In this episode, we discuss:
- Why you need to eat fat
- Why the Paleo diet template makes sense
- Where these misguided ideas about fat came from
- The Seven Countries Study
- Zoё Harcombe’s research on fat
- Why you should be skeptical of some news headlines
- Why dietary guidelines don’t work
- The epidemiological evidence
- Conclusions about saturated fat
- The Obesity Epidemic: What Caused It? How Can We Stop It?, by Zoё Harcombe
- “Evidence from Randomised Controlled Trials Did Not Support the Introduction of Dietary Fat Guidelines in 1977 and 1983: A Systematic Review and Meta-Analysis,” by Zoё Harcombe
- “The Nitrate and Nitrite Myth: Another Reason Not to Fear Bacon,” by Chris Kresser
- USDA Food Composition Databases
- “Re-evaluation of the Traditional Diet-Heart Hypothesis: Analysis of Recovered Data from Minnesota Coronary Experiment (1968–73),” by Christopher Ramsden
- “The Challenge of Reforming Nutritional Epidemiologic Research,” by John Ioannidis
Hey, everybody, Chris Kresser here. For the past 50 years we’ve been told that eating meat, saturated fat, and cholesterol is unhealthy. Recently, a growing number of people are turning to a vegetarian or vegan diet with the goal of improving their health.
But is it really true that meat and fat are bad for us? And are vegetarian and vegan diets a good choice for optimizing health and extending lifespan? If not, what is the optimal human diet? Join me on the Joe Rogan Experience on Thursday, September 27th, as I debate these questions with vegan doctor Joel Kahn. You can tune in live at 1pm PT (***Updated time!***) at JoeRogan.live or, you can catch the recording at podcasts.joerogan.net, on YouTube, or in iTunes or Stitcher.
If you’d like to receive updates about the debate, including links to the recording and new articles and information I’ve prepared on this topic, go to Kresser.co/Rogan. That’s Kresser.co/Rogan and put your email in the box.
Okay, now onto the show.
Welcome to another episode of Revolution Health Radio. This week I’m very excited to welcome Dr. Zoë Harcombe as the guest on the podcast.
Dr. Harcombe is a Cambridge University graduate with a BA and MA in economics and math. Zoë enjoyed a successful career in blue-chip organizations before leaving corporate life in 2008 to pursue her passion. Her early career involved international roles and management consultancy, manufacturing, and marketing in global organizations from FMCG to telecoms before specializing in personnel and organization. At the peak of her career, Zoë was vice president for human resources for Europe, Middle East, and Africa. Having written three books between 2004 and 2007 while being head of people, Zoë left employment to research obesity full time. This culminated in the publication in 2010 of The Obesity Epidemic: What Caused It? How Can We Stop It?
Zoë returned to full-time education in 2012 to complete a PhD in public health nutrition, which was awarded in March 2016. Her PhD thesis was entitled “An Examination of the Randomised Controlled Trial and Epidemiological Evidence for the Introduction of Dietary Fat Recommendations in 1977 and 1983: A Systematic Review and Meta-Analysis.” A number of peer-reviewed articles have emanated from this work, and the first was the 64th most impactful paper in any discipline in the year 2015. Zoë lives with her husband and rescue animals in the Welsh countryside surrounded by food, a.k.a. sheep, hens, and cows.
Now I’m really excited to talk with Zoë because she recently published a paper critiquing the U.S. dietary guidelines and the U.K. dietary guidelines for the lack of evidence behind their recommendation against eating saturated fat or limiting it to less than 10 percent of calories in the diet. And she, as I just suggested with her bio, has probably spent more time looking at this than anybody else. She wrote her PhD thesis, as the title suggests, on the evidence, or lack of evidence, rather, behind the dietary guidelines around total fat and saturated fat. And she went all the way back to the late 70s and early 80s to look at the studies that were used to create the original dietary guidelines. And then the second half of her paper looked at all of the research that has been published since then through 2016.
And as we’ll discuss in the show, the conclusion is that the dietary guidelines never really had any meaningful evidence behind them to justify restricting saturated fat back in the late 70s and the early 80s. And the same is true today in 2018. So I hope you enjoy this interview as much as I did, and let’s dive in.
Chris Kresser: Zoë, it is such a pleasure to have you on the show. We were just chatting before the show, and I can’t believe we haven’t connected by now. We walk in many of the same circles, and I’ve followed your work for some time. So I’m really, really grateful that you’re able to join us.
Zoë Harcombe: Oh, I thank you so much for having me. I’ve just followed you for so long and your “nitrates in bacon” is my just absolute go-to blog. Stop, people, worrying about bacon.
Chris Kresser: All right. Well, we’re going to talk a lot more about that and there’s so many things we could talk about today. But the main reason that I wanted to have you on the show is to talk about your recent paper critiquing the dietary guidelines both in the U.S. and in the U.K. related to total fat and particularly saturated fat. And everyone who’s listening to this knows that for many, many years, really, I guess about 40 years now, right? It goes back to about 40, 41 years now, we’ve been told that fat in general, although that’s maybe slightly changing in the public perception recently and even in some of the dietary guidelines that fat in general is bad, and particularly saturated fat is terrible.
But as we’re going to discuss in the show, your research has shown that that’s maybe not what the evidence actually says. So before we dive into that, why don’t you tell everyone a little bit about your background and how you came to this work.
Zoë Harcombe: Okay, I’ll do a really quick one because I know you’re not sort of a three-hour podcast man. So I’ll give you a composite history. First fascination came when I was studying at Cambridge University in the late 80s, early 90s and started seeing obesity growing around me. And it was just a fascination to me because it’s the last thing that people want to be. People do not want to be overweight, let alone obese, and it was just starting to explode and had already exploded in the U.S. And I was just fascinated.
So I wanted to understand, why do we have an obesity epidemic? When you start looking at why, you go back to looking at when, and you can’t help but see, particularly on the U.S. graph, that it just takes off like an airplane at about 1976 to 1980, that pivot point in the NHANES data. And of course, therefore, you go back to look at that period of time. Did anything particularly happen? Did we suddenly start eating 10,000 calories a day and sitting around on our backsides? Well, actually, no we didn’t.
Did you grow up hearing that saturated fat would give you a heart attack? You’re not alone. Check out this episode of RHR for an in-depth look at the science surrounding saturated fat with researcher and author Zoё Harcombe.
And the UK data was particularly interesting. We seem to be eating fewer calories nowadays than we did back in the 1970s, when we were much slimmer. Barely any obesity in the UK by about 1972. And you then start looking at an event called the dietary guidelines, which came in with the Senator McGovern committee in 1977, and of course these were then embedded in the US Dietary Guidelines for Americans in 1980 and then every five years since. And there is debate.
There are people who will say the introduction of the dietary guidelines has nothing to do with the rise in obesity and type 2 diabetes and more beta conditions related to diet and lifestyle. But it at least needs to be looked at. I mean, I say it coincided with the epidemics in obesity and diabetes, and we need to therefore explore was that a coincidence or was it a factor that was material in the changes in our health. And I am one of the people who thinks it is material and that our shift to basing our meals on starchy foods, grains, fruits, vegetables, largely carbohydrates, being encouraged to consume as high as 60 percent of our diet in the form of carbohydrate, the one macronutrient we don’t actually need, I do think it has made a difference. And I’m not alone in that view.
And then of course you look at guidelines and say, well, what were they about? And of course, they were about restricting total fat to no more than 30 percent of calories and saturated fat to no more than 10 percent. And because they were only three things that we ate and because protein is in everything other than sucrose and oil, so it tends to be fairly constant in any natural diet, and the peer study showed this beautifully, nice evidence for this, protein tends to stay constant around 15 percent. So as soon as you set the fat guideline, you’ve automatically set a carbohydrate intake minimum of 55 percent. And that’s what we did.
So I wanted to understand why did we set that total fat guideline. If that was the thing that started everything, why did we do that? And did we get it right or did we get it wrong?
Chris Kresser: So you have a BA and an MA from Cambridge in economics and math, and then in 2016 you got a PhD in public health nutrition. And what I find really fascinating is what you … tell us a little bit about your PhD thesis.
Zoë Harcombe: Yeah, so the PhD thesis was using the relatively modern techniques, and they have been around since the 1970s, but we’re really using them a lot now. And that’s the systematic review and meta-analysis. And when we pulled together evidence from randomized controlled trials, ideally, if not from cohort studies, it’s considered to be the pinnacle of the evidence that we could examine. So I approached looking at the dietary fat guidelines in four ways. And one was to say, what was the RCT evidence at the time available to the committee? Had they looked at it back in …
Chris Kresser: That’s “randomized controlled trials,” for those who are not familiar.
Zoë Harcombe: Yes. Yeah, so the randomized controlled trial evidence available to the US committee in 1977, and then the UK committee deliberated in 1983, and that allowed one more study available to the UK committee that wasn’t available to the US committee, and that was the Woodhill Sydney Diet Heart Study. And also to look at the epidemiological evidence available, had the committee chosen to look it at the time the guidelines were set? So that was the first two papers, the first half of the PhD. And then the second half was to bring it up to date and to say if the committees were deliberating again today and they had all the RCT, randomized controlled trial, evidence available and all the epidemiological evidence available we have today, what would the conclusions be, looking at it in an up-to-date scenario. And that was the four parts.
Chris Kresser: Yeah, so you’ve spent, how many years did it take for you to get your PhD?
Zoë Harcombe: Yeah, three and a half. I did it full time. I just decided to stop everything else I was doing and do it full time. And as anyone who’s ever done one meta-analysis knows, to try and do four …
Chris Kresser: Yeah, that’s a lot of work. But the upshot here is that you have a PhD in public health nutrition. You spent the better part of four years examining the evidence base for the last 40 years connecting fat and saturated fat to health and disease. And this is exactly why I wanted to have you on the show to talk about this topic because it’s one thing if you have a kind of armchair critic who’s cherry picking one or two studies to make their point, which often happens on the internet, right?
Zoë Harcombe: Yes.
Chris Kresser: It’s another thing to have someone who’s trained at the level that you’ve been trained at who spent four years objectively looking at this evidence and then publishing on it and showing where it doesn’t add up. So let’s dive into that now.
I want to start by talking about some basics because I think they’re really important. I love how you did this in your recent dietary guidelines paper. Just a few facts about fat that maybe not everyone is aware of or has thought about much.
Zoë Harcombe: Okay, so I shared these in conference presentations and I was really pleased that when I did this, there’s peer-reviewed study. They didn’t get edited out because I thought they might be a bit chatty, if you know what I mean, for a peer-reviewed paper.
Chris Kresser: Yeah, yeah.
Zoë Harcombe: But I actually had a couple of nice comments in the margin of people saying, “Oh, good point, I hadn’t thought about that.”
Chris Kresser: Yeah, yeah. I had the same reaction.
Why You Need to Eat Fat
Zoë Harcombe: Oh, thank you. So the most important one is that we must consume fat. Human beings must consume fat. We die without consuming dietary fat. We must consume essential fatty acids, that’s why they’re called essential.
“Essential” in nutrition means something that we must consume, not just something that the body needs. And of course we have the fat-soluble vitamins A, D, E, and K, and they come in foods with fat. And they need fat to be absorbed. So that’s pretty vital. When people demonize fat to the extent that they do, they always give the impression that we could get away without eating this stuff. And we couldn’t. I don’t know how quickly we’d die, but we would. So we need it.
Number two, again, that people seem to realize particularly when they demonize saturated fat is that every single food that contains fat, and it’s actually quite difficult to find a food that doesn’t contain fat, sucrose doesn’t. But not much else doesn’t. So every food that contains fat contains all three fats. That’s:
- Saturated fat
- Monounsaturated fat
- Polyunsaturated fat
And only the proportions vary. So again, people talk as if we can avoid saturated fat and only eat unsaturated fat, and that is completely impossible unless you’re in a lab and you’re trying to create single fats. It is impossible if you’re going to eat food, which I recommend that all people do.
And then the other interesting factoid, I love playing around on the USDA all foods database and just looking at things that add up and things that don’t add up. And it was a real surprise to me when I first started looking at foods that when it comes to food groups, there is only one food group that has more saturated than unsaturated fat, and that is dairy products. So your struggle to find a meat, and I have not yet found one, that has more saturated than unsaturated fat, typically the main fat in meat is monounsaturated fat. And that goes for lamb or steak or chicken.
Chris Kresser: Even pork.
Zoë Harcombe: Absolutely. And therefore lard, which I just love, because people just think lard is pure disgusting.
Chris Kresser: Saturated fat, yeah.
Zoë Harcombe: Saturated fat.
Chris Kresser: If you were to put lard or coconut oil together, people would say lard would be the unhealthy choice from a saturated fat perspective. But of course, coconut oil is 97 percent saturated, I think.
Zoë Harcombe: Yes.
Chris Kresser: Something like that.
Zoë Harcombe: And lard is 39 percent.
Chris Kresser: Right.
Zoë Harcombe: So nowhere near as bad. And that’s not saying that saturated fat is worse than unsaturated fat. It’s just stating a nutritional fact. The only food group that has more is dairy products. And then of course you’re getting to, “Well, are dairy products bad for us?” And it’s really difficult to think that they are when you look at the nutritional profile of dairy products and the bone nutrients calcium, phosphorus, vitamin D. Look at any profile of any dairy product and you can’t help but think …
Chris Kresser: And the evidence.
Zoë Harcombe: Yes.
Chris Kresser: I mean there’s a ton of evidence suggesting that full-fat, but not nonfat or low-fat dairy are beneficial for both cardiovascular and metabolic health. And there was actually a new study recently published, I’m not sure if you saw it. I am less persuaded by it. Or you mentioned it, the PURE Study.
Zoë Harcombe: Oh, yes.
Chris Kresser: Because it has within-country or between-country comparisons, which I think we’ll be talking about later, is problematic. That was an issue with the Seven Countries Study. But it does align with many of the other studies that have been done on this topic previously, showing that when people eat more full-fat dairy, that’s associated with lower body weight, with lower blood sugar and better glycemic control, and with lower risk of heart disease based on cardiovascular markers. And that’s actually the reason that the full-fat dairy works better than the low-fat or the nonfat dairy in that regard because some of the nutrients that are thought to be beneficial for cardiovascular and metabolic health are in the fat.
Zoë Harcombe: Yes, amazing.
Chris Kresser: So if you take out the fat, you take out the benefit.
Zoë Harcombe: Yeah, yeah.
Chris Kresser: Yeah.
Zoë Harcombe: Absolutely, I’m a huge fan of dairy, personally.
Chris Kresser: Me too. I mean, of course if someone is lactose intolerant or they’re intolerant of the proteins, it needs to be avoided. But for people who are not, what I always say is at least the evidence that we have suggests that it’s healthy when it’s well tolerated by the individual.
Zoë Harcombe: Yes.
Why the Paleo Diet Template Makes Sense
Chris Kresser: So given this, given that fat is essential, that all foods contain all fats and that saturated fat is not even the highest percentage of fat in any food except for dairy, this leads us to some pretty interesting conclusions. You mentioned in your paper, which I loved, and I loved that they kept it in here too, it’s illogical that the same natural food would be both helpful and harmful. Like you can’t eat a steak and eat it so that you’re only eating the unsaturated fats and not the saturated fats.
Zoë Harcombe: Yeah, yeah. There’s no other way of putting that. It just, it doesn’t make sense. Whether your belief system is in God or nature, food is provided around us on this planet, and it makes no sense that in that same food that we need to thrive and survive, something has been put that is trying to kill us at the same time as all the things being there that are trying to save us and enable us to live. And we’ve evolved of course over—I’m reading Sapiens at the moment, so there’s an argument over our heritage—but, I mean, we’ve certainly been around potentially since Australopithecus, Lucy, two-and-a-half, maybe three-and-a-half million years ago. And we’ve done pretty well eating anything we can forage or hunt around us.
Chris Kresser: That’s right.
Zoë Harcombe: The idea that they came up with in the last 40 years that this stuff is trying to kill us, it’s just so stupid.
Chris Kresser: It doesn’t add up at all.
Zoë Harcombe: Yeah.
Chris Kresser: It doesn’t add up. So another example you used which I love because it really turns nutrition-dominant paradigm ideas on their head is the olive oil versus pork chop example. Tell us about that.
Zoë Harcombe: Yeah. So I have a little postcard that I leave on the chairs at conferences, as well, so there’ll be many around the world. And I put up a picture of a sirloin steak, mackerel, olive oil, and mention the pork chop. And a couple of interesting, fun factoids. One is that the mackerel has twice the total fat and one-and-a-half times the saturated fat as the sirloin steak, which isn’t a problem because both of them are great foods. But of course we’re told not to have red meat and we’re told not to have oily fish. So that’s illogical.
And then olive oil has 14 percent saturated fat versus a typical pork chop might be only sort of one to two grams, but then people say, “Oh, you wouldn’t consume 100 grams of olive oil.” No, but a tablespoon of olive oil has more saturated fat than a 100-gram pork chop. And again, we can make a mockery of nutritional advice when you know something about food.
Chris Kresser: Yeah, I mean, and it’s easy to see how you could have a salad, if you have a big salad with a couple tablespoons of olive oil and dressing versus a 200-gram pork chop, you’re still eating more saturated fat there.
Where These Misguided Ideas about Saturated Fat Came From
So, I mean, this is so obvious when you look at it this way, and it makes you wonder, how did we get the idea that saturated fat is bad in the first place? I know there are some political and social roots of this, and feel free to talk a little bit about that if you want. But in particular, how did this arise from the evidence? What was your sense of that as you did your PhD and looked deeply at all of this?
Zoë Harcombe: Yeah, so one of the most important chapters in the PhD is the review of the literature. And you go back in the review of the literature, and of course in this topic area, you’ve got to go back to the Russian pathologists in the early part of the 19th century, when they noticed the cholesterol deposits in the arteries of the autopsies that they were doing. So they started to hypothesize, had these cholesterol deposits actually caused the death of this relatively young person that they were performing an autopsy on, and could they come to any conclusions about those sort of cholesterol stores of fatty deposits?
And many people know this, it’s been said in conference presentations, that at the time they then started experiments on rabbits, feeding them foods containing cholesterol, feeding them purified cholesterol, to try to see if they could mimic the impact that they thought food might be having on the human body. And of course, as some people have worked out, rabbits are herbivores and the only foods that contained artery cholesterol are foods of animal origins. No exception. So you find dietary cholesterol only in meat, fish, eggs, and dairy, which are things that rabbits can’t tolerate.
Chris Kresser: Yeah, it’s strange. In a certain way, this almost supports what we were just saying. Eat a species-appropriate diet. The message there is not “don’t eat cholesterol.” It’s “don’t feed cholesterol to an animal that’s not supposed to eat it and don’t feed humans foods that we’re not supposed to eat.”
Zoë Harcombe: Absolutely. And very interesting. When they fed purified cholesterol not in animal foods to the rabbits, they didn’t have any problems. And when they fed cholesterol foods to dogs, they didn’t have any problems because dogs are omnivores.
The Seven Countries Study
So we then wind forward to the 1950s, and Ancel Keys gets a bit of a bad rap in our world. I like to look on him in quite a more balanced way because he did some brilliant work, like the research starvation experiment. But he did kind of fall by the wayside a little bit on the fat thing.
So his first exploration was with the Russian experiments in mind to try to see if dietary cholesterol impacted blood cholesterol. And he concluded it did not, and he never deviated from that view. And the best quotation I found on that was from the 1954 symposium on atherosclerosis, and he said, “Cholesterol in food has no impact on either cholesterol in the blood or the development of atherosclerosis in man.” Which was brilliant because he had actually exonerated animal foods. But he didn’t make that connection at the time. Maybe his nutritional knowledge just wasn’t good enough and he just hadn’t quite worked out, “If I’m finding nothing when I’m feeding human subjects,” because you could do that then with ethics, “human subjects massive amounts of dietary cholesterol via loads of animal products, they don’t develop any blood cholesterol problems and they don’t develop any signs of atherosclerosis,” he should’ve concluded, “I therefore just exonerated what I’ve been feeding them.” Which would be:
- Possibly fish
But most likely meat, cheese, and eggs. But he didn’t. For some reason he was convinced that fat was the bad guy. If it wasn’t cholesterol in food, then it had to be fat in food. And yet again, having given his human subjects animal foods, he should’ve said, “What are the macronutrients in those animal foods? Okay, so it’s fat and protein. Dairy products have got a little, little bit of carbohydrate, but essentially what I’ve just fed them is fat and protein. So I should turn my attention to the one thing I haven’t fed them, which is carbohydrate.” But he didn’t do that.
So he was convinced that total fat was the problem, and of course we then had the Mount Sinai presentation in 1953, which gave us that famous Six Countries Graph, which has nothing to do with the Seven Countries Study. And then of course there were a number of countries that he’d left out. And Yerushalmy and Hilleboe found this out and unfortunately published a little bit too late, in 1957, saying, “Hey, hang on, you left out all of these other countries. And if you put them all on there it looks a bit like a spider scatter, that the pattern has gone. But the Seven Countries Study had already started in 1956. And Keys seemed pretty determined that he was going to come to the end of the Seven Countries Study and find fat guilty.
Now interestingly, and this is not terribly widely known, he could not find anything against total fat. So when, as part of my PhD, I pulled the epidemiological studies that were available at the time the guidelines were introduced, and of course the Seven Countries Study was one of those, and you’ve got Framingham and Honolulu, Puerto Rico, the London bank and bus study, and the Western Electric study being the others, none of those six found any relationship between coronary heart disease and total fat. So Keys acted. He went in with the total fat hypothesis. He accepted that it was not total fat. Now he had spent so much time and money on this study, he needed to find something. And he could find an association between saturated fat in the different cohorts, and coronary heart disease in the different cohorts. But at the same time he claimed, and this is in the summary paper, “I found no issue with weight, obesity, I found no issue with sedentary behavior activity, I found no issue with smoking.”
So things that we now know he was wrong about, we give him the benefit of the doubt on the one thing that he did find, which was saturated fat. And the other five peer studies, the ones I’ve just mentioned, did not find anything against saturated fat. And of course, they were all in country studies. So they were right, they were in community studies.
So you take Framingham. It’s a small town, it’s looking at people who eat a certain level of total fat or saturated fat versus people who don’t. So you’ve got all the other factors, or many of them, constant. You’ve got the same GDP, the same politics, the same community, the same access to healthcare. Go to Japan in the 1950s versus the US in the 1950s, you’re comparing efficiently.
Chris Kresser: Completely different.
Zoë Harcombe: Exactly.
Chris Kresser: Not even apples and oranges. We’re not even in fruit category there. I just want to pause here and just highlight this for people who are less familiar with research and methodologies. What Zoë’s saying is that if you … the problem with comparing groups of people between countries is that there’s so many factors that vary from country to country and lifestyle, physical activity, the type of foods they eat. Saturated fat comes in lots of different types of food. So what kinds of foods are people eating in the US versus in Japan, where saturated fat would be found in totally different type of food? So comparing between countries just makes the possibility of confounding factors and all of the other issues of epidemiological research, it just amplifies them and makes them even more likely. So typically, especially today, those between-country studies are often discounted or taken with a large grain of salt because it’s so hard to control for factors even within the population, much less between different populations.
Zoë Harcombe: Absolutely. Absolutely yes.
Chris Kresser: Okay, so, I mean, this is … the crazy thing to me about this, Zoë, and I’m sure this struck you at many intervals throughout your PhD, is just how much of a house of cards the whole evidence base is behind the idea that saturated fat is bad for us. There’s this illusory truth fallacy that we were chatting about before we hit the record button, as well, which is the idea that if you hear something repeated enough times, you just start to believe that it’s true, whether it has any basis in fact or not. And we think maybe that researchers and scientists are immune to this illusion. But the fact is, they’re not.
John Ioannides, one of the most famous epidemiologists in the world, one of my favorite quotes of his is, “Claimed research findings may often be simply accurate measures of the prevailing bias.” So, I mean, that sums it. He has all these pithy quotes that just sum it up in, like, 10 words. Which basically means that once you have a certain idea and it’s out there because of groupthink and confirmation bias, that idea will often just be perpetuated, even if it was never based in fact in the first place. Because someone will link to that original study that turned out to be erroneous as proof, then someone does a later study and you link to that second study. And then it just becomes a chain of references that all point back to that original study that then it was later shown to be invalid. So it’s crazy to me that 40 years of dietary policy has been based on such flimsy evidence.
Zoë Harcombe: I should declare my own bias, actually, going in, because up until 2010 I’d been a vegetarian for about 20 years. Then my own bias going in was that fat was bad, saturated fat was bad, saturated fat equaled animal fat, which of course I now know absolutely that it doesn’t. All fats are in all foods, especially coconut oil, which is purely vegan. And I believed what I’d been taught at school, that we should be eating low-fat foods and healthy whole grains and plenty of fruit and vegetables. And I believed it too. And I was at a dinner party just a couple of weeks ago, and there were a couple of young people who were engaged and full of life and full of news and full of opinions.
And as we sat down to dinner, they were reliably informing me and my husband that they didn’t eat much meat because it was full of saturated fat, which of course it isn’t, and saturated fat is bad for you, which of course it isn’t. And I said, “You guys work in the finance industry. How did you pick up, how did you become authorities on dietary fat at your tender young years?” We had done a superb marketing job on fat and cholesterol worldwide and people have fallen for it.
Chris Kresser: Absolutely. Yeah, yeah, and it’s deep. It’s really a form of conditioning. At least, I’m not sure what’s happening with with kids now, but I grew up certainly at a time where butter and eggs and all those foods were really demonized. And it becomes kind of part of your cultural conditioning, and it’s so deeply hardwired in the brain, it can be really hard to let go of it. I, as my listeners know, I was a macrobiotic vegan at one point. So I took it about as far as you could go.
Zoë Harcombe: Wow.
Chris Kresser: And I remember in high school, I was an athlete and the whole carb loading paradigm. I was eating, like, bagels with nothing on them, like dry bagels and breakfast cereal with nonfat milk for breakfast, and eating pasta and pancakes before my basketball games because the thinking was that would be good for athletic performance and also good for my health.
So I can be pretty extreme when I go for something. I took it to the extreme and when I started to figure out, I mean, it took a very serious chronic illness for me to snap out of that. And even with that, I remember when I was first starting to eat more fat, I had this distinct feeling, like I was doing something wrong or I should do it behind closed doors, or that something bad was going to happen to me. And it took quite a while for that to unwind. So I think there’s that kind of deeper psychological influence happening here too.
Zoë Harcombe: I’ve read your stuff on that. You write so, like it just happened yesterday. I mean you just describe it, and you just did it then. I could see you running around the track with your bagel. It sort of stays in your mind, doesn’t it, how we felt and what we thought we were doing when we did all of that stuff.
Zoё Harcombe’s Research on Fat
Chris Kresser: Absolutely. So let’s talk a little bit more. Let’s kind of dive in with a little more of a fine-toothed comb on your thesis and your review of the RCTs from 1977, the randomized controlled trials. Which again, if we’re looking at a hierarchy of evidence, it’s not that RCTs are perfect or they don’t have potential issues, but certainly when compared to epidemiological issues and all of the problems there, which we’ll discuss a little later, they are more reliable. So what did you find in your review of RCTs related to saturated fat and either death from all causes and death from heart disease?
Zoë Harcombe: So this paper came out in February 2015, and it went nuts. And if you Google it, it was front page in New Zealand and in the UK papers. And I spent the whole day when it came out. The phone was ringing the second I put it on in the morning and it was the BBC, could I come in? And I ended up doing about 20 or 30 interviews that day, just back to back. And it just went nuts. So I think it went nuts because it was the unique part of the PhD that was looking at the evidence at the time. And so people were picking up on the idea that we’ve been eating low fat for 40 years and the evidence wasn’t there at the time to back up the call to do that.
So the major findings from that paper were first of all that there were only six studies, six randomized controlled trials that were available to the UK committee. Only five were available to the US committee, and they’ll be pretty well-known to people. It’s like the Rose corn oil, olive oil trial, the low-fat diet; the Leren Oslo Diet-Heart Study; the MRC soybean study; the Sydney Diet Heart Study; and the LA Vet study. And you pull all of those together, there is no difference whatsoever. Not even to just leak the significance, it was actually the exact same number of deaths in the controlled side as in the intervention side. There was no significant difference in coronary heart disease mortality, it wasn’t quite an identical number, but it was something like 221 versus 219, or something. It was so close. It was virtually identical, again.
A really interesting finding, and this just massively undermined the diet–heart hypothesis and was not a finding that we expected to come across. It just came out. We were able to measure the … Across the polled studies, there was a significant difference in cholesterol being lowered in the intervention studies. But of course that made no difference whatsoever between mortality or coronary heart disease mortality. And I then went on to try to understand why it may have been the case that cholesterol had been lowered by the intervention and not made any difference to health benefits apart from the fact that cholesterol is not bad for us. But why didn’t intervention diets lower cholesterol?
And I think it’s because the main intervention was to swap out saturated fat and to swap in polyunsaturated fat. And a lot of the polyunsaturated fats that they were putting in, corn oil, soybean oil, vegetable oils, contain plant sterols. And plant sterol is effectively plant cholesterol, and it competes in the human gut with the human cholesterol and it replaces it, to an extent. So if you take plant sterols in margarines or spreads or in vegetable oils or indeed in some grain plant products, or some people take them from tablets from the health food shop, which is a really crazy … they will replace your own cholesterol to an extent and lower your blood cholesterol. But I’ve looked at the evidence for the end outcomes on heart disease. I’ve got another paper on that that was published in an editorial, and that shows that actually the overall benefit is not there. It’s actually overall harm of administering plant sterols in the end outcomes of heart disease.
But I think that’s why they lowered cholesterol and perhaps the studies weren’t long enough for the harm from that replacement to actually manifest itself in a difference in outcomes. And I would then expect the interventions to have more deaths from heart disease and more deaths therefore from all-cause mortality.
One of the other really big aspects I think that grabbed the media is the point that we made at the end of the paper, saying that these six studies, when he pulled them together, amounted for fewer than two-and-a-half thousand men, not one single woman had been studied, and not one of those men was healthy. They had all had a heart attack already.
Why You Should Be Skeptical of Some News Headlines
Chris Kresser: So this is just really key here. You cannot generalize, even if the results were consistent across all these studies, which they weren’t, implicating … Or it sounds like they were consistent in the opposite direction that people thought they were. But even if they had implicated saturated fat as increasing total and CHD mortality, coronary heart disease, that would only be applicable to men.
Zoë Harcombe: Sick.
Chris Kresser: With pre-existing, yeah, sick men.
Zoë Harcombe: Yeah.
Chris Kresser: Not women and not men that are not sick.
Zoë Harcombe: Yeah, absolutely, yeah. And interestingly not one study called for change, and studies at the time were far more ethical, I think, than they are today. Far less media orientated, far less trying to get a press release. They would just say things how they were. And a couple of them were a bit nervous about potential toxicity of the fish oil that we’d administered, and they’re the ones that were a bit worried about … the corn oil study had more deaths in the intervention group, and said, “We’re worried about the potential harm from the fish oil intervention.” And the low-fat diet study, the last sentence of that study just cracks me up. It just says, “A low-fat diet has no place in the treatment of myocardial infarction,” which is heart attack.
Chris Kresser: It’s interesting to me what you just said that how much the, both the reporting on studies has changed in the media and also even the way that researchers talk about their findings themselves to the media. I think I was reading an article in Science that was published in 1993, and they were talking about relative risks, which we can get into more detail when we talk about epidemiological evidence. But this is the percentage increase in risk from a given intervention, and they were outside of nutrition, still today in any other field, epidemiologists would consider anything below a 200 to 400 percent increase in risk to be indistinguishable from noise, meaning they would consider anything less than a 200 percent increase in risk to be not significant statistically. And in this article, Marcia Angell, who is a former editor of the New England Journal of Medicine, was quoted as saying that, “They typically didn’t really accept a paper unless it had a relative risk ratio of over three for nutrition.” And that just blew me away because today, like IARC’s panel about red meat and processed meat causing increased risk of cancer, the percentage increase is 18 percent.
Zoë Harcombe: Yeah.
Chris Kresser: That’s not even remotely close to the 200 percent which is the lower end of the threshold. And yet the media headlines are not saying indistinguishable increase in risk observed in people eating more red meat. They come out and just claim causality. They say, “Red meat and eating red meat and processed meat is going to kill you.”
Zoë Harcombe: And as Bradford Hill would say, “There’s nine criteria and that double is just one of them.”
Chris Kresser: Yeah.
Zoë Harcombe: “So hit the double and then you can look at the other eight.” But none of them hit the double, none today get anywhere close.
Chris Kresser: Nowhere near and yet, and I think this is partly an artifact of the world we live in, just with, like, proliferation of the internet and so many headlines. Everyone’s vying to get attention and so you have to … a headline that said, almost insignificant increase in risk observed in people who eat more red meat than other people. But of course there are other diet and lifestyle factors that we’re not considering. That’s not to make a good headline, right? Nobody’s going to click on that. And so people want the flashy, clickbait headline saying low-carb diet will shorten your lifespan or eating red meat will give you a heart attack. Even though I would hope that the researchers themselves somewhere deep down know that that’s a gross exaggeration of their findings.
Zoë Harcombe: Yeah.
Chris Kresser: And as for the media, I guess it’s just that we don’t have science journalists anymore.
Zoë Harcombe: We don’t. I do a note every Monday where I look at a paper from the previous week and dissect it. And you can tell, mostly the ones that get into the media have had a press release. And if you look at the press release and you look at the media article, the media have just taken the press release almost verbatim. The press release provides a couple of quotes, they end up in every single article. Completely lazy journalism. Occasionally they might call in the UK me or Dr. Aseem Malhotra or Dr. Malcolm Kendrick and just say, “Do you want to give an opposing quote?” and occasionally they’ll stick it in.
Mostly they’ll just run off the press release, and the researchers should be challenging the press release. I mean, our paper in February 2015 was press released, and I remember having a few toings and froings because I wanted it to be scientific. It’s a big enough claim in itself to say we only studied two-and-a-half thousand sick men, and then we introduced these guidelines for 250 million Americans and 50 million Brits. That’s okay, enough. We don’t need to sensationalize it anymore than that.
Chris Kresser: Absolutely.
Zoë Harcombe: So I tried to get it down to the facts that we found and not to put any spin on them.
Chris Kresser: Yeah. Yeah, to your credit, I mean, that’s so hard to do in this crazy media environment that we live in now. And to be fair, there are definitely researchers that make an effort to do that. And you still will see that in reports where, I was reading one on, I can’t remember what it was, but it actually stood out to me because I don’t see it as often as I do. I was impressed by both at what the researchers were saying and that the author of the article. Because they went out of their way to say this is just an association or correlation. It doesn’t prove causality, and here are the reasons why it might not be a causal relationship, and why we need more research. But my sense of that is it’s almost like when you watch, if you see a commercial for a drug and then you have like the 20 seconds of side effects after the 10 seconds of the commercial. People have heard that so many times they just kind of tune that out and they’re only really still paying attention to the headline.
Zoë Harcombe: Yeah.
Chris Kresser: So let’s talk. So you went back, you reviewed the RCTs from the late 70s that were responsible for creating dietary guidelines that, as you said, applied to hundreds of millions of people around the world and probably affected many more even just by osmosis. Those ideas becoming firmly entrenched in industrialized society, even if they weren’t part of formal dietary guidelines. And then you went back and analyzed all of the research that had been done from, was it from the late 70s to 2016?
Zoë Harcombe: Yes, so we then took it up to date. And I actually said in the recent paper that’s just been published by the BJSM, the one on is saturated fat a nutrient of concern, and that’s because the USDA is now looking at it again for the next dietary guidelines. And I actually put in that paper that the day that the paper is saying there was no evidence at the time came out, I was astonished that Public Health England came out almost immediately on the day. I said, “Okay, so maybe there wasn’t evidence at the time, but we’ve got plenty of evidence today.”
Chris Kresser: Right.
Zoë Harcombe: I was surprised that they were prepared to concede. I thought they said, “No, no, no, this is ridiculous. The Seven Countries Study was marvelous and that’s all we need. And we can ignore everything else.” But they didn’t. They said, “Okay, there was no evidence, but there is plenty nowadays.” And of course it takes so long to get papers published that with my supervisory team, we’d already moved on to the next step, which was looking at the evidence available today. So we had that paper pretty much ready to go. And of course you keep in the original six studies, then you just add in any other randomized controlled trials that have looked at coronary heart disease, mortality, and total mortality. Those were our two outcome criteria so we wouldn’t lose some RCTs that only looked at events, for example. But then that then brought in the Women’s Health Initiative, the DART study, the STARS study, and the very well-known Minnesota Coronary Survey study.
Chris Kresser: Zoë, before we go on I want to pause there. Let’s talk about why you chose total and coronary mortality as an endpoint and why that’s important—to focus on the mortality endpoints versus just the events.
Zoë Harcombe: Yeah, I’ve got to credit Dr. Malcolm Kendrick with this, and I am such a Dr. Malcolm Kendrick fan, it’s just not true.
Chris Kresser: Yeah, me too.
Zoë Harcombe: And I just remember, I mean, I’m fortunate enough to know him and consider him a friend and to meet him on occasions. And just every time I meet with him he says stuff and I’m just, why is this not just the only thing that’s being taught in medical school because it’s so sensible? So he’ll say, “I can guarantee that you won’t die from heart disease by pushing you off a cliff.” And it just, it then sticks in your head. Okay, so the important thing is total mortality because there is no point to reducing heart disease if you increase deaths of something else. So all this stuff going on with statins. Oh, we think we can reduce some events. We could have a whole different program on statins. But would there be any point in making any benefit anyway, even if they could, if they, for example, as they might do, increase your risk of cancer or dementia or mind health, etc., etc.? So it has to be total mortality. The only thing that matters is are you going to help people to live longer, to die later?
Chris Kresser: Absolutely.
Zoë Harcombe: That’s what we’re trying to do with health interventions. And so we’ve got to have all-cause mortality in there and then we’ve got to have heart disease mortality and not just events. Because that’s where the dietary fat guidelines came about. They were issued in the name of trying to stop deaths, particularly in men at the time, younger men at the time, from coronary heart disease. So if they’re not going to achieve that, then they’re not even going to achieve what they were introduced for. So why on earth were they introduced?
Chris Kresser: Yeah, thanks for clarifying, and sorry to interrupt. But I’m banging on this drum all the time. I just want to make sure that people understand it because it’s a crucial distinction. You frequently see headlines like “XYZ intervention reduces the risk of heart attack by 20 percent,” which again, as we just said, in an epidemiological study, that’s meaningless. We can’t distinguish that from chance anyways.
But even if it’s an RCT, then the first thing I’d do is go look at the table to see if they even measured total mortality. Which previously, that was less common. It’s more common now, I’m finding. But then when you look at total mortality, there is often no difference. So that’s where the disease substitution is happening that you were just talking about. The risk of death.
Zoë Harcombe: A bit of gossip. Malcolm Kendrick wants to die from a heart attack.
Chris Kresser: Rather than cancer?
Zoë Harcombe: Exactly. Rather than cancer.
Chris Kresser: Yeah, that’s what I tell people too.
Zoë Harcombe: Yeah, he doesn’t want to go early, don’t get me wrong. He probably wants to go at sort of 98 drinking a glass of red wine, playing with his grandchildren when he gets them.
Chris Kresser: Yeah. You just have a heart attack in your sleep overnight. You don’t wake up one morning. That sounds a lot better to me than dementia or Alzheimer’s or cancer.
Zoë Harcombe: Yeah.
Chris Kresser: So you don’t need to belabor this, but it’s really important to point out because I think it’s something that people who are less familiar with research may not have thought of. So okay, so you chose total and CHT mortality, and I believe you ended up with 10 RCTs?
Zoë Harcombe: Yeah. So the original six and then the Women’s Health Initiative, DART, STARS, and the Minnesota Coronary Survey, pull them all together, there’s no difference in all-cause mortality. There’s no difference in coronary heart disease mortality. Again, there was a significant reduction in cholesterol in the interventions that did not meet any difference in coronary heart disease mortality or all-cause mortality.
So essentially, all we did by adding in the former recent studies was that we increased the number of people studied quite dramatically. It came up into the tens of thousands, not least because the women’s health initiative alone brings along tens of thousands of people to the party. And of course it then became more female than male because of all the women in the Women’s Health Initiative. But we still in those 10 only ended up with one study including both men and women that would be a primary prevention study, so people who had not already had a heart attack, and that was of course the Minnesota Coronary Survey. And this in itself found no significant results at the time of publication and of course we then had that brilliant paper where … it should be on the tip of my tongue, the person who went back to look at this, Christopher, I’m thinking. You know the person I mean, who went back to look at the Minnesota Coronary Survey and also went back to look at the Sydney.
Chris Kresser: Was it Hibbeln?
Zoë Harcombe: No.
Chris Kresser: No that’s Joseph Hibbeln and Christopher, they’re both the guys who have done a lot of the critique of the polyunsaturated fat research, or am I thinking of someone different?
Zoë Harcombe: Oh, I’ll be kicking myself and don’t worry. Stick it in your show notes. But it’s a very well-known team that went back to look at both of those studies and even thought there was no evidence found against the dietary intervention at the time, they found that there was some unpublished data. And it just made it even more robust that we had been demonizing fat at the time. So all the RCTs as of 2016, and there haven’t been any since, and there’s still no more evidence than we had at the time the guidelines were introduced.
Chris Kresser: Wow, it’s just, it’s really kind of remarkable, actually. And it’s again just going back to this idea that a lot of this evidence is really based on a house of cards. And as an example of the fallibility of these guidelines, the US in 2015 for the first time removed their advisory that we should not be eating dietary cholesterol. Because they finally acknowledged the cholesterol in the diet does not have any relationship with heart disease. And that was kind of like a pretty major thing that just, like, slipped through.
There weren’t really big announcements or any fanfare around that. Like, “Hey, everybody, we’ve been really wrong about this for the last 30 or 40 years and we just want to bring that to your attention.” And I even remember reading editorials written by scientists who were kind of still anti-saturated fat and cholesterol, and were saying things like, “We can’t really make too much of this because the public is going to lose faith in our ability to guide them with diet.” And I have a sense that the same thing is going to happen with saturated fat in the next few years. And maybe already people know this, but they’re just not willing to do it yet because if they do, people will absolutely lose faith in the diet guidelines.
Zoë Harcombe: Yeah, and they need to.
Chris Kresser: Yeah.
Zoë Harcombe: I mean, let’s face it, they need to lose faith. The best thing that they could do, the dietary guideline committees, would be to come out and say, “Guys, we were wrong. I’m sorry, we were wrong. And we’re going to get it right from now on.” And that would be the only way that we would start having trust in them again. But all this surreptitious slipping things out, then slipping things in, anyone who’s working in this field is just finding this completely unprofessional and noncredible.
Why Dietary Guidelines Don’t Work
Chris Kresser: I mean, this is a whole other discussion, but it’s worth pointing out that the idea of top-down, one-size-fits-all dietary guidelines that will apply to everyone is really not consistent with our understanding, our modern understanding of human biology, biochemistry, and physiology. And I think that this, the whole idea of dietary guidelines that would apply to everyone needs to just die. Because that has led to this reductionist approach, which one researcher calls nutritionism, I like that idea, which is that a nutrient is a nutrient is a nutrient no matter what it’s found in.
Saturated fat in candy or pizza or junk food will have the same impact as saturated fat found in a steak or another whole food. And it’s led to this extreme focus on macronutrients and isolated food components rather than looking at the whole context of the diet. And that’s starting to change slowly. There have been some pretty good studies in the last couple years. There was one, I’m sure you know which one I mean. It was looking mostly at weight loss and they compared, they designed a study that was comparing the effects of a healthier low-carb versus a healthier low-fat diet. And they found that both were actually pretty effective compared to the standard junk food diet that most people eat.
And we need more studies like that, and if we let go of this kind of one-size-fits-all approach, we might actually be able to start looking at the context of foods we’re eating, and then where maybe one person does need to eat more fat and fewer carbs and another person might do better eating a little bit less fat and more carbs from whole foods relative to that other person. So to me that’s one of the biggest assumptions behind the dietary guidelines that’s not mentioned.
Zoë Harcombe: Yeah, that was the Gardner study, wasn’t it?
Chris Kresser: Yeah.
Zoë Harcombe: I corresponded that enough.
Chris Kresser: Yes, the Gardner study.
Zoë Harcombe: Yeah, very good study, yeah.
Like what you’re reading? Get my free newsletter, recipes, eBooks, product recommendations, and more!
The Epidemiological Evidence
Chris Kresser: So let’s just briefly touch on epidemiological evidence. I mean, there’s so many issues with observational nutrition studies. I don’t know if you saw John Ioannidis’s recent review. It was published in JAMA and I’m going to pull up a couple choice … It was called “The Challenge of Reforming Nutritional Epidemiological Research.” And I’m going to read the first two sentences because they’re classic Ioannidis in how pithy and direct they are. It says, “Some nutrition scientists and much of the public often consider epidemiologic associations of nutritional factors to represent causal effects that can inform public health policy and guidelines. However, the emerging picture of nutritional epidemiology is difficult to reconcile with good scientific principles. The field needs radical reform.”
Zoë Harcombe: Oh, I couldn’t agree more. I just couldn’t agree, I mean, I had the privilege of seeing John present at the Food for Thought conference in Zurich, which was arranged by the British Medical Journal and Swiss Re, a reinsurance company, and he gave the, I guess you’d call it the keynote, after-dinner speech by videoconference into the conference hall where we were in Zurich. And it was uncomfortable, shall we say, for some of the audience.
Chris Kresser: I could imagine. In an audience full of nutritional epidemiologists, probably didn’t like what he had to say.
Zoë Harcombe: Yeah, largely. I mean, I sat near Nina Teicholz, and we were absolutely loving it. But I won’t mention any names, but a couple of nutritional epidemiologists did walk out.
Chris Kresser: Yeah, yeah, I’m not surprised. It’s hard to consider, and there are ways that nutritional epidemiology can be done better. We can have more advanced data collection methods and an application of Bradford Hill criteria, which you mentioned, to increase the chance that the relationship between variables is causal. But the way it is now … there’s another critique that I love by Archer and he says, “For results to be scientific, data must be, number one, independently observable. Number two, measurable. Number three, falsifiable, Number four, valid, and number five, reliable. And these criteria distinguish scientific research from mere data collection and pseudoscience.”
And when you look at nutritional epidemiology, they do not satisfy those basic criteria for science because they’re relying on data collection methods like food frequency questionnaires, which are just a joke. I mean, they’ve been so thoroughly debunked as a reliable way of assessing what someone is eating. We know that human memory is not an accurate reproduction of past events. It’s just basically a highly edited anecdote regarding what we ate. And we know that these approaches that are used to assess what people are eating in these studies are really not accurate and not reliable and don’t fulfill the basic criteria of science.
So, I mean, we could talk a lot more about the problems with epidemiology, but I think let’s, given the time constraint, let’s just go on and talk with those caveats, those huge caveats. What did the epidemiological evidence suggest if it had been included in the original analysis that you looked at and then also since then?
Zoë Harcombe: So we covered in some depth at the time that essentially it was just the Seven Countries Study that found anything. None of the six studies found anything against total fat, and then just the Seven Countries Study alone found something against saturated fat. When you bring the epidemiology up to date, and I actually did it, I had to do it in a different way in the fourth part of the PhD because they didn’t have data on current epidemiology and total mortality or coronary heart disease mortality. So there was going to be nothing that I could actually update the original studies with.
So I had to look at different measures of looking at any relationship that I could find with deaths and total fat or deaths and saturated fat separately. Of course they’re not interventions anymore. So you are into just this epidemiological base of looking at the fat intake in different regions or in different studies. So it was slightly different to the other three that were looked at, and they were completely different studies, and probably studies therefore that are less well known to people. They were certainly less well known to me. So things like the Ireland-Boston study, Kushi, the US Health Professionals, Lipid Research study that’s very well known. The Pietinen Finnish counts study, a UK health survey by a couple of people called Boniface and Tefft. She’s not very well known, this new heart study, and then the Gardner Japanese study, which is probably not too badly known within the field. And when you separate it out, look in it, coronary heart disease deaths, so we couldn’t get the total mortality anymore.
But we could at least get the heart deaths and align those to either the total fat, where it was examined, or the saturated fat. There was, again, no significant difference for coronary heart disease deaths and total fat or saturated fat consumption. We were back to a limitation of the pooled studies from those seven that I’ve mentioned being almost entirely male. So 94 percent of the people involved in those studies were male. They were at least mostly healthy. Almost all of them had not already had a heart attack, but there was still no relationship for coronary heart disease deaths and total or saturated fat. So there was then a fifth paper that I published with the BJSM that wrapped up the four studies.
So it went through essentially what we’ve gone through now, which is, what did I do, looking at RCTs then, RCTs now, epidemiology then, epidemiology now? What was found? What wasn’t found, which was far, far more. And then an era that I suggest we’ll probably be heading into quite soon was to put what I’d looked at in context of other meta-analyses that had been done. Because I’m a PhD researcher, I was not straight out of finishing my degree. But I’m still just a PhD researcher looking at this evidence fresh in a systematic way.
A number of other people have also looked at the data in this field either for mortality or for events or for interventions or for epidemiology. And I therefore wanted to look at what everybody else had done to say have I found something different. Has everybody found this? Because you have to do that. You can’t come to the end of your PhD and say, “I find if I might drop the toast buttered 100 times out of 100 it falls on the butter on the floor,” if everybody else has found more (audio cuts out 59:06) it doesn’t fall with the butter on the floor. You’ve got to put your own research in context.
Chris Kresser: That’s another core principle of science. Shapiro, an epidemiologist, said, “We should never forget that good science is skeptical science, and science works by experiments that can be repeated. When they’re repeated they must give the same answer.” So this is another core principle. So what did you find when you looked at these other meta-analyses?
Zoë Harcombe: So the main ones that were pulled together, and there’s a great table in the paper five, which is one from 2016. I think it’s called “Dietary fat guidelines have no evidence base: Where next for public health nutritional advice?”
Chris Kresser: That’s a pretty straightforward title.
Zoë Harcombe: Yeah, it is pretty straightforward, actually yeah.
Chris Kresser: Not beating around the bush.
Zoë Harcombe: They’re pretty good, actually. They help you with titles. So they come up with catchy ones.
Chris Kresser: Yeah.
Zoë Harcombe: And I then went through, for example, I got the tape in front of me, so you’ve got Skeaff and Miller from 2009 who looked at RCTs and epidemiological studies. And they looked at mortality and events for total fat. You’ve got a fairly well-known study with Siri Tarino and colleagues from 2010 looking at epidemiology of fatal and nonfatal coronary heart disease and cardiovascular disease. Mozaffarian, who I had the pleasure of meeting in Zurich over the summer, and his paper looking at just events, not mortality. We might come back to that one. And then of course you’ve got the two well-known Hooper studies, that’s the Cochrane research, which should be the gold standard, but we can take a bit of a better look at that one. And then you’ve got Schwing, Jacqueline, Hoffman from 2014, RCTs, and my own study. And then of course you’ve got the Chowdhury study that looked very interestingly at the four different types of fats, saturated fat, monounsaturated fat, polyunsaturated fat, and they included trans fats in their research looking at coronary disease for both randomized controlled trials and epidemiological studies.
Chris Kresser: Right, and I’d like to read the conclusion of that one: “Current evidence does not clearly support cardiovascular guidelines that encourage high consumption of polyunsaturated fatty acids and low consumption of total saturated fats.”
Zoë Harcombe: Here, here.
Chris Kresser: So that was the Chowdhury, and then, so there were, I think, 39 total reports.
Zoë Harcombe: Yeah, 35 non-significant. And we don’t shout that often enough.
Chris Kresser: Yeah.
Zoë Harcombe: If you stop—and I’ve done this for a Welsh TV program that I was working on—we went to Cardiff, which is the capital city of Wales, and we walked down the main street in Wales and we said to people, “What do you think about fat? What do you think about fruit?” And people would tell you, “We need to eat five a day and fat is bad for us.” So people have got the messages. But what we’re not telling them is when you actually look at all the evidence, 35 out of 39 results were non-significant. No findings. And that has to be the most significant thing that has been found, that we didn’t find much. Why don’t we look at that more often? That’s so more powerful to me than the four findings. And if we just whiz through those. In the Chowdhury study that you mentioned, the one finding that they did make was against trans fats, and I don’t think you or I would give them any argument over that one.
Chris Kresser: No.
Zoë Harcombe: Mozaffarian, I really liked him in Zurich. But I was involved in a paper that critiqued his 2010 study which said, “You should replace saturated fats and polyunsaturated fats because there’s an impact on CHD events.” And our paper criticized that paper for excluding two studies that were not favorable to polyunsaturated fats, which was the Rose corn oil study and also the Sydney Diet Heart Study. And including, and it’s all bad studies, the Finnish Mental Hospital Study, which was not randomized, not controlled, crossover trial. I mean, just the worst possible trial to try and slip in to pretend it’s an RCT. So we critiqued that paper. I like to think he wouldn’t publish that paper if he had the opportunity tomorrow. I can’t speak for him, that wasn’t right.
And that’s, of course, exactly what the Sax paper did last year, the American Heart Association paper. Again left out two unfavorable studies, the same two. Included the Finnish Mental Hospital Study, they shouldn’t have done. And so basically, there were only two findings, and they boil down to one because it was the same research team, Hooper and the Cochrane team, working out of the east of England in the UK. One paper was from 2011 and the other was from 2015. And among 11 known findings for CVD, mortality, total mortality by modified fat, reduced fat, any kind of variation of fat, the only finding they could come up with was for CVD events when they looked at all RCTs for saturated fat reduction intervention. And we can get into that.
Chris Kresser: Yeah, I mean, I think we can stop here at least in terms of the actual studies because it’s, just to review what we’ve discussed, the randomized controlled trials that have been done since the late 1970s to today have not supported the idea that saturated fat increases the risk of death from heart disease, or any cause. The epidemiological evidence that has been done throughout that period does not support that hypothesis either. And even if it did, we’d still have all of the issues that epidemiological research, that make it problematic, like the healthy user bias and inaccurate methods of data collection, small risk ratios.
We talked about that earlier, how the increase in relative risk is so low that it doesn’t really meet the threshold for assuming a causal relationship in any kind of epidemiology outside of nutrition, and even in nutrition 20 or 30 years ago. Other people who’ve meta-analyzed these data have come to a similar, if not the same, conclusion as you did in your research and your PhD thesis. And I just want to highlight something that you said about how the Finnish study, which is really not a good study at all, has been included in a number of analyses. And you might wonder why that would happen if the researcher is aware of its limitations and that it’s not a valid study to draw any inferences from on this topic, why would it be included?
And again I’d like to turn to a John Ioannidis quote, and he says, “Consequently, meta-analyses become weighted averages of expert opinions. In an inverse sequence, instead of carefully conducting primary studies informing guidelines, expert-driven guidelines shaped by advocates dictate what primary studies should report.”
Zoë Harcombe: Yeah.
Chris Kresser: Doesn’t that sum it up?
Zoë Harcombe: He’s so brilliant, isn’t he?
Chris Kresser: So in other words, we start out with, the way that science should proceed is by doing experiments, and then if any guidelines are made, to make them based on these objective experiments. But the way it actually happens, a lot is we start out with a certain agenda and then we design studies that will return results that support that agenda. And anyone who’s worked with data in any capacity knows how easy that is to do.
Zoë Harcombe: Yeah.
Chris Kresser: And it’s not even conscious all the time. This is where confirmation bias comes in. It’s very difficult to guard against. I have to, and I’m not 100 percent, I’m not saying I’m 100 percent successful. But I can watch even myself. If I go and look for research on a particular topic and I have a certain idea, it’s very easy to just skip the studies that don’t support that idea. And that happens among scientists. It’s a real cognitive bias that is very difficult to guard against. And I think it has a huge effect on research.
Here’s another quote, this is from Casazza: “Confirmation bias may prevent us from seeking data that might refute propositions we have already intuitively accepted as true, because they seem obvious. For example, the value of realistic weight loss goals. Moreover, we may be swayed by persuasive yet fallacious arguments.” So again I come back to psychology. It’s a real … we’re human beings doing this work, right? All researchers and scientists are not infallible. They’re human beings as well. Many of them have their own ideas and preferences about diet and nutrition. They’ve been influenced by many of the same things that we as laypeople or myself have been influenced by. And it really, really does affect the outcomes of this research.
Zoë Harcombe: Yeah, yeah. I would agree with you. I mean I went in with a vegetarian, fat is bad, carbs are good bias. I am aware that I was so shocked by the things that I found when I started researching in this field, even before the PhD, you do then get quite skeptical and quite angry. And I now almost trust nothing. So every week I’m taking a paper from any kind of field, though mostly typically nutrition, and the low-carb study was one that I did recently, and then there was a weight loss drug that came out. And then I looked at red meat, the evidence because that was topical for something. And when I’m going in, I just assume that there’s going to be errors, and I’ve yet to find a published paper that doesn’t have something that you can point out as being really quite seriously wrong or disingenuous or open to interpretation.
Chris Kresser: Yeah.
Zoë Harcombe: I mean it’s really shocking. I saw on Twitter just a couple of days ago a guy, an academic, got so fed up with all the emails that you do get saying, “Oh, I really enjoyed this paper. Please can you write some papers for us.” Because papers are big spinners for the journals. He got so fed up with all these spam emails that he made up a complete nonsense study using Latin words that made something look really impressive, but it was basically saying something like, “If you do this with excrement, this happens.” I mean it was just, it really was, he was really taking the mickey and it got through peer review and he put it on Twitter. And he said, “I’m delighted to say that my complete nonsense article has just been published by this complete nonsense journal.”
Chris Kresser: Yeah, and I’ve heard other experiments like that that have been done where a lot of stuff like that has made it through peer review. And there have been a lot of critiques of peer review and why it’s broken and the links to the money in the research industry. Marcia Angell, who I mentioned earlier, in the context of the relative risk who was the former editor of New England Journal of Medicine, has famously said that some, I’m going to paraphrase, but “I now no longer think we can believe any published research,” is basically what she said.
So yeah, I mean, I think it doesn’t mean that research is not valuable. It doesn’t mean that we have to just become, I think you can go too far with this where we just say, “Oh, forget it. There’s no point in even trying. Let’s just discount all research equally.” Because there are differences in the quality of research and there are still studies that are done well even if they’re not done perfectly. And it’s the best tool that we have, that we’ve discovered today to try to answer some of these questions, at least on a population-wide scale.
Conclusions about Saturated Fat
So I want to close by just kind of going through some conclusions here. One, we’ve talked throughout that the evidence against total saturated fat is incredibly weak, if not nonexistent. But something we touched on briefly but I want to highlight here is that even if saturated fat were harmful, you have to consider the source of it in the diet. Get away from this reductionist approach where we think that saturated fat coming in different forms is going to have the identical effect. Because we don’t eat nutrients. We eat foods that have nutrients in them.
And I love how in your paper you pointed out that pizza, desserts, candy, potato chips, pasta, tortillas, burritos, and tacos accounted for 33 percent of saturated fat consumed in the diets of US citizens. A further 24.5 percent was unaccounted for and collated as “all other food categories,” which is almost certainly processed food. And so as a result, only 43 percent of saturated fat came from natural foods like dairy products, nuts and seeds, and burgers and sausages. Although I’m guessing that the burgers and sausages had highly processed buns and sugary ketchup and other stuff on them too. So how can we even look at those things as being anywhere remotely similar, much less the same?
Zoë Harcombe: Yeah. I mean I actually took out the processed meat in that section. So I got the natural foods listed down to cheese, nut butter, nuts, and seeds. And collectively together they accounted for 20.8 percent of saturated fat intake. But then I actually made the point it would’ve been ideal for the unprocessed chicken, beef, and eggs to have been separated from the processed meals because they always just lump them together. And they will always put, whenever there’s a study damning red meat, particularly in the US, it will always include hamburgers, which they are very firmly processed food in the UK, but for some reason seems to be considered as some sort of Paleo food in America.
Chris Kresser: Right, right.
Zoë Harcombe: And meat-type dishes, or something, which might be a curry.
Chris Kresser: KFC.
Zoë Harcombe: Yeah, exactly, Kentucky Fried Chicken.
Chris Kresser: That’s the way most people eat chicken, KFC or chicken nuggets at McDonald’s. There’s your chicken.
Zoë Harcombe: Absolutely, yeah. Or maybe a curry takeaway ready meal that you pop in the microwave and if it’s got a couple of percent of meat in it, you’ve done well. You bought a more expensive one than the average, which doesn’t even bother putting the meat in.
Chris Kresser: I mean, I’m … yeah, go ahead.
Zoë Harcombe: No, and that’s the crazy thing because I want us to be able to have heated agreement. I don’t want to be fighting anyone out there. The whole world is much better if we get on and we’re in harmony than when we’re fighting over anything, whether that’s territory or politics or nutrition or anything else. So I would love for us to find a way forward. And I think there could be a way forward by saying, “Guys, can we have a heated agreement that we can demonize processed food?”
And I know there’s industry conflicts all over the place and we have to expose the industry conflicts with the guidelines set in committee in the UK, they are completely dominated by the processed food industry and grocery retail. I mean, it’s just horrific. If somehow we could get the conflicts out of setting health guidelines, which please, for goodness’ sake, must be objective. We must tell people honestly what is healthy. Not tell them what the food industry wants them to believe is healthy. That’s got to be step one.
Chris Kresser: Absolutely.
Zoë Harcombe: If we can get all of that nonsense out, surely then we could agree that real food has got to be better than processed food. And there might be some debate what’s real food. But if it’s found in a field, it’s found growing on a tree, it’s found in the natural environment … I said to my niece when she was five years old, fish swim in sea, fish fingers don’t. Breaded fish.
Chris Kresser: If it comes in a bag or box, you probably shouldn’t be eating it.
Zoë Harcombe: Yeah.
Chris Kresser: I mean, of course, there are exceptions. Butter usually comes in a box, but yeah, that’s a general guideline, right?
Zoë Harcombe: Yeah, we know what real food is. It’s the best-quality meat, eggs, dairy products, fish you can get a hold of, it’s vegetables, seeds, nuts, fruits in season. There will be some debate over legumes and whole grains and how recently they’ve been part of our consideration set, and I agree with your point, there’s not one diet fits all because some people just cannot tolerate grains and legumes and fibrous products, suffering from irritable bowel syndrome or celiac or other digestive conditions. But somewhere within that real food, total consideration set, surely we ought to be able to set some principles that people can follow that are not based on advice from the processed food industry.
Chris Kresser: Absolutely, and I mean, I’ve said this so many times that Sean Croxton, who used to write in the health space, he came up with a diet advice that was JERF, he called it. J-E-R-F, just eat real food, which is, like, look, okay, we can debate about is it better, like you said, whole grains or legumes, in or out, saturated fat higher or lower, carbs higher or lower within this context of a whole-foods diet. But is there any doubt that if everyone ate real foods, we would decrease the burden of chronic disease and early mortality by something like 60 to 80 percent? I mean, I have no doubt of that.
And that’s again where this reductionist focus on nutrients completely isolated from the context of the foods that they come in has been such a disservice. Because imagine if we spent the last 30 or 40 years just hammering home the message that eating real, whole, nutrient-dense foods is really, like, if you want to simplify it for public health, like, that’s the message. Don’t even worry about those other finer points. And we would not, well there’s a whole other discussion about whether people will actually follow that advice if you give it to them and given the influence of our brains with highly rewarding and palatable foods in the food industry and all of that. But there’s no doubt that if people really did follow that advice, we probably wouldn’t even be having this discussion right now.
The other thing about that is it is possible at least in theory to, like, if we really wanted to answer the question of is saturated fat harmful, the way we would need to do that is we would need to take two groups and they would both have to have the same baseline healthy diet that we’re talking about. Just eating real, whole foods, right? And then in one group, they would eat more saturated fat. And then we would, this is to be a randomized controlled trial, we’d lock them up in a metabolic ward so that we could control all of the variables that we know can influence health, or at least most of them, and then we’d follow them for about 15 or 20 years and see what happens. And the problem is that’s never going to be done. I mean, that study would be hundreds of millions of dollars, if not billions, and no, Coca-Cola’s not going to pay for it, right? I mean none of the, no drug company is going to pay for that study. So unfortunately, that study is unlikely to ever happen.
Zoë Harcombe: Yeah, but “just eat real food” would work as a message until.
Chris Kresser: Exactly.
Zoë Harcombe: Yeah.
Chris Kresser: Just eat real food, and then we can use other mechanistic studies and other lines of evidence and maybe even shorter trials to try to answer some of the finer points. And those trials should also include individual, should also pay attention to individual factors or genetic or non-genetic factors that might bear on the answer to that question, so that we don’t then extrapolate the findings to everybody instead of just one group of people, for example. We know there are genetic polymorphisms that make some people hyper-responders to saturated fat and that can lead to an increase in LDL particle number. And the clinical significance of that is still controversial and debatable. But we know pretty certainly that that does happen.
So, but then if you were to extrapolate those results to someone that didn’t have those genetic polymorphisms, that would not be a valid inference. So yeah, it’s just disappointing that, I mean, we know this and yet we still go on doing the same things over and over again. And I have to throw in one last Ioannidis quote which—from that more recent, or I think from one of his previous papers, and I’m going to paraphrase this one because I don’t, let me see if I can find it—yeah, “Definitive solutions won’t come from another million observational papers or small randomized trials.” In other words, that was from a paper he wrote called “Implausible Results in Human Nutrition Research.” So in other words, doing the same thing over and over and expecting a different result is the definition of insanity, right?
Zoë Harcombe: Absolutely.
Chris Kresser: So, Zoë, thank you so much for joining me, and I know we went a little long, everybody, but I hope you enjoyed it and got a lot out of it. And I just, I wanted to have one podcast that we could direct people to to really answer this question and look at all the evidence on saturated fat in particular and its relationship with mortality and heart disease mortality. And I think we didn’t cover everything, but I think we did a pretty good job of getting the most of it out there. So thank you so much.
Zoë Harcombe: Oh, thank you. Can I just add one thing, because I think we just about completely nailed everything.
Chris Kresser: Please, of course. Yeah.
Zoë Harcombe: When we ran through the 39 results and found that only four were significant, and we dismissed Mozaffarian and we agreed with Chowdhury, and then the two Hooper results, which were just on the CBD events, we can actually put those to bed as well because aside from the fact that they’re events and they’re not mortality and we both agree that mortality is best, the thing that you have to then look at is why did Hooper find something different to all the other people? And when I went in detail, Hooper had actually managed to include four studies which involved only 646 people that were not about cardiovascular disease. And she’d asked the study authors if they happened to have data on cardiovascular disease events. So this was non-peer-reviewed data. That was the first thing.
When, and I owe Dr. Trudi Deakin in the UK, I always credit her for this finding, she spotted in the Hooper paper that when Hooper actually did as she should do, the sensitivity test on that one single finding, it was no longer significant. So the test that had to be done was not just which studies intended to reduce saturated fat or which studies actually did reduce saturated fat.
Chris Kresser: A key distinction there.
Zoë Harcombe: That’s really, really important, yeah. So Trudi looked at this and found that it is declared in the paper, but it’s tucked away on sort of page 158, or something.
Chris Kresser: Right.
Zoë Harcombe: That when the ones that were tested did actually reduce saturated fat only were included, there was no statistical significance and it was not generalizable because again in the whole of the evidence that was looked at by Hooper in either of those two papers, there was no single study of healthy men and women. But I think sensitivity tests apart from non-peer-reviewed data and apart from events, I think we can actually put that one to bed as well. So when you do that, because that’s the one that the other sites still try to hang onto. That’s the one that came up in the Professor Noakes trial when that’s down there for him as an expert witness. They tried to wave that in front of us and said, “Oh, see saturated fat is bad.”
Chris Kresser: Right.
Zoë Harcombe: So we hit them back with an, “Oh, no it isn’t.”
Chris Kresser: What’s the data?
Zoë Harcombe: Yeah, we kind of went in on the data. So there just is no evidence, and knowing the facts about fat, it would make no sense if there were.
Chris Kresser: Exactly. And that reminds me of the recent low-carb study which you and I both critiqued on our blogs. It wasn’t a low-carb study. The people were eating something like 40 percent of calories is carbohydrate, not to mention the fact that they reported a calorie intake that was basically at starvation level, which would invalidate the entire data set. So you don’t even need to go any further. I mean we did, but, like, that would’ve been enough, right? And all it takes is one major error like that, and it casts doubt on the entire data set and makes any kind of inferences that you would draw from it invalid. And I don’t think people understand that enough.
Zoë Harcombe: Yeah, I think critiques, I think the word you used there was quite polite. Actually, I think we both annihilated that study.
Chris Kresser: Yeah, well it was.
Zoë Harcombe: And a few other people as well, Nina Teicholz and Georgia Ede.
Chris Kresser: Several, yeah. I mean, it was frankly like shooting fish in a barrel.
Zoë Harcombe: Yeah, it wasn’t hard, that one, was it?
Chris Kresser: Yeah, it was not hard. I’m just actually kind of shocked that that kind of study makes it through peer review and gets published, given all of those issues. So anyways, yes, thank you so much for doing all of the work that you do, Zoë. It’s such a pleasure to meet you, virtually, at least, and to be able to really just concisely and comprehensively go through all of these deficiencies in the evidence and to just make it clear for people that this, despite the fact that they’ve heard this probably for 30 or 40 years, depending on how old they are, and despite the fact that it still forms the basis of our dietary guidelines, there really is no evidence to support it.
Zoë Harcombe: Yep.
Chris Kresser: Fantastic. Well, where can people find more about your work, Zoë?
Zoë Harcombe: Just ZoeHarcombe.com. And as I say, I blog every week. That’s my sort of business model. So if anyone wants to sign up and get the newsletter, there is lots of stuff on open view. But if you do that, then you support what I do and you help me to stay independent because I don’t take any money from anyone for anything in any circumstance.
Chris Kresser: Key. Absolutely.
Zoë Harcombe: I just work away and come up with what I want to find. And I know some people are on the email list who are quite fans of whole grains, for example. And I know every time I write a newsletter saying I looked at this whole grains study and it really didn’t stack up, I know that I’m upsetting some people who are subscribing, but I have to go with where the evidence takes me and I have to report as I find. So that’s what I do.
Chris Kresser: Yeah, yeah. I’m disappointing my readers all the time with my opinions and it’s important, I think, to stay true to what the data is showing and be as objective as possible about it. You’re one of the few people that I do follow regularly. I love reading your stuff, so everyone who’s listening to this, go check out the blog. It’s one of the most thorough and insightful sources on all of these topics that we discussed today.
And Zoë, we didn’t get a chance to go into much detail on red meat above and beyond its saturated fat content, which as we know is less than its polyunsaturated fat content. But Zoë has recently tackled that, the evidence behind red meat being associated with high risk of heart disease and death. And i’d really recommend checking that out too, because that’s another persistent myth that continues to this day.
Zoë Harcombe: Indeed.
Chris Kresser: Okay, everybody, thanks for listening. Continue to send in your questions at chriskresser.com/podcastquestion, and we’ll talk to you next time.
Better supplementation. Fewer supplements.
Close the nutrient gap to feel and perform your best.
A daily stack of supplements designed to meet your most critical needs.
I am not sure how much sense it makes to proclaim that saturated fat is not harmful and then go on to say, “We know there are genetic polymorphisms that make some people hyper-responders to saturated fat and that can lead to an increase in LDL particle number. ” How would one of those people know that saturated fat IS harmful to them? It doesn’t make sense to me to try to change the global view of saturated fat just because it applies to most – BUT NOT ALL. Perhaps the recommendation to keep it low was so that none would be harmed?