This Will Change Your Mind on Ultra-Processed Foods (Harvard MD PhDs Discuss)

The term ultra-processed foods, it's effectively us putting blinders on and ignoring the mechanisms that cause overeating.

To me, that's proof that overeating doesn't drive obesity. It's not the overeating that's driving the weight gain so much. It's the metabolic effects of the food. Focus on the drivers. You lower your body weight set point, then you're just naturally less hungry.

Do you want to be the sheep that follows or do you want to lead and think critically?

It is really shocking to me. How uncritical so many of the leaders in our field are around this notion.

All right everybody, my guest today is somebody very special to me. This is Professor David Ludwig, MD, PhD. For his academic bio, it's quite lengthy. I'll try to keep it concise. He is an endocrinologist and researcher at Boston Children's Hospital. He is a full professor of pediatrics at Harvard Medical School, founder of the Optimal Weight Life Program, OWL Program, and director of the New Balance Foundation for Obesity. And Prevention Center. He's perhaps best known as the thought leader and originator of the carbohydrate-insulin model. And he's also a treasured mentor of mine throughout my time at, at Harvard Med and even before. And I just wanna add a sticky note, one of the very few senior professors who I've met in my career who is consistently retaining his humility, his open-mindedness, his thoughtfulness, his commitment to doing what's right for others, even at personal cost. That's a very rare quality. We're not going to delve into his altruistic acts behind the scenes, but I did want to pin that because I think it's something pretty special. Today we are going to be talking with Professor Ludwig about a popular, controversial, and really thoughtful article he wrote on the topic of processed foods and ultra-processed foods for the prestigious New England Journal of Medicine. I would encourage everybody to keep an open mind on this. It might not go in the direction you expect. Anyway, Professor Ludwig, thank you for joining me.

Thanks. Now that you've totally embarrassed me with that introduction, um, let's dig in. Excellent.

So, uh, I want to get right into the juicy part. So you published, uh, a perspective piece in the New England Journal, which if people don't know is I think the highest impact clinical journal in the world. And one of the lines from it is the available evidence suggests that a focus in national nutrition policy on reducing ultra-processed food consumption won't make Americans any healthier. Now, I know this quote's going to set off some fire alarms for members of the audience, and I know you and that you're not defending Twinkies. So can you unpack this quote a little bit for us? How do you defend the statement?

Sure. Let's begin by asking, what is this term, ultra-processed foods? It's, it's become an iconic label of all the things that are wrong with the food supply. And, you know, who could oppose reducing ultra-processed foods? All of these industrial products that we intuitively know can't be doing us any health favors. But for any public health policy, we need a functional definition. So let's review the concept. So this comes originally, although the term had been used since even the 1950s, it has been popularized by a Brazilian researcher who in 2009 proposed a system that was subsequently labeled NOVA. NOVA doesn't stand for anything, it just means new. So the NOVA system reacted to the fact that a focus on nutrients, so what are nutrients? You know, protein, fat, carbohydrates, saturated fat, sugar, vitamins, minerals, that traditional focus on nutrients might have been fine for addressing nutrient deficiency diseases, pellagra and rickets, or protein deficiency, but it wasn't working for chronic diseases. Chronic diseases kept going up— obesity, diabetes, heart disease, cancer. And, uh, so this Brazilian researcher suggested that we instead focus on food processing and that the less processed a food is Basically, the better. Although he also acknowledged that some processing has been both traditional and goes back hundreds of thousands of years. Fire is a processing system. And so this NOVA system has 4 categories. The first is whole or minimally processed foods, you know, things like vegetables, fruits, minimally processed grains like oat groats or wheat berries, beans, nuts, eggs, milk, fish, and meat. So these are category 1. Category 2 are traditional culinary, uh, ingredients, but that includes things like white flour, refined sugar, um, salt, and different kinds of fats. Group 3 is still okay. This is processed foods in which you take these minimally processed or unprocessed foods and you combine them with these culinary ingredients like oil, fat, or salt. So far, so good. But then there's Category 4, which is ultra-processed, and this is considered these industrial monstrosities. The problem is the definition that to make something ultra-processed by this definition, it's got to be a, a commercial product, something that you don't make in your kitchen, and it has to have only one ingredient or main food substance that wouldn't traditionally be used in the kitchen. So there are all sorts of things that we use and the food industry uses that are benign or even helpful— amino acids, various kinds of additives that improve the taste of foods or serve to make them more convenient. But they they have no adverse health effects. So this ultra-processed category includes a range of ingredients, and we can come up with some examples, from perfectly fine foods that will be needed if we're going to feed the world. Vegetarians are going to eat plant-based meat alternatives, for example, to foods that are demonstrably unhealthful. But when you have something that is that imprecise, this basket of wastebasket, in effect, I'm arguing that neglects critical mechanisms, then the risk in public health policy is that we create more harm than good.

So what I hear you saying almost is that the ultra-processed food rallying cry is it's problematically vague and imprecise. And although it might work as a general guiding principle for your average person, because like, you know, you can categorize certain things as very clearly bad ultra-processed foods like You know, candy, Twinkies. What we're looking for here in guiding nutrition policy is something that's scientific and easy to define. Ultra-processed foods doesn't meet that definition. So it's kind of like something that people might be familiar with is the idea of like blaming the, you know, burger for what the bun did. We have a big category here. We're not actually getting specific about, okay, what forms of processing are creating the health issues. And without that precision, we can't actually faithfully conclude that this is actually going to be helpful in addressing the obesity epidemic. Is that a fair summation?

Yeah. And to use the metaphor, I use this in my blog, which I believe you'll, you'll link, you know, the metaphor that— or is it an analogy that is oftentimes used? Let's say you're— but with a little twist here. Let's say you and a you and your buddy want to go out and shoot arrows at a target in the woods for practice. And these arrows are very expensive. You head out. Unfortunately, your buddy has very imprecise aim. So maybe an occasional target arrow will hit somewhere in the target, not the bullseye, but you might get somewhere. Other arrows miss the target entirely and get lost in the forest. Maybe even one arrow veers off wildly and hits you in the leg, even though you thought you were standing at a safe distance. So this is the problem. This is the threefold problem with imprecise public health policy. It can miss the mark and not accomplish much. You know, it's not hitting that bullseye often. It wastes time and energy. You know, we've got those expensive arrows that are now lost and can't be used for other things. And it can cause actual harms. The big problem driving Precision is that processing has radically different effects based on the macronutrients— carbohydrate versus protein and fat. So by abandoning nutrients entirely, which this system does, it's, you know, says nutrients hasn't helped us in the past. Let's completely forget about it and let's consider, you know, the processing of wheat berries into white bread, or basically fruit into sugar, as equivalent as the processing of steak into hamburger, or soybeans into tofu for the vegetarians. Now, this makes no sense because processing rapidly accelerates the digestion rate of carbohydrates. And the only thing keeping carbohydrates from being absorbed rapidly is the intact food structure of an apple where the sugar is, uh, sequestered into the cells of the apple or the fiber in the food form of a grain kernel. You know, you process that and then the carbohydrates can be digested within literally minutes. Blood sugar and insulin rise excessively, and that drives all sorts of metabolic problems, including weight gain. Such as in the carbohydrate insulin model that you've mentioned. But, you know, you don't have to believe in that model, you know, think that model's right to understand the problems with raising insulin. I mean, this is a unifying factor in what was called the insulin resistance syndrome. We now know it as the metabolic syndrome. So by failing to distinguish between the processing of carbohydrates, which is harmful, and the processing of fats and proteins, which is neutral, Other examples: olives to olive oil. Olive oil is a single macronutrient extraction of olives in the same way that a sugary beverage is a single macronutrient extraction of cane or high fructose corn syrup— of corn. But they have radically different health effects. And my paper and my blog go through many, many examples of this. So by abandoning an understanding of mechanisms, we throw the baby out with the bathwater. I realize there's been a lot of images here, but yeah, no, no, hopefully you got the picture.

Honestly, more analogies, the better. You just have the one that sticks with someone. But I hear what you're saying. You're saying that, you know, like I come back to this idea of a virtue signal because I think it's, it's, it's quite reflective. Like it sounds bad, ultra-processed food, but when you start to unpack it, there's a lot of different types of processing and I don't think it should be controversial to say, yeah, a different form of processing is going to have different effects on metabolism. And if we're going to guide nutrition policy, we need to understand, okay, what type of processing are we talking about? Otherwise, yeah, you're going to end up lumping things like, you know, cold-pressed olive oil with, you know, refined sugar because they go into the same processing category. In fact, when you start to go through the examples, it's pretty appalling. I mean, like, tinned sardines in brine can classify as NOVA 3, or certain cheeses. Whereas like you said, you know, pure syrups and sugar can be classified as Class 2. So like on that—

Okay, let's, let's take the example that you included, I think in the post, you compared maple syrup with tinned sardines. So you're right, maple syrup, if you simply tap a tree, that's a traditional method and you boil it down without adding anything, that's a form of minimal processing. That's allowed in Category 1. So maple syrup, although there's certainly ultra-processed forms of it, but traditional maple syrup would be potentially 1 because it's a, you know, it's like a whole— it's a minimally processed food. Maybe it's a 2. That's another problem because there's a lot of confusion about how things get categorized. So tinned sardines are at least nova-3, because any time you, you know, you don't just put the sardines in the tin and leave it be, you know, you're going to add at least salt, probably some kind of oil, and that makes it nova-3. And then if you have any additive that, you know, adds flavor, but it wasn't a traditional— traditionally used in culinary practice or something that helps enhance the appeal of it. It still, you know, could be a very healthful food, but now we bumped it up to Nova 4.

No, I mean, those examples I think highlight the issues. I wanna move on and ask, because I think people will be wondering, there are studies like observational and randomized controlled trials where they say, hey, you know, we are gonna compare ultra-processed food diet to another diet, and the ultra-processed food diet inevitably has a worse outcome. So can you unpack some of the core issues with the existing literature on ultra-processed food that You know, maybe the headlines would present as rigorous science, but might actually be flawed science.

So you're talking about the clinical trials?

Clinical trials.

Okay. So how long do we have today to unpack these problems?

Consider this your elevator pitch on the topic, and we'll link all— you will link your blog, we'll link the New England Journal piece. Yeah.

Okay. So there are a few key problems. First off, One can bias. So these— let's start by saying, what are these studies? These are now basically 3 main ones. 2 have come out this summer. One came out in 2019 and has been highly cited and a reason for the rush to incorporate ultra-processing, the notion, into national health policy. So these range from 2 weeks. One is 2 weeks, one is 3 weeks, and the other is 8 weeks. So we've got a problem right here. In fact, the 8-week paper found— compared ultra-processed to minimally processed diets. They fed people these different diets, although in this particular case they were free-living, and they found that actually both groups lost weight. So that was one huge inconsistency with prediction because the prediction is that all ultra-processed foods make you overeat. And the headlines still say overeating, but if you're losing weight, you're not overeating. And Gary Taubes has talked a lot about this. He has a blog out actually today on this. But the people in the minimally processed group lost about 2 pounds more according to their statistical models. So the first problem is that if you look at the diets, they're radically different. There's ways of designing these diets in which the ultra-processed and the less processed foods are fairly compared. Here they're radically different, as is the case in all of these studies. In fact, in my blog I highlight a typical meal, uh, where they— in one case the ultra-processed food is a burger and fries, and the minimally or unprocessed meal is salmon with this massive portion of green beans. Now really, is this, is this fairly testing? How do we know it's the ultra-processing and not all of the other things that we know can affect eating habits? The familiarity of the food, the tastefulness, the energy density. These are all issues. Going back to that 8-week study, there were 55 people who were randomized and it was just a crossover study, 8 weeks on each, and they had 12 dropouts. Out of 55 for 8 weeks. That's a lot. And which— they all came from which one group? And which was that? That was the unminimally processed group that lost an extra 2 pounds by their models. So we've got a lot of problems here already. First of all, if so many people drop out of one group, it's telling us that they don't like the foods, they're finding it difficult to eat. So is it That's surprising that people might eat a few hundred calories less and lose a couple of pounds over 8 weeks from foods they don't like. We could design all sorts of untasty diets that have nothing to do with ultra-processing. They're going to make people lose a couple of pounds. If you take a, like, a lot of garlic and add it to the diet, everything, you know, including the ice cream, you're going to find that Many people don't really like it, even though some people might be able to tolerate garlic. Some people will really hate it and drop out of the study. But we're not recommending garlic as the answer to obesity. Maybe werewolves, but it's not the answer to chronic disease. And so this study has the same problem. And one more key issue here is that when you have so many dropouts, as has been known for decades, the— your confidence in statistical estimates go down. Because you don't know what the dropouts really did if they had stayed with the study. Those might be the people who perhaps gained more weight. They certainly differed from the others. So the assumptions in the statistical models, technically something called data missing at random, it's a technical approach where you can predict what people would have done based on other characteristics, that these assumptions aren't correct. So the 2 pounds could easily have been 1 pound or nothing. And yet this study, like these other 3, are being held up as definitive proof of the benefits of ultra-processing. When I think people are simply getting distracted by all the bells and whistles of these feeding studies. Just because you do a lot of things rigorously, you make the meals, you deliver them, you monitor food intake, you monitor outcomes carefully, does not mean that the data are valid. If there's one key flaw. Every home chef knows this. If you get one ingredient seriously wrong, the cake isn't going to— the low-carb cake isn't going to taste good.

Your logic is usually pretty watertight, but I have to point out you screwed up your mythology. I think garlic's vampires, right? Not werewolves. But anyway, you heard it here first. Yeah, that's my only critique. But in all seriousness, On the garlic point.

Guilty as charged. Yeah, it's, it's, it's the silver bullet.

The silver bullet is the werewolves.

Is, is for the, which is what, let's, we can complete this then. It does seem to many people like ultra processing is the silver bullet.

But all right. All right. You completed the metaphor. You heard it here first. David Ludwig abandoning the carbohydrate insulin model for the vampire diet. But in all seriousness, on palatability. Since you bring it up, I think proponents of, you know, the standard energy balance model idea that ultra-processed foods cause obesity will lean towards this idea that ultra-processed foods are designed to be hyperpalatable and that this leads to overeating. You alluded to some of Gary Taubes' ideas on this, but can we unpack that concept of hyperpalatability and why similar to ultra-processed foods, I think we probably both agree it's not scientifically rigorous useful?

The notion of ultra— of hyperpalatability gives me serious indigestion, and it is really shocking to me how uncritical so many of the leader— leaders in our field are around this notion. It is a tautology, basically. So it sounds good— hyperpalatable— you eat, you know, so much you can't control yourself, and you You know, somehow the food industry is able to manipulate these foods so that it undermines your metabolism, overwhelms your self-control, as if humans are automotoms. I mean, first off, adults control themselves all sorts of ways through conscious and subconscious temptations, you know, drugs, smoking, sex. You know, we, we modify our behavior. We're talking about a tiny energy balance difference, which is the result— which results in obesity. I mean, one extra slice of bread, um, a day, you know, would lead to 10 pounds weight gain over, uh, 3 years, if all things being equal. Um, so clearly adults have the capacity to control themselves, even if there are subconscious drives here. But the notion itself is circular. What are foods That are hyperpalatable? Well, they're the ones that make us overeat. How do we know that they're over— hyperpalatable? Well, we observe people overeating on these foods and therefore we define them as hyperpalatable. Leads to comparisons that don't pass the taste test. If this were Coke or Pepsi, these relatively bland processed carbohydrates popcorn, baked potato chips, sugary beverages. I mean, you know, they're sweet and appealing, but they're pretty bland. These are all hyperpalatable foods, whereas a fine French cuisine or Italian food isn't. Does that really make sense that the dry popcorn is tastier than the bouillabaisse?

Yeah. I mean, and maybe these terms will go over people's heads, but the simplest way I could put it, and maybe this is oversimplifying it, but in science you have an independent variable and a dependent variable. The independent variable you change, the dependent variable is the outcome. They cannot be the same thing. And effectively, overeating and then by association hyperpalatability, they are the same thing. You're saying you overate. How do we know you overate? Because you gained weight.

No, it's so it's just a classic tautology. And yet there's no, you know, I'm, I'm still awaiting an operational definition of hyperpalatability that everybody can agree on. Then we can, we make a definition and then we can test it and find out if it's, you know, meaningful or there are exceptions or if it's worthless. It has become one of these terms, as tautologies are, that confuse rather than elucidate thinking.

Yeah, it's so hard to—

and now that doesn't mean that— that doesn't mean that foods don't have an ability to drive a positive energy balance, drive fat storage, promote disease. That's— but is it due to the inherent palatability, hyperpalatability of these foods, or is it due to the metabolic effects? We did a study more than a decade ago published in AJCN. We gave— this is one of the few blinded trials on food. It's hard to blind people to what they're eating, mask the foods. But in this case, we made milkshakes. They had— in this case, they had identical macronutrients, same protein, fat, and carbohydrate. We just used different carbohydrates in one or the other. It was a crossover study. You know, one got a slow-digesting carbohydrate, cornstarch, uncooked cornstarch. And the other got corn syrup, which is much faster digesting. So it came from the same species and we controlled the sweetness and the taste so that they were comparable. We gave these milkshakes and then did functional magnetic resonance imaging of the brain. So same milkshakes, basically the same sweetness. And then 4 hours later, when the subjects would have had no conscious recollection of the tastes in their mouth. We saw that after the fast-digesting carbohydrate milkshake, the nucleus accumbens— that's considered ground zero in the brain for cravings, addiction-like behavior, rewards— that lit up in every participant after the fast-digesting milkshake associated with a surge and crash in blood sugar. After the slow-digesting milkshake, the nucleus accumbens remained con. That could provide a basis for understanding why people would crave these foods in a way that has nothing to do with the taste. It's the post-absorptive or metabolic effects of these foods, and that gets completely disregarded in the conventional view of obesity.

You said a few things. There are two in particular that I want to highlight because they're so important. One's kind of been operating in the background. And this idea of like a mechanism and a concrete scientific explanation that is distinct from the heuristics we use to navigate around in the world. I know when people hear this conversation and hear like, okay, so we're going to have to define heuristics because it's a, it's a template, it's a guide, it's a vibe, it's a feel.

So like, it's best demonstrated through a learning, a learning opportunity, like a way of—

you're, you're, you're— as, as humans, our brains build like best guesses to navigate around the world. They automatically do this. What's going to be right most of the time that's going to help me with my survival and understanding community connection, you know, has to do with also your human experience. So this, this concept of ultra-processed food, because we have certain images about what this is, gets internalized. And it's really hard to unlearn. And it's the same thing for other concepts like eat a balanced diet, which I also think is kind of hand-wavy and not well-defined. Um, you can go through a litany of examples for energy balance. The idea that, oh yeah, you must overeat to gain weight, therefore overeating must cause you to gain weight and be fat. It's just we've been exposed to them so much they become internalized in our understanding definition of the world. Whereas when you try to parse it scientifically, it's actually very anti-scientific. And distinguishing between those two things is, I think, very important for your average person to do, but also very difficult just because of the society we've grown up in and what we've been exposed to. So I want to just challenge people first and foremost to kind of partition between those things.

Let me— let's, let's just pause for a sec. So I'd like to pick up an example of why these, as you're calling them, heuristics can be misleading. So let's go back to your example and then we'll bring it back to ultra-processed foods. Let's go back to your example. People who are gaining weight are eating more. If you eat more, then you have this positive energy balance and therefore overeating is the cause of obesity. So what's wrong with that? It seems to make sense, right? Well, let's do a metaphor. Let's say you're a parent and you notice your teenage son or daughter is suddenly eating 1,000 calories more. And boy, are they growing fast. They're eating a lot more. They've started to grow fast. So the overeating is making them grow, right? Wrong. Wish it's the opposite. It's— well, that's right, because because you can't overeat yourself to any more inches than you've already got. That's— so in this case, it's the opposite. And that's, that's intuitively obvious to us. We know that a rapidly growing boy is going to be— or girl is going to be hungrier and they're going to overeat relative to how they were eating to support the calorie needs of putting down the new tissue. Of course, if they didn't, then they wouldn't grow as well. But it's not the overeating that's making them grow. It's not the mother's overeating during pregnancy that makes the fetus grow. And so from that perspective, the alternative perspective, the alternative view paradigm, carbohydrate insulin model, fuel partitioning model, is that it's not the overeating that's driving the weight gain so much. It's the metabolic effects of the food. Food that are causing fat cells to grow. When the fat cells are growing, they're sucking in too many calories. There aren't enough for the brain, the liver, the rest of the body, and we get hungry to replace those calories being sequestered in fat. So if that's the case, then it's opposite the way we typically think. And if that's the case, then our focus on calories is misleading and ineffective, and that we need to focus on the problem. Okay, so that's an example. It's an ongoing debate, but it's an example of why these, as you call them, heuristics can mislead us. And so coming back to ultra-processing, ultra-processed food is brilliant marketing. I mean, really, I have to say that without that term, the NOVA system, you know, would be one of many others. By the way, there are many others, a dozen or more food processing-based systems. But NOVA has exploded in popularity and influence and is being discussed as a basis for the new dietary guidelines. And I think part, part of this is just marketing because that term is so— has become iconic. But we need to ask whether the term is sufficiently precise to do more good than harm. In public health.

So what I hear you saying, and then, you know, bringing it back to just what is the core problem, I think what we're mostly talking about here is the obesity epidemic and associated metabolic health epidemic. And to bring it back to this, you know, original misleading heuristic around obesity, it's not that calories push you to grow, push you to become obese, is that the metabolic hormonal milieu creates a vacuum which then you eat to fill. And so the term ultra-processed foods, while it's a great slogan, it's effectively us putting blinders on and ignoring the question, well, what are the mechanisms that lead to the hormonal milieu that cause overeating, quote unquote overeating?

Yeah.

Yeah.

So let me jump in again before we go past these. So this brings up the key— another key problem with these trials. They're short. You can— there are all sorts of biological systems that are controlled unconsciously. Breathing, body temperature, you know, growth, all these things. But we can affect them with our— through our behavior temporarily. Like, we all know that you don't will your body temperature to be one way or the other, but you can throw off your clothes or jump into cold water, as I'm sure you do. And living part-time in Denmark, I do. And you can lower your temperature, but it's temporary. You can't lower your temperature permanently. Nobody has that willpower. It's too difficult. Biology is too powerful. In the case of body weight, we can do it. The time frame is different. You can— most people can change their body weight consciously by restricting calories for days, weeks, a few months. You know, very few people can maintain that longer. And we have extensive research, meta-analysis going back a century that, you know, if you look at the curves of weight loss by calorie restriction, 2, 3, 4 months, and then weight regain, usually back to where you start by a year. That's not what we want. We want to change the system so you don't have to constantly work through calorie restriction and fight your biology. And that's where the carbohydrate insulin model offers an alternative. That you focus on the drivers, you change that, you lower your body weight set point, and then you're just naturally less hungry and your metabolism is faster. And that happens without conscious struggle. So this is the problem with the, you know, the ultra-processed food studies. They're all too short. You know, in fact, in the one that kicked it all off, 2019 in Cell Metabolism from the NIH, there was a difference overall of about 500 calories a day. Again, wildly biased differences in food energy density, no washout, a lot of other problems. But if you look at the curve, that difference in calorie intake between groups was weakening rapidly even during those 2 weeks. The p-value for that weakening effect was 0.0001, less than that. So if you extend that for another 2 weeks, it's gone. It's completely extinguished. All sorts of things can affect body weight again over the short term. We don't base public health policy on it, and we need to relearn these sort of clinical trials 101 issues. No drug study would ever— would be— you'd be laughed out of the journal, any journal, if you, you know, any big journal, if you proposed to make definitive conclusions on a weight loss drug from 2-week or 2-month data.

Yeah. So I want to highlight this because I think this is one of the most difficult things for people to get. When the idea of calories cause obesity is challenged, the example that's always brought up is, well, what if you take people, you know, and put them in a room for a month and like force-feed them calories? They will gain weight. But what you're saying and what I agree with, or what I hear you saying, is that The short-term trials with the bells and whistles, like, you know, the feeding trial that's in The Metabolic War that is kind of short-term, they sound compelling because they have the bells and whistles. But extrapolating from that scenario to what is causing chronic obesity in our population is kind of like akin to me jumping in the air and saying, hey, if I take the like freeze frame of me at my peak, oh look, I'm flying. Like you only are getting a very misleading picture rather than what is actually like the driving metabolic cause of chronic obesity. Maybe that makes sense to people.

Yeah, yeah, exactly. I hadn't thought about the jumping metaphor, but let's specifically take the overfeeding studies. Yes, for more than 70 years, researchers have taken subjects, volunteers, and said, we'll give you a lot of money if you'll come into our metabolic ward and we're going to feed you an extra 1,000 calories a day of really tasty foods, cakes, cookies, all sorts of delicious things. And people say, really? That's great. First day, great. Second day, not so great. By the third day, they're probably miserable. They are so stuffed that, you know, they feel like we do after Thanksgiving dinner. Their metabolism, they're sweaty. They have— they suffer as much as the people in the starvation studies. So if overfeeding— if overfeeding were driving obesity, then once they finish the study, those calories should stay, right? Because they should just go back to their usual way of eating. But no, there's compensation. They wind up eating less for weeks until their weight comes right down back to where it started. And these studies are very clear, and I can provide the references. You know, 3 months later, the— 3 months later, they are the same weight as they started. To me, that's proof that overeating doesn't drive obesity. But yet somehow people continue to use your, you know, snapshot of what happens before Gravity takes—

no, I've done these experiments myself, sometimes intentionally for that purpose and other reasons. So my Oreo experiment, for example, pure Oreo edition. I can tell you, day 1 Oreos, tasty.

Yeah.

By day 5, I was like, this is the worst thing ever. And by 3 weeks after the Oreo study, I had dropped below my baseline weight. I just rebounded.

We should acknowledge there's, there's going to be individual variation, and among overall people come right back, there are some people who maintain some of that weight. Nobody maintains all of it, but there are some people who maintain some of it, and that might be enough to drive weight gain. And so I'm not saying that you can't alter the system by overeating to some degree, and I actually have a paper with a Danish guy named Thorkild Sørensen that suggests there might be a way of integrating the carbohydrate insulin model with the energy balance model. But I think that without understanding that if you force calories into fat by overeating, they tend not to stay there unless you also have metabolic effects that serve to lock those calories into fat. That's insulin. And that's why it's much harder to overeat on steak than sugary sodas.

But I actually think this is a great example of where it's not obvious what the alternative hypothesis might be. I'll elaborate, which is, you know, you said individual variation and yes, some people when they overeat maybe are forced to overeat. But I think about it over more of a chronic time and they develop obesity, then it's harder to lose weight. But when you look at some of the studies and they've done human studies on this and people like, say, post-bariatric surgery, you can actually see the epigenetic fingerprint of past obesity, like in the epigenome of the fat cells where they're primed to be more inflammatory.

Which—

and I bring this up because it's not as simple a matter of, oh, there was just more energy and it expanded the balloon that is the fat cell, now the fat cell wants to retain it. It's like, no, there are epigenetic changes, there are microbiome changes, so many things that are changing that are complex. And from a scientific perspective, if we want to address the problem, we need to understand what these are, not just wave our hands and say they overate.

The, the irony is that the most ardent opponents of the ones who are doing at least, you know, the basic science of the carbohydrate insulin model will be the first to recognize that body weight is influenced importantly by metabolic, genetic, hormonal factors. I mean, you can knock out a gene which has a variety of— in a mouse which has a variety of metabolic and hormonal effects in the brain, but also in the body. And they see that it gains weight. And some of that is not just overeating, even though, as we've talked about, the fat can drive the brain just as the brain can drive the fat. But so you can pair feed them and these knockout models still will gain more weight on the same number of calories. So the basic scientists will accept this easily if you're talking about genes, if you're talking about drugs. But somehow when you come to diet, all bets are off and these mechanisms can't apply, even though we know that diet is one of the most potent modifiers of our hormonal response.

So you're saying you can uncouple calorie intake from fat gain very easily. You can even have systems where your calorie intake goes down, you gain more fat. And obviously there's—

that's, that's very clear. It's very clear that you can even—

so you do the most controlled experiments in mice. So people will just be like, oh, it's a mouse study. But it's like, no, no, no. Your concept is that it's the calories that drive fat gain as a law of physics. It should apply to all mammals. If you can like prove this even in an animal, you can uncouple calorie intake from fat gain, then yeah, it should generalize to human beings as well.

Anyway, one key, one key footnote here. It's not— we're not violating the first law of thermodynamics. We're not saying that Somehow we can uncouple energy balance from fat storage. We're uncoupling the food intake, which means that if you're losing weight with more calories or gaining weight with fewer calories going in, it means your metabolism is changing. Your metabolism is either speeding up or slowing down. We have a BMJ paper that shows that, but that in humans, but, but this is a given. In animal models. We know that calorie— that metabolism is influenced in these metabolic models of obesity. We're just having trouble getting the basic scientists to take the next step and say, you know, if diet can affect genes, if diet can affect hormones, hormones can affect genes, genes can affect metabolism, why not diet can affect metabolism And then let's use that for the benefit of chronic weight management.

It's all just about critical thinking. And I think that's so difficult for— well, it's difficult for the average person. But what we're really saying here is because of what we've been exposed to and internalized, it's also difficult for most scientists in the field. And that's the problem.

Well, you know, the most— the scientists are fine critical thinkers, but I think there is a, as you say, as there oftentimes is. The existing paradigm or heuristic, you know, that can be very useful to guide experiments, but it can be a blinder for something that's sitting right here. You know, it's, it's like an answer that's sitting just here, but the blinder is keeping you from seeing that. Unfortunately, that's the nature of the human mind, especially when we build careers on certain perspectives. And I'm not leaving myself out here. I, I clearly have my own blinders, which is why ideally we want to have productive dialogues with our opponents because the opponent is motivated to tell you what you did wrong. And ideally you should say, thank you, that'll help me design a better study. And we do that both ways. Um, that's why I meticulously avoid ad hominem on my social media and my writings. I can write strongly, but I don't attack people personally. And, um, I'm surprised at how many of my colleagues have forgotten this sort of like what I learned in kindergarten lesson.

Yeah. Well, I won't name names, all of a sudden come to mind, but I, I do wanna wrap up and leave this somewhere pragmatic for people, which is, you know, we've challenged people to think—

so bring it back to ultra processing.

We challenge people to think critically, but now, you know, what should the average person keep in mind when making daily food choices? Because we're not saying go buy a Twinkie and, you know, Eat them. Yeah.

Well, again, let's, let's maybe I could wind it up this way that the NOVA system, I was one of the early supporters of the system. I wrote a paper in JAMA acknowledging it and complimenting this Brazilian researcher in 2011, and he just proposed it in 2009. So it's a very useful concept for helping us to understand how processing might influence food quality independent of nutrients, but it's not ready for primetime, and it's way too imprecise. So we need to bring back an understanding of nutrients. So my take-home lesson here is that the processing of carbohydrates, I mean, minimal processing is all right. You know, you don't want to eat the husk of a, you know, of a grain, minimal processing, but keep your carbohydrates as minimally processed as possible. If you're not on a very low carbohydrate diet, then it doesn't matter. Keep— whereas for processing of fats and proteins, it's much different. So then we get away from some of these very odd comparisons in the NOVA system, which like says it's fine to eat cornflakes if you add sugar at home. But you can't buy a store-bought brand that is high protein, low sugar, and high fiber because it's an ultra-processed food. So that, you know, or you can eat steak at, you know, like if these days it's $20 a pound, but you can't eat most commercial burgers at a third the price. And then for the vegans who are going to need the plant-based meat alternatives, I mean, they're all ultra-processed, even many versions. And I looked this up very carefully. I didn't want to make any mistakes. Many versions of tofu that are commercial are ultra-processed. So are we supposed to make tofu from soybeans at home through this very laborious process? You know, we need something better.

So when it comes to ultra-processing, if we're going to make it a little bit of a better quote unquote heuristic, although not a perfect one, probably be more cautious around ultra-processed carbohydrates, fats and proteins. And then, you know, but then we don't need the term ultra-processed.

Then we can just use the term, then we can use the term highly processed. That's a more, much more specific term, but it doesn't have a clickbait value. Well, you know, remember, remember the low-fat diet era? Only, you know, a little before you were born. But everybody was convinced if you don't want fat on your body or in your coronary arteries, don't put it in your mouth. That was pretty strong heuristic. And even though there's still a remnant of that in dietary policy and in people's the public opinion. Mostly we've abandoned that. And so, you know, these simplistic, even if they're compelling terms, can be— yeah, you know, it can be confronting.

I think what we're saying here, I'm going to do my call to action. I'm going to give you your chance to give your call to action is, you know, thinking critically, you're often the one that's herding the sheep or redirecting the flock. And that can be very uncomfortable. So the question I ask for people is like, you know, these discussions can be uncomfortable. They can cut across the grain. Do you want to be the sheep that follows or do you want to, you know, lead and think critically? So that's my open question to people. I do want to end by asking you to tell me how my audience can support this work in this dialogue. Yeah.

Just one final footnote. It doesn't mean that, you know, eating less on an individual basis, eating less ultra-processed foods, most people will get healthier. The point is that it's too imprecise to base public health policy on, and it distracts from the key mechanisms, which is the processing of carbs are bad, the processing of fats and proteins for the most part are not. Although we have— fats can be susceptible to chemical damage, not physical but chemical damage like trans fats or oxidation. But mainly that's the key point. In terms of supporting— what was the question?

How can people support you? I will link down the New England Journal and your Medium blog post, but I'm guessing it's—

yeah, I think what we really want to do is just have a dispassionate debate about the topic before we embark on a massive new public health campaign. We saw how the low-fat diet, based on similarly confounded and epi and weak trials, how badly that worked out. That caused a lot. We can't afford another public health blunder. So maybe I'm wrong and maybe the ultra focus on ultra-processing is the right way to proceed now? I don't think so. But we need to— we need to debate this. And I'm disappointed at how polarized, you know, people are. In fact, when I send, you know, my recent New England Journal of Medicine paper to some of the more ardent supporters of ultra-processed foods, they don't engage on the merits. They say, yes, I know I know how you feel. And but, but I'm making specific arguments. I'm making it in a pretty good journal. Other people— I'm not alone here. There are many food scientists who are raising these. Let's debate it. I mean, you know, I understand that there's a desire to just do something, but just doing something can cause more harm than good. And most importantly, it can take the focus off of an actionable and immediately actionable target. The processed carbs.

Target processed carbs. Talk about mechanism. Have high-level dialogue. Read the New England Journal piece if you can access it. Read Professor Ludwig's blog. We'll link it below. Thank you for joining me.

Great talking to you as usual.