Optimize Your Diet with Walter Willett Part 1

>> Here at Integrative Nutrition, we strive to give you the most up-to-date information on health and nutrition in the most unbiased way. Nutrition is a unique field as there are always going to be a lot of opposing views. Although you may not agree with everything that you're exposed to in this field, we encourage you to hear all the facts and then make your own judgment to determine what works best for you. In this class, Walter Willett aims to show you the importance of preventive nutrition. He highlights a lot of important research gathered over the last few decades to show you what we have learned about diet and health. We understand that all of this information can be overwhelming. As you watch, focus on the underlying message, the important role that diet has in both development and prevention of disease. Keep in mind, you do not need to know everything to get out there and coach clients. It's about understanding and awareness, and it's up to you to decide how you may want to use this information and put it into action in your life and your clients'. >> I'm going to try to cover a few of the highlights of things we've learned about in the last 10, 20 years that are most important for people to understand if they want to be healthier and also to weave into this, a little bit about how we learn about diet and health. And that I think is really important, how do we gather the evidence that we need to understand, what is really the smartest best choice when we have so many things we could choose in our daily eating habits. The problem is of course for many people and you'll hear this, "I'm confused," that if you look at the web or pick up almost anything to read about health and nutrition, you can find for any food or any meal positions that this is the most toxic thing you could possibly eat all the way to the other end that this will save you from cancer, heart disease, and you'll live forever and be happy all the way. So how do you sort out this huge amount of mixed information and somehow come up to understand what are the healthiest choices? So I'm going to look at the kinds of evidence that we really do need if we want to put the odds in our favor when we sit down at a meal in terms of our likelihood of being healthy over the years. I think it's helpful to go back in time, about 50 or 60 years, to look at one of the early kinds of studies that provided powerful evidence that most of the diseases of western civilization are not inevitable. And this was the Seven Countries Study lead by Ancel Keys, we call this an ecological or correlational study. Pretty simple, he basically looked at 14 different populations in 7 different countries, about 1000 men each, and then over a decade, carefully monitored their rates of coronary heart disease or heart attacks, the number one cause of death in United States and now the number one cause of death in almost every country in the world. And what he found was this that there was about a tenfold gradient and risk from the highest country, which was Finland at that time, to the lowest region, which was Crete in Greece. And other epidemiologists, about the same time, were doing what we call migrant studies. They were looking at people who moved from low incidence countries like Crete or the Japanese villages that are in the lower left-hand corner and they found that those migrants who moved to the United States after not too many years adopted the rates of heart disease as European Americans living in the United States. And so that said very powerfully that these very large differences in heart disease rates were not due to genetic factors. And I think if you read the news now, you'll hear all about genetics, genetics, genetics. They play a minor role actually in determining who will get and who will not get heart disease if you look around the world. This says that something about diet or lifestyle is much more important, much more powerful than genetics. And of course, the next step is if we could really understand what it was, then we could maybe modify our risk, so everybody could be at low risk like people living in Crete. There were some clues from this study that saturated fat might be an important part of the problem because of this correlation that you see there. But we know that many other things are different in the high-risk versus the low-risk countries, differences in physical activity, other aspects of diet, maybe sleep, see that people in Crete took a siesta for example, and that has been shown to be related to lower risk of heart disease. Smoking, lots of things could be alternative explanations for these differences in rates and therefore what we would call confounding variables. So this was a clue, but we couldn't really conclude cause and effect that saturated fat was a major determinant of heart disease. There were other epidemiologists at that time looking at cancer rates around the world. And they found a similar picture, large differences in rates of breast cancer, in colon cancer around the world. And if we looked at people who moved after some time, they adopted the rates of the new environment. Again, playing into something that was potentially modifiable, you know, might be diet, might be animal fat in the diet, but we shouldn't really conclude that this was causal, in terms of breast cancer, shown here, reproductive factors play a very powerful role. And we know that that might explain some of these differences, breast feeding, things like that. So you really again should not draw conclusions from that. But those correlations were so tempting that many people did draw the conclusion that fat was the major determinant of cancer, breast cancer, colon cancer, and many other cancers in western countries, even though the evidence in fact was very shaky. So this conclusion really did permeate dietary guidelines in the late 1980s, 1990s, up until quite recently. And this is the 1992 US Food Guide Pyramid. Right at the top, it says all types of fats and oil should be used sparingly. And then of course, you have to eat something, so if you're not gonna eat fat, by default, you will end up eating large amounts of carbohydrate, which they put at the base of the pyramid. We should have up to 11 servings a day of things like Rice Krispies, and Wonder Bread, and crackers, and things like that. That was supposed to be good for us, even though there was not a shred of information that eating large amounts of those foods was actually good for us. And if that wasn't enough, carbohydrate in the diet, they put potatoes there in the vegetables, so you could have up to 13 servings of starchy foods per day. And there are some other curious things in this pyramid that they put together red meat, and chicken, and fish, and nuts, and legumes all into one group and said you should have two to three servings a day, even though there was some suggestion at that time that it made a big difference which of those you were choosing and they didn't distinguish amongst them. And then of course, dairy has an interesting position that it's the only food or milk is the only food that is said to be essential. There's lot of options in these other categories. I grew up in Wisconsin and Michigan, so I knew that you were supposed to have three or four servings a day. But curiously, if you look around the world, most people don't drink milk as an adult and their bones are not falling apart. And we'll come back and talk about that. Well, I began to be a bit worried that this push to reduce all types of fat in the diet and eat more carbohydrate might not be such a good thing. And this was, I think, first suggested by another kind of study. What we call the controlled feeding study done by some colleagues in the Netherlands. And this is where you take a few dozen people and you actually give them food to eat, so you totally control their diet for a couple of weeks. You just need a few people, a few dozen people and a few weeks to do this kind of study. And you look at things like changes in blood cholesterol level or changes in blood pressure. And what they showed was the lower fat, higher carbohydrate diet, which was like the diet being recommended, HDL-cholesterol went down and in the bottom part of the panel, triglycerides in the blood go up. And we know from lots of other studies that low HDL and high triglycerides are related to higher and not lower risk of heart disease. So this was a bit of a red flag that you would think all of us being equal, you might actually be better off on the higher fat diet where they used olive oil compared to the lower fat diet which was being recommended. And this study was repeated over and over again and it's very clear that if you do replace fat, pretty much any type of fat with carbohydrate, HDL goes down, triglycerides go up, and that is a little bit worrisome. Now another player came on the scene, although it had been there for about 100 years and almost nobody was paying attention to it, and that's trans fat in the diet. These are 60 pound blocks of partially hydrogenated soybean oil that I bought at a restaurant supply store just a mile or two from our School of Public Health in Boston, and this is what our fast food industry was using. They were putting it in their fryer, later just turning up to 400 degrees and saying cooked in vegetable oil, but here at room temperature, you can do lots of things, interesting things with these blocks of trans fat. You can build buildings, you can do sculpture with it. But of course, we were interested in what it might do for our coronary arteries. And again, my colleagues in Netherlands, doctors Mensink and Katan led this line of investigation with trans fat. And what they found was if they compared saturated fat to trans fat, it looked like you actually got more of an elevation in total cholesterol levels with saturated fat in the diet. But we've also come to understand that total cholesterol doesn't mean too much. It's really the cholesterol fractions that are important and we want to lower our LDL and raise our HDL cholesterol and the ratio of LDL to HDL gives the best prediction for heart disease. And as you can see on the bottom, trans fat was almost twice as bad on a gram for gram basis as well as saturated fat in the diet. Again, many people repeated this study, found the same thing. And replication is an important part of a scientific process. I'm sure most of you know. And so this did suggest that trans fat could be a serious problem even though people had been eating it and people were in fact being told it was good for them. They didn't say trans fat but they were told to be consuming Crisco instead of lard and margarine instead of butter. And especially in the 1960s, '70s, '80s, the amount of trans fat was very high in those products. However, one of the other concepts that we've come to appreciate over the last decade or two is that there are many pathways leading from diet to coronary heart disease or CHD here. And I've only talked about how a diet can lead to changes in blood lipids and then on to coronary heart disease, but we've come to appreciate that diet can act through affects on blood pressure, thrombotic tendency, meaning tendency for clots to form in the coronary arteries, it can operate through affecting insulin resistance, oxidation, homocysteine levels in the blood, and very importantly, inflammation and endothelial dysfunction, and ventricular irritability and arrhythmia, And actually that last is really important because people usually die of a heart attack because of not just enough blood, but the thing that precipitates death is sudden cardiac death, an acute arrhythmia, and diet can influence the likelihood of that happening. So the point is that if you look just at the effect of diet, like cholesterol fractions like we've talked about already. You could be misled because there could be counterbalancing benefits through other pathways or there could be effects in the same direction or synergy among these various pathways. And for example, we've come to understand that trans fat can make inflammation increase. And that's been documented in many controlled feeding studies now. So probably trans fat would be even worse than we might have expected looking just at the effects on blood cholesterol fractions. The important point is that we want to look directly at heart disease as well as looking at the effects on just one pathway at a time because by looking at heart disease directly, we integrate all these possible pathways. So our group has, for the last 40 years, spent most of our effort conducting a large prospective studies where we look at large populations of people, we have three large studies that include almost 300,000 men and women among whom we had been collecting diet, physical activity, smoking, other lifestyle factors in detail and updating that every two to four years as we go along and then carefully documenting who gets breast cancer, who gets a heart attack, who gets diabetes, etcetera. We were looking at almost every outcome. And then if we wanted to focus in on one dietary factor, we can control statistically for other aspects in diet and lifestyle and focus in on our particular research question. So these studies are rather complex, but as an example, this is what we saw when we looked at types of fat in the diet and risk of coronary heart disease. This is after about 10 years of follow up, almost 1,000 of them had died of a heart attack or been hospitalized for an acute myocardial infraction. And here we're looking at specific types of fat in the diet, each compared to the same number of calories and carbohydrates and controlling one for each other, and in everything else, I'll show you, we're adjusting for smoking and physical activity and these other potential confounding factors. And what we saw was, maybe not surprisingly given what I've said already, trans fat was by far the most strongly related to risk of coronary heart disease. Saturated fat was very weakly related to heart disease compared to carbohydrate but mono and polyunsaturated especially were related to lower risk of heart disease. So total fat was not related to heart disease, it was really the type of fat that was important. And again, these findings have been replicated in multiple other studies. Now in principle, the ideal study is a randomized controlled trial where you randomly assign people to a different diet, a different dietary factor, it might be high trans fat or low trans fat, and then you follow them to see who gets a heart attack and who doesn't and you evaluate your hypothesis in that way. By randomization, you guarantee that the groups are virtually identical. The problem is that it's really hard to keep thousands of people on a diet for many years. And this is one study that did try to do that, the Women's Health Initiative, most expensive study ever done, and this is looking at the cumulative incidence of cardiovascular disease. As you can see the two groups, the low-fat and the high-fat groups are essentially identical there. There was no difference in heart disease rates between low-fat and high-fat groups. Now you might say that was consistent with the data I've shown you already from the Nurse's Health Study in our cohorts, but the problem was, in this study, several years later, they published data on the cholesterol fractions in the two groups and there wasn't any difference. And we know that if you go on a low-fat diet, it does change your cholesterol fractions, HDL goes down, triglycerides go up. Again, there wasn't any difference. So it really meant that they didn't test the hypothesis that people, there was virtually no difference in low-fat or high-fat groups in terms of their fat content and really shows how difficult it is to do that kind of study. There was one study though, a randomized trial, that did show a difference in the treatment groups. And this was a randomized trial of a Mediterranean diet conducted in Spain. They had a low-fat controlled diet and then a Mediterranean diet that was higher in fat where they gave people nuts. And then in the third group, Mediterranean diet where they give them extra virgin olive oil. And as you can see, the rates of heart disease were lower in the two Mediterranean diet groups compared to the controlled group. So why they were more successful in keeping people under diets is an interesting question in itself here. It did show all the benefits of the Mediterranean diet and of the nuts and olive oil, and probably, there was a greater degree of motivation because they actually gave people, on a regular basis, nuts and olive oil, which made it easier and provided some incentive to stay in the study. So to summarize this little section on types of fat and heart disease, coronary heart disease rates can be dramatically reduced by nutritional means, but this will not be achieved by replacing saturated fat with carbohydrate. That's pretty much a wash. We should avoid recommendations regarding percent of energy from fat and avoid pejorative references to fat or fatty foods. Fat is not bad. You really have to focus on the type of fat. And advice about dietary fat should focus on replacement of saturated fat and trans fats. Try to eliminate the phrase low-fat from your vocabulary because it really is scientifically meaningless and actually can be confusing and misleading to many people's detriment. Now back to cancer and breast cancer in particular. There was another randomized trail that tried to look at low-fat diets and breast cancer risk. And this one did show some change in HDL cholesterol. And if anything, the breast cancer rates were a little bit higher on the low-fat diet. There was not a statistically significant difference but it was running in the opposite direction. So the data for breast cancer and fat in the diet are also not supported by randomized trials. Now one major additional problem with randomized trials, when we were looking at cancer or heart disease, is that no-one's quite sure how long you would really need to run the study before you should see, in effect, if there is a benefit or harm. And the randomized trials, again, are really challenged by maintaining people on separate diets. So we've looked at this and this is one advantage of long-term prospective studies where we, we're not telling people to change their diet but we're tracking their diet so we can look at people who have been consistently high-fat or low-fat over a long period. So this is after 20 years. And looking at different intakes of fat in the Nurse's Health Study, by that time, there were over 3,500 cases among postmenopausal women, and as you can see, just no difference in breast cancer incidence across a very wide range of fat intake. Now that's fat in the diet, and of course, that's totally different than fat in the body. And obesity and overweight are huge problems, I think you are totally aware of that and also very much aware that this is an epidemic in United States and in most countries around the world like Mexico going up faster than we are and actually surpassing us in obesity. It's really a tragedy seeing what's unfolding before our eyes in so many countries. There has been some confusing literature in this area, claiming that it's actually better to be overweight. Have some of you heard that? Yeah, and if you read the New York Times, don't read the New York Times if you're interested in learning anything about nutrition or science, unfortunately, they have been pushing, Gina Kolata in the New York Times have been pushing this story that it's good to be overweight. And it just doesn't make any sense, you know, blood pressure goes up, your lipids deteriorate, glucose tolerance deteriorates if you're overweight. And why should that be that good for you? There would have to be some magic going on. The problem is, of course, is I think anyone of you who is talking care of patients is that we don't just go along at our weight and drop dead, most people get sick, lose weight, and then die. And sometimes they can, for many conditions, lose weight for quite a while before they actually die. And also, as we get into our, you know, 80s and 90s, many people develop frailty, meaning we lose weight, especially lean mass, and that's a vicious circle because then we exercise less, and then we lose more lean mass and then we... There's a high probability. It's really a downward spiral that physicians have appreciated for hundreds of years. And basic data are that there is increased risk of total mortality, diabetes, heart disease, many other conditions, anything above a BMI of 25 and definitely when you get to the obese range of BMI above 30, therefore individuals, there will be some differences in a person's optimal weight. And that's one reason why I like to keep an eye on three numbers, waist circumference, which your fat's growing, that tells you that's not a big muscle we're building up down around our waist, so that's unfortunately... And we need to do something about it. And also keep in mind what we weighed at an age about 20 because most people were at a pretty good weight then. Of course, that's changing now and a lot of people are arriving at age 20 already overweight but gains in weight after age 20 are almost entirely going to be fat mass, unless somebody is doing a really unusual muscle building program. And of course, that's obvious that they've been doing that if that's the case. So looking at weight gain, waist circumference and BMI gives you three numbers that will help an individual know when they need to be paying attention to weight, most of us do in reality. So it is really important to maintain a healthy weight. And for most people that means not gaining or gaining as little weight as possible as we go through our adult life. Now what is the best diet to help maintain weight? Until fairly recently, there was surprisingly little good evidence on that topic and it's so important, it was really surprising. But in the last 15 years, there have been many studies that have looked at different dietary factors in weight control. I just don't have time to go into those. That's a big topic in itself. You can study this by randomized trials because you need just a few 100 people and following people for a year or two will give you good information, less than a year is just too short because people on almost any diet lose a few pounds and then regain it. Just a change in the diet will usually be worth a few pounds of weight loss. So to, I think, summarize this quickly, the bottom-line is that the percentage of calories from fat in the diet is, again, not an important factor. If anything, the data in some studies tend to show low carbohydrate diets, and under some circumstances, being a little more effective for weight loss. This was probably the best study, the POUNDS LOST study showing no difference for a high-protein, low-protein, high-fat, low-fat, or high-fat, low-carbohydrate diet. This was done in about 800 people followed for two years. And the fat or macronutrient composition of the diet really made little difference. Now some people took away from that, "Oh, diet isn't effective and your dietary choices are ineffective in controlling weight." And that is the wrong conclusion. It's just that the macronutrient mix is not very important but other aspects of diet quality do seem to be importantly related to weight control. And there is no single factor, there is no magic bullet out there, or some, you know, you can't find a new magic bullet of the month almost out there for weight control. There is no single factor but adding together the elements of a healthy diet can make an important difference. So this is an analysis we did and published a couple of years ago in the New England Journal of Medicine, Dariush Mozaffarian led this work. The three different colors stand for three different cohorts. And without going into all of the details, this produced a massive amount of data because we started with about 80,000 people who were not obese and were healthy basically. And we're looking at weight gain over a 20-year period in 4-year segments. And the picture here is, actually if you look at the right, the biggest problem is sugar-sweetened beverages. And this is showing the estimated effect per serving. So this would be per one 12 ounce serving of a sugar-sweetened beverage, the problem is that many people take many servings a day of sugar-sweetened beverages. In low-income American groups, we just analyzed, the average was three servings of sugar-sweetened beverages per day. And many people, about 10%, were taking five or six servings of sugar-sweetened beverages per day. This is a huge metabolic problem and a threat to health. This I'll talk about a little more. Fruit juice also is related to weight gain but dairy products, on the right, were pretty neutral. If we look at foods on the left, potatoes, red meat showed up, refined grains is being related to more weight gain. Again, dairy products were sort of neutral. And then at the left, fruits, vegetables, whole grains, and interestingly, yogurt in all three cohorts was related to the least weight gain. And there's a lot of interest in the microbiome and all of the bacteria in yogurt being beneficial. So it is interesting if you look at the fluids that are related to less weight gain that pretty much describes the Mediterranean type diet. And quite, also, interestingly, one of our colleagues in Israel did what I think is one of the best studies in the diet and weight control field. They randomized three groups and they did this at a work site where they had a more control of diet than just anybody off the street because many of the foods were prepared and eaten at work and they could control that. So the shaded part shows the first two years of the study, they had a low-fat group and a relatively high-fat group in a Mediterranean type diet that also had about the same fat composition, fat content as the higher fat group. And all three groups lost weight over the first two years. The low-fat group lost the least weight, but then the study stopped and everyone went back and worked on other things. But after another four years, the investigators went back and weighed people and collect their blood samples, so six years into the study, four years without any dietary intervention. And the Mediterranean diet group had maintained most of the weight loss and they were better off metabolically. The low-fat group had essentially regained everything and the higher fat group were sort of in between. But definitely, I think people were able to stay with the Mediterranean diet because there was a lot of variety, it was foods that were enjoyable and metabolically had an important benefit. So I think that's sort of the bottom-line of what we know about diet and weight control that what's really important is the long-term, not what you weigh after a couple of months. Yes, that may get you into the swimming suit in the summer, but if you're thinking about health, that's a year-round issue.