 13 years or so, there was a significantly better cognitive function in the men on beta-carotene compared to placebo, which I think is a really important finding. Again, nobody's paying any attention. To that, but, uh, part of that illustrated the benefit of carotenoids, but also the fact that you really need long-term studies to look at, uh, cognitive function. So, this is in our health professionals follow-up study of men. We're doing this in women as well. And we have, uh, it started in 1986, repeated measurements of diet and covariance and then we're looking at cognitive function, uh, assessed by a simple, uh, self-administered test. Here, I won't go into the details, but we take the average of these two measurements of cognitive function out of 2008, 2012, and then look at many years later, uh, many years before that. Diet. Uh, and we put in a lag here to deal with the fact that, uh, the possibility that cognitive function might be affecting diet or reverse causation. So we put in a big lag here. And we do see that uh men who were adhering to a more Mediterranean type dietary pattern had about a 40% lower risk of poor cognitive function. Uh, and a pretty nice dose response relationship had a significant trend. We've seen the same thing in women as well, and now we're digging down because we really have big numbers here, uh, and what is it about the Mediterranean diet that might be important and it particularly seems to be vegetable consumption and digging down further that looks like the carotenoid rich, uh, fruits and vegetables are particularly important for maintaining cognitive function. So that's a work in progress. We have published a few papers on that, but uh we're, we're doing a lot more on it right now. Now, back to obesity. Again, one of the huge drivers of this downturn and plateauing and then it looks like downturn in life expectancy. Uh, the blue line here is, uh, healthy people who never smoked and uh we've done that because, looked at that group because smoking is an important confounder when we're looking at weight. Smokers tend to be thin, but because they smoke, have a lot of uh increased mortality. So the blue line is really giving us the, the best relationship. And it's very clear that The lowest mortality here is under a body mass index of 25, and even with slight increases just in the mild overweight range, we see significant slightly higher levels of BMI, significant increases in overall mortality, and then it goes way up in the obese range. Yeah, so, contrary to some of the headlines you might have seen in the New York Times, it's not good to be overweight or obese. We, we really, uh, uh, wanna keep our, uh, fat mass low. Uh, and A lot of their work has been on childhood obesity, and we, we need to do even more there, but we've pretty much ignored weight gain in adults, uh, which turns out to be equally important. And the average weight gain in adults is about 1 pound a year, which sounds like very little, but by the time someone's reached 50, they put on 30 pounds, and that makes a huge difference in risk of many diseases. This is a composite of major chronic disease. And this is, you can see here, even at a weight gain of 5 kg, there's a significant increase in risk of major chronic disease and this is about average weight gain here and for Americans, uh, and uh quite substantial increase in overall burden of chronic disease and again, then it goes up from there. So, what is it that affects weight? Um, There's no single magic bullet here, I think as everyone would understands. Uh, but the percentage of calories from fat is really not important. This is a recent, very well done randomized trial by, uh, uh, Chris Gardner at Stanford, and, uh, this is on a low-fat diet and, uh, a low carbohydrate diet, uh, virtually identical weight loss here. And we've done a meta-analysis of, of over 50 study randomized trials that are going on for 1 year or more. And if anything, the people on low-fat diets tended to gain a little more or or had less weight loss than the people on Uh, uh, lower carbohydrate diets. So that fat in the diet is not really the explanation for our obesity epidemic or the solution to, uh, weight change, uh, our solution to obesity. Uh, what is it about diet, uh, that's important to, uh, it looks like, uh, again, it's a composite, uh, many different foods are contributing a little bit to whether we're more likely to gain weight or maintain weight. And the foods, uh, this is a major analysis we published a few years ago combining all three of our cohorts. Uh, Dari Mozaffarian was the first author on this. And uh some of the foods were related to more weight gain, potato chips, potato, uh, uh, mashed potatoes and any potatoes in any form, processed meats, unprocessed meats, butter, sweets and desserts, refined grains, uh, cheese about neutral, and then vegetables, nuts, whole grains, fruits, and especially yogurt were related to lower amounts of weight gain. Uh, and, uh, fruit, uh, and sugar-sweetened beverages, uh, really the most serious, uh, form of calories, and this is, this is per serving per day, and the really serious problem is that many people have 2345 servings a day of sugar-sweetened beverages. And so that is the single most important contributor to weight gain. Uh, uh, fruit juice is also, uh, positively related to more weight gain and then the dairy products sort of neutral. Uh, but it happens that this pretty much describes a Mediterranean diet if you look at these foods related to less weight gain, and that's been supported by some randomized trials that have looked at randomized, I looked at Mediterranean diet, uh, compared to low fat diets or low carbohydrate diets. The Mediterranean diet looked as though it was most successful over the long run. So that's a quick uh uh run over a lot of different topics. We've done some analysis looking at how much we, uh, how much major disease uh might we prevent if we put these all together in a package of low-risk, uh, lifestyle factors. And of course, that would be non-smoker. This is our definition of low-risk body mass syndex. Less than 25, uh, exercise, half an hour a day of brisk walking or more, and a good diet score. We uh defined that low trans fat, high polyunsaturated to saturated fat ratio, low glycemic load, higher cereal fiber, uh, fish twice a week, adequate intake of folate, alcohol, moderate amount, and Uh, uh, somewhat to our surprise, even in this population, this is the nurses' health study of health professionals, only 3.1% of the women fell into this low-risk category. But if everybody had done that, we could estimate that about 82% of heart disease would have been prevented. And we've updated that analysis. Uh, a nurses' health study to see basically the same findings. So there's a huge potential for reducing coronary heart disease, but only 3.1% of women, even in this health-conscious group were taking advantage of that full package. We've also looked at a similar analysis for type 2 diabetes, and there we estimate about 92% of type two diabetes could be prevented by A healthy diet and lifestyle package, but again, very few, uh, a very small percent of the participants were following that healthy lifestyle package. So, I was asked to uh say a little bit what can physi, how can physicians help promote healthy eating and uh this is just a quick menu of uh thoughts about that. Uh, first of all, I would put right at the top of the list, practice healthy eating ourselves, uh, because basically, uh, physicians who interact with patients are not very good about counseling on health, healthy. Behaviors if we don't practice them ourselves. We've done a great job on smoking. We need to do the same on diet. Uh, and we need to track our patients' BMI for, of course, pediatric populations. We have these growth charts. We do that well. We really need to be tracking adult BMI as well. And, uh, looking at weight change since age 20. Uh, that weight gain during adult life is very important, uh, uh, to uh minimize. Uh, and then, uh, we should try to assess our patients' diets even if crudely, and there we do need to give physicians, uh, some simple better tools for doing that. But even, uh, pointing out to patients that we're concerned about their diets is important. And then we need to develop and offer a simple menu of options for weight control and improvement in diet quality that uh and studies have been done, even when a patient comes into a physician's office and the patient is obese, about half the time, the physician says nothing about it, and that's almost a certification that I went to see my doc, no problem, didn't even mention my weight. Uh, so, uh, saying nothing is actually a problem. Um, and most of us don't have time to Uh, give detailed counseling, but, and that's why we need to provide a, a number of options and different people respond to different kinds of, uh, programs for weight control, sometimes just providing them with information and Uh, this is a, a book actually I, uh, put together with the help of Pat Skarrett, who's a gifted science writer, uh, for the general public. I drink to be healthy. Uh, disclaimer, I don't get any royalties from this, but this is a new edition and this is pretty much up to date if you want any more information on this. And it's written for the For the general public, but most physicians got almost no nutrition background in medical school. I, I didn't really, uh, and that unfortunately hasn't changed. Uh, so physicians find that pretty useful to us. It's, it's pretty easy read. That's, for some people, that's enough and they're, they're motivated, they can really make a big difference. Uh, other people need something like Weight Watchers group support, uh, others, individual counseling with a dietitian, but, uh, to give, uh, patients some option, uh, letting them know this is an important issue, and, and giving them some options as to how they would like to best take on a healthy weight control. Uh, some physicians are finding it useful to get more involved in dietary counseling, uh, and cooking and really, uh, engaging with patients more directly. Uh, that's not for everybody. Uh, unfortunately, we're all too often rushed and need to have some help in, uh, guiding patients. Uh, taking advantage of teachable moments, uh, when somebody develops hypercholesterolemia or hypertension, uh, unfortunately, too often it's just a prescription, uh, but, uh, nutrition and lifestyle are, are the major determinants. These are the real causes for most of the people who are developing those conditions, uh, and yet we're not taking advantage of that. And then, uh, I, on the list, I put, uh, not, maybe not for everybody, but consider expanding your influence in your own institution or beyond, uh, that, uh, unfortunately, hospitals and the healthcare system in general have really not been, uh, stepped up to the plate, not being part of the solution here as I go around seeing the Harvard hospitals even. Uh, it, it's, it, it's not a pretty picture. Um, you don't need data. You can look at our staff and at all levels. Uh, uh, many are seriously overweight and, uh, look at the cafeteria, the food services. Uh, I hate to say it, look outside, those breakfasts are not healthy and we shouldn't be damaging our own health or, and we should be setting a good example. We have on our website at the School of Public Health, healthy muffins for, uh, breakfast occasions. So I invite whoever. Or orders the food to, um, that, that we could have healthy muffins that we serve. When we look at uh Our meetings, we're often, uh, uh, group meetings, we're often serving pizza and, uh, soda, uh, just setting a horrible example as well as damaging the health of, of people who are consuming that. We really should be setting the best possible examples and, and we're not. Uh, and, uh, unfortunately, I often felt uh that Physicians and others have nihilism about dietary change. It's hard to change, but don't be discouraged. That was the same thing we heard about tobacco smoking too, and yet we've had a dramatic effect there that the changes we have won't happen overnight. They won't, are talking to patients about healthy diets and weight isn't going to. Uh, uh, change everything overnight, but over, uh, we really need cultural change and bit by bit, we, we do see that diets do change. They, they've changed a lot actually over the last few decades, mostly in the wrong direction, but some, in some good directions too, uh, sugar-sweetened beverages are down about 25% now over the last 10 years or so, and that's, that's really important change. So that's a quick spin over a lot of topics and you may have a minute or two, I think left for questions to set. Right. Um, Doctor Willett, um, thank you for that great talk. Um, You and I spoke, uh, before the talk and, uh, even 25 years ago at the School of Public Health, uh, you were doing great work in teaching us. You didn't mention caloric restriction. Yeah, uh, caloric restriction in principle is what we'd like to do, but it's really hard to do that. Uh, the, uh, the trials that, that have attempted caloric restriction over a period of a few years have just not been able to achieve much at all in caloric restriction. Uh, so what seems to work best is changing diet quality, which, uh, indirectly will Leave people more satisfied and have, uh, uh, it looks like better effects on long-term energy balance, which is really what we're trying to achieve, while at the same time having a lot of benefits independent of the, uh, independent of weight control as well. So, I understand the, uh, the point about sustainability. If it's not sustainable, then it's not gonna work. But, um, is this more hype than reality that about the benefits of caloric restriction, if you can maintain it or achieve it? Yeah, I, I think if people really could restrict their energy intake, uh, it will show up as weight change if we really do it. But again, it, uh, and almost if that really would result in if they could do it, people would either lose weight or at least not continue to gain weight and that would be good. Now, of course, it depends on how you're restricting it. You can restrict it by, by limiting healthy foods and, and that would be not good. So dietary composition is really important. Restricting it in a healthy weight would be good, but Somehow we're obviously wired to, uh, uh, to eat calories and, uh, and consciously restricting caloric, calories is not, it's, it's very, very difficult. In the trials that attempted that, they, uh, people could do it for a short run but just couldn't maintain it. So, it doesn't look like that's gonna be a very successful way just focusing on caloric restriction. Yes. Thank you. It was fascinating. I, um, It seems like amongst a certain part of the population, this message is getting through. You've, you, you've made your case. Um, it's not as obvious and easy for the entire population to understand is, uh, that it's bad to eat a poorly composed diet than it is to put your mouth on an exhaust pipe and inhale it. So, and it took decades to get people to stop smoking, but that's a yes or no phenomenon, and we could mandate it. You cannot smoke in our facility. And it was very difficult to get people to get over the habit and get over the addiction. But we have to eat, so it's more subtle to change what you eat, and much of what we eat is based on taste and carb addiction, and that, that, that taste and addiction has happened because we were fed cereal when we were little kids, uh, and it's hard to break. So For example, you talk about all the muffins out here, which are, in terms of institutional change, I, I think about that. Um, those healthy muffins that are probably on your website. Don't taste good to most Americans, right? No, we try to and try to. We, we developed those recipes with the Culinary Institute of America, and they're actually pretty good, right? So, so, so that's the question for some of us, right, who think about it, who are educated, who can afford it. Like, I'm not doing prepared meals that are mailed to my house that are healthy, and, and, and how do we get it to the point where, uh, I mean, even in hospitals, we're serving the same processed grain. Um, how do we get us that people actually like and get used to having things that have the proper dietary composition? We should ban muffin. I mean, I don't wanna be paternalistic, but we should have an alternative. We can't just say no. It was easy to say no to cigarette smoking in a hospital. It's hard to say no to bagels, but we have to provide an option that's easy and tastes good. Well, I think, I think we should say no to bagels, and, but if a healthy attractive option out there is is. Uh, yeah, the point you're making is, uh, uh, that we could have another few hours of discussion about that. Uh, and there's no simple answer to it, but, uh, I think it's very important that you point out that part of the population is taking advantage of this information and they're doing very well. It's sort of the Whole Foods crowd and it's strongly related to income and education. Um, and we really need to find ways to help everybody come along. Uh, but the fact that some people are doing it and really seeing, uh, the, what's happening in the US, we are on two different paths. Part of the population is getting very much healthier. Another part of the population is stagnated or getting worse, uh, and so that's spending that life expectancy downward. So it's, it's not inevitable. I think that's a very important point that we all just go off in a bad direction. Um, I did, I was asked to, a few years ago, chair a task force for New England to look at how we could, uh, develop healthy weights. Uh, and, uh, we, you have to break this down. It's so complex. So we thought we could break it down into about 8 different sectors. Uh, it was schools and that's quite a bit of work is being done there. And again, it's the well-off communities that can hire a dietitian and chefs. They're doing some fantastic work. Low-income communities are usually being left behind. But it can be done and we have to do, everybody, all kids should have that, uh, uh, that experience. Uh, it's healthcare system and we're not, we're, we're not doing a good job in the healthcare system. That, um, all right, we're going to all stories, but, uh, but there's some good examples. Mass General, uh, their cafeteria for their staff is actually really good. I don't know if you've been over there, but if you haven't, take a look around. It has lots of information and lots of good options, and they've documented that they can influence choice. Not everybody takes the healthiest choice, but they are, uh, shifting the balance of toward healthier foods there. What we serve to our patients, we should be dealing with. Every wall, every opportunity should be messages and examples of healthy foods, but most of all, setting good examples and Going around hospitals, seeing at nursing stations, you know, pizzas and soda and stuff that we just should not, that we, we, we can't be doing that. It's, it's damaging our employees, it's damaging, it's setting a horrible example, uh, as well. Healthcare system, work sites, another sector. Actually, a lot of companies are getting it because it's a double win for the employees and for their bottom line as well. The whole physic, the whole food environment and media are very important issues. We're not gonna win. Uh, entirely unless we can control the information that kids are getting better, that in one survey, kids, kids saw about 2000 advertisements in a year for food, zero for fruits and vegetables. I mean most of it was junk. Uh, and, um, the physical environment making it safe to ride a bike and walk and exercise. We're doing something about that. We, we need to do a lot more. Uh, we need to, uh, the economics of this need to be dealt with as well. And then there's some uh details on surveillance and, and, and monitoring this. But if we sort of break it down by sectors and I, I find it really helpful. It's so complex as you say, there's so many influences and uh if we do something right in just one of these sectors, that's not gonna solve the problem. We need to be working on all of them, but Uh, every one of them is necessary too. So wherever we are, we, we should be working on what we have, where we have influence. Uh, and, uh, I think those of us in the health system can do a lot, uh, right here. In some ways, we should be the leaders, uh, in, in shifting this balance, but we're not. Uh, uh, and, uh, again, we do have examples of communities, and I haven't seen the numbers myself, but apparently in Massachusetts for the last couple of years, we, first of all, our life expectancy has not gone down like the rest of the country and we have Uh, at least he stable on the weight issue, and, and we, we, I think, uh, we have been putting more resources into this in Massachusetts. Uh, we, we need to do a lot more, but it does, it does look like we are starting to see some of the benefits. We have a huge amount of sort of uh turning back at the time, uh, but it's not, uh, inevitable that we just keep climbing, uh, uh, in terms of overweight prevalence. Yes. We can talk to stuff we'll put apples out there next week. OK, great. Right. Yeah. Alright thanks. I I. Oh. Oh. I. it was really.