The Truth About Nutrition Science: Is The Government Getting it Wrong?

David: 
These become self-fulfilling prophecies. We say that people can’t follow them. We leave them with inadequate support in our modern food environment that doesn’t support a low carb diet. And then we say, see, nobody follows it.

Bret: 
Welcome to the Metabolic Mind Podcast. I’m your host, Dr. Bret Scher. Metabolic Mind is a nonprofit initiative of Baszucki Group where we’re providing information about the intersection of metabolic health and mental health and metabolic therapies such as nutritional ketosis as therapies for mental illness.

Thank you for joining us. Although our podcast is for informational purposes only, and we aren’t giving medical advice, we hope you will learn from our content and it will help facilitate discussions with your healthcare providers to see if you could benefit from exploring the connection between metabolic and mental health.

Do you find yourself sometimes confused by nutrition science? Unclear what it really says we should or should not be eating? Well, you’re definitely not alone. And nutrition science is a very complicated field and I think you could say whatever it’s done up to this point certainly hasn’t cured or prevented the diabetes or obesity epidemic.

But why is that? What is it about nutrition science that’s so challenging? And how can one study, which recently or you know, in the past few years, came out? Some people say it’s one of the greatest studies done, and some people say it’s the worst study and is misleading and should be retracted. The same study, but two completely different reactions.

How can that be? And do we have a new opportunity now with changes in the way research is being funded within the US government to allow for maybe a different approach that can better help us understand what causes this diabetes and obesity epidemic and how to get past it and how to help people get healthier?

I’m joined by two of the most prominent voices in this field with Gary Taubes and Dr. David Ludwig to discuss these topics and come up with some potential solutions of what we can and should be doing differently.

All right. Gary Taubes and David Ludwig, thank you so much for joining me again here at Metabolic Mind. 

Gary: 
Great to be with you. Thank you, and Bret, great to be here. 

Bret: 
Yeah. Now, I’m sure most people watching know who you both are, but I would like to take just 30 seconds each to give a quick reminder of who you are in case this is someone’s first time hearing you.

Dr. David Ludwig, let’s start with you. 

David: 
Sure, I’m a pediatrician and endocrinologist. My home base is Boston Children’s Hospital and Harvard Medical School. I’m spending some time in Denmark as a consultant for Steno Diabetes Center, which is a major patient care and research institute here.

The focus of my work has been trying to understand how diet independent of calories affects our hormone metabolism, even the expression of our genes in ways that affect risk for obesity, diabetes, and a range of chronic diseases, which have of course become the focus of the current administration and the health and human services.

Bret: 
Very good. All right, and Gary Taubes. 

Gary: 
And I’m a journalist and author. I’ve been writing about nutrition, chronic disease issues for 25 years now, most notably. And some infamous New York Times Magazine cover stories. And then a series of books that began with Good Calories, Bad Calories in 2007, and Rethinking Diabetes last year.

And I have a Substack on which I’ve been writing about these issues called Uncertainty Principles. 

Bret: 
Yeah, and and that’s part of why we’re here because of a recent Substack article from you, Gary, about NAH has a nutrition problem. And David based on a BMJ article you published plus a Medium article that you published about the concerns with nutrition research and the way it’s been done.

And David, you, alluded to the change in leadership in for the MNIH with Jay Bhattacharya now leading NIH, which will likely change sort of the focus on the way NIH research is being done. But let’s rewind for a second and actually, let’s take a broader scope. Let’s start with the broad picture.

What should the focus of NIH-funded nutrition research be? If we were just to describe what should the goals of that research be? Can we sum it up? Can you two sum it up without taking hours to describe it? Because there should be so many details. So, Gary, what should the, how would you sum it up?

Gary: 
I was going to defer to David, but to me it’s pretty simple. We have an epidemic of obesity and diabetes related chronic diseases, depending on how you could argue about the numbers. But over the past century, the, prevalence incidence of obesity and diabetes has exploded.

And with it, you get a whole host of associated chronic diseases. And like any epidemic, you want to identify the fundamental causes of the epidemic. If it was a infectious disease, what’s the agent? If it’s lung cancer, we had a lung cancer epidemic once, it was cigarette smoking and tobacco. So what’s it driving?

This epidemic of the disease. And if you can identify the cause of it, you can prevent and treat it. So that’s what I think the goal we always have to keep in mind. 

Bret: 
All right. So David, now you’ve done the research. You’ve been doing the research, and you’ve been analyzing the research for decades.

So tell us, what you think the goal should be? 

David: 
I think the first point to emphasize is that nutrition isn’t a single specialty. Heart disease or cancer or neurological conditions, nutrition affects virtually every chronic and acute and many acute diseases throughout the purview of the NIH.

And this actually relates to a debate that’s been going on. Should you have a specific institute for nutrition or should nutrition be divvied up across all of the institutes that focus on the different diseases? But, I think getting the research right on nutrition is critical. And nutrition has been a sort of a a poor stepchild for many years, severely underfunded.

Compare it to a single phase three drug study. If we want to create, and this is oftentimes industry funded, one drug for just one chronic disease condition, a new statin or a new blood pressure medicine, that process, bringing that drug from the laboratory to the patient’s office could cost a billion dollars.

And yet, most nutrition studies struggle for a few pennies on the dollar in that comparison, when in fact, nutrition research is more complicated than drug studies because we have to get a lot of things right, including behavior. So I think that it’s time that nutrition be recognized as underpinning, a range of chronic diseases and that it be taken seriously. 

Bret: 
Yeah, and when we look at NIH funding, they give $30 billion in funding with $2.2 going to nutrition. And on the one hand you could say, $2.2 billion, that’s a lot of money. But, on the other hand, when you see how much that is spread out among studies and how much each study gets is actually pretty poultry compared to the drug studies that you just mentioned.

But in your article and, Gary and your Substack, you both make the point of the way the research money has been spent and on the types of studies it has been spent. And you’ve even used the term is unethical, which is, pretty strong, especially when you talk about these studies, especially the food study from Kevin Hall or the low carb, or the sorry, the low fat versus the low carb study that he did.

Some people are putting such accolades on those studies as being some of the best nutrition studies done. And yet after reading your articles, they seem like some of the worst nutrition studies done. So why is there such this wild disconnect between how people interpret the quality of nutrition studies?

Gary: 
Yeah, one of the problems we’ve confronted in the whole nutrition field is there were certain belief systems that have existed since, I’m going to argue since the second World War, about the causes of the chronic diseases that affect us. So we get fat because we eat too much. Heart diseases caused by the saturated fat in the diet and eating too much fat, and we should all eat in effect, mostly plant diet, the more fruits and vegetables and the diet, the more whole grains and legumes, the healthier will be. 

And so you have these belief systems and we happen to have a field of research, I hesitate to call it science, where people consider a study to be well done if it agrees with their belief system. So if you do a study and you find exactly what the community wants you to find, your study is held as a success and, landmark research, depending on what you’ve done, and it’s published in one of the best journals in medicine.

And if you do a study and you find something that seems to contradict this conventional thinking on disease, diet-disease relationship, then it’s much harder to get it published. The journalism tend to be much less prestigious, or influential, and one fundamental problem, it’s not just the design of the studies, it’s people will welcome studies as extraordinarily important if they confirm their beliefs.

They will ignore them if they don’t. And this is the kind of bias that, arguably, the scientific method was designed to minimize for the last 400 years. And this is the kind of struggle that we’re dealing with today. 

Bret: 
Yeah. And that’s very pessimistic to hear. That’s very sad to hear. 

Gary: 
I’m notoriously pessimistic, yeah, 

David: 
I would be a little more optimistic. The problem is the truth ultimately prevails. It just doesn’t necessarily prevail along our timeline. Or sometimes in our lifetimes, our at least in our professional lifetimes. But I think a big problem is confirmation bias.

And that can happen on all sides. And so we have to recognize, that to the degree that Gary and I share certain beliefs about, our opinions about low carbohydrate diets are such that they’re also subject to confirmation bias. But it is impressive to me how some fundamental flaws, we’re talking stats 101-type flaws that would get a pharmacological study laughed out of a journal, are disregarded by some of my most esteemed colleagues because they comport with their mindset. 

And I think there’s a special problem when it comes to these very short feeding studies, which we’ll get into a little later, but you alluded to them. So they have all the bells and whistles. You bring people in and you control everything they eat and you maintain constant observation and you measure their metabolism with advanced technologies it’s really impressive.

These studies, very few places can do them. So we tend to rely on those results even when there’s a fatal flaw. It’s like having a really fancy car with all the bells and whistles, but the engine’s broken. It’s not going to get you very far scientifically. 

So the problem, which we’ll talk more about is that these short, very short term studies, these two week trials that are claiming to inform nutritional chronic disease understandings, are inspiring a whole family of children as part of, for example, the nutrition for precision health study. 

Now that in some ways has some advantages over the other studies we’ll talk about. But it’s this whole mindset. The paradigm that we can learn something about chronic disease, which takes years to develop from two week dietary trials. That’s, a problem. 

Gary: 
And let me give the rationale here, which is, as David just alluded to, chronic disease take years to develop. And if he wants to study the diet-disease relationship and truly understand that you need clinical trials that take years and follow patients or at least many months. So you could get some realistic idea of what a diet might be doing to some individual over the long term, not the short term. But the longer a trial has to go, the more expensive it is and the more difficult it becomes to do because you have to keep people on diets for months, years.

And this has been a failing of every major clinical trial, probably ever done in nutrition. And it’s led the nutrition researchers to believe that nobody ever adheres to a diet. If they argue against doing this kind of research, that’s why. But nutrition research in this sense, when you’re looking at the effects on chronic diseases, have these very unique, extraordinary challenges that other sciences tend not to have.

And that those challenges make doing trials correct. They can be exorbitantly expensive. And so the community has decided that the way to deal with this is to pretend that they could do trials that are, as David says, have stamp 101 errors in them and decide those are good enough. Because that’s the best they can do.

And in a functional science, you don’t say this is the best we can do. Therefore we’re going to believe the results are meaningful. You say, if this is the best we can do, we have to rethink what we’re doing. 

David: 
Yeah, just let’s just not set the bar too high and we don’t have to do a study for 10 years to have fundamental insights into the way that diet affects her metabolism and chronic disease risk.

We do need to go more than two weeks. And the reason is that the process of adaptation to a change in diet takes more than two weeks. Now, this is, we’ve known this in substantial degree for many years, but the specifics of this have become clarified in the last five years or so. That when you shift from, say, a high carbohydrate diet where your brain is dependent upon glucose that’s required, and any brief interruption in the glucose supply would cause loss of consciousness, seizure death, so the brain is adapted to that when you want to go to a low carbohydrate diet. You need a new fuel, the body has one. It’s evolutionarily ancient and it’s called ketones.

That’s why we call these very low carbohydrate diets, ketogenic diets. But to make this conversion from using gluten, one fuel to another fuel, it’s like you have an oven that’s burning gas and you want to convert to electricity. You’ve got to get the plumber out and they’ve got to take apart the pipes and rewire everything and bring in the electricity.

That doesn’t happen overnight. It takes several weeks. And so we know what’s going to happen for the first two weeks on a low-carb diet. It’s going to look bad. You’ve cut off the glucose supply. But the ketones, which take several weeks to reach their steady state, haven’t gotten there yet. And so that’s why people on a low carb diet, when they first start, describe what’s called the keto flu. 

They feel tired, hungry, fatigued. And if you study them for that first two weeks, you learn nothing about chronic disease. You just learn about the process of changing metabolism. And this is a profound Confusion. 

Bret: 
Yeah, so let’s briefly get into this study to talk about the specifics. So in this study, in a metabolic ward, or sorry, in yeah, in a confined controlled space like you were talking about, Gary, where everything’s controlled and all the food is provided, you know exactly what somebody’s eating.

Either started either on two weeks on a low carb diet or two weeks on a low fat diet, and then immediately crossed over those who were on the low carb diet, then crossed to the low fat, those who are low fat, crossed to low carb. And, Gary, in your Substack, you show this amazing graph. It is, it’s very well done breaking down the graph. 

Gary: 
It wasn’t mine. It was Kevin Hall’s, the author of the study. 

Bret: 
But you broke it down to show it. 

Gary: 
Kevin gets full credit. 

Bret: 
Yeah, so Kevin’s graph from his study, but you broke it down to show what happens before the crossover, in which case, essentially everybody’s eating about the same number of calories, whichever diet they were on the first week.

The second week it actually starts to separate where the low carb diet looks like they’re starting to eat fewer calories, but then when they cross over, that’s when all of a sudden the calorie difference is dramatically different with the low carb dieters eating more calories. But as time goes on, again, it starts to come down.

So, whereas initially the first week, if you switch to the low carb diet, you’re eating many more calories. The second week it started to come down, but the conclusion of the study was, this is clear, you eat low carb, you eat more calories. And it was widely accepted despite the faults. And going back to your comments about people’s biases and confirmation bias, it took someone who didn’t agree with the results to take a look at it and say, no, this isn’t correct.

Nobody who agreed with the results was dissecting the paper to look at it. 

David: 
Actually, let me just give you a little correction of the process here. The first paper, the trial, it was published in Nature Medicine in 2021. And it claimed that it targeted the carbohydrate-insulin model, which Gary and I have advanced in the first and last sentence of the abstract. 

So its intention was very clear. And, of course, we have a dog in this fight because we created the model Anyway. The paper claimed that there was no carryover effect. Carryover effect is what you were just talking about. It’s when one diet… 

Gary: 
Can we back up and just give a little bit more description of what the study did since for people who haven’t read the Substack of the BMJ articles? Just fill us in on exactly as you put it… 

So, who’s the interviewer here? 

Bret: 
We all are, this is just a discussion.

Gary: 
We got a journalist on the other side. I can’t help myself. We’re missing a little bit of the context, which is just, okay, I’ll give it. In effect, it’s a comparison between a plant-based vegan, excuse me, plant-based low fat diet compared with an animal-based ketogenic diet. So this is a low fat, high carb diet that’s plant-based versus a low carb, high fat diet that is animal-based.

And Kevin Hall, the principal investigators at the National Institutes of Health at the time, although he very recently retired, Kevin compares, houses his inpatient, his patients, his inpatient, his subjects, his inpatients in a metabolic ward in Bethesda at the NIH. And he feeds them one diet for two weeks, followed by the other diet by two weeks in random order.

And he hypothesizes that this carbohydrate-insulin model that David and I have advanced is predicts that people will eat fewer calories on the low carb diet because the carb-insulin model predicts that people will, that carbohydrates are fattening. And so if people are going to lose weight by Kevin’s thinking, they’re going to eat less on that diet.

So, two weeks on one diet randomize and two weeks on the other and ran the mortar. If this carb-insulin model is right, people eat less on a low carb diet than on the low fat diet. Now I’ll give it back to David. 

David: 
Yeah, that’s great. Alright, so we have these two diets that are competing head to head.

And in a standard trial, they would do so in parallel, it would be a horse race. They both leave the gate at the same time and run, and you see which gets to the end of the track first. But in these fancy metabolic ward feeding studies, it’s very expensive and it’s very hard to get enough participants.

So there’s this desire to increase power and you do this crossover. And the crossover is when you switch from one diet to another. In this case, there was no washout period where the diets could settle and then people could start the new diet fresh. And, so in the original paper, the investigative team said we checked and there’s no washout, but there’s no carryover effect. Leaving aside for the moment that there’s really no good way to rule that out because the tests aren’t sensitive enough and statisticians have warned for decades. 

Bret: 
So the conclusion is that this four week trial, two weeks on each diet, has disproven the carbohydrate-insulin model.

So it gets to the question of data science studies and how they inform or misinform us. And you’ve both made the case that bad studies are worse than no studies. So tell us why this study is so misleading. 

David: 
So when the paper first came out, it did so with the impression that it undermined the carbohydrate-insulin model people were eating.

Overall much more on the low carb diet than the low fat diet, which is the opposite of prediction. We recognize that this is maybe too short, but it looked okay. Statistically, the investigators said there was no carryover effect, based on their initial analysis. And that’s pretty much where the field left it for a couple of years, two years later, in 2023.

The main team at the NIH reported that, you know what, there was a carryover effect. They seemed, they didn’t quite say it this way, but they clearly calculated it the wrong way initially. And then they discovered that there was a massive carryover effect, maybe the largest ever reported in the field of nutrition, about 2000 calories a day.

That’s a carryover effect. That’s not even the treatment effect. That’s as much as some people eat all day. And so what that means is that the, so this was published in 2023, and then my team saw that and repeated the analysis and confirmed that was the case. But the fight has been over the significance of that carryover effect. The NIH group saw it as curiosity. 

They published it and they said, oh, this is interesting, and there are all these reasons why that might be. We looked at it and we were aghast. We said this totally invalidates the original trial, that the conclusions that a low carb diet are worse from that trial are flat out invalid, and the paper needs to be corrected.

Over the last two years, there’s been back and forth and back and forth, and we seem to be talking past each other. We have yet to get a clear explanation for why any trial, let alone this one, would be valid in the presence of such a massive carryover effect. 

Gary: 
Let me say, you have this study that’s published in a very prestigious journal, Nature Medicine.

David and I were going back and forth because I made a mistake in my Substack. I kept saying, Nature of Metabolism, and it’s Nature Medicine. And they’re both prestigious, but I still can’t believe that any such, an article would ever get published in that kind of journal because it’s A: it’s so poorly done.

And the answer is you think it got published because it confirmed what people want to believe, which is that a low fat, plant-based diet is a better option, is healthier than this animal-based ketogenic diet. But the study comes out, it’s cited 140 times since then and it’s, the argument is remember that if you eat this low fat diet, you’ll eat 700 calories less a day. And then they publish this 2023 paper that says, oh, wait a minute, if you actually only look at the first two weeks you’re on the diet. 

So it doesn’t matter whether you randomize to the low carb diet first or the low fat diet, you eat roughly the same calories. And if anything in the second week, you eat fewer calories on the low carb diet and then you get transitioned to the second diet. And suddenly the result of the transition, this is what David’s referring to as a carryover effect, is that now you eat 2000 calories less on the low fat diet than the low carb diet. 

So not only does the direction of the benefits shift with this carryover effect, but as David said, you get this enormous effect. And Kevin Hall and his collaborators report this in their paper and it goes back and forth with David.

And the original article was still out there in the literature being cited as evidence for the benefits of this low fat diet. 

David: 
I’d like to be slightly less cynical than Gary, which I admittedly is not maybe not hard to do. I don’t think that the, look, yes, I think two weeks studies, one has to view with a great level of caution, but I don’t fault the journal for publishing the paper.

I don’t fault the reviewers for accepting the paper for recommending acceptance, which presumably they did. The paper said that there was no carryover effect and it’s a subtle point that the level of confidence that you can have in these post hoc tests for carryover.

Yes, the convention, the state-of-the-art thinking in stats is, you can’t do this, but it’s a relatively subtle point. I think the big problem happened when the investigators at the NIH realized that they had goofed, they had miscalculated the carryover effect. They published a paper, they created a new term that nobody I know of has ever heard of to explain why they originally said there was no carrier effect. It gets a little technical, but this new term is “within subject diet order effects,” which doesn’t make sense to any of us. But the problem happens when once that’s happened, and then we said, and others said, wait a second, this carrier effect invalidates the trial. We’ve yet to have any productive responsiveness from the team to explain really what they mean by this new term “within subject effects” to explain why the trial would be valid in the first place in this case. And ideally, just to say, all right, we got it wrong.

We’re going to do the responsible thing. We’re going to retract the paper, redo the analysis properly, and republish with the right conclusions. That’s what scientists should do. That’s what’s in the obligation if we are actually thinking about what’s best for public health and patient care because they’re both based on having an accurate scientific evidence base.

Bret: 
This study is invalid. So let’s go let’s transition from this study to another two week study looking at high ultra-processed food or a low ultra-processed food diet. So very similar setup. Two weeks on each, and the conclusion was you eat more on the high ultra-processed food diet, something that pretty much everybody in the world agrees with or wants to agree with, unless you’re making ultra-processed foods, unless you’re a business that makes ultra-processed food. Everybody wants to believe ultra-processed foods is the enemy. 

Gary: 
I’m going to, I’m going to correct Bret there. If you’re a business that makes ultra-processed food, you want people to eat more of your ultra-processed food.

That’s why you ultra process. That’s one of the reasons why you ultra process it is to get people to eat more of your food than other foods. 

Bret: 
But you don’t want, you don’t want a study pointing to that as the cause of obesity. That’s what you don’t want. No, but yes, you have a very good point and you, but in this one. 

David: 
Oh, let’s just, let me push back on that setup.

Bret: 
Okay. Because I think here we go. 

David: 
We don’t want to, is we want to get the science right. Why? Because let’s think about, we’ve been down the road before where we base public health policy on short term studies and epidemiology. It was called the Low Fat Diet Craze, 50 years ago, and we know how that turned out.

So I think what we all ideally want is to know whether the concept of ultra-processed foods is robust and vigorous, or whether it needs to be improved in some way, or whether it’s actually not at all useful, that other things might be more useful. After all, this is a concept that was coined by one team in Brazil in 2019, so that’s very recent.

Scientifically, it’s never been revised and updated and taken apart and reassembled by other teams. There are, in fact, dozens of systems for classifying foods based on processing. Is this the best? Maybe, but I think what we all want to know, or should want to know, is whether this is the right system to base our public health policy on.

Bret: 
Yeah, and that’s the point I was leading towards. So, that’s exactly right. Just because we all want to agree in a way with that conclusion or do agree with it doesn’t make the science any better and that’s still a problem. Which points out the challenges with nutrition research.

if you ask any doctor, any clinician, any health researcher, they would probably all agree that yes, ultra-processed foods are terrible. They don’t have any protein, they don’t have fiber, they don’t have nutrition. They’re, they’re calorie dense, go on and on.

But how strong is the nutrition research showing that ultra-processed foods are harmful? 

David: 
But, Bret, that’s actually, that’s not, that’s a misconception. Processing affects the carbohydrates a great deal. If you take traditional, farrow wheated berries, that’s a slow digesting food.

It doesn’t raise insulin very much if you process it into white bread, or if you take whole food and you process it into juice, it changes a lot. But the processing of fats is pretty neutral. Olives to olive oil, we hear that olive oil is one of the all healthiest things you can eat. Dark chocolate is a highly processed, and in most cases, ultra-processed food.

If you take protein. A steak may be tastier and more expensive than hamburger. If we’re not going to be elitist, why should we make everybody eat steak if they want meat? It’s much less expensive, but there’s no difference in their health effects. And so the key question is whether all ultra-processed foods are really unhealthy or is this notion going to demonize processed foods that we will need and take the focus off of the really important drivers of chronic disease. 

Gary: 
Can I give you an example? Because one of the hypothesis is the ultra-processed foods cause the chronic disease epidemic, obesity, diabetes, et cetera. And ultra-processed foods is a very poorly defined concept.

It’s the amount of additives and the, if you can’t make it in your own kitchen with the ingredients in your kitchen, it’s an ultra-processed food. So Coca-Cola, for instance, is an ultra-processed food, but I am assuming if you make lemonade or your kid does and sells in a lemonade stand with lemons water and sugar, it’s not.

And then the question becomes is Coca-Cola with an equivalent sugar content? But all these chemicals also inherently more harmful than the lemonade. And we don’t actually know the answer. And one way to think about it, the New York Times about four months ago, did an article where they quiz, where you could, you were supposed to identify the ultra-processed foods in the supermarket.

And so one of the questions was you were presented with four ice creams and one of them was Häagen-Dazs. They were all vanilla and one was Häagen-Dazs, and I forget what the other three were. And the other three all had numerous ingredients and binding agents and gums. And they were ultra-processed.

And Häagen-Dazs only had four ingredients, maybe three, cream sugar and vanilla. Häagen-Dazs was not an ultra-processed food, but this ultra-processed food hypothesis says that people get fat because they would eat, if they ate the other ice creams, but not if they ate Häagen-Dazs. And there’s another way, the way we used to look at it is to say, let’s look at the macronutrient composition.

The carbohydrates and proteins and fats and the type of carbohydrates and proteins and fats. And maybe it’s just the sugar content in the high fat diet, maybe ice, you could eat all the ice cream you want, and it’s just the calories. So there are simpler questions that you can ask, but the ultra-processed food question says, look, if there’s 15 ingredients in the ice cream, it’s ultra-processed and therefore it’s unhealthy.

And if there’s four like Häagen-Dazs ice cream, it’s fine. Eat it all you want. And a study to begin with should be able to elucidate is it the number of ingredients or specific ingredients? And if it’s specific ingredients, is it the dyes? Is it the preservatives, is it the gums or is it the macronutrients and the processing of the carbs versus the fats?

And when this concept of ultra-processed foods was introduced, I think it was 2015, it just exploded. 

David: 
2009, 2009. 

Gary: 
Was it 2009. Okay. And it’s been accepted, as David said, with no critical discussion. There are critiques of it in the literature. 

David: 
No, there are plenty of critiques.

Gary: 
But what I mean is the acceptance has come without any critical discussion. If you think of how much discussion was had by the nutrition community back in the 1960s when they first raised the issue of dietary fats causing heart disease, massive panels of nutrition experts were put together just to discuss.

And we have the reports from those that are, today it’s ultra-processed foods are the problem. And there’s as much effort going into targeting the dyes and the chemical ingredients as there are maybe just the sugars and the refinement of the carbohydrates. And it’s as though you skipped one major step in the science and decided we’re just going to go to this ultra-processed food thing.

In general, we’re going to assume that’s meaningful. And it’s right now we’re going to try and establish why that causes obesity, but we’re going to ignore all these macronutrient issues because that’s that’s 19th, 20th century science and we’re in the 21st century. 

David: 
Let me just summarize that if I might.

Very simply, the ultra-processed food tends to be defined based on three things. The processing, the additives, and where it’s, where you make it. So Gary raised the question with whether some things, sugary beverage is made at home or by, in a big industrial food company, whether that makes any difference, that’s one issue.

The issue about additives, there’s some that are positive. There’s some additives that are completely innocuous or helpful. There are others that are clearly toxic like emulsifiers, which can disrupt the gut lining. So do we need to distinguish those? And then lastly, the nature of the processing, as we talked about before. Are processed carbohydrates, the main problem more than processed fats and protein?

And if that’s the case, we should focus on it and not get distracted by an overbroad concept. So that’s the debate. And that kind of leads us to the evidence. And there’s two lines of evidence. One is, which is 99% of it is epidemiology without a doubt, people eating more ultra-processed foods are unhealthier in every way that you look at them.

You got their weight, diabetes risk, heart disease, cancer mortality, everything looks bad. But what you have to do is you look at the Table 1 of all of these papers. The Table 1 has the baseline covariates. It tells you who these people are who are eating little bits, moderate amounts, or lots of ultra-processed foods.

And overwhelmingly, every time, these are fundamentally less healthy people. They have lower income, they have lower pursuit of other health behaviors, less exercise, more smoking, all of these things that’s called confounding. We try to take into account in experimental models, but it’s hard to completely eliminate it.

So you want clinical trials and, unfortunately, the only trial we have just two trials. One is two weeks, and the other is one week long. And that’s, I think, the topic that you wanted to get into. 

Bret: 
So we, we’ve clearly, I think, outlined how it’s very difficult to study this and study this appropriately with something that’s going to give us meaningful information that we can apply clinically.

So how do we do it differently? We have an opportunity potentially in front of us as a country now to do things differently with a change in leadership. And it’s clear what we’ve been doing has not helped the obesity and diabetes epidemic. It’s probably created more confusion than less confusion in certain and more certainty in terms of nutrition research. There are probably a million different things you both would want to do differently, but what are the key things that we can do to study this better and understand this better? 

David: 
Yeah, I think there’s one that tops the list. It’s just, it just begs out and it’s so simple. You have different, so I’ve proposed it’s an idea and I think Gary’s pretty much in agreement, that the problem is more focused with the carbohydrate processing than the proteins and fats. This is totally testable. We have a lot of evidence in favor of a low carb diet for weight loss, for diabetes. So let’s put these head to head. Does a low carbohydrate diet versus a low ultra-processed foods diet maybe versus some control, which maybe is a low fat diet.

You line ’em up, you send off three groups following each. You have enough people, you have enough intensity, treatment intensity to help people actually follow these diets. So they get to meet with the nutritionist regularly. They get some in-home counseling and maybe they get some meals provided, which takes money.

Again, it’s less than a drug trial and you let them go for a year or 18 months, and you ask, does the focus on ultra-processed foods produce a substantially better effect than other approaches? And if that’s the case, we have actionable public health data. But what if it turns out that focusing on a low carb diet blows away the ultra-processed results?

if we don’t do the study, we’ll never know that. And we’ll again, miss the boat. 

Bret: 
So how long would a study like that need to be and how much would it cost, do you think? 

David: 
I think if we had all the bells and whistles, it’s still going to cost a fraction of the phase three drug trial, like it gets done all the time.

I think if we, most, like the NIH has a cap on standard awards of $500,000 a year for five years, which usually gets cut back. So maybe at most you get $2 million on a standard award. I think to do this kind of a trial, we’d like $20 million, but that’s still the 10th the price of a drug trial.

So you do this with a few hundred people per group. You provide them with substantial support, including helping them, provide them with foods so that we increase the likelihood that they really follow it. And you do it for a minimum of a year, ideally two years, is two years too long to wait?

I’d rather have the right answer in two years than the wrong answer in two weeks. 

Bret: 
That’s great. Great point. 

Gary: 
How would this differ from the DIETFITS study that Christopher Gardner did? 

David: 
True that’s a fine study. His conclusion was that, there wasn’t much difference between the low fat and the low carb diet.

We reanalyzed the data published in AJCN, and we found that there was actually a significant difference favoring the low carbohydrate diet. We just used different approaches, and, in fact, a key hypothesis of the carbohydrate-insulin model proved true that the people who make a lot of insulin are most sensitive to the carbohydrates.

So those are the people who gained the most weight, the high carb diet, the problem with that is that he reduced what’s called the, he reduced the sugar and the glycemic load in both groups. So the contrasts that he produced were small, and that probably obscured a lot of the differences. So what we would want to do is really separate these groups in the trial as much as possible.

We want the low carb group to be eating substantially low carb. It doesn’t have to be a ketogenic diet. You want to get the carbs down to maybe 25% or less. Totally in my view, sustainable. You can have some fruits and occasional grain dish, but you get the carbs pretty far down. You get the ultra-processed group to really focus on those ultra-processed fruits so they can have plenty of carbs in the form of potatoes and lots of grains and all sorts of treats, but they just have to make their desserts at home.

And, then maybe you have a third group, which follows a traditional low fat diet. And you have them really differentiated. And again, supported so that we can get a meaningful answer. 

Bret: 
And the other part about the DIETFITS study, Gary, is that it started off as a true low carb diet, but then they liberalized the carbs that it went.

So by the end of the study, they were eating 130, 140 grams of carbs per day, which is not how it started. So it really was two different diets depending on the timing. But they put that into the methods specifically because it would be hard to maintain long term. Like they just made the assumption people would not be able to maintain the low carb diet long term.

So they built in this “add carbs as you go cycle,” which kind of leads me to my next question. We are approaching this from the scientific perspective of a randomized control trial, as what we need to know the difference. It’s what you do in drug trials, it’s the gold standard. But when we’re talking about nutrition, people need to go home and choose what to eat and they have to enjoy their food and it has to keep them full and not hungry and not craving.

And when they get emotional, they have to be able to stick to the diet. Many more factors than just taking a pill. So are we even barking up the wrong tree to say we need to do this in a randomized controlled trial versus a self-selected people choose the diet they want to choose, which clearly is not the same high level of science, but is it more appropriate for nutrition science?

How do you weigh those factors? 

Gary: 
Okay, first of all, if you do a study like that, the people who choose one diet are going to be very different than the people who choose another diet. So when you say it’s not up to the sort of typical scientific standards, what you’re saying is, we’re going to get a result that can’t be interpreted because we won’t know why the people on one diet are healthier than the other.

And it’ll also depend, for instance, if you did the study in Texas where people I assume like meat and barbecue more than they like than they do here where I live in Berkeley, California, you’re going to get an entirely different result because the people who choose one diet versus the other are going to be very different.

The only way, again, one of the critiques I’ve been making of nutrition science since I got into this business is that, and I said this in the beginning, you sacrifice scientific rigor for these vague real world, implications or the reality of how expensive and challenging it is to do the study.

So instead of doing the science right so you can actually get a reliable answer that you can trust. And the question becomes, you know, every experiment asks a question and you want to ask the right question. What we want to know, like I said, is on one level, what are the cause of these epidemics?

We’ve known, for instance, in diabetes since 1797, that restricting the carbohydrate content of the diet is the single best way to treat type two diabetes diet. The dietary treatment for type two diabetes, just feed people and affect ketogenic diets that, so that’s been known for 228 years, and yet the American Diabetes Association would still recommend, for the most part, restricting the fat content and the whole grains, fruits, vegetables.

David: 
No, they’ve updated that. 

Gary: 
They have become updated a little, but not tremendous. 

David: 
They’ve become officially macronutrient agnostic, although, I guess you’re right, there is an implicit bias toward high carb because they suggest you start with something like 40 grams of carbs per meal or something like that.

That does tend to push people in that direction. 

Gary: 
What I was going to ask if, again, taking the journalist perspective here, and what I just said is we’ve known for several hundred years at these diets or the best treatment, dietary treatment for type two diabetes. it’s been clear since the 1960s that they’re extraordinarily useful for obesity, which is related to type two diabetes.

It’s been clear since the 1920s. It’s extraordinarily useful for pediatric epilepsy. We now have research suggesting they’re very extraordinarily useful for other cognitive disorders, but we haven’t and we’ve made progress in that a far greater number of people are aware of the diet, are doing a diet today, but we have not shifted public opinion.

This is where I get the institutional opinion. This is where I’m pessimistic and David is more optimistic, or I’m cynical and David is more, the sun will come out tomorrow, whatever. We’re going to call that terminology. The but how do we, the trial, you’re describing is, would be a very useful trial.

And let’s say you get the result you expect. And it’s published in an ideal world, The New England Journal of Medicine or JAMA, because they’re very influential medical journalists. How do we get people to care, especially when they say nobody wants to be on a diet anyway, and now we have these drugs?

David: 
Yeah. there’s a lot there, Gary, but I think the first point to emphasize is that unlike the problem right now I don’t see is with the public. The problem is with the research agenda. We’ve had dozens of, not just multimillion, like multi $10 million diet trials, that are all focused on low fat.

And the one notable one was look ahead, which was a low fat diet among people with diabetes, type two, to reduce heart disease. And it was closed early for futility when there was no evidence that the low fat diet would ever, no matter how long you ran the trial at millions of dollars a year, would show any benefit.

So we’d have all of these in pediatrics, in adult, we have not one large multi-centered trial of equal scope focused on low carb so from sponsored by the NIH. So let’s get a few trials. Let’s do them right so that we can see the true magnitude of effect that you can get when people are truly supported to do a low carb study.

See these become self-fulfilling prophecy. We say that people can’t follow them. We leave them with inadequate support in our modern food environment that doesn’t support a low carb diet. And then we say, see, nobody follows it. But that’s, we can break that if we have data. I think the people who are struggling with obesity, type two diabetes, other chronic diseases and don’t want to take drugs, will view them as real options.

And the last thing I’ll say is that I’ve seen thousands, tens of thousands, of patients and recommended for the most part modified low-carb diets. Not extreme ones, but modified ones. And, oftentimes it’s a self, it’s a, victorious cycle that happens. People get on a diet like this, they feel less hungry.

They can start losing weight without the same struggle. And the shame that they’ve experienced around their in effectiveness, those lack of self efficacy, that goes away and then they’re embrace the diet and make the changes willingly. I’ve seen that many times, not necessarily for everybody, but why shouldn’t this be an option for the public?

it’s a shame that the fundamental research that would cost a fraction, like one day of the cost of type two diabetes would pay for this. 

Bret: 
Yeah. So again, that was the point where I was going towards that you said not everybody does well on that diet. And so when we’re talking about nutrition, we have to recognize that not everybody’s going to fit one bucket.

Even if that bucket has been proven to be the healthiest and the best diet, not everybody’s going to want to follow that diet. And that was part of my question about self-selecting the diet. But you bring up a perfect point is that if we’re not talking about low carb diets as an option because they haven’t been studied the same way, then we’re leaving out that entire group of people who are going to thrive and do well on the low carb diets.

So if we can’t afford to do the randomized controlled trial for two years, then we have to do these smaller studies, but we’re trying to hold them to the same standard as the bigger study. I mean with this whole discussion as a background, are you both more optimistic that now we are in a place where we are going to fund the studies to help us better understand whether, and how low carb benefits individuals that we can start talking about it more?

Not that it’s the one and the only diet, but that it has to be promoted as an option. Are you both optimistic? 

Gary: 
Can I, and just this one and only diet concept, but again, if you think of what we’re trying to figure out as the cause of this epidemic, so massive explosion in obesity and diabetes rates over the past century, something’s causing that.

And it’s probably the same thing in every person for the most part and every country, because you see these epidemics play out in very similar ways, and there might be many factors involved in every person including genetics and epigenetics. But it’s probably very a dietary effect and it’s probably some aspect of the diet that can be identified.

The carbohydrate-insulin model that we were talking about proposes that the problem is the refining of the carbohydrates. So if people want to prevent this, not just treat obesity and diabetes through a dietary approach, but prevent it on some level, they have to remove the cause of the disorder, which is either the refined carbohydrates specifically or, quite likely and regrettably, virtually all carbohydrates, whether refined or not, except those in green leafy vegetables. 

So even for the argument would be the conventional wisdom today is that a diet that works for weight loss, a diet that gets you to restrict calories, the alternative argument is a diet that works for weight loss is a diet that gets you to restrict the carbohydrates that caused this excess fat accumulation.

When we mentioned the DIETFITS study earlier, which the not-for-profit I co-founded and helped fund, one of the issues was Christopher Gardner, who did it, had said, I think eating sugar and refined flour is bad for anyone. So I’m going to tell both groups not to eat sugar and refined carbs.

So it’s quite possible that a diet that works, whether it’s vegetarian, mostly planned vegan, carnivores, a diet that restricts first and foremost sugar and refined carbs. But that has to be established scientifically before you could really communicate it widely. And it’s like identifying cigarettes as a cause of lung cancer.

Once you did that, you can tell people, if you don’t want to get lung cancer, don’t smoke. But before that, you had no way to do it. And fortunately, cigarettes and lung cancer was an issue that could be pretty much established reliably with epidemiological studies. These nutrition issues aren’t. 

Bret: 
That’s a great point.

Gary: 
That’s problem with the one diet for everyone issue. 

Bret: 
That’s a great point. When Chris designed the study, he made the assumption that refined sugar and refined flour is bad for everybody. But you’re saying we don’t actually have the evidence to support that, which I think is just fascinating.

Gary: 
Yeah. You just would need a better study to establish that scientifically so you could communicate it as, because the world is full of people, you could have Mike Jacobson from the Center of Science Public Interest on, and he’ll tell you, here’s all the reasons why we shouldn’t be eating meat or animal products, and he’ll be just as convincing as we are.

And he’ll lack science completely. So on some level, you need to have reliable, rigorous science to establish public health recommendations because otherwise it’s all about who do you listen to and who you think is authoritative? 

David: 
Coming back to your question, you have on one hand public health, which is we identify the main drivers, whether it’s smoking, we’re postulating that processed carbs.

Not all carbs are a problem for obesity, but that you identify the fundamental drivers and you make the recommendation across the population. At the other extreme is the fashion of precision medicine where one size doesn’t fit all and that everybody’s unique. And through at least in this new, big national study, $170 million substantial proportion of the NIH nutrition research budget.

That the idea is that artificial intelligence, AI, can identify based on omic elaborate measurements of our metabolism, what diet would work best for everybody. And that every diet, every person needs a different diet. I think the truth is probably in the middle, and I’m going to say closer to the public health side, that there are some fundamental consistencies in how much of us, many of us respond but there are important individual differences, and that’s actually built into the carbohydrate-insulin model.

The people who are high insulin secretors. How do you know if you’re high insulin secretor. You had your fat around your midsection. You look like an apple. Those people are going to just, those are the ones who gained a massive amount of weight during the low fat years. If you’re a low insulin secretor, you may have a wider range of different kinds of diets that could work.

So I think that we do want to have both perspectives, but I do worry that we’re going to get lost in this personalized nutrition world and lose sight of the forest for the trees. 

Bret: 
Yeah, it comes to the world where there’s no judgment. There’s no right or wrong. Everybody can eat whatever’s right for them and figure it out for themselves, which really leaves no guidance. and no understanding of where people can start. But since this is really about the nutrition science I want to bring up one other point here is that, if we’re spending $170 million of government money to do a study, you expect that to be the top quality study that’s going to really answer a question.

And if we compare that to a self-funded study of just a hundred people and a topic that’s never really been studied before as a pilot study to, hopefully, spark some interest in future studies. Maybe you hold it to a different standard. That’s how I would see it, that you can hold those to different standards.

That one is to stimulate discretion and for the research. And one, if it’s $170 million of government money, better answer a question. Do you see that? Am I hearing echoes of, ketos CTA here? Potentially, yes. I think it deserves to be addressed. Just the difference in the widespread difference in funding and different structures.

So my question is, do you see that as the same? Because you made the analogy, sometimes you need to pull the weeds more than plant the seeds. You need to get rid of the bad science. So how do you see that difference? 

David: 
We started this conversation by saying that these two by two crossover trials without a wash on period, that’s the one that we’re really, that Gary Taubes highlighted in his recent Substack and we focused on in our recent BMJ paper.

You know that these two by two, two week trials are inherently like they’re designed to fail. You can’t, there are going to be crossover effects. There’s no way you can analyze this and they’re too short. You’re not going to learn anything useful in nutrition. For the most part, the nutrition for Precision Health Initiative, that $170 million study is a little bit better than that for a couple of reasons.

They do have washout periods of a minimum of two weeks. Two weeks isn’t long enough, but it’s better than nothing. And the other advantage is it’s a three by three. What does that mean? There are three different diets being tested. And they’re being tested in each of three different periods rather than two diets in two periods for complicated statistical reasons.

You can, you have a tool to try to extract the diet effect from the carryover effect in a way that you can’t with a two by two. So it’s a little bit better. But what I rather have $10, $17 million long-term trials rather than one 170 million trial of two weeks? I think you know the answer to that question.

Gary: 
One of the things that mystifies me about this trial is you imagine we get the results that they want. So they have 8,000 subjects and you’re taking all these, as David put it, these omics, all these measurements you I’m doing. Sure they’re doing DNA sequencing. And the thinking is at the end of this, you have all this information about how people respond to these diets in two weeks.

So you’ve gone from eating whatever your baseline diet is to some diet for two weeks and now the it compared to these three other diets and now we have AI grind through all this data. And AI, by the time this is done, will be brilliant beyond our imagination. And that will be able to tell each of 8,000 different types of people what diet will maximize their health in theory and maximize their chance of living a long, healthy life from whatever they saw in two weeks.

I don’t know how AI’s going to do this. I can’t imagine it, but I’m not AI. Now everybody in the country has to go in and find out which type of person they are to find out which type of individualized diet they should be on. So somebody like David, who is sitting in his clinic in Boston, and he is, his population is full of, his waiting room is full of, he had an pediatric obesity clinic, right?

It’s full of obese children suffering from obesity. Many of them lower socioeconomic class, I would guess most of them. And the idea is David is going to DNA type them and he’s going to test their gut biome and then he is going to be able to tell this child, you’re on diet, one sub a category X, and this child is on diet two, sub B, category Y.

And to think that is better than if, maybe saying, look, kids, I know you love carbs. We all love carbs, but they’re making you fat, which could be true. It may not be. 

David: 
But we probably wouldn’t put it in exactly those words in the clinic. 

Gary: 
But it might benefit if it did. Maybe we should. 

Bret: 
Yeah, maybe we should be in that point.

Gary: 
Like a poker comes in, you don’t say cigarettes may or may not give you lung cancer. You don’t say, look, 9 out of 10 of you guys ain’t going to get lung cancer. But so you might want to consider smoke quitting because you’re the one in 10 that will. Your patient comes in smoking, you say, look, if you want to be healthy, long-term health, you can’t smoke.

It’s just, it’s a way. But you have to have the knowledge to say that. 

David: 
I think that’s what we all agree on, that we need the definitive knowledge. It is deplorable that we have such poor database on such fundamental questions that have been debated for a century. We can answer these questions with, we know how to get the answer.

We just need to make nutrition research a priority and stop settling. Therefore, such demonstrably poor designed trials, which would not make it out of the starting gate in most other fields. 

Bret: 
Wow. I think there’s been a whirlwind of a discussion about nutrition research, about how what we have been doing has not been working, but hopefully with new opportunities ahead of us.

So as things transpire, hopefully, we can have you both on to talk more about the progress. But in the meantime, David, where can, where can people follow you to learn more about you and your work? 

David: 
I’m on X, the ex-Twitter. I’m on Bluesky and Facebook. My handle is, for the most part, David Ludwig, MD.

And I’m also easily reached through Boston Children’s Hospital. Let me end with an appeal that relates to the topic of this conversation. While we hope that the NIH is going to be funding these definitive trials because, this is pennywise protection for a pound of disease, there’s a lot of philanthropists who are interested in getting effective answers quickly.

we only need, 10 or 20 million to do a study that will, be in the history books, contact me. Very good. 

Bret: 
And Gary, how about you? 

Gary: 
Yeah, I am also on X and Facebook and LinkedIn, Gary Taubes and my Substack is Uncertainty Principles, because I’m fascinated, I think we are always dealing with uncertainty and how we try and minimize it is. The key to good science and I suppose good health.

Bret: 
So, very good. Thank you both for joining me. I really appreciate. Thank you guys. I want to take a brief moment to let our practitioners know about a couple of fantastic free CME courses developed in partnership with Baszucki Group by Dr. Georgia Ede and Dr. Chris Palmer. Both of these free CME sessions provide excellent insight on incorporating metabolic therapies for mental illness into your practice.

They are approved for a MA category one credits, CNE nursing credit hours, and continuing education credits for psychologists. And they’re completely free of charge on mycme.com, there’s a link in the description. I highly recommend you check them both out. Thanks for listening to the Metabolic Mind Podcast.

If you found this episode helpful, please leave a rating and comment as we’d love to hear from you. And please click the subscribe button so you won’t miss any of our future episodes. And you can see full video episodes on our YouTube page at Metabolic Mind. Lastly, if you know someone who may benefit from this information, please share it as our goal is to spread this information to help as many people as possible.

Thanks again for listening, and we’ll see you here next time at The Metabolic Mind Podcast.