Kevin Hall: What Should We Eat?

A leader for conducting rigorous randomized trials of humans along with animal models for understanding nutrition and metabolism, Dr. Kevin Hall is a Senior Investigator at the National Institutes of Health, and Section Chief of the Integrative Physiology Section, NIDDK. In this podcast, we reviewed his prolific body of research a recent publications. The timing of optimizing our diet and nutrition seems apropos, now that we’re in in the midst of the holiday season!

Below is a video snippet of our conversation on his ultra-processed food randomized trial.

Full videos of all Ground Truths podcasts can be seen on YouTube here. The current one is here. If you like the YouTube format, please subscribe! The audios are also available on Apple and Spotify.

Note: I’ll be doing a Ground Truths Live Chat on December 11th at 12 N EST, 9 AM PST, so please mark your calendar and join!

Transcript with links to publications and audio

Eric Topol (00:05):

Well, hello. This is Eric Topol with Ground Truths, and I'm really delighted to have with me today, Dr. Kevin Hall from the NIH. I think everybody knows that nutrition is so important and Kevin is a leader in doing rigorous randomized trials, which is not like what we usually see with large epidemiologic studies of nutrition that rely on food diaries and the memory of participants. So Kevin, it's really terrific to have you here.

Kevin Hall (00:34):

Thanks so much for the invitation.

Ultra-Processed Foods

Eric Topol (00:36):

Yeah. Well, you've been prolific and certainly one of the leaders in nutrition science who I look to. And what I thought we could do is go through some of your seminal papers. There are many, but I picked a few and I thought we'd first go back to the one that you published in Cell Metabolism. This is ultra-processed diets cause excessive caloric intake and weight gain. (Main results in graph below.) So maybe you can take us through the principle findings from that trial.

Kevin Hall (01:10):

Yeah, sure. So that was a really interesting study because it's the first randomized control trial that's investigated the role of ultra-processed foods in potentially causing obesity. So we've got, as you mentioned, lots and lots of epidemiological data that have made these associations between people who consume diets that are very high in ultra-processed foods as having greater risk for obesity. But those trials are not demonstrating causation. I mean, they suggest a strong link. And in fact, the idea of ultra-processed foods is kind of a new idea. It's really sort of appeared on the nutrition science stage probably most prominently in the past 10 years or so. And I first learned about this idea of ultra-processed foods, which is really kind of antithetical to the way most nutrition scientists think about foods. We often think about foods as nutrient delivery vehicles, and we kind of view foods as being the fraction of carbohydrates versus fats in them or how much sodium or fiber is in the foods.

Kevin Hall (02:17):

And along came this group in Brazil who introduced this new way of classifying foods that completely ignores the nutrient composition and says what we should be doing is classifying foods based on the extent and purpose of processing of foods. And so, they categorize these four different categories. And in the fourth category of this so-called NOVA classification scheme (see graphic below) , they identified something called ultra-processed foods. There's a long formal definition and it's evolved a little bit over the years and continues to evolve. But the basic ideas that these are foods that are manufactured by industries that contain a lot of purified ingredients made from relatively cheap agricultural commodity products that basically undergo a variety of processes and include additives and ingredients that are not typically found in home kitchens, but are typically exclusively in manufactured products to create the wide variety of mostly packaged goods that we see in our supermarkets.

Kevin Hall (03:22):

And so, I was really skeptical that there was much more about the effects of these foods. Other than that they typically have high amounts of sugar and saturated fat and salt, and they're pretty low in fiber. And so, the purpose of this study was to say, okay, well if there's something more about the foods themselves that is causing people to overconsume calories and gain weight and eventually get obesity, then we should do a study that's trying to test for two diets that are matched for these various nutrients of concern. So they should be matched for the macronutrients, they should be matched for the sugar content, the fat, the sodium, the fiber, and people should just be allowed to eat whatever they want and they shouldn't be trying to change their weight in any way. And so, the way that we did this was, as you mentioned, we can't just ask people to report what they're eating.

Kevin Hall (04:19):

So what we did was we admitted these folks to the NIH Clinical Center and to our metabolic ward, and it's a very artificial environment, but it's an environment that we can control very carefully. And so, what we basically did is take control over their food environment and we gave them three meals a day and snacks, and basically for a two-week period, they had access to meals that were more than 80% of calories coming from ultra-processed foods. And then in random order, they either received that diet first and give them simple instructions, eat as much as little as you want. We're going to measure lots of stuff. You shouldn't be trying to change your weight or weight that gave them a diet that had no calories from ultra-processed foods. In fact, 80% from minimally processed foods. But again, both of these two sort of food environments were matched for these nutrients that we typically think of as playing a major role in how many calories people choose to eat.

Kevin Hall (05:13):

And so, the basic idea was, okay, well let's measure what these folks eat. We gave them more than double the calories that they would require to maintain their weight, and what they didn't know was that in the basement of the clinical center where the metabolic kitchen is, we had all of our really talented nutrition staff measuring the leftovers to see what it was that they didn't eat. So we knew exactly what we provided to them and all the foods had to be in our nutrition database and when we compute what they actually ate by difference, so we have a very precise estimate about not only what foods they chose to ate, but also how many calories they chose to eat, as well as the nutrient composition.

And the main upshot of all that was that when these folks were exposed to this highly ultra-processed food environment, they spontaneously chose to eat about 500 calories per day more over the two-week period they were in that environment then when the same folks were in the environment that had no ultra-processed foods, but just minimally processed foods. They not surprisingly gained weight during the ultra-processed food environment and lost weight and lost body fat during the minimally processed food environment. And because those diets were overall matched for these different nutrients, it didn't seem to be that those were the things that were driving this big effect. So I think there's a couple of big take homes here. One is that the food environment really does have a profound effect on just the biology of how our food intake is controlled at least over relatively short periods of time, like the two-week periods that we were looking at. And secondly, that there's something about ultra-processed foods that seem to be driving this excess calorie intake that we now know has been linked with increased risk of obesity, and now we're starting to put some of the causal pieces together that really there might be something in this ultra-processed food environment that's driving the increased rates of obesity that we've seen over the past many decades.

Eric Topol (07:18):

Yeah, I mean I think the epidemiologic studies that make the link between ultra-processed foods and higher risk of cancer, cardiovascular disease, type 2 diabetes, neurodegenerative disease. They're pretty darn strong and they're backed up by this very rigorous study. Now you mentioned it short term, do you have any reason to think that adding 500 calories a day by eating these bad foods, which by the way in the American diet is about 60% or more of the average American diet, do you have any inkling that it would change after a few weeks?

Kevin Hall (07:54):

Well, I don't know about after a few weeks, but I think that one of the things that we do know about body weight regulation and how it changes in body weight impact both metabolism, how many calories were burning as well as our appetite. We would expect some degree of moderation of that effect eventually settling in at a new steady state, that's probably going to take months and years to achieve. And so the question is, I certainly don't believe that it would be a 500 calorie a day difference indefinitely. The question is when would that difference converge and how much weight would've been gained or lost when people eventually reached that new plateau? And so, that's I think a really interesting question. Some folks have suggested that maybe if you extrapolated the lines a little bit, you could predict when those two curves might eventually converge. That's an interesting thought experiment, but I think we do need some longer studies to investigate how persistent are these effects. Can that fully explain the rise in average body weight and obesity rates that have occurred over the past several decades? Those are open questions.

Eric Topol (09:03):

Yeah. Well, I mean, I had the chance to interview Chris van Tulleken who wrote the book, Ultra-Processed Peopleand I think you might remember in the book he talked about how he went on an ultra-processed diet and gained some 20, 30 pounds in a short time in a month. And his brother, his identical twin brother gained 50, 60 pounds, and so it doesn't look good. Do you look at all the labels and avoid all this junk and ultra-processed food now or are you still thinking that maybe it's not as bad as it looks?

Kevin Hall (09:38):

Well, I mean I think that I certainly learned a lot from our studies, and we are continuing to follow this up to try to figure out what are the mechanisms by which this happen. But at the same time, I don't think we can throw out everything else we know about nutrition science. So just because we match these various nutrients in this particular study, I think one of the dangers here is that as you mentioned, there's 60% of the food environment in the US and Great Britain and other places consist of these foods, and so they're unavoidable to some extent, right? Unless you're one of these privileged folks who have your backyard garden and your personal chef who can make all of your foods, I'm certainly not one of those people, but for the vast majority of us, we're going to have to incorporate some degree of ultra-processed foods in our day-to-day diet.

Kevin Hall (10:24):

The way I sort of view it is, we really need to understand the mechanisms and before we understand the mechanisms, we have to make good choices based on what we already know about nutrition science, that we should avoid the foods that have a lot of sugar in them. We should avoid foods that have a lot of saturated fat and sodium. We should try to choose products that contain lots of whole grains and legumes and fruits and vegetables and things like that. And there's some of those, even in the ultra-processed food category. I pretty regularly consume a microwavable ready meal for lunch. It tends to be pretty high in whole grains and legumes and low in saturated fat and sugar and things like that. But to engineer a food that can heat up properly in a microwave in four minutes has some ultra-processing technology involved there. I would be pretty skeptical that that's going to cause me to have really poor health consequences as compared to if I had the means to eat homemade French fries every day in tallow. But that's the kind of comparison that we have to think about.

Eric Topol (11:36):

But I think what you're touching on and maybe inadvertently is in that NOVA class four, the bad ultra-processed foods, there's a long, long list of course, and some of those may be worse than others, and we haven't seen an individual ranking of these constituents. So as you're alluding to what's in that microwave lunch probably could be much less concerning than what's in these packaged snacks that are eaten widely. But I would certainly agree that we don't know everything about this, but your study is one of the most quoted studies ever in the ultra-processed food world. Now, let me move on to another trial that was really important. This was published in Nature Medicine and it's about a plant-based diet, which is of course a very interesting diet, low-fat versus an animal-based ketogenic diet. Also looking at energy intake. Can you take us through that trial?

Plant-Based, Low Fat Diet vs Animal-Based, Low Carbohydrate Ketogenic Diet

Kevin Hall (12:33):

Sure. So it's actually interesting to consider that trial in the context of the trial we just talked about because both of these diets that we tested in this trial were relatively low in ultra-processed foods, and so both of them contained more than a kilogram of non-starchy vegetables as a base for designing these, again, two different food environments. Very similar overall study design where people again were exposed to either diets that were vegan plant-based diet that was really high in starches and was designed to kind of cause big insulin increases in the blood after eating the meals. And the other diet had very, very few carbohydrates of less than 10% in total, and we built on that kind of non-starchy vegetable base, a lot of animal-based products to kind of get a pretty high amount of fat and having very low carbohydrates. Both diets in this case, like I mentioned, were pretty low in ultra-processed foods, but what we were really interested in here was testing this idea that has come to prominence recently, that high carbohydrate diets that lead to really large glucose excursions after meals that cause very high insulin levels after meals are particularly obesogenic and should cause you to be hungrier than compared to a diet that doesn't lead to those large swings in glucose and insulin and the prototypical case being one that's very low in carbohydrate and might increase the level of ketones that are floating around in your blood, which are hypothesized to be an appetite suppressant. Same sort of design, these minimally processed diets that one was very high in carbs and causes large swings in insulin and the other that's very low in carbs and causes increases in ketones.

Kevin Hall (14:22):

We ask people, again, while you're in one food environment or the other, don't be trying to gain weight or lose weight, eat as much or as little as you'd like, and we're going to basically measure a lot of things. They again, don't know what the primary outcome of the study is. We're measuring their leftovers afterwards. And so, the surprise in this particular case was that the diet that caused the big swings in glucose and insulin did not lead to more calorie consumption. In fact, it led to about 700 calories per day less than when the same people were exposed to the ketogenic diet. Interestingly, both food environments caused people to lose weight, so it wasn't that we didn't see the effect of people over consuming calories on either diet, so they were reading fewer calories in general than they were when they came in, right. They're probably eating a pretty ultra-processed food diet when they came in. We put them on these two diets that varied very much in terms of the macronutrients that they were eating, but both were pretty minimally processed. They lost weight. They ended up losing more body fat on the very low-fat high carb diet than the ketogenic diet, but actually more weight on the ketogenic diet than the low-fat diet. So there's a little bit of a dissociation between body fat loss and weight loss in this study, which was kind of interesting.

Eric Topol (15:49):

Interesting. Yeah, I thought that was a fascinating trial because plant-based diet, they both have their kind of camps, you know.

Kevin Hall (15:57):

Right. No, exactly.

Immune System Signatures for Vegan vs Ketogenic Diets

Eric Topol (15:58):

There are people who aren't giving up on ketogenic diet. Of course, there's some risks and some benefits and there's a lot of interest of course with the plant-based diet. So it was really interesting and potentially the additive effects of plant-based with avoidance or lowering of ultra-processed food. Now, the more recent trial that you did also was very interesting, and of course I'm only selecting ones that I think are particularly, there are a lot of trials you've done, but this one is more recent in this year where you looked at vegan versus ketogenic diets for the immune signature, immune response, which is really important. It's underplayed as its effect, and so maybe you can take us through that one.

[Link to a recent Nature feature on this topic, citing Dr. Hall’s work]

Kevin Hall (16:43):

Yeah, so just to be clear, it's actually the same study, the one that we just talked about. This is a secondary sort of analysis from a collaboration we had with some folks at NIAID here at the NIH to try to evaluate immune systems signatures in these same folks who wonder what these two changes in their food environment. One is vegan, high carbohydrate low-fat diet and the other, the animal-based ketogenic diet. And again, it was pretty interesting to me that we were able to see really substantial changes in how the immune system was responding. First of all, both diets again seem to have improved immune function, both adaptive and innate immune function as compared to their baseline measurements when they came into the study. So when they're reading their habitual diet, whatever that is typically high in ultra-processed foods, they switched to both of these diets.

Kevin Hall (17:39):

We saw market changes in their immune system even compared to baseline. But when we then went and compared the two diets, they were actually divergent also, in other words, the vegan diet seemed to stimulate the innate immune system and the ketogenic diet seemed to stimulate the adaptive immune system. So these are the innate immune system can be thought of. Again, I'm not an immunologist. My understanding is that this is the first line defense against pathogens. It happens very quickly and then obviously the adaptive immune system then adapts to a specific pathogen over time. And so, this ability of our diet to change the immune system is intriguing and how much of that has to do with influencing the gut microbiota, which obviously the gut plays a huge role in steering our immune system in one direction versus another. I think those are some really intriguing mechanistic questions that are really good fodder for future research.

Eric Topol (18:42):

Yeah, I think it may have implications for treatment of autoimmune diseases. You may want to comment about that.

Kevin Hall (18:51):

Yeah, it's fascinating to think about that the idea that you could change your diet and manipulate your microbiota and manipulate your gut function in a way to influence your immune system to steer you away from a response that may actually be causing your body damage in your typical diet. It's a fascinating area of science and we're really interested to follow that up. I mean, it kind of supports these more anecdotal reports of people with lupus, for example, who've reported that when they try to clean up their diet for a period of time and eliminate certain foods and eliminate perhaps even ultra-processed food products, that they feel so much better that their symptoms alleviate at least for some period of time. Obviously, it doesn't take the place of the therapeutics that they need to take, but yeah, we're really interested in following this up to see what this interaction might be.

Eric Topol (19:46):

Yeah, it's fascinating. It also gets to the fact that certain people have interesting responses. For example, those with epilepsy can respond very well to a ketogenic diet. There's also been diet proposed for cancer. In fact, I think there's some even ongoing trials for cancer of specific diets. Any comments about that?

Kevin Hall (20:10):

Yeah, again, it's a really fascinating area. I mean, I think we kind of underappreciate and view diet in this lens of weight loss, which is not surprising because that's kind of where it's been popularized. But I think the role of nutrition and how you can manipulate your diet and still you can have a very healthy version of a ketogenic diet. You can have a very healthy version of a low-fat, high carb diet and how they can be used in individual cases to kind of manipulate factors that might be of concern. So for example, if you're concerned about blood glucose levels, clearly a ketogenic diet is moderating those glucose levels over time, reducing insulin levels, and that might have some positive downstream consequences and there's some potential downsides. Your apoB levels might go up. So, you have to kind of tune these things to the problems and the situations that individuals may face. And similarly, if you have issues with blood glucose control, maybe a high carbohydrate diet might not be for you, but if that's not an issue and you want to reduce apoB levels, it seems like that is a relatively effective way to do that, although it does tend to increase fasting triglyceride levels.

Kevin Hall (21:27):

So again, there's all of these things to consider, and then when you open the door beyond traditional metabolic health markers to things like inflammation and autoimmune disease as well as some of these other things like moderating how cancer therapeutics might work inside the body. I think it's a really fascinating and interesting area to pursue.

Eric Topol (21:55):

No question about it. And that also brings in the dimension of the gut microbiome, which obviously your diet has a big influence, and it has an influence on your brain, brain-gut axis, and the immune system. It's all very intricate, a lot of feedback loops and interactions that are not so easy to dissect, right?

Kevin Hall (22:16):

Absolutely. Yeah, especially in humans. That's why we rely on our basic science colleagues to kind of figure out these individual steps in these chains. And of course, we do need human experiments and carefully controlled experiments to see how much of that really translates to humans, so we need this close sort of translational partnership.

On the Pathogenesis of Obesity, Calories In and Calories Out

Eric Topol (22:35):

Yeah. Now, you've also written with colleagues, other experts in the field about understanding the mechanisms of pathogenesis of obesity and papers that we'll link to. We're going to link to everything for what we've been discussing about calories in, calories out, and that's been the longstanding adage about this. Can you enlighten us, what is really driving obesity and calories story?

Kevin Hall (23:05):

Well, I co-organized a meeting for the Royal Society, I guess about a year and a half ago, and we got together all these experts from around the world, and the basic message is that we have lots of competing theories about what is driving obesity. There's a few things that we all agree on. One is that there is a genetic component. That adiposity in a given environment is somewhere between 40% to 70% heritable, so our genes play a huge role. It seems like there's certain genes that can play a major role. Like if you have a mutation in leptin, for example, or the leptin receptor, then this can have a monogenic cause of obesity, but that's very, very rare. What seems to be the case is that it's a highly polygenic disease with individual gene variants contributing a very, very small amount to increased adiposity. But our genes have not changed that much as obesity prevalence has increased over the past 50 years. And so, something in the environment has been driving that, and that's where the real debates sort of starts, right?

Kevin Hall (24:14):

I happen to be in the camp that thinks that the food environment is probably one of the major drivers and our food have changed substantially, and we're trying to better understand, for example, how ultra-processed foods which have risen kind of in parallel with the increased prevalence of obesity. What is it about ultra-processed foods that tend to drive us to overconsume calories? Other folks focus maybe more on what signals from the body have been altered by the foods that we're eating. They might say that the adipose tissue because of excess insulin secretion for example, is basically driven into a storage mode and that sends downstream signals that are eventually sensed by the brain to change our appetite and things like that. There's a lot of debate about that, but again, I think that these are complementary hypotheses that are important to sort out for sure and important to design experiments to try to figure out what is more likely. But there is a lot of agreement on the idea that there's something in our environment has changed.

Kevin Hall (25:17):

I think there's even maybe a little bit less agreement of exactly what that is. I think that there's probably a little bit more emphasis on the food environment as opposed to there are other folks who think increased pollution might be driving some of this, especially endocrine disrupting chemicals that have increased in prevalence. I think that's a viable hypothesis. I think we have to try to rank order what we think are the most likely and largest contributors. They could all be contributing to some extent and maybe more so in some people rather than others, but our goal is to try to, maybe that's a little simple minded, but let's take the what I think is the most important thing and let's figure out the mechanisms of that most important thing and we'll, number one, determine if it is the most important thing. In my case, I think something about ultra-processed foods that are driving much of what we're seeing. If we could better understand that, then we could both advise consumers to avoid certain kinds of foods because of certain mechanisms and still be able to consume some degree of ultra-processed foods. They are convenient and tasty and relatively inexpensive and don't require a lot of skill and equipment to prepare. But then if we focus on the true bad guys in that category because we really understand the mechanisms, then I think that would be a major step forward. But that's just my hypothesis.

Eric Topol (26:43):

Well, I’m with you actually. Everything I've read, everything I've reviewed on ultra-processed food is highly incriminating, and I also get frustrated that nothing is getting done about it, at least in this country. But on the other hand, it doesn't have to be either or, right? It could be both these, the glycemic index story also playing a role. Now, when you think about this and you're trying to sort out calories in and calories out, and let's say it's one of your classic experiments where you have isocaloric proteins and fat and carbohydrate exactly nailed in the different diets you're examining. Is it really about calories or is it really about what is comprising the calorie?

Kevin Hall (27:29):

Yeah, so I think this is the amazing thing, even in our ultra-processed food study, if we asked the question across those people, did the people who ate more calories even in the ultra-processed diet, did they gain more weight? The answer is yes.

Kevin Hall (27:44):

There's a very strong linear correlation between calorie intake and weight change. I tend to think that I started my career in this space focusing more on the metabolism side of the equation, how the body's using the calories and how much does energy expenditure change when you vary the proportion of carbs versus fat, for example. The effect size is there, they might be there, but they're really tiny of the order of a hundred calories per day. What really struck me is that when we just kind of changed people's food environments, the magnitude of the effects are like we mentioned, 500 to 700 calories per day differences. So I think that the real trick is to figure out how is it that the brain is regulating our body weight in some way that we are beginning to understand from a molecular perspective? What I think is less well understood is, how is that food intake control system altered by the food environment that we find ourselves in?

The Brain and GLP-1 Drugs

Kevin Hall (28:42):

There are a few studies now in mice that are beginning to look at how pathways in the brain that have been believed to be related to reward and not necessarily homeostatic control of food intake. They talk to the regions of the brain that are related to homeostatic control of food intake, and it's a reciprocal sort of feedback loop there, and we're beginning to understand that. And I think if we get more details about what it is in our foods that are modulating that system, then we'll have a better understanding of what's really driving obesity and is it different in different people? Are there subcategories of obesity where certain aspects of the food environment are more important than others, and that might be completely flipped in another person. I don't know the answer to that question yet, but it seems like there are certain common factors that might be driving overall changes in obesity prevalence and how they impact this reward versus homeostatic control systems in the brain, I think are really fascinating questions.

Eric Topol (29:43):

And I think we're getting much more insight about this circuit of the reward in the brain with the food intake, things like optogenetics, many ways that we're getting at this. And so, it's fascinating. Now, that gets me to the miracle drug class GLP-1, which obviously has a big interaction with obesity, but of course much more than that. And you've written about this as well regarding this topic of sarcopenic obesity whereby you lose a lot of weight, but do you lose muscle mass or as you referred to earlier, you lose body fat and maybe not so much muscle mass. Can you comment about your views about the GLP-1 family of drugs and also about this concern of muscle mass loss?

Kevin Hall (30:34):

Yeah, so I think it's a really fascinating question, and we've been trying to develop mathematical models about how our body composition changes with weight gain and weight loss for decades now. And this has been a long topic, one of the things that many people may not realize is that people with obesity don't just have elevated adiposity, they also have elevated muscle mass and lean tissue mass overall. So when folks with obesity lose weight, and this was initially a pretty big concern with bariatric surgery, which has been the grandfather of ways that people have lost a lot of weight. The question has been is there a real concern about people losing too much weight and thereby becoming what you call sarcopenic? They have too little muscle mass and then they have difficulties moving around. And of course, there are probably some people like that, but I think what people need to realize is that folks with obesity tend to start with much higher amounts of lean tissue mass as well as adiposity, and they start off with about 50% of your fat-free mass, and the non-fat component of your body is skeletal muscle.

Kevin Hall (31:45):

So you're already starting off with quite a lot. And so, the question then is when you lose a lot of weight with the GLP-1 receptor agonist or with bariatric surgery, how much of that weight loss is coming from fat-free mass and skeletal muscle versus fat mass? And so, we've been trying to simulate that using what we've known about bariatric surgery and what we've known about just intentional weight loss or weight gain over the years. And one of the things that we found was that our sort of expectations for what's expected for the loss of fat-free mass with these different drugs as well as bariatric surgery, for the most part, they match our expectations. In other words, the expected amount of fat loss and fat free mass loss. The one outlier interestingly, was the semaglutide study, and in that case, they lost more fat-free mass than would be expected.

Kevin Hall (32:44):

Now, again, that's just raising a little bit of a flag that for whatever reason, from a body composition perspective, it's about a hundred people underwent these repeated DEXA scans in that study sponsored by Novo Nordisk. So it's not a huge number of people, but it's enough to really get a good estimate about the proportion of weight loss. Whether or not that has functional consequences, I think is the open question. There's not a lot of reports of people losing weight with semaglutide saying, you know what? I'm really having trouble actually physically moving around. I feel like I've lost a lot of strength. In fact, it seems to be the opposite, right, that the quality of the muscle there seems to be improved. They seem to have more physical mobility because they've lost so much more weight, that weight had been inhibiting their physical movement in the past.

Kevin Hall (33:38):

So it's something to keep an eye on. It's an open question whether or not we need additional therapies in certain categories of patients, whether that be pharmacological, there are drugs that are interesting that tend to increase muscle mass. There's also other things that we know increase muscle mass, right? Resistance exercise training, increase this muscle mass. And so, if you're really concerned about this, I certainly, I'm not a physician, but I think it's something to consider that if you go on one of these drugs, you might want to think about increasing your resistance exercise training, maybe increasing the protein content of your diet, which then can support that muscle building. But I think it's a really interesting open question about what the consequences of this might be in certain patient populations, especially over longer periods of time.

Dietary Protein, Resistance Exercise, DEXA Scans

Eric Topol (34:30):

Yeah, you've just emphasized some really key points here. Firstly, that resistance exercise is good for you anyway. And get on one of these drugs, why don't you amp it up or get it going? The second is about the protein diet, which it'd be interesting to get your thoughts on that, but we generally have too low of a protein diet, but then there are some who are advocating very high protein diets like one gram per pound, not just one gram per kilogram. And there have been studies to suggest that that very high protein diet could be harmful, but amping up the protein diet, that would be a countering thing. But the other thing you mentioned is a DEXA scan, which can be obtained very inexpensively, and because there's a variability in this muscle mass loss if it's occurring, I wonder if that's a prudent thing or if you just empirically would just do the things that you mentioned. Do you have any thoughts about that?

Kevin Hall (35:32):

Yeah, that's really a clinical question that I don't deal with on a day-to-day basis. And yeah, I think there's probably better people suited to that. DEXA scans, they're relatively inexpensive, but they're not readily accessible to everyone. I certainly wouldn't want to scare people away from using drugs that are now known to be very effective for weight loss and pretty darn safe as far as we can tell, just because they don't have access to a DEXA scanner or something like that.

Eric Topol (36:00):

Sure. No, that makes a lot of sense. I mean, the only reason I thought it might be useful is if you're concerned about this and you want to track, for example, how much is that resistant training doing?

Kevin Hall (36:13):

But I think for people who have the means to do that, sure. I can't see any harm in it for sure.

Continuous Glucose Sensors?

Eric Topol (36:19):

Yeah. That gets me to another metric that you've written about, which is continuous glucose tracking. As you know, this is getting used, I think much more routinely in type one insulin diabetics and people with type 2 that are taking insulin or difficult to manage. And now in recent months there have been consumer approved that is no prescription needed, just go to the drugstore and pick up your continuous glucose sensor. And you've written about that as well. Can you summarize your thoughts on it?

Kevin Hall (36:57):

Yeah, sure. I mean, yeah, first of all, these tools have been amazing for people with diabetes and who obviously are diagnosed as having a relative inability to regulate their glucose levels. And so, these are critical tools for people in that population. I think the question is are they useful for people who don't have diabetes and is having this one metric and where you target all this energy into this one thing that you can now measure, is that really a viable way to kind of modulate your lifestyle and your diet? And how reliable are these CGM measurements anyway? In other words, do they give the same response to the same meal on repeated occasions? Does one monitor give the same response as another monitor? And those are the kinds of experiments that we've done. Again, secondary analysis, these trials that we talked about before, we have people wearing continuous glucose monitors all the time and we know exactly what they ate.

Kevin Hall (37:59):

And so, in a previous publication several years ago, we basically had two different monitors. One basically is on the arm, which is the manufacturer's recommendation, the other is on the abdomen, which is the manufacturer's recommendation. They're wearing them simultaneously. And we decided just to compare what were the responses to the same meals in simultaneous measurements. And they were correlated with each other thankfully, but they weren't as well predictive as you might expect. In other words, one device might give a very high glucose reading to consuming one meal and the other might barely budge, whereas the reverse might happen for a different meal. And so, we asked the question, if we were to rank the glucose spikes by one meal, so we have all these meals, let's rank them according to the glucose spikes of one device. Let's do the simultaneous measurements with the other device.

Kevin Hall (38:53):

Do we get a different set of rankings? And again, they're related to each other, but they're not overlapping. They're somewhat discordant. And so, then the question becomes, okay, well if I was basically using this one metric to kind of make my food decisions by one device, I actually start making different decisions compared to if I happen to have been wearing a different device. So what does this really mean? And I think this sort of foundational research on how much of a difference you would need to make a meaningful assessment about, yeah, this is actionable from a lifestyle perspective, even if that is the one metric that you're interested in. That sort of foundational research I don't think has really been done yet. More recently, we asked the question, okay, let's ignore the two different devices. Let's stick to the one where we put it on our arm, and let's ask the question.

Kevin Hall (39:43):

We've got repeated meals and we've got them in this very highly regimented and controlled environment, so we know exactly what people ate previously. We know the timing of the meals, we know when they did their exercise, we know how much they were moving around, how well they slept the night before. All of these factors we could kind of control. And the question that we asked in that study was, do people respond similarly to the same meal on repeated occasions? Is that better than when you actually give them very different meals? But they match overall for macronutrient content, for example. And the answer to that was surprisingly no. We had as much variability in the glucose response to the same person consuming the same meal on two occasions as a whole bunch of different meals. Which suggests again, that there's enough variability that it makes it difficult to then recommend on for just two repeats of a meal that this is going to be a meal that's going to cause your blood glucose to be moderate or blood glucose to be very high. You're going to have to potentially do this on many, many different occasions to kind of figure out what's the reliable response of these measurements. And again, that foundational research is typically not done. And I think if we're really going to use this metric as something that is going to change our lifestyles and make us choose some meals other than others, then I think we need that foundational research. And all we know now is that two repeats of the same meal is not going to do it.

Eric Topol (41:21):

Well, were you using the current biosensors of 2024 or were you using ones from years ago on that?

Kevin Hall (41:27):

No, we were using ones from several years ago when these studies were completed. But interestingly, the variability in the venous measurements to meal tests is also very, very different. So it's probably not the devices per se that are highly variable. It's that we don't really know on average how to predict these glucose responses unless there's huge differences in the glycemic load. So glycemic load is a very old concept that when you have very big differences in glycemic load, yeah, you can on average predict that one kind of meal is going to give rise to a much larger glucose excursion than another. But typically these kind of comparisons are now being made within a particular person. And we're comparing meals that might have quite similar glycemic loads with the claim that there's something specific about that person that causes them to have a much bigger glucose spike than another person. And that we can assess that with a couple different meals.

Eric Topol (42:31):

But also, we know that the spikes or the glucose regulation, it's very much affected by so many things like stress, like sleep, like exercise. And so, it wouldn't be at all surprising that if you had the exact same food, but all these other factors were modulated that it might not have the same response. But the other thing, just to get your comment on. Multiple groups, particularly starting in Israel, the Weizmann Institute, Eran Segal and his colleagues, and many subsequent have shown that if you give the exact same amount of that food, the exact same time to a person, they eat the exact same amount. Their glucose response is highly heterogeneous and variable between people. Do you think that that's true? That in fact that our metabolism varies considerably and that the glucose in some will spike with certain food and some won’t.

Kevin Hall (43:29):

Well, of course that's been known for a long time that there's varying degrees of glucose tolerance. Just oral glucose tolerance tests that we've been doing for decades and decades we know is actually diagnostic, that we use variability in that response as diagnostic of type 2 diabetes.

Eric Topol (43:49):

I'm talking about within healthy people.

Kevin Hall (43:53):

But again, it's not too surprising that varying people. I mean, first of all, we have a huge increase in pre-diabetes, right? So there's various degrees of glucose tolerance that are being observed. But yeah, that is important physiology. I think the question then is within a given person, what kind of advice do we give to somebody about their lifestyle that is going to modulate those glucose responses? And if that's the only thing that you look at, then it seems like what ends up happening, even in the trials that use continuous glucose monitors, well big surprise, they end up recommending low carbohydrate diets, right? So that's the precision sort of nutrition advice because if that's the main metric that's being used, then of course we've all known for a very long time that lower carbohydrate diets lead to a moderated glucose response compared to higher carbohydrate diets. I think the real question is when you kind of ask the issue of if you normalize for glycemic load of these different diets, and there are some people that respond very differently to the same glycemic load meal compared to another person, is that consistent number one within that person?

Kevin Hall (45:05):

And our data suggests that you're going to have to repeat that same test multiple times to kind of get a consistent response and be able to make a sensible recommendation about that person should eat that meal in the future or not eat that meal in the future. And then second, what are you missing when that becomes your only metric, right? If you're very narrowly focused on that, then you're going to drive everybody to consume a very low carbohydrate diet. And as we know, that might be great for a huge number of people, but there are those that actually have some deleterious effects of that kind of diet. And if you're not measuring those other things or not considering those other things and put so much emphasis on the glucose side of the equation, I worry that there could be people that are being negatively impacted. Not to mention what if that one occasion, they ate their favorite food and they happen to get this huge glucose spike and they never eat it again, their life is worse. It might've been a complete aberration.

Eric Topol (46:05):

I think your practical impact point, it's excellent. And I think one of the, I don't know if you agree, Kevin, but one of the missing links here is we see these glucose spikes in healthy people, not just pre-diabetic, but people with no evidence of glucose dysregulation. And we don't know, they could be up to 180, 200, they could be prolonged. We don't know if the health significance of that, and I guess someday we'll learn about it. Right?

Kevin Hall (46:36):

Well, I mean that's the one nice thing is that now that we have these devices to measure these things, we can start to make these correlations. We can start to do real science to say, what a lot of people now presume is the case that these spikes can't be good for you. They must lead to increased risk of diabetes. It's certainly a plausible hypothesis, but that's what it is. We actually need good data to actually analyze that. And at least that's now on the table.

Eric Topol (47:04):

I think you're absolutely right on that. Well, Kevin, this has been a fun discussion. You've been just a great leader in nutrition science. I hope you'll keep up your momentum because it's pretty profound and I think we touched on a lot of the uncertainties. Is there anything that I didn't ask you that you wish I did?

Kevin Hall (47:23):

I mean, we could go on for hours, I'm sure, Eric, but this has been a fascinating conversation. I really appreciate your interest. Thank you.

Eric Topol (47:30):

Alright, well keep up the great stuff. We'll be following all your work in the years ahead, and thanks for joining us on Ground Truths today.

**************************************

Footnote, Stay Tuned: Julia Belluz and Kevin Hall have a book coming out next September titled “WHY WE EAT?

Thank you for reading, listening and subscribing to Ground Truths.

If you found this fun and informative please share it!

All content on Ground Truths—its newsletters, analyses, and podcasts, are free, open-access.

Paid subscriptions are voluntary. All proceeds from them go to support Scripps Research. Many thanks to those who have contributed—they have greatly helped fund our summer internship programs for the past two years. I welcome all comments from paid subscribers and will do my best to respond to them and any questions.

Thanks to my producer Jessica Nguyen and to Sinjun Balabanoff for audio and video support at Scripps Research.

Note on Mass Exodus from X/twitter:

Many of you have abandoned the X platform for reasons that I fully understand. While I intend to continue to post there because of its reach to the biomedical community, I will post anything material here in the Notes section of Ground Truths on a daily basis and cover important topics in the newsletter/analyses. You can also find my posts at Bluesky: @erictopol.bsky.social, which is emerging as an outstanding platform for sharing life science.

Get full access to Ground Truths at erictopol.substack.com/subscribe