The Peter Attia Drive - #368 ‒ The protein debate： optimal intake, limitations of the RDA, whether high-protein intake is harmful, and how to think about processed foods ｜ David Allison, Ph.D.

00:00:00.000 Hey, everyone. Welcome to the drive podcast. I'm your host, Peter Atiyah. This podcast,

00:00:16.580 my website, and my weekly newsletter all focus on the goal of translating the science of longevity

00:00:21.580 into something accessible for everyone. Our goal is to provide the best content in health and

00:00:26.780 wellness, and we've established a great team of analysts to make this happen. It is extremely

00:00:31.720 important to me to provide all of this content without relying on paid ads to do this. Our work

00:00:37.000 is made entirely possible by our members. And in return, we offer exclusive member only content

00:00:42.760 and benefits above and beyond what is available for free. If you want to take your knowledge of

00:00:47.980 this space to the next level, it's our goal to ensure members get back much more than the price

00:00:53.260 of the subscription. If you want to learn more about the benefits of our premium membership,

00:00:58.080 head over to peteratiyahmd.com forward slash subscribe.

00:01:04.240 My guest this week is David Allison. David, returning for his third conversation on the drive,

00:01:10.020 is a world-renowned scientist, an award-winning scientific writer who has been at the forefront

00:01:15.380 of obesity research for the last 20 years, and is currently the director of the Children's

00:01:20.480 Nutrition Research Center at Baylor College of Medicine. I wanted to have David on because

00:01:26.220 protein has become one of the most contentious and confusing topics in nutrition today. What was

00:01:31.180 once a fairly straightforward subject has now turned into a debate full of conflicting claims,

00:01:36.680 dogma, unnecessary controversy, and a whole lot of name-calling. David brings both a deep understanding

00:01:43.800 of the science and a clear-eyed perspective on how to separate evidence from opinion. This is part one

00:01:49.900 of a two-part deep dive on protein, and next week I'll be joined by Rhonda Patrick for part two,

00:01:55.560 after which we'll put this protein discussion to rest once and for all. In this episode, we discuss

00:02:01.400 the historical cycle of demonizing macronutrients and why protein has recently become the focus,

00:02:07.640 the origins and limitations for the RDA for protein, and what the evidence suggests about optimal intake

00:02:15.060 for health, longevity, and performance. Conflicts of interest in nutrition science and why transparency

00:02:20.180 around data, methods, and logic matter much more than funding sources. The challenges of conducting

00:02:26.280 high-quality nutrition studies, including the debate over crossover designs, the limits of

00:02:31.640 epidemiology, and the underfunding of rigorous trials compared to pharmaceutical trials. What the

00:02:37.180 evidence really says about higher protein intake, muscle protein synthesis, and whether concerns about harm

00:02:43.400 are supported by actual data. How to think about processed and ultra-processed foods, including

00:02:49.260 definitions, heuristics, and the question of whether they're inherently harmful or simply a convenient

00:02:55.080 villain. And finally, the difficulty of tackling obesity through public health, the limits of current

00:03:01.880 approaches, and whether future solutions may rely more on drugs like GLP-1 agonists or broader societal

00:03:09.060 changes. So without further delay, please enjoy my conversation with David Allison.

00:03:20.160 Hey David, thanks for coming back. This was probably the shortest trip you've made here, right? You've

00:03:24.240 got a new home? Yeah, I got a great new gig at the Children's Nutrition Research Center in Houston,

00:03:29.500 Texas, and Baylor College of Medicine in Texas Children's Hospital. I'm having a great time.

00:03:34.000 So it was just a nice easy car ride over here. Very well. Well, we're going to actually start by

00:03:39.060 talking about something that, believe it or not, I don't want to talk about because I'm kind of sick

00:03:44.220 and tired of talking about it. And I'm going to apologize in advance to all the listeners, because

00:03:48.940 if they've been listening at all, they're probably sick and tired of hearing about this. But unfortunately,

00:03:54.840 this is a topic that has gone from being what I would consider pretty straightforward to somewhat

00:04:03.260 contentious. And I'm going to do my best to refrain from speculating on the reasons why it's become

00:04:09.480 contentious. Although I have many views on that and many views on the people who choose to make it

00:04:14.300 contentious, which I'll also refrain from. But let's just try to dive into the arguments around

00:04:19.620 the macronutrient that is more in the crosshairs than any other today, which is protein. And that's

00:04:24.680 kind of interesting when you consider the arc of your career. It was certainly easy to understand how

00:04:31.220 people demonized fat, and then they demonized carbs. And here we are today, come in full circle,

00:04:37.760 we're demonizing protein. Where do you place that in the arc of the historical lens of nutrition?

00:04:43.480 It shows many things, but at some level, it's almost perfectly predictable, which is we all eat.

00:04:50.460 We eat every day. Eating is part of our sustenance, but it's part of culture, family,

00:04:54.900 certainly part of economics, identity, social class, religion, and so on. And so it's fun to

00:05:02.600 talk about. And there's lots of motivations, motivations people recognize and motivations

00:05:08.080 they may not recognize. And that leads always to this attention on it. That attention drives

00:05:13.820 a big economic engine of food sales. So there's lots of interest in this and lots of stakes in this,

00:05:19.860 if you'll pardon the pun. So people shift and it's not even just the macronutrients.

00:05:24.780 This week it's seed oils and next week it's phytoestrogens and soybean feminizing youth. And

00:05:29.980 it's one thing after another where people look for the villains and the heroes and the angels and the

00:05:34.120 demons in food. Only three macronutrients, so they keep looping around. Protein has become in the last

00:05:41.500 few years almost a fever pitch of enthusiasm and excitement from one part of the community and

00:05:50.160 it's driving sales and it's driving behavior and some people like me are having fun with it.

00:05:55.680 And then I think there's always that group that sees other people having fun or making money and

00:06:01.600 heaven forbid doing something that might be seen as the easy way out or the contrived or constructed

00:06:08.340 way out as opposed to the so-called natural way out of doing something. And then that upsets them,

00:06:16.160 that offends them. You're not being prudent. You're not taking the natural course. You're not

00:06:21.000 taking the old-fashioned course. You're having too much fun. You're trying too hard to achieve big

00:06:26.660 things. We don't like that. So we're going to try to poo-poo it or shut it down or minimize it.

00:06:32.260 And I think that's where things are.

00:06:33.580 Yeah. So presumably at some point this will no longer be relevant and no one will talk about

00:06:38.640 it and we'll move back to fats being the bad thing. Although I guess we're kind of there with 1.00

00:06:42.080 seed oils to your point. But let's talk a little bit about protein. So a lot of the consternation

00:06:47.780 stems, I think, from a debate around the RDA and the recommendation of 0.8 grams per kilogram of

00:06:55.680 body weight. Do you want to tell folks a little bit about where that came from? Where does this

00:07:01.620 ubiquitous recommendation that we consume 0.8 grams of protein per kilogram of body weight? So again,

00:07:08.120 just to put that in perspective, I weigh 180 pounds. So that is probably 82 kilos. So I should be eating

00:07:17.100 about 60 to 65 grams of protein according to the RDA. It's the middle of the day. I've had 60 grams of

00:07:24.840 protein today. So I could basically stop eating protein for the rest of the day now, right?

00:07:28.420 Yeah. Some people would say that. Now others, I don't know if I would go so far as to say at the

00:07:33.700 other extreme, but a more active proponent of the importance of not only greater amounts of protein,

00:07:41.100 but particular types and distributions and so on would be somebody like Don Lehman.

00:07:46.520 And he might say, I hope I'm not inappropriately putting words in his mouth, but I think he might say,

00:07:50.840 Peter, you actually need to be eating protein at least three or four times a day.

00:07:54.760 Yeah. Because that 60 was in one sitting.

00:07:57.240 Right. And at each of those sittings, you probably want to hit about 30. I emphasize that word about,

00:08:04.260 it's not like there's some mathematical proof that there's some hard threshold and certainly not for

00:08:09.620 you who is 20% bigger than me, that your hard threshold would be the same as my hard threshold,

00:08:15.040 but in the neighborhood of 30. Right off the bat, if you follow that advice, you'd need to be having

00:08:20.200 double that amount of protein and you'd have to have it distributed differently. And if you haven't

00:08:24.620 distributed the way you currently did, it'd be more than double. So I think there are a lot of

00:08:28.540 people who disagree with that. The history is people at some point recognized that we needed protein to

00:08:34.480 live. The key indicator of that was nitrogen. And people looked at nitrogen balance. How much did

00:08:40.080 you take in? How much did you excrete? And they found that people could achieve nitrogen balance,

00:08:45.380 or at least sort of ordinary normal people of the time at about that level. No one ever proved,

00:08:53.240 demonstrated, or I think even claimed that that was the best amount or the upper limit,

00:09:00.660 or it was just, that's probably enough to maintain nitrogen balance, which probably means it's

00:09:06.180 compatible with survival. And if you think about the early days, that was pretty important. Europe

00:09:12.260 couldn't reliably feed its population until two things were entered into Europe. One was guano,

00:09:18.920 bat poop, fertilizer, and the other was potatoes. And when after the Columbus hit Hispaniola,

00:09:25.900 and eventually what came back was, among other things, guano and potatoes, suddenly Europe could

00:09:31.840 feed its population. But it took a while for the potatoes to catch on. Louis XV actually wore a potato

00:09:37.740 boutonniere to get people to think that they were kind of safe. Two Polish scientists in 1928 published

00:09:44.220 a paper in which they had two young people. They never said they were the young people, but I kind

00:09:50.100 of suspected perhaps they were. One man, one woman. And for six months, they fed them nothing but

00:09:55.620 potatoes, a little bit of fat to cook the potatoes in, and a little bit of fruit to avoid deficiencies.

00:10:01.180 So the only source of protein for practical purposes was the potato. And what they showed

00:10:06.920 was nitrogen balance was perfectly fine. And despite sort of this quote unquote bad carbohydrate,

00:10:12.860 at least some people think it's a bad carbohydrate, no one got diabetes. They didn't gain weight.

00:10:17.320 How many calories did they eat a day?

00:10:19.360 Don't remember that, but they were sort of roughly normal weight thin people of the time. So probably,

00:10:25.680 I would guess a little more than 2000.

00:10:27.020 8,000, 2,500.

00:10:28.140 Somewhere in that neighborhood. As I said, they did nitrogen balance studies and they were perfectly

00:10:31.860 fine. But that's entirely different than saying, what if they were older people? Or what if they

00:10:39.640 were pregnant? Or what if they were recovering from a bicep tendon tear? Or what if they were

00:10:45.840 bodybuilding? Those are all very different things.

00:10:47.760 Or what if they were active?

00:10:49.460 Yeah.

00:10:49.700 So when we go back and look at some of the USDA-based studies on this topic, and we wrote

00:10:56.320 about this. So the subjects for this study were, if my memory serves correctly, lean, inactive,

00:11:04.860 sedentary young men.

00:11:06.400 Who were, I think you said, about 150 pounds, if I recall.

00:11:09.940 Yep. Yeah. Maybe even a bit lighter, but yeah, about that. So these are guys that weighed 65 to 70

00:11:16.820 kilos, very inactive, and nitrogen balance was demonstrated in them, that you could achieve

00:11:23.780 it at 0.8 grams of protein per kilogram. So again, let's just do the math, make it easy.

00:11:29.340 Say 50 grams of protein was able to keep these folks in nitrogen balance. 1.00

00:11:34.260 It's interesting that everything you said, this is true of, I think, clinical research.

00:11:38.840 You have to look at the population that is studied and ask the question, how do I differ

00:11:44.040 from that population? Am I bigger? Am I training? Do I have a more ambitious goal than not dying

00:11:52.480 or not wasting away? And I don't want to minimize those goals because you alluded to the fact that

00:11:57.760 for the vast majority of human history, not dying was an amazing goal. Living to the next day,

00:12:04.420 to the next harvest, to the next season was essential for 99.99% of human history.

00:12:12.900 So this idea that being optimal or thriving, it's a very, very modern luxury we have.

00:12:21.340 It's both modern and ancient all at once. So yes, you are correct. But also the idea that there are

00:12:26.500 different kinds of goals that we optimize is actually the very nature of evolution. What's being selected?

00:12:33.700 And Steve Simpson and David Robbenheimer from down under, from Australia, study something called the protein

00:12:40.540 leverage hypothesis. We may want to come back to at some point. What they've shown is that what at least

00:12:46.380 some animals do in experimental settings they're able to set up is they consume enough protein to optimize

00:12:53.500 their genetic fitness, meaning how many of their genes they're able to transmit to the next generation,

00:13:00.400 which is if you want to win the evolutionary game, that's your goal, which is different than living

00:13:05.740 a lot longer. That's winning the personal game, but not your genes games. So yes, it matters. And it may

00:13:13.040 be that we now want to shift a little from one toward the other. And it may even change during the course

00:13:17.920 of life. It may be that at one point in life, my goal is to be as perhaps big and strong as I can.

00:13:23.880 Maybe at another point in life, it's too slow aging. I may have different strategies for those

00:13:29.720 two different things. Before we go any further, I should have done this at the outset, but I realized

00:13:35.480 that some people are going to be watching us on video and they won't hear the introduction where I'll

00:13:39.920 have made this point. But we should both disclose that we're involved with a company called David

00:13:46.000 Protein that makes protein bars. Because we're going to be talking about protein today and unrelated,

00:13:52.020 but related, we're also going to talk about processed food, which I think is a very interesting

00:13:55.140 topic. This is a company that makes high protein bars, which are by definition processed. And so

00:14:01.680 I just want to make sure everybody listening understands that I'm involved in that company.

00:14:05.660 You're an advisor as well. And I'd like to hear your thoughts to the argument that says, well,

00:14:11.060 Peter and David, you guys can't really have an open and honest discussion about this topic

00:14:14.940 because you have this conflict of interest. I mean, I have my own thoughts on an argument like that,

00:14:19.620 but I'd like to hear yours. Sure. Just a couple of factual points

00:14:22.900 of clarification. So yes, I have a grant from and have been a paid advisor to the David Protein

00:14:28.900 company. Second, the David and David Protein is not my David. I don't own the company. I like the

00:14:34.260 bars. I eat them. And every time I take one out at a meeting or something, someone will look at me and

00:14:38.100 say, how vain are you? You have your own personalized bars with David printed and big left. I said,

00:14:44.460 no, it's not me. It's Michelangelo's David. The idea is you eat the bar, you'll look like that David.

00:14:49.040 But I've also, we had a protein conference we ran about six months ago. It was amazingly successful.

00:14:54.520 The degree of interest from academia and industry and others are tremendous. We must have had

00:15:00.220 somewhere on the order of 50 different companies contribute. So I'll disclose that as well.

00:15:04.740 So I have a lot of interest in this. I've had funding from National Cattlemen's Beef Association,

00:15:09.080 pork producers, and other groups with interests in protein. In terms of the idea of, does that make

00:15:14.540 us trustworthy or not? I distinguish that from trusted. Whether it makes us trusted, that's

00:15:19.640 somebody else's judgment. Trust me or don't trust me, however much you want, that's up to you.

00:15:24.920 Trustworthy, I think, has to do with the processes. And my colleagues and I, we have a saying we've sort

00:15:30.320 of coined and we like to use a lot. And we say, in science, three things matter. The data,

00:15:35.660 the methods used to collect the data, which give them their probative value, which shows what they

00:15:40.720 mean, and the logic connecting the data to conclusions. And everything else is tangential.

00:15:47.980 And so some people who don't have, quote unquote, the goods on an argument will resort to other

00:15:53.740 things. They'll resort to ad hominem attacks. They'll resort to innuendo. They'll resort to quips.

00:15:59.220 Quips are great. Innuendo and ad hominem attacks, not so great, in my view. But none of those are

00:16:04.700 dispositive. And when you think about things, we can really declare things known or not known.

00:16:10.100 No one needs to argue about your conflicts of interest if you say that, I can prove that there's

00:16:16.920 a greatest prime number. And people say, well, no, Euclid proved there isn't. And I don't have to

00:16:22.580 say maybe you're paid for by the prime number company or something. I can just say, here's the

00:16:27.160 proof and you're wrong and there's no point in discussing anything else.

00:16:29.900 Prime number company. Think of the value, David, of prime numbers if they were finite.

00:16:35.860 Can you imagine how much the value of 3, 5, 7, 11, 13, like those numbers would increase in value

00:16:45.000 so much more if they became finite?

00:16:47.180 Think today of-

00:16:48.100 Like Bitcoin.

00:16:49.020 Decoding and encoding Bitcoin.

00:16:50.560 Yeah. That's a very elegant explanation. And I think it's worth reiterating that point,

00:16:56.980 which is at the end of the day, the three things that matter are what are the data? How were the

00:17:03.140 data collected? What were the methods used to collect them? And then what is the string of logic

00:17:07.960 that connects those data to their conclusions? And all of these things should be quite transparent.

00:17:13.160 Now, you've chosen a career, a field of inquiry in which it can be more difficult to do all of the

00:17:22.320 above than in, say, genetics or biochemistry or particle physics, where one of those steps is,

00:17:33.380 in your case, particularly difficult. And that is the manner in which data are collected. In other

00:17:38.180 words, I don't think nutrition scientists are at a loss for logic, but where I think they struggle,

00:17:44.480 if they're studying humans at least, is collecting these data can be really challenging, really

00:17:49.340 expensive. The species of interest is not amenable to close quarters for long periods of time, which

00:17:56.440 is how you would obviously run a controlled experiment in a biological setting. So this is

00:18:01.460 maybe more of a philosophical question, but what is the future of nutrition science? I mean,

00:18:04.940 we're going to come back to the main topic, but this is just such an interesting tangent.

00:18:08.180 Do you believe that there is a much brighter future, a step function and improvement in the

00:18:14.140 quality of nutrition science that lies ahead with AI synthetic data collection or creation rather?

00:18:20.580 Is there something that could fundamentally change nutrition science in terms of how we go about

00:18:25.660 gathering data so that we can be potentially less reliant on epidemiology, which I'm sure we will

00:18:29.960 discuss the shortfalls of today? I think the answer is things will get better. Whether it's a step

00:18:34.320 function or not, I'm not so sure. I want to expand or add to the branch you've thrown in and add a

00:18:41.060 parallel branch, which is, I think there are two reasons why nutrition science is so fraught.

00:18:47.440 One you've pointed out is the methodologic challenge. Can we collect the kind of data

00:18:52.860 we really want with the kind of methods we really want? The other part is the social aspect,

00:18:57.700 which you've hinted at until we've gotten to this point, which is why is it so emotional? Why do

00:19:02.940 people attack each other? Why do people go beyond the data? And I think I see that in any area,

00:19:10.420 the more that area of inquiry is related to economics, religion, social values, personal experiences,

00:19:19.860 the more you get emotion and the deviation from logic and so on. And we see it in whatever people

00:19:28.740 study, child rearing, same-sex marriage, anything that has that emotional valence,

00:19:34.920 that everyday experience and so on leads to more bringing in of non-scientific points of view.

00:19:42.020 So I think that's something we have very strongly. And then I think the other we have is the

00:19:46.240 methodologic challenge of collecting the data. I see benefits on both fronts, but I think both

00:19:51.920 will be slow. I don't think in many cases it's going to be a simple thing of if we could just

00:19:57.640 fix that, if we could just figure out how to measure food intake and free-living people well,

00:20:03.300 and we're on the horizon, then everything will be okay. That's important. I hope we do figure out how

00:20:08.480 to measure food intake well and free-living people, but that alone will not be a solution,

00:20:13.320 a sufficient solution. What I think on the front of the emotional piece is it's going to come slow.

00:20:20.340 When you look at the arc of history of much of human endeavor, at least from my point of view

00:20:25.560 and the point of view of, I think, people like Stephen Levitt and so on, you look over the long

00:20:30.000 haul, things are always getting better. You smooth the function a little bit. Murder rates are way down.

00:20:35.600 Violence rates are way down. Education rates are way up. Lifespan is up, et cetera, et cetera.

00:20:41.640 Freedom is up. But there are lots of ebb and flow. So we may be in a little bit of an ebb and flow now,

00:20:48.600 but I do hope that things will get better as they always have and some more and more rationality.

00:20:54.220 And that's something that as a scientist, I feel very strongly about. As scientists, we focus too much

00:21:00.540 on immediate trust in science and saying, we need to get more trust in science and on immediate issues.

00:21:06.820 How do we get trust in this issue about vaccines or drugs or what have you? Instead of saying,

00:21:13.780 how do we get trust in the scientific process? How do we maybe risk losing a battle? Maybe I'm

00:21:20.280 not going to convince people today that what I think about food additives or vaccines or protein or

00:21:25.340 something is, quote unquote, the right answer. And I'll have to live with that. But if I can convince

00:21:29.980 them that I'm an honest broker and here's how science works and we can work together through science,

00:21:34.360 in the long run, I think that's better. And I think that's coming and something we need to focus

00:21:38.440 on. In terms of the nutrition per se, we have so many challenges. But as a methodologist, I like

00:21:45.280 challenges. Job security, and it's fun. I like figuring things out. So we can't randomize to

00:21:50.700 everything. So how do we get causal inference? We can't blind every aspect of diet. And if we could,

00:21:57.420 we'd be missing some of what we're trying to study because some of the effects of diet

00:22:00.520 involve the effects of perceiving what you're eating. Some of the effects of this drink I'm

00:22:06.680 drinking is how it tastes. And if you blind me to it, then you've taken away that potential effect.

00:22:12.420 There are issues of measurement. How do you know what I really ate? There are issues of adherence.

00:22:17.460 If you tell me to drink one of these every day, do I actually do it? Do I drink one and only one,

00:22:22.640 as you've instructed? There are issues of duration. Could you get me to drink it now and measure

00:22:28.000 something in me 15 minutes later? Sure. Could you get me to drink one every day for the next 20 years

00:22:33.280 and measure stuff? Difficult. There are issues of the model organism. One thing is, we're going to

00:22:38.320 talk about longevity a little bit. Somebody once said to me, you never want to study longevity in

00:22:42.660 an organism that lives as long as you do. It's a bit of a challenge. If you and I at our age were

00:22:48.340 to start, especially mine, I'm a little older than you, were to start a big study and say,

00:22:53.080 I want to study 20-year-olds and give them different nutrition and see who lives longer.

00:22:59.220 And that'll help me figure out for myself what to eat. I'll be dead long before the study is in

00:23:04.160 and not be able to benefit from it and also not be able to find the answer. And maybe we want answers

00:23:10.020 before 60 years from now. So those are just a few, and I could go on and on. Those are just a few of the

00:23:16.820 many challenges we have in nutrition. And I think we're going to chip away at them,

00:23:21.120 but a lot of it's going to have to be settling for various rough inferences to say,

00:23:28.100 this information, I need to recognize its limits. I need to be honest with the public about the

00:23:34.620 limits. I need to say, I haven't shown this unequivocally, but it looks like this is the

00:23:39.520 most reasonable answer now or the most supported answer now. And I'm willing to accept that. But

00:23:45.860 let's be honest, it's not demonstrated. Great example of that, you know, you're seeing the

00:23:50.660 feud in the literature now between Kevin Hall and David Ludwig on the use of crossover designs as an

00:23:56.060 example. And crossover designs in which you give person, let's say diet A followed by diet B,

00:24:02.440 and you give other people a randomized to diet B followed by diet A, they have an Achilles heel

00:24:07.400 that is not there for, let's say, parallel groups where it's just some get diet A and some get diet

00:24:12.680 B. And that is you can have what's called carryover effects. And it turns out, I've started to study

00:24:17.840 this, tip my hat to David Ludwig who pointed the issues out to me and I hadn't fully comprehended

00:24:22.920 them before that. There's almost no way around it.

00:24:26.060 Even with a washout between the crossover?

00:24:28.040 Not an absolute way. You get into what I call argument land. You can say, come on, David,

00:24:33.440 it was a blinded drug I gave and I know the kinetics of it and I know it's out of the system

00:24:39.160 by this date. How could it possibly be having this? And I could say, well, Peter, you're probably right.

00:24:44.480 It's a good argument, but the validity still depends on your argument. It's not an absolute

00:24:50.640 a priori proof. And if I said, well, maybe what the drug did is it permanently changed something in that

00:24:56.120 person? And we should just explain for listeners why this is an important discussion in science,

00:25:02.720 especially in human trials and especially in human trials with nutrition, because there's a real

00:25:07.960 statistical power. And I don't use that in the beta sense of the word, but a statistical gift

00:25:13.400 that comes from being able to do a crossover in that you can now leverage a student t-test, for example,

00:25:19.520 as a very powerful statistical tool that allows you to use fewer subjects and therefore a fraction of

00:25:26.000 the cost. So I'm guessing, I don't follow this debate by the way, but I'm going to guess that

00:25:30.380 you're going to say Kevin Hall favors a crossover, not to speak for Kevin, but I would bet the reason

00:25:35.320 Kevin favors it is because the type of work Kevin does is insanely expensive. He's putting patients

00:25:41.500 in metabolic chambers and therefore the fewer patients that he needs to do that with, the easier

00:25:48.240 he can do his work. Is that basically the argument?

00:25:50.780 You're absolutely correct. And when you said statistics in the, or power rather,

00:25:54.440 in the beta sense, that is what it is. It is statistical power, the probability of rejecting

00:25:59.800 the null hypothesis. If the null hypothesis is false, which is one minus beta, the type two error

00:26:05.460 rate, there's actually a little bit more, but I think 90, 90 plus percent of the motivation is what

00:26:10.140 you've described. The other, as actually Kevin pointed out to me in some dialogue we were having

00:26:14.720 was sometimes it's just throughput. Even if I had infinite money, I can't put people through

00:26:21.760 the procedure fast enough because there's only so many chambers. Someone else pointed out to me,

00:26:27.800 it could be patient availability. If you said, I'm studying this in this rare population,

00:26:32.400 I can't get a thousand people even if I had the money because they don't exist.

00:26:36.160 I can only get 10 or 20 people. So all of those things strongly favor the crossover,

00:26:41.400 which in almost all circumstances will be much more statistically powerful, meaning you can get the

00:26:47.300 same amount of precision out of many fewer subjects. But the problem is that you have this

00:26:52.420 carryover. And so that the difference between the two groups who get treatment A and treatment B

00:26:58.400 in the second period could be a function of the true effect of treatment A versus treatment B in

00:27:04.680 the second period, or it could be an effect of what treatment A and treatment B did in the first

00:27:10.580 period carrying over. And now you don't have a clean estimate of the effect of treatment A or

00:27:16.300 treatment B anymore. Now, what you can do is what you said is a washout. If you said, I understand how

00:27:21.860 this thing works and it's a molecular effect. It's not a social effect. It's not learning. It's not a

00:27:27.520 surgical thing. I didn't cut a piece of their anatomy out that doesn't reverse. This thing is

00:27:32.520 completely reversible. It's completely blinded. I gave a long enough washout. Then by that argument,

00:27:39.460 you can say, I rule it out. But you can never say, absolutely rule it out. If you're dealing

00:27:45.340 with things other than that, where you can't have a long enough washout, or it might be a psychosocial

00:27:49.740 effect or a learning effect or some permanent effect, anything from bariatric surgery to something

00:27:56.560 like an allergen, which may permanently sensitize the body. I know people say mRNA vaccines. Maybe the

00:28:03.140 mRNA hangs around for a while. Is that a permanent effect? Or a vaccine itself is a, we hope in some

00:28:08.500 cases is a permanent effect, like a measles vaccine. All of those things, they're not going

00:28:13.380 to wash out. So those are things where the crossover has a limit. And the question is,

00:28:21.140 does that mean you try to do what David Ludwig is, I think, arguing? Again, I hope I'm not

00:28:26.140 inappropriately putting words in his mouth, is saying they're just invalid. Just either don't use them

00:28:30.760 at all. Or you can only use them in this way. And if you get what looks like it might be a period by

00:28:36.720 treatment interaction, then discard the study. It's invalid, so on. Or do you say, which I think

00:28:43.360 Kevin would say, and I would side with, is accept the limitations of the study. It doesn't mean it's

00:28:49.840 a flawed or incorrect study. It means it's a limited study. And as long as you point those limitations

00:28:55.220 out, you may or may not want to accept it. In the same sense, if you have an observational

00:28:59.060 epidemiologic study, I think most of us, again, there are exceptions, but most of us would not say

00:29:03.640 never do one again. They have zero value. What we'd say is they're not invalid studies in and of

00:29:10.340 themselves. What they do is they leave open alternative explanations for findings other than

00:29:15.900 causation. Could be some bias of measurement error. It could be some bias of sampling. It could be some

00:29:20.720 reporting bias. It could be some confounding, et cetera, et cetera. And we say, as long as you

00:29:25.780 acknowledge those, then the study shows what it study shows. It's weak inference, but it's not

00:29:31.260 nothing. And I think the same thing is with crossover. And we may have to accept that with

00:29:35.120 lots of stuff. That's the idea of nutrition. We may have to accept there are these limitations

00:29:39.900 of our knowledge. So, for example, if you do a study of cheddar cheese made in Wisconsin

00:29:45.060 versus some appropriate control, and you say, look, this is what the effects I get of this,

00:29:51.580 and so therefore cheese consumption has this effect. I say, well, Peter, are you sure it's

00:29:56.600 cheese consumption in general, or is it just cheddar cheese? Is it just cheddar cheese made

00:30:01.580 in Wisconsin, or is it any cheddar cheese? Is it only cheddar cheese when eaten with these other

00:30:07.080 things, or these? And I think that's something that's so hard to control in nutrition that will wind

00:30:12.760 up with these statements of saying, it looks like it's cheddar cheese in general, and maybe we can

00:30:17.220 do some other studies to suss it out, or it looks like it's cheese in general. But we'll always have

00:30:21.900 this sort of saying, this looks like what it is. Here's a good recommendation, but it's not absolute

00:30:27.980 knowledge. Well, with that, let's go back to where we started, which is the RDA for protein

00:30:34.580 consumption. Now, many folks, Don Lehman and others, have argued that the RDA is insufficient if

00:30:41.200 you're actually trying to optimize health, and if you're actually in pursuit of another agenda,

00:30:45.660 which might be avoiding sarcopenia later in life, achieving your peak in physical performance.

00:30:51.720 That's not the peak of physical performance, right? But if your objective is beyond survival,

00:30:58.600 you might want to have more. And the numbers for what more tends to be seem to converge in the

00:31:05.900 ballpark of 1.2 to 1.6 grams per kilogram, if you're trying to go for a minimum effective dose,

00:31:14.380 but easily up to two. So what is your best aggregation of the data on where you start to reach

00:31:24.480 diminishing returns? I hesitate to talk this way because it's the way I think, but I know it's not

00:31:29.560 the way others do. I always think about the concavity of a curve. So the more concave down it is,

00:31:35.200 the more negative the second derivative, the quicker you get to that point of diminishing

00:31:39.440 return. It just means the curve is shaped that way. And so most things in biology work that way.

00:31:45.700 They're not positive second derivatives where the more you do, the better it gets,

00:31:50.640 and the rate at which it gets better goes up. That's very rare. So what is your gestalt on

00:31:57.600 the optimal zone based on all of the above?

00:32:02.480 Yeah, I think it's important to distinguish between concave downward or a curve that doesn't

00:32:08.820 keep accelerating, a decelerating curve, curve with a negative second derivative versus non-monotonic.

00:32:15.540 Those can be non-monotonic, but it's not the same thing. So one is the diminishing returns.

00:32:20.600 It keeps going up, but it goes up ever more slowly. And there may be a point at which it

00:32:27.260 never reaches. And that point that it never reaches may not be 100% of whatever it is you're

00:32:32.620 thinking about. That's different than saying it actually goes down at some point. So a lot of

00:32:37.540 things go down at some point, right? Too little will kill you, too much will kill you.

00:32:41.360 And that's actually more common in biology, right? So too much thyroid hormone, too little thyroid

00:32:46.160 hormone, a very bad thing.

00:32:48.160 Too few calories, too many calories.

00:32:49.620 In the case of protein, if we're thinking about it not in an absolute sense, but as a percent of

00:32:56.200 calories, let's just say, and let's assuming your calories are at an appropriate level for whatever

00:33:01.140 it is you want to your life, what I think we can say with pretty good confidence is there's some

00:33:07.160 level that's too low. I think we can also say with pretty good confidence that there's some level

00:33:12.420 above the RDA that with very, very rare exceptions, perhaps, is at most not worse than, and in the vast

00:33:21.900 majority of cases, likely better than the RDA level. You might think of it as the real base level, like

00:33:30.000 the higher end good rental. And then there's something further out there where I would say there's less

00:33:41.120 certainty, and that's where probably a little more of the debate is. So there's not no debate between the

00:33:46.320 base level and what I'm saying is closer to the, perhaps, I don't want to use the word optimal because

00:33:51.580 that implies a dip, but superior, known superior level, little debate in between, but not so much in my view.

00:34:00.300 And then there's more debate about beyond the known superior level and a lot of uncertainty there.

00:34:05.940 I think the evidence is very clear that when you go from the RDA level of 0.4-ish grams per pound or 0.8-ish

00:34:14.760 grams per kilogram up into roughly double or even a little bit more of that, roughly two.

00:34:20.240 I know no evidence of harm in any group, other than perhaps, again, the very rarest folks. And

00:34:26.440 even that would typically not be protein in general. It would be specific types. So if you said to me,

00:34:32.200 phenylketonuric can't have phenylalanine, okay, fine. Someone who's got allergy to whey protein can't

00:34:38.700 have whey protein, but that does mean protein in general. I know of no evidence for harm, even in

00:34:44.700 people with chronic kidney disease or anything else. I think there's lots and lots of evidence

00:34:51.200 for benefit in at least medium-term observable phenomena like body weight, like appetite control,

00:35:00.500 like bone strength, muscle, and so on for people to be consuming more. And I think it's especially true

00:35:09.360 in people who are recovering from injury, bodybuilding, looking for performance in athletics,

00:35:15.340 looking for strength, who are older, who are growing, all of those things.

00:35:20.460 I mean, sometimes, David, I think it's easier to flip it and just say,

00:35:23.700 why don't we just identify all the people who don't benefit from a higher amount of protein than

00:35:28.400 the RDA? Because I worry that when we start to carve out a bunch of categories and say,

00:35:33.740 if you're in one of these categories, you should be consuming more, you're always going to kind of

00:35:38.080 miss something or someone might not identify. So for example, you've used the term bodybuilder.

00:35:42.600 Most people would never identify themselves as a bodybuilder because when they think bodybuilder,

00:35:47.700 they think of the caricature type bodybuilder we see on a magazine cover when you're at the airport

00:35:54.500 that doesn't actually look anything like what you want to look like. But the truth of the matter is,

00:35:59.080 it's hard for me to imagine, I don't know that I can think of one of my patients, let me just start

00:36:04.300 with that as a sample size, one of my patients who doesn't need to, at a minimum, work hard to

00:36:10.260 maintaining their muscle mass. And many of my patients are working hard to add muscle mass.

00:36:16.500 And so what differentiates them from a bodybuilder? The difference is the bodybuilder, of course,

00:36:21.180 has the benefit of using super physiologic doses of androgens, consuming, basically training all day

00:36:28.620 and doing nothing but optimizing around that. But the reality of it is, we're all sort of

00:36:33.700 bodybuilders. Everybody's a bodybuilder if they're really thinking about it in the lens of what are

00:36:40.260 we on this earth for? We're on this earth to create the most robust body we can have. And it doesn't

00:36:45.920 have to look like it's bulging with muscles. First of all, most of us couldn't achieve that if we

00:36:50.500 wanted to, notwithstanding the fact that most of us could never achieve it anyway.

00:36:54.400 Okay. Let's think through who should be consuming the RDA.

00:37:00.480 I would say with rare exceptions, the answer is probably no one.

00:37:04.620 Okay. I mean, that's a very important statement.

00:37:06.600 There are rare exceptions. It goes back to goals. You might say you might find people,

00:37:11.920 in fact, not might, I think you almost certainly would, who say, I don't care about any of those

00:37:16.320 things you just mentioned. The only thing I care about is saving the planet. And my thing is I should

00:37:24.660 eat as little as possible and as little protein as possible and as little animal product as possible

00:37:30.180 for that. Somebody else could say, my goal is to be nearer to God. And this gets me nearer to my God

00:37:36.100 in my way. Someone else in my goal is an aesthetic. I want to look like a heroin addict in a doorway in

00:37:42.840 Manhattan in 1970. That's your aesthetic. But those people are rare. 0.57

00:37:47.400 I think it's really important for people to understand that this argument around the RDA is

00:37:54.240 adequate and that's what you need to eat. And anybody who is suggesting you eat more than that

00:37:59.320 is wrong. We have to actually flip the question and say, okay, who is best served by eating at the RDA

00:38:06.320 versus say 2X the RDA at 1.6 grams, which for me would put me at 150 to 160 grams of protein per day

00:38:15.960 instead of 60. I think here's where we get into that issue of recognizing the limits of our knowledge

00:38:21.740 and then being able to wrangle with them rationally as opposed to irrationally

00:38:26.080 state the limits of our knowledge. So I put up a LinkedIn post and I put up many about protein in

00:38:31.120 the last 12 or more months. Yeah. We'll link to them all in the show notes here so that people can

00:38:35.960 kind of, I don't want to make this a discussion where we're plowing through papers because that's

00:38:40.620 already been done and we'll link to all those things. But yes, I'll make sure that everyone goes

00:38:45.040 back to them. So feel free to reference specific ones so they'll have access to them immediately.

00:38:49.040 And one of them, I included a quotation was from one of my old mentors in graduate school,

00:38:53.860 Harold Euchre, who famously said, I'm a data nut. Students gave a t-shirt that said that.

00:38:59.420 And he would say, show me the data. It was the time I think of that Jerry Maguire movie being

00:39:04.220 popular, show me the money. He was like, show me the data. So what I sort of said is people keep

00:39:08.640 raising these questions of harms. I said, show me the data. Can anybody send me? And it was sincere.

00:39:14.380 I said, this is an open call. And I sent it to like some of the top people in the world,

00:39:18.260 including those who are a little bit hesitant on protein intake or denigrated. And I said,

00:39:23.820 can anybody send me one or more papers that are intervention studies, not observational ones,

00:39:30.620 that are in humans, ideally randomized, but I'll take an intervention even if it's not randomized,

00:39:35.800 but it's got to be controlled. Controlled intervention study in humans feeding different

00:39:40.840 levels of protein in which the different levels of protein intake are separable from other effects

00:39:46.480 that show deleterious effects on a clinically or intrinsically meaningful endpoint. Don't show me

00:39:54.720 that this molecule changed or this gut microbiota changed. What do I do with that? Why do I care that

00:39:59.760 that gut microbiota changed? Nobody cares about those things intrinsically. We care about them only

00:40:05.820 in so far as if they give us heart attacks or strokes or earlier or later death or greater strength or make us

00:40:12.200 better looking. We care about how long we live, how good we look, how we feel, our strength, what we can do.

00:40:19.040 We don't intrinsically care about whether this molecule in our body is higher than that molecule

00:40:24.280 or this gut microbe. And I said, can anybody send me one? Nobody.

00:40:29.740 I was on part of that thread, David. And wasn't there one paper that did surface that was a TPN trial,

00:40:38.180 total parenteral nutrition in patients in the ICU?

00:40:42.220 I think that there were some, I think that Dudley Laming sent, and Dudley's a great scientist and a good friend

00:40:47.840 who studies protein mainly in mice and other things, but also a little bit in humans. And I think he cited a couple

00:40:54.260 of references. And I think they were from Luigi Fontana, if I recall correctly. And there was some

00:40:59.840 short-term trials in patients with cancer. And I don't remember all the details.

00:41:04.760 Yes. I'm working my hardest to give an honest look because I think what you did is the right thing to

00:41:10.560 do, which is, look guys, we're having religious debates on social media where people are using

00:41:15.280 their Twitter platforms to like lambast people they disagree with and call them names and do all this

00:41:21.240 sort of nasty stuff. Why don't we just do this like grownups and show me the data? And so the data is

00:41:26.720 show me human clinical trial intervention studies that demonstrate the deleterious effects of quote

00:41:33.240 unquote high protein. And yeah, the only thing that I saw was you took these patients who were very,

00:41:41.600 very sick in the ICU, so sick that they can't consume enteral nutrition, which means they can't

00:41:48.460 eat because they're probably ventilated and their guts aren't even working. So you can't actually put

00:41:53.300 feeding tubes in them. So you use a central line, you put a intravenous catheter into one of the major

00:41:59.040 central veins in their body and you give them all of their nutrition through that conduit, which is

00:42:04.520 called parenteral nutrition. And with total parenteral nutrition, you are chemically crafting the exact

00:42:12.700 composition of what they consume. Exactly how much glucose, exactly how much fat, what type of fat,

00:42:18.280 how much protein, what type of protein, what micronutrients, et cetera, et cetera. And again,

00:42:22.920 I don't remember the exact study, but it showed that there was no benefit to a higher protein diet.

00:42:30.100 And this was counterintuitive. Now that I'm remembering it, I believe the study sought to ask

00:42:35.040 the question, wouldn't patients in the intensive care unit benefit from a higher protein diet? Because

00:42:40.660 they're very catabolic. The body is incredibly catabolic in that setting. And so I think the surprising

00:42:46.320 outcome of that study was that the patients on the higher dose of protein did know better.

00:42:51.240 If my memory serves me correctly, they didn't do worse. Am I remembering that correctly?

00:42:57.640 If my memory also serves me correctly, there was no statistically significant effect

00:43:02.520 on what I would call an intrinsically clinically important outcome. There wasn't a statistically

00:43:09.980 significant effect on lifespan. Right. And in that patient population,

00:43:14.000 death, mortality is the most important outcome. There are other things, days in the ICU, days on

00:43:20.140 the ventilator, those things all matter tremendously. But you're looking at such a sick population

00:43:24.440 that's on the precipice of death that when you look at ACM or all-cause mortality, you're going to get

00:43:28.680 some interesting and valuable insights. And we're also going to have to really dig into this idea of

00:43:34.840 when we've got different sources of data, and none of which are the data we really want.

00:43:39.920 What we really want is the randomized controlled trial in tens of thousands of people, so we can

00:43:48.140 look at subgroups, and we'd have a lot of power. We want perfect adherence. We want it free living.

00:43:55.200 We want it in people eating foods under the circumstances about which we're going to make claims.

00:44:00.220 Most of us are not asking, well, if I am on TPN and unconscious and being tube-fed by a surgeon,

00:44:08.540 then what? Most of us are saying, when I go to the grocery store and decide what I want to bring

00:44:13.580 home for dinner tonight, then what? And those are not the same thing. So we won't have that.

00:44:19.120 We'll have lousy epidemiologic studies with lousy self-reported data following people, large groups

00:44:25.560 for long periods of time, and causal inference will be fraught. We'll have mouse studies in which we're

00:44:30.460 not sure we can generalize from the mouse to the human. We'll have short-term studies of people

00:44:36.300 being tube-fed. We want to talk about long-term studies of people eating, quote-unquote, ordinary

00:44:41.800 foods in ordinary ways. And we're going to have to sort of think about if they all line up perfectly.

00:44:48.080 Smoking is an example of that, where almost everything lines up. You may not have the perfect

00:44:52.000 study you want, but the cell stuff, the mouse stuff, the animal stuff, the epidemiology,

00:44:57.840 the clinical trials would get people to stop smoking. They all line up to say,

00:45:01.720 smoke is really bad. Don't smoke. If they all line up, great, then it's easy. If they don't

00:45:07.040 all line up, we also have to start to talk about how strong is each piece of evidence,

00:45:11.180 both in terms of its generalizability to what we really want, as well as in and of itself,

00:45:15.860 is it strong? And I think those studies that we've just described are not especially dispositive to me.

00:45:20.980 Yeah, I think another type of inquiry that can be misinterpreted, but there's a great analogy for

00:45:26.820 it, is around the dose response curve for muscle protein synthesis as protein dose increases.

00:45:34.980 There was a study that was referenced that looked at, as you went from 0.8 to 1 to 1.2 to 1.4 to 1.6

00:45:43.920 grams of protein per kilogram of body weight, what did the rate of muscle protein synthesis do? In other

00:45:50.380 words, where did you start to achieve the plateau and beyond which you were not going to get more?

00:45:55.560 And what I think this study demonstrated was that in less trained individuals, you will achieve

00:46:02.920 higher levels of MPS for lower amounts of amino acids. But again, two things stand out here. The

00:46:09.740 first is be careful what patient population you're looking at in the study and make sure it applies

00:46:14.500 to you. So in an individual who's training an hour a day, I don't think they can compare themselves to

00:46:22.300 someone who went from sitting on a couch to training 90 minutes a week. Very different.

00:46:27.380 The other place I would say that there's a perfect parallel there, and I want to make sure everybody

00:46:31.480 who's trying to figure out what bucket they belong in can sort of do the mental gymnastics here.

00:46:36.600 If you take an untrained individual, a person who is 100% sedentary, which sadly is the majority of

00:46:45.000 people in the United States, and you put them on a fitness regimen of three 30-minute whole body

00:46:52.260 workouts a week, take that sedentary person, I take them into a gym, and I get them to push around

00:46:58.200 weights 30 minutes three times a week. Not to failure, not to profound exhaustion.

00:47:05.280 There is no desire to maim them or make it such that they can't get out of bed the next day.

00:47:12.160 Will they achieve a training benefit? Will they achieve some benefit? And the answer is

00:47:15.380 unbelievably. Unbelievable benefit. David, if you put me into their workouts, would I achieve any

00:47:23.760 benefit? I would argue virtually none. Why? Why the difference? Well, because you're already sort of

00:47:30.080 on the asymptote, right? It's like anything else. Take someone who's never had any

00:47:34.840 value of it and you get a big benefit. Give a mouse that has no leptin, just a tiny little bit

00:47:40.340 of leptin, and it immediately starts to slim down quite a lot. Give a mouse that has a normal amount

00:47:46.280 of leptin a little bit more, you're going to observe almost nothing. Give a kid who's been

00:47:50.900 studying algebra for an hour a day vigorously and diligently an extra 10 minutes of studying

00:47:57.900 algebra, you're probably not going to get that much benefit. Give a kid who's never been exposed to

00:48:02.180 algebra at all, 10 minutes a day of tutoring algebra, probably start to get some real benefits

00:48:06.520 soon. By the way, that's a beautiful example. When I hear people say you don't need much more

00:48:12.100 protein than 0.8 because of that study, you know what it makes me think? That's like telling a kid

00:48:16.980 they only need to study algebra 10 minutes a day if they want to master it. Absolutely. One of the

00:48:21.860 things that irks me about the field of nutrition, and I don't know the solution to this, it goes back a

00:48:25.880 little bit to the challenges we have, but this is really an economic challenge, not an intrinsic

00:48:29.980 challenge. Look at the sample sizes of studies in nutrition in general, randomized controlled trials

00:48:37.080 of nutrition in general, especially of things like protein intake. And then look at the sample sizes

00:48:44.340 of studies of statins, randomized controlled studies of statins, GLP-1 agonists, vaccines, etc. They're

00:48:53.240 different by multiple orders of magnitude. It's not uncommon to read these studies saying, well,

00:48:58.620 we're interested in the effects of protein consumption on African-American women over 0.98

00:49:03.080 age 50 with diabetes and without. So it's six in each group. You're like, six in each group.

00:49:09.080 You had 60,000 over there in that pharma study. And so we have really weak data on this. And so it's

00:49:16.780 not surprising that often we don't show these big effects. This is important, David. I'm sorry to

00:49:22.180 interrupt. I want to tie this back to the discussion we had around disclosures. There's sort of a reason

00:49:28.820 that virtually everybody in nutrition science is taking some money from food industry. Now,

00:49:37.420 there are someone who works at the NIH who is funded entirely at the NIH. Kevin obviously doesn't need

00:49:42.080 to. Is Kevin back at the NIH, by the way? Not to my knowledge.

00:49:44.180 Okay. Anyway. But most everybody who's at a university is cobbling together money from both

00:49:51.260 the government and industry. And that's demonstrated by the belt and suspenders bootstrapping approach

00:49:59.440 that comes into nutrition science studies, which are not well funded. So maybe just explain why is it

00:50:06.660 that it's easy to fund a pharma study with 60,000 people in it, and it's hard to get the funding to

00:50:14.560 study 600 people in a nutrition study? The first question is very simple. It's the economic model

00:50:20.820 of it. Pharmaceuticals are patentable. And the way our country works is, in general, you cannot market

00:50:27.960 a pharmaceutical without the FDA's approval. And the FDA will generally not give approval unless you've

00:50:34.560 met their bar for having demonstrated as the act of Congress for the FDA's structure mandates,

00:50:41.580 they must have a reasonable basis for concluding that the benefits outweigh the harms under proposed

00:50:50.240 conditions of use. And so the FDA says, this is what it's going to take to convince us. And it's going

00:50:54.660 to be these big randomized control trials, as well as a few other things. And so the companies say,

00:50:59.460 we've got to do it. Then there's the economic model that makes it feasible for them to do it in most cases.

00:51:04.380 Sometimes the reason we don't have certain drugs is not because they can't be made. It's because

00:51:08.820 the pharma companies say, it's not worth it for us because we won't make enough money to offset

00:51:15.360 the development platform. That's a problem there. But in any case, in other situations like a GLP-1

00:51:21.600 agonist, it does. So they develop them. They spend hundreds of millions of dollars. They do them to the

00:51:27.280 utmost rigor. They're probably the most rigorous human health studies done on the planet these days.

00:51:34.140 To your point, if you're willing to spend, let's say it's 2 billion today, 3 billion, whatever the

00:51:39.380 number is, 10 years and a couple billion dollars is an enormous investment. But if you can recoup it,

00:51:45.360 it's worth it. How do you recoup that in nutrition science?

00:51:48.280 You don't. Their margins in the food industry are much lower. They often can't patent stuff

00:51:55.040 quite so easily. It's hard to patent a grapefruit, right? So if you're the grapefruit sellers and you

00:51:59.940 want to do some study, you're not going to patent grapefruit even though you may have some benefit

00:52:04.340 from it. So that creates a problem which sometimes leads to why people want supplements and things so

00:52:09.620 that maybe they can get some patent protection. But even that can be limited at times. You don't have

00:52:14.640 the economic model for it and so they just don't. I don't have the numbers at my fingertips.

00:52:20.760 I don't know if anybody does. But if you said to me, we're going to take all the money spent on

00:52:27.860 research looking at the effects of food, not how do you make food, not how do you make a better

00:52:33.480 chocolate bar or a better macaroni and cheese or something, but what are the effects of eating that

00:52:38.700 chocolate bar or a macaroni and cheese? And you added it up across every single commodity group,

00:52:45.160 the dairy council, the egg board, every single food company, every dietary supplement company,

00:52:51.720 and you added all of what they spend on research. I would be very surprised if it exceeds a billion

00:52:56.800 dollars across the entire country. So there's very little money, relatively speaking, there.

00:53:03.600 They don't have the economic wherewithal to do it. They also don't have the mandate to do it.

00:53:07.700 My group, 20 plus years ago, we did the first randomized controlled trial

00:53:11.760 ever commissioned of their products by the Frito-Lay company. And my gosh, were they scared

00:53:18.240 about this. Didn't know what they're getting into. It turned out we compared chips fried in corn oil to

00:53:23.660 low-fat chips and cookies and crackers and things to traditional chips and cookies and crackers that

00:53:29.600 had more saturated fat, trans fat. And the idea was, is low fat better than corn oil? And the answer

00:53:37.040 was, no. Assuming you can control your calories, you're better off eating the full fat corn oil chips.

00:53:43.580 What about the saturated, the high saturated fat and the trans fat?

00:53:46.360 They were the worst. Yeah.

00:53:47.760 What were the outcomes?

00:53:49.040 CVT type outcomes.

00:53:50.520 So biomarkers, obviously.

00:53:51.580 With the exception, I think, of triglycerides. My recollection is the traditional trans fat,

00:53:56.620 sat fat were worse. I think the low fat, high carb was worse for triglycerides. And the high fat

00:54:02.900 corn oil type stuff was better for everything. That's published in AJCN. Marie-Pierre Saint-Ange was

00:54:08.560 the first author of that. And it was interesting. The criticism we got, again, it was Marion Nestle was

00:54:14.100 one of the few critics. And it was typical. Never touched the science. They never said,

00:54:18.840 well, the design was wrong or the measurements were wrong. It was like, well, they were funded

00:54:22.800 by industry. And I call this, this was her words, I call this a calorie distractor. If you have

00:54:28.380 something to say about the science, why don't you stick to the science and do that instead of quips and

00:54:33.460 ad hominem attacks? But that's all we got. Anyhow, so the food industry traditionally has not done a lot

00:54:39.900 of this. These studies are expensive. That study we did, I don't remember the exact number, but it's

00:54:45.140 probably in the neighborhood, especially if we inflated it to today's costs, might be getting

00:54:49.900 close to a million dollars or something. But that's nothing compared to what we talked about with

00:54:54.280 pharma. The other issue is the other big funder is NIH. And NIH, I think both have seen often these

00:55:01.180 things as something the industry should fund. Well, I mean, if it's they're selling it, let them fund it.

00:55:06.520 That's one. Sometimes it's not seen as really deep, big science. Yes, it's practically interesting,

00:55:13.220 Dr. Allison, but where's the big, deep scientific hypothesis? And then the last thing is, in my

00:55:20.060 opinion, here's clearly an opinion, the misprioritization of funding. And this is

00:55:25.160 something that now we see the NIH addressing very vigorously. Again, I would agree with some of what

00:55:30.320 they're doing. I disagree with some of what they're doing, but Jay Bhattacharya, he's a real smart guy.

00:55:34.560 And he, with others, are trying to say, let's repurpose some of the funding. Less here, more here.

00:55:41.240 And I look at the observational epidemiology, which can be very expensive in attrition.

00:55:46.640 And I think the new information yield, and I underline that phrase, new information, is often

00:55:52.660 very low. If somebody else comes out with a new study tomorrow and says, we did another one, and it

00:55:57.740 was a million subjects. And we measured the food intake as carefully as we can with self-report in

00:56:03.600 this population. We had a couple of biomarkers, and there's something interesting about this. And

00:56:08.220 here's what we showed with protein intake and longevity. I don't care which directions. You

00:56:13.240 showed greater longevity, you showed lesser longevity. I'm like, I already knew it was a

00:56:17.000 question. You haven't answered the question for me. You gave me a little thing to scratch my head

00:56:22.040 about maybe, but it didn't really move the needle. Let's talk a bit about the epidemiology in this

00:56:26.400 space. So I think everybody listening to this podcast knows what epidemiology is, and we've

00:56:31.500 talked a lot about the limitations of it and what a healthy user bias is. But give us the landscape of

00:56:37.340 how epidemiology has looked specifically at this question of the relationship between protein intake

00:56:44.660 and outcomes of health. What are some of the near unique or particular circumstances of

00:56:53.340 epidemiology that lend itself to confusion here? I think the most important thing that I want to say

00:56:58.660 is a sort of a, maybe a little bit of a weird kind of left turn on this, but I think the greatest limit

00:57:04.400 or problem with the nutrition epidemiology in a context like we're discussing now, or in this exact

00:57:09.320 context of protein consumption and things like longevity and long-term major health, is the opportunity

00:57:15.720 cost. It's that we're spending the money and often a non-trivial amount of money. These big epidemiology

00:57:21.260 studies can be very expensive, and we're not spending it therefore in the big randomized, well-done,

00:57:27.520 controlled clinical trials. That's, I think, the biggest problem. If you said to me, well, we did some

00:57:31.820 epidemiology, there's something we can glean, and it's interesting and fun, and there was no cost. Okay.

00:57:36.940 But if you say, we did that, but we could have done one really good or maybe 10 medium-sized

00:57:42.760 randomized controlled trials for that, that was a big loss. Now, in and of themselves, if we just

00:57:48.360 stick to the epidemiology, you and many others frequently and correctly point out the issue of

00:57:53.620 confounding. We all say correlation is not necessarily causation. The example of ice cream

00:57:59.320 consumption and murder rates is trotted out. We say, you know, more ice cream consumption associated

00:58:04.820 with more murder rates. Guess what? It's heat. People eat more ice cream when it's hot. They

00:58:08.380 murder more when it's hot. All true. All fine. But that's actually just a tiny piece of it.

00:58:13.760 There's so much more. There's the measurement problem, and that measurement is not random.

00:58:18.200 And even if it was random, it's usually not taken into account. If it was random and we knew it was

00:58:22.700 random and we took it into account statistically, we could make the problem kind of go away. But it's

00:58:27.960 often not random. It may be correlated. People who eat more of this may systematically bias their

00:58:34.660 reporting down than people who eat less of that. There's selection bias. People may choose to be in

00:58:40.080 the study or not choose to be in the study, and that may affect things. There's what's called

00:58:45.180 collider bias. I control for something, and I think I'm doing a good thing by that, but in fact,

00:58:50.960 I've created an inadvertent association. I could go on and on with statistical obscura.

00:58:57.020 What do you think are the three most important biases that impact this particular question

00:59:03.340 when asked through an epidemiologic lens? I think it's the confounding, particularly,

00:59:09.740 but not only by culture and socioeconomic status and social class education. Second would be the

00:59:17.280 measurement error, in particular, again, the non-random measurement error. And then I would

00:59:24.500 probably say some selection biases that are a little hard to specify. Miguel Hernan at Harvard

00:59:32.000 is perhaps one of the most thoughtful people on talking about the other ways these biases can creep

00:59:37.800 in in terms of when people start the study, who gets in the study, and when we consider their

00:59:43.900 exposure as occurring, he does some ways to try to correct that. He thinks that's more important

00:59:48.620 than confounding. So I think those are probably the three biggest ones. Those are the intrinsic

00:59:53.120 issues. I think the other one, which is very big, but not intrinsic, is I'm going to say honesty,

01:00:00.280 and maybe that's a little too strong, but I think it's the honesty or maybe the sincerity,

01:00:05.620 perhaps is a better way of saying it, the sincerity of reporting by the investigators.

01:00:10.280 I don't think many investigators are lying in an explicit sense, but I do think there are both

01:00:16.760 intentional and unintentional efforts at distorting. That is, people want to tell a story and they

01:00:23.180 emphasize some things and de-emphasize others. They hide some things and don't hide others.

01:00:27.940 So it's not an explicit lie, but there's a manipulation of the information.

01:00:33.220 And why do you think the editors at journals are unable to address that in the review process?

01:00:40.280 For the majority of editors, it's lack of ability, lack of resources, and lack of courage. For a

01:00:47.840 minority of editors, it's limitations on the ability to really go in and suss it all out.

01:00:56.540 So the New England journals of the world, the sciences of the world, the jammers of the world,

01:01:02.020 they have the resources within reason and the sophistication and to sort of go after this,

01:01:07.960 but they can't get everything. Often when I think about the idea of peer reviewing, but also to some

01:01:13.500 extent, the editorial review, which has a little more teeth than the peer reviewers themselves.

01:01:17.820 I look at it like restaurant reviews. If Zagat or whoever goes in to review a restaurant or

01:01:22.780 sort of give them Michelin stars, they can tell you, did they like the offerings on the menu? Was the

01:01:27.540 food tasty? Did it look good? Was the service good? They're not going back and doing a microbial count

01:01:33.240 in the kitchen. They're not checking how often the chef washed his or her hands. Those are things you

01:01:39.300 need a health inspector who's got some authority, who can do spot checking, surprise visits,

01:01:44.540 who's got equipment and so on. That's what you need there. And I think peer reviewers are like

01:01:50.040 restaurant critics. Now, does it look good? Is it interesting? As a peer reviewer, I can't go back

01:01:54.440 and look at everybody's raw data. In some cases, we'll see something that looks funny. And then through

01:02:00.180 the journal, we're doing a couple of these now, we will get the raw data from people. And then we

01:02:04.340 often see things that are quite funny. And we often get a lot of fights with review authors who don't

01:02:09.540 want to let us look at their raw data and kind of tells you something.

01:02:12.920 Where is AI in this process? I mean, why are we not, or are we using LLMs to serve as peer editors?

01:02:20.860 Short answer is we are. Long answer is we're at the stage of infancy and amateurishness with it. So

01:02:26.440 it's coming. It'll get better and better. But we can do some very simple things now. People look for

01:02:31.600 these so-called tortured phrases. When you find these phrases that kind of look like a word salad,

01:02:36.580 that's sort of a hint often that you've got something plagiarized or just fabricated. We can

01:02:42.180 look in some circumstances for things that literally don't add up. So some people made something called

01:02:48.560 the Grimm test. I forget what it's an acronym, but it's basically when you know, let's say you have

01:02:53.380 a Likert scale and you know, you have a certain number of subjects, then the mean of that scale

01:02:58.720 can only have certain values. And if you say, Hey, it doesn't have one of those values, something must be

01:03:03.940 wrong. Those are examples.

01:03:06.180 And do we know if these AI agents, these peer review agents, I'll call them, are being trained on known

01:03:13.040 fraudulent manuscripts? Because we certainly have an abundance of things that were demonstrated to be

01:03:19.120 frauds. So it would be, I assume, a reasonable thing to do to start training these AI agents on

01:03:24.100 that to start identifying the patterns? Again, the answer is yes, but very much in its infancy.

01:03:28.640 Okay. Who's leading the charge on this?

01:03:30.920 I don't know if there's one person who's leading it. I think James Heathers, who's now got a position,

01:03:36.360 part of the challenge with a lot of these so-called data sleuths is it's hard to get paid to do that.

01:03:41.260 So I get paid to be, I was a paid dean and now I'm a paid center director and kind of in my spare

01:03:46.440 time, I do a little sleuthing. People like James Heathers, where it's more his full-time gig,

01:03:51.520 it's hard to get paid. But Retraction Watch set something up for him. See, he's one. Tracy

01:03:56.180 Weisberger over in Europe is another. There's a woman whose name I'm fortunately going to mispronounce, 1.00

01:04:02.160 but who's a Dutch scientist who came up with something called StatCheck. And StatCheck,

01:04:07.920 if the statistics are reported in the format of the American Psychological Association, which is a

01:04:12.220 very clear format, has software that will specifically go and make sure those all check.

01:04:17.120 So those are some examples, but there are many others that people are working on.

01:04:21.040 What do you think is the most compelling piece of epidemiologic data against the idea of exceeding

01:04:27.260 the RDA? I don't think there are any compelling observational epidemiologic data. When I think of

01:04:33.820 the things where you've got relatively hard endpoints, relatively large sample sizes, usually they've not

01:04:39.720 looked at hard, big thresholds. And even if they have, they've looked more continuously at protein

01:04:45.040 intake. Even if they have, I can show you studies in either direction, studies that make it look like

01:04:50.500 more protein is beneficial and studies that look the other way. And I don't think any of them are

01:04:55.020 dispositive. I think protein intake is highly confounded with social class as well as type of

01:05:01.520 protein intake. Let's flip the question now again, which is, okay, I'm going to argue that lower

01:05:07.520 protein is better because I've just demonstrated you are not going to starve to death at, let's just

01:05:14.300 round up and call it one gram per kilogram of body weight. Peter, you should be eating 85 grams because

01:05:21.460 I know that that is safe. But I don't know that if you double that, that you're not going to get

01:05:27.900 cancer. So what can we say about quote unquote high protein diets and cancer or heart disease?

01:05:34.920 So what I think we can say is absolute knowledge is beyond us in this context, but we can have

01:05:41.880 reasonable degrees of certainty, certainty for practical purposes, as long as we need to be open

01:05:47.240 with them. You know, one of the things that I often get frustrated with the nutrition science and

01:05:51.360 some other elements of scientific community is we say science is always evolving and the public needs

01:05:57.700 to understand that it's not something wrong when we, in their view, flip-flop and we change our mind.

01:06:03.480 We change our mind when new data become available. Well, that's true. But then the implication is we

01:06:09.420 were honest with the public all along and saying, this is what we think today. It's not absolutely

01:06:13.800 certain. And the one thing that always just stands out in my mind is early 1990s, Michael Jacobson,

01:06:19.700 Dr. Michael Jacobson, PhD, nutrition scientist, director of something called Center for Science in the

01:06:25.920 Public Interest, some word science in it, coming out on national TV on a camera, holding a plate of

01:06:32.020 fettuccine Alfredo outside an Italian restaurant after a new report had come out on composition of

01:06:36.980 what people ate in Italian restaurants, got huge press. And as he holds it out to the camera, he says,

01:06:43.440 this is a heart attack on a plate. This is a heart attack on a plate is not a tempered statement that

01:06:50.140 says, this is what I think we know today, but it could change later. So I think we need to be more

01:06:54.780 honest about that. What was he referring to that was causing the heart attack in that?

01:06:59.080 I think the implication was the saturated fat or maybe the saturated fat and sodium. So in any case,

01:07:04.380 I think we need to do better with that. I know of no compelling evidence for harm. That doesn't

01:07:13.120 mean there couldn't be any, but I know of no compelling evidence for harm. I know of no studies

01:07:17.840 showing that humans get more cancer with this. I think we need to be skeptical of the mouse studies,

01:07:23.600 the epidemiologic studies for reasons I've indicated. I think at some point we need to

01:07:29.080 recognize as people did when I put up that LinkedIn post and I said, show me the data. Can anybody send

01:07:35.100 me one? Many people responded with something akin to, I can't show you that study, David,

01:07:41.920 but can you show me the compliment or the opposite study that meets all your criteria and shows there

01:07:48.480 isn't harm? And the answer may be no, I'm not sure. But what I can say is at some point when you've

01:07:56.900 looked enough and you failed to see harm, where's the burden of proof? So if you don't have a really

01:08:03.540 strong a priori rationale, now we can argue about that. There probably is legitimate debate.

01:08:08.640 You might legitimately say, I have a strong a priori rationale, there isn't harm. And Dudley

01:08:12.940 Lambing might say, I have a strong a priori rationale, there is harm. You're entitled to your opinion.

01:08:17.220 Those are opinions. And then we can both look and say, well, neither one of you has the absolute

01:08:21.900 definitive data. At what point, where's the burden of proof? And I think right now the notion is the

01:08:28.220 burden of proof is on those who want to argue for higher protein intake. Well, the evidence isn't

01:08:32.940 clear enough to show that the RDA is too low, to which I would reply, well, the evidence isn't clear

01:08:38.080 that it isn't. Now, where's the needle point? Where does prudence point? And I don't think the status

01:08:45.300 quo is necessarily what's always prudent. Saying leave everything the way it is, is always the

01:08:49.980 best way. Sometimes it is. But in this case, I think we've left it alone long enough. I think

01:08:54.420 it's time to say, we've looked and looked and looked. If I look at a thousand swans and I cannot

01:08:59.680 find a black swan, does that mean that no black swans exist? Of course not. But if I looked at 10,000

01:09:06.840 black swans, if I sent a helicopter in the air and we scoured and we looked with binoculars,

01:09:13.360 if I sent teams of undergraduate students to walk around ponds and measure, look for every black 1.00

01:09:19.940 swan, if I put drones with cameras out and I have failed time after time after time to find a black 1.00

01:09:26.880 swan, at one point say, maybe for practical purposes, I can say black swans probably are not

01:09:33.100 something we need to worry about. Yeah. I think that is probably the most logical way to frame this,

01:09:40.040 which is, I do not believe we're ever going to get a dose toxicity study for protein the way we do

01:09:47.460 for figuring out what the LD50 of a drug is, where you push the toxicity and you figure out, okay,

01:09:54.280 at this dose, we will kill 50% of people. So the real question becomes, what would be the bracket

01:10:02.380 you would put around for 90% of the population to exist within this range is an appropriate way 1.00

01:10:11.460 to interact at the grocery store? To your point, what's the problem we're trying to solve here?

01:10:17.120 The people listening to us don't care about most of what we've said today. I mean, we had to say it,

01:10:23.320 but they actually just want to know, look, man, I'm really confused because I'm reading people

01:10:28.800 who have a lot of piss and vinegar in what they're saying on this topic. How much protein should I be

01:10:34.840 eating? What should my family be eating? Should I be avoiding protein in my kids? Should they not be

01:10:40.100 eating snacks with protein? Whatever the argument is. So our guidance to our patients is pretty

01:10:45.760 straightforward and it varies based on their preference. We have some patients who are vegetarians,

01:10:52.140 who have been lifelong vegetarians, who can't stand the feel of meat. This is not like a belief

01:10:59.900 that they have that meat is bad for them, but they genuinely don't like meat. We have other patients

01:11:03.660 who won't eat any animal products. So they're going to have a harder time reaching the upper limits of

01:11:09.740 protein consumption. And so with those people, we're just trying to nudge them as high as we can get

01:11:14.340 them. And that's probably not going to get much higher than about 1.2 grams per kilogram of body

01:11:19.080 weight. But the guidance we're giving people is we really like to see you at about 1.6 to 2. It's

01:11:24.700 an easy heuristic. It's easy to remember too, because you're basically consuming almost a gram

01:11:29.320 per pound of body weight. So how would you advise people based on the, tell me the, for all intents

01:11:34.920 and purposes, you don't need to worry about black swans argument. So let's separate the idea of

01:11:40.560 what I think people ought to do if they have these goals that most of us have, like live longer,

01:11:46.540 be stronger, be healthy, et cetera. How do I think we should tell them to achieve that? The latter one,

01:11:52.660 I'll loop back to in a minute, but I'm not really the expert on that. But the former one, what I would

01:11:58.140 say is, I think you and I are largely aligned on this. I would say, if you just want to survive, or if

01:12:05.640 you just want to survive, the RDA is probably okay for most people. But if you want to thrive in these

01:12:12.460 goals that most of us share, then I would aim for in the neighborhood of two grams per kilogram

01:12:19.400 per day, per person, spaced out throughout the day.

01:12:24.280 So for example, if you want to pass, you should study this many hours per day and you will get a C.

01:12:32.020 Right.

01:12:32.740 But if you want to have the best shot at getting into the best college you could get into,

01:12:37.880 because you want to study engineering, you should probably study this much and you're probably

01:12:43.700 going to need to try to get A's.

01:12:45.640 Great analogy. My uncle, my great uncle, who was a professor from the old country and the old school

01:12:52.400 and a philosophy professor, when his kids would come home in America from school, he would say,

01:12:59.300 good, now playtime's over. Now you sit down and we study Latin and math and logic and you do the real

01:13:06.400 work. And then there's probably yet another example. You may know that in South Korea,

01:13:11.740 they actually had to pass some laws recently or chose to pass some laws recently that I forget

01:13:15.520 the details of them, but that students could only study so much because they're really starting to

01:13:19.900 get worried of students just going too far. And I think those are good examples. To me,

01:13:24.140 it's sort of like the RDA for protein intake is like what my uncle saw the American schools as.

01:13:29.920 It's a, yes, it's just enough to pass high school and do something.

01:13:33.380 The next level up is what my uncle did, which was we're going to really do some work and get

01:13:39.040 you closer to an optimum level. Maybe what some students do is maybe go too far. But I also think

01:13:44.480 it's interesting to see in what way were they going too far? Did anybody ever say or show that

01:13:50.760 if you study algebra or anything else, 12 hours a day instead of two hours a day, it directly causes

01:13:58.800 harm. And I don't know of anything like that. Now, if you said to me, if you study algebra or

01:14:06.820 whatever for 12 hours a day, ipso facto, you are not exercising, you are not socializing,

01:14:13.820 you may be not sleeping enough, et cetera, then you have a problem. It's not the direct effect of

01:14:19.680 the studying.

01:14:20.960 It's the substitution effect.

01:14:22.040 Right. So if I literally drank, I handed you a drink when we came in, one of my favorite playtime

01:14:28.140 things, I love to play with these different protein products and have fun with them. And this

01:14:32.100 is a drink that has 20 grams of protein and 90 calories, almost pure protein. And there are other

01:14:38.820 things like that. If you or I were to say, I just want as much protein as I can, so I'm just going to

01:14:43.780 live on that stuff. There would be no, to my knowledge, no direct harm. I'm not worried about my

01:14:49.600 kidneys shutting down or something. But I would say, well, wait a minute, did you have any vitamins

01:14:53.960 and minerals? Did you have any pleasure from eating other foods? Did you have enough energy

01:14:58.800 to work out hard by not having any carbohydrate? Maybe a little carbohydrate would make you work,

01:15:03.740 able to work a little harder. So I think there are other losses, but beyond those other losses,

01:15:09.340 or if you get a fatty acid deficiency because you haven't had any linoleic acid, beyond the losses,

01:15:14.100 I don't know of any risk. So to me, I look at it and say, life support, basic maintenance,

01:15:19.600 RDA. Strong evidence for thriving, two-ish. Above two, probably more benefit, but you're starting to

01:15:28.180 hit the asymptote. I don't know that it's going to come down. I don't know of anything that says that

01:15:32.600 it's not going to be monotonic, that's going to turn around, but diminishing benefit and then

01:15:38.000 starting to get into more costs of economic costs of buying fancy products, of the costs of your time

01:15:44.600 and attention on it, the costs of not eating something else you might like. I know you and

01:15:49.100 I have both, I don't know, struggled with is the right word, but wrangled with fruit. We both like

01:15:54.320 fruit and we've both played with diets at times where we've minimized the consumption of fruit for

01:15:59.140 other goals. And yet both perhaps come back to it a little bit and said, I don't want to give up

01:16:03.520 fruit. And those are examples, I think, where there's some optimization, but I see no harm.

01:16:07.640 And I think the more refined your goals are, then the more it's reasonable to push it a little bit.

01:16:14.060 If I wanted to win the Olympics, I'm more motivated to push it. For me, if I can lift

01:16:19.780 one more pound of weight on the bench press, who cares? But if I'm trying to win the Olympics,

01:16:24.740 then it matters. So let's pivot from here into the extension, the logical extension, I think,

01:16:30.500 of where we're going, which is not just as it pertains to protein, but to a broader discussion

01:16:34.600 around processed foods. So this is also a very, very hot topic today. So I want to just talk about

01:16:40.120 it broadly, and then we can talk about protein as a subset of this, because obviously a lot of

01:16:44.760 processed foods are optimized around protein and not just pleasure. But I will say I have read

01:16:51.160 quite a bit on this topic, and I've read some pretty compelling arguments on all sides,

01:16:57.060 that if you just took processed and ultra-processed foods off the market, people would be better.

01:17:03.420 You would force a change in the system that would lead to healthier outcomes. And of course,

01:17:08.160 I've read other very compelling arguments that say, look, if you actually correct for caloric

01:17:12.840 intake, there's nothing per se that is wrong with a processed food, at least to the first order,

01:17:20.460 second order, potentially. Let's maybe start with just some of the definitional stuff.

01:17:24.680 What separates a processed food from an ultra-processed food?

01:17:27.520 There is not a single accepted definition. The most commonly used classification system

01:17:33.040 is called NOVA. It's very controversial for many reasons, some social, some scientific,

01:17:39.020 some linguistic or definitional. It has to do with degrees of steps and the types of steps

01:17:45.260 and so on involved. These are very popular topics because they give us a new demon. We talked earlier

01:17:51.820 about the importance of demons. Kelly Brownell, who as much as anybody has been one of the leading

01:17:57.580 thinkers in the field of obesity research in the last half of the 20th century.

01:18:03.160 Is he at Yale still?

01:18:04.420 No, no. He moved to Duke years ago, and I think he may be semi-retired now. But anyway,

01:18:09.560 really good thinker. And I can remember being in a meeting with Kelly. He and I were both speakers.

01:18:15.040 The meeting was convened by leaders in the food industry. They were inviting him,

01:18:18.980 and this is just as he was ramping up the rhetoric of toxic environment, toxic food environment,

01:18:24.900 the epidemiologic environment, the public health environment argument. This was when it had just

01:18:30.000 begun emerging. The NHANES-3 and then the later NHANES data was suddenly this wake-up call for the

01:18:37.040 country. Yeah, we knew there was obesity and we knew it was getting worse. We didn't realize how rapidly

01:18:42.420 it had started to get worse in the last couple of decades. And now there was this hype, this panic.

01:18:50.000 And Kelly spoke and he said, we need social movement here. He said, we can't do it on our

01:18:56.100 own. I've been the behavioral psychologist type. He himself struggles with obesity. It's not education.

01:19:01.440 The guy's brilliant. He's enormously educated, but he still struggled, as did many other people

01:19:06.340 of a similar degree of education and expertise. And he said, we're treating this too much as an

01:19:11.920 individual problem. He said, you've got companies like McDonald's who have a goal of nobody should

01:19:17.540 ever be more than X minutes away from McDonald's if they're driving in the United States. And he said,

01:19:22.020 that's a problem. It's a problem when I'm driving down the street and I'm being assaulted by all the

01:19:26.120 signage and so on. And he said, we need to change this. We need a social movement.

01:19:31.320 Now he's looking at all these executives from the food industry. And he says, history shows us over

01:19:38.440 and over that social change happens when there's a villain. He says, we need a villain. And he says,

01:19:44.900 guess who it is? You. So that was the start of this villainization.

01:19:50.180 What year was that?

01:19:51.120 I don't remember exactly, but I would say it was mid nineties. And it was Michael Mudd,

01:19:55.780 who was from Kraft at the time, convened that meeting. Good meeting.

01:19:59.500 And he's like, I invited this guy.

01:20:02.320 Michael had courage. Anyway, so that was that start of it. And now ultra process is just a

01:20:07.200 great way to demonize stuff. It's just another demon. We're off of soybean oil, maybe phytoestrogens.

01:20:13.500 We dealt with fat for a while. We dealt with sugar for a while. Those are still floating around.

01:20:18.380 Let's give folks a couple of examples of processed versus ultra process. So like dried fruit is

01:20:23.600 processed.

01:20:24.220 Sure. Virtually everything we eat is processed.

01:20:26.160 Yeah. Almost everything you eat is processed. But how do you then cross the chasm to ultra

01:20:30.200 processed?

01:20:31.540 Even if you say there's a single definition like Nova, I can't recite for you the exact criteria,

01:20:36.340 but it's the number of steps. It's the degrees of steps. It's the types of steps.

01:20:40.500 Most things that come in a package are now viewed as ultra processed.

01:20:44.540 That's right. Number of greens, number of steps. But you're right. Dried fruit,

01:20:48.580 even fruit that's cut up is processed. Wine is processed. Cheese is processed. Milk is homogenized.

01:20:54.920 Sometimes it's pasteurized, hopefully. Almost everything we eat is processed. And for that

01:21:00.920 matter, I don't want to eat a lot of certain unprocessed foods. There's lots of evidence that

01:21:04.880 unprocessed dairy products cause a lot of harm. So processing is good.

01:21:10.880 Now let's talk about the state of the evidence. Because again, a good story goes a long way.

01:21:16.240 I mean, our ancestors, and to be clear, my view on this is that the story is much more nuanced than

01:21:22.140 the one I'm about to lay out. But the one I'm about to lay out makes sense. And I get it.

01:21:25.960 We didn't evolve eating ultra processed foods, let alone processed foods. And ultra processed foods are

01:21:33.560 engineered to be highly palatable. And part of that engineering process also makes them calorie dense

01:21:42.560 because part of making things hyper palatable is putting in a lot of sugar and a lot of fat.

01:21:48.040 So it's not that the food companies who make these things are trying to make us fat. They're not

01:21:53.960 trying to hurt us. Marlboro is not trying to hurt you with cigarettes. They just want you to smoke

01:21:59.260 them forever. It's an unfortunate consequence of the product. So they really want to create something

01:22:04.800 that tastes remarkable that you just want to keep buying over and over again. And the problem is

01:22:09.660 you're going to end up eating more of it in terms of calories because of the very nature of the

01:22:14.040 product they're trying to sell you. So clearly this is a problem. We can't have these foods around

01:22:19.660 because we can't eat them in moderation. We're going to overeat them. This is one thing we can

01:22:25.980 agree on. Overconsumption of calories always leads to bad things relative to what your needs are. So

01:22:31.840 again, that number is variable by individual, but for any given individual, eating more than they require

01:22:37.320 leads to physiologic harm. So by that logic, why are we having this discussion? Why don't we just get

01:22:43.160 rid of ultra-processed foods? Wow, you've given me so much to work with. One thread I just want to

01:22:48.500 throw out quickly, we may or may not have time and interest in coming back to it, is the idea that

01:22:54.520 fewer calories are better, at least to a point. To a point is true, but one of the things particularly

01:22:59.880 with protein is we look at some of these studies that people are citing. So I realize I'm sort of

01:23:05.640 looping back to our prior conversation, but these are from the mouse looking at longevity. And it's

01:23:10.820 interesting, they're very dependent on ambient temperature. So the studies of caloric restriction

01:23:15.720 in general, and protein restriction in particular, which show benefit, are much more present at 22

01:23:23.280 degrees Celsius, which for a mouse is a thermogenic challenge, thermoregulatory challenge, as opposed to

01:23:29.500 thermoneutral conditions, roughly 27 to 30 degrees Celsius. We don't live our own lives in the chronic cold.

01:23:38.240 So we'll leave that out there and say it depends on the conditions.

01:23:43.200 There's several threads you've allowed me to pick up here. One has to do with the idea of categories.

01:23:49.080 All categories are social constructs. We often hear that said about things like race, and it's true, 1.00

01:23:54.240 race is a social construct. So is furniture, so is vaccine, so is medicine, etc. What you choose to

01:24:02.320 use as a category and how you choose to define membership in that category is a judgment call.

01:24:08.340 And it's not intrinsically written in stone what that should be. They're useful or not useful for

01:24:13.660 your purposes. That's the first thing to point out. So same thing as ultra process. There's not a correct

01:24:18.340 definition of ultra process. You can define it any way you want. The words mean what we say they mean.

01:24:23.460 Once you've defined it clearly, then we can say, what's its value? How can we use it? What's its utility?

01:24:28.400 The second thing I would say is, as with any category, if it permits a lot of variability in

01:24:34.100 that category, which ultra process certainly does, it starts to get a little silly to talk about the

01:24:39.220 overall category because you say, so wait a minute, you want me to consider a meal replacement shake

01:24:46.800 in the same category as I consider a big gulp and a 7-Eleven. In the same category, I consider a

01:24:54.240 chocolate bar. In the same category, I consider a TPN nutrition. These are all arguably ultra process

01:25:02.060 and very different things. And I think we want to talk about the things. The next thing I want to bring

01:25:08.200 up is what are we trying to do with it? And this is probably intellectually most important. If you say

01:25:14.620 to me, David, what I am looking for is a heuristic, and this loops back to something I said we'd address

01:25:19.860 later, or earlier I mentioned, I said we'd come back to it. The distinction between how much protein

01:25:25.180 do I think somebody ought to eat for a certain goal versus how would I tell them to eat it or

01:25:31.400 what would I tell them about it? And those are two different things. If you said, David, I just want

01:25:36.160 something I can tell people so that when I tell them this, it has a beneficial effect. Does telling

01:25:42.280 them not to eat ultra process foods or to eat as few of them as possible, defining ultra process in

01:25:47.980 this way, help? And the answer is, it might help a lot. It might vary a little bit depending on how

01:25:53.720 you said it. We need more studies. We need to figure out, my guess is as with everything else,

01:25:58.340 it won't last long. So you could tell them not to drink sugar-sweetened beverages. You could tell

01:26:02.500 them not to eat fettuccine Alfredo because it's a heart attack on a plate. All these things have a

01:26:07.940 small effect for a while and usually not so much for a long term, but it might help. We need to study

01:26:13.380 that. That says nothing, however, about the effects of the food per se. If you said, no, David,

01:26:20.320 my goal is to tell people about the foods they should eat if they actually ate them, what the

01:26:28.260 effects would be. Or my goal is to determine the effects of foods. Then I would go back to a

01:26:34.260 statement from a wonderful book called A Fly in the Ointment by someone named Joe Schwartz. And Dr.

01:26:41.800 Schwartz, who's a food scientist, says something like, there is a motto. Repeat after me. The effect

01:26:49.980 of substances in the body depends on their molecular structure, not their ancestry. So if you give me

01:26:59.280 this molecule or a collection of them to eat and you extracted that molecule from some berry and it's

01:27:08.160 natural or you synthesize that molecule in a laboratory, but in the end, it's the same molecule. We agree. I mean,

01:27:16.420 you could try to synthesize, it might have some slight difference, but let's assume it's literally the same

01:27:20.820 molecule and it's the same structure. You give it to me in liquid form and liquid form or gaseous form or whatever

01:27:26.440 it is. If you say these are going to have different effects because of where they came from, seems to me we're in

01:27:32.500 homeopathy now. This makes no sense. My favorite example of this is natural sugar versus processed

01:27:40.900 sugar, which is, I don't know if the people who say this are ignorant of what fructose and glucose are

01:27:49.060 or if it's deliberate marketing shenanigans. I think it's a mixture of deliberate marketing

01:27:55.100 shenanigans, but it's also the marketing of ideas, not just food products, by people with particular

01:28:01.780 philosophical and other bents who are anywhere from the interest to make themselves famous to push a 0.98

01:28:09.140 philosophical thing, to push an anti-industry thing, whatever it is. But yes, I agree with that.

01:28:14.980 And there's so many others, natural vanillin versus synthetic vanillin. If it's still vanillin,

01:28:20.820 it's vanillin. There's a wonderful book by Alan Levinowitz called Natural, in which he will blow away

01:28:26.900 every common conception of what natural is and what its value is. And he talks about the idea of,

01:28:33.580 you know, lots of people want the natural vanilla. Most of the vanilla flavoring we get in this country

01:28:38.820 is not so-called natural from the vanilla plant. But if it was, it would probably have a much worse

01:28:45.340 environmental impact than the other sources. So it's not always so simple. In any case, I think Schwartz

01:28:52.900 is right. It's not the ancestry. So whether you give me the molecule and it was locally grown

01:28:58.080 or not locally grown, organic or not organic, ultra-processed or not ultra-processed, it's the molecules

01:29:06.220 and their structure that matter. Conditional upon the molecules and the structure, doesn't matter where

01:29:12.340 it came from.

01:29:13.660 But in defense of the argument that ultra-processed foods must be worse, if you look at the ingredient

01:29:20.520 list, David, the sheer number of molecules there would suggest, we're playing Russian roulette here,

01:29:28.000 I don't recognize half the names of the things on the bag of Doritos. I'm making that up. I haven't

01:29:35.640 looked at a bag of Doritos in a while.

01:29:37.100 Corn, corn oil, salt and spices. Maybe not a great example.

01:29:41.000 You could certainly find an ultra-processed food at the grocery store in which you cannot

01:29:45.180 comprehend 50% of what's in it.

01:29:47.520 That's right.

01:29:47.860 And you don't really know the dose either because the only thing that the FDA requires is that you

01:29:53.320 list them in order of abundance. But it could be that the first one represents 99% of it and the

01:30:00.500 other 12 represent 1% of it. And even amongst that, there's an uneven distribution, et cetera,

01:30:05.880 et cetera. And it could be that these are just preservatives and color additives and they have no

01:30:09.800 physical barrier, but you just don't know. Point is, when I eat an apple, or even if I eat a

01:30:14.880 processed apple in the sense that it's been pre-cut up or it's just a dried apple that's

01:30:19.260 dried apple chips, where I can look at the ingredient list and it says, apples, I got to be safer than if

01:30:24.420 I'm eating, come on, 20 things of which I can't pronounce 13 of them.

01:30:28.880 If that's the logic of your thinking, I have a great product I'd like to sell you.

01:30:32.440 It's gluten-free, it's seed oil-free, it's not ultra-processed, it's free of any harmful,

01:30:39.800 thing. It's chemical-free, it has no chemicals in it. I call it vacuum. And I would like to sell

01:30:45.440 you this vacuum. And basically, that means nothing, because we are chemicals. As my friend

01:30:50.680 Ferg Clydesdale, who's the former head of the nutrition food science department at UMass Amherst,

01:30:57.660 used to say, the whole purpose of eating is to get chemicals into the body, to replace the chemicals

01:31:03.500 the body loses through the process of living. All food is chemicals, we are chemicals. When you eat

01:31:09.360 that apple and you say, I understand it, that's apple, unless you have a lot of chemical knowledge

01:31:14.900 I don't have and most people don't have, you don't understand that any better than you understand

01:31:19.520 something else that says benzoate phosphate or what have you in it. You just think you do because

01:31:25.800 you think you understand what an apple is at a chemical level. You understand what it is at a

01:31:30.460 fruit level, maybe, but not at a chemical level. There are many chemicals in an apple and an orange

01:31:36.420 that you or I couldn't pronounce. And that if someone wrote out what the chemicals are,

01:31:41.200 we would say, what is that scary sounding thing? And we also know that things that we think of as

01:31:46.980 natural can be just as harmful. Fox glove, hemlock, Socrates killed by being forced to drink all

01:31:54.600 natural hemlock. It was all natural, very harmful. So poisons, drugs often come from natural things.

01:32:02.040 There's also a misperception that what we think of as natural is somehow has been around for

01:32:08.420 thousands of years. And in some cases it's true. In many cases it's not. So the oranges and the apples

01:32:15.440 and the grains that you're eating today were largely not around years ago. They've been bred,

01:32:23.160 even if it's not transgenic, they've been bred by ordinary breeding things. The cows, the chickens,

01:32:29.080 the pigs have been bred to be different. They are not indigenous species. All the chicken,

01:32:35.780 all the cow, all the pig that we eat in this country, the soybeans, none of that's indigenous. 1.00

01:32:42.520 They're all invasive species. Turkey, that's indigenous to North America. Pecans and black

01:32:47.840 walnuts, but not the other walnuts that they love so much in the Mediterranean diet over there.

01:32:51.680 They're not American. I think this is all silliness. The only thing that's artificial here is our 1.00

01:32:56.460 creation of these categories. And we should just recognize that we create the categories.

01:33:00.840 Let's make them meaningful and useful. So back to ultra processed foods. If you said to me,

01:33:06.180 I'm just looking to give my patients or my friends or my mom or myself a hint, a heuristic.

01:33:13.580 There are lots of heuristics that don't- Don't require that classification.

01:33:17.160 Right. But there are also lots of heuristics that work, even though there may be some illogic

01:33:23.040 embedded in them or some incompleteness or some variability. If you said to me, for example,

01:33:28.580 don't talk to people on the street who look like this. If I was your kid and you said, look,

01:33:33.500 you're walking down the street, you're in a big city. Someone comes up, asks you if you have a match,

01:33:37.660 just keep walking. It's probably a good heuristic. They probably don't really want to match,

01:33:41.400 but some might, and you might miss some, but it's not a bad way to stay safe. Similarly,

01:33:46.960 if I said to you, you know what? Don't eat anything from the center aisles at the grocery

01:33:52.260 store or as little as you can. Get as much as you can from the periphery, the dairy, the produce,

01:33:58.080 the meats, the fish, and so on, you might be better off. You will wind up eating less ultra

01:34:02.640 processed foods. You'll wind up eating less energy dense foods in many cases, et cetera, et cetera.

01:34:08.480 But that doesn't mean there's something intrinsic about the periphery of the aisles.

01:34:12.700 It doesn't mean if I took Twinkies and put them in the periphery and I took fat-free Greek yogurt

01:34:17.500 and I put it in the middle, that suddenly it becomes bad and Twinkies become good. It's a

01:34:22.820 heuristic and it might work for you to the extent you can stick to it and to the extent it's correlated

01:34:27.380 with these things, it might work. But it says nothing about the causal effect of being in the

01:34:32.460 periphery or the central aisle. And I think as long as we take ultra processed foods at that level,

01:34:37.040 hey, power to you. In the same sense, as if we were talking to that person back with the protein,

01:34:42.720 you said, how do you explain it to them? Maybe you don't say eat two grams per kilogram

01:34:46.680 because that's too hard. You have to use the metric system and Americans hate the metric system and

01:34:51.260 you have to do math and so on and you have to count stuff. But if you just said, make sure you have

01:34:56.580 two servings of lean fish per week and one serving of lean this and eat some egg whites and so on,

01:35:04.180 maybe that works. And we all have our own foods. During World War II, they tried to get Margaret 0.98

01:35:09.040 Mead, they hired Margaret Mead, the federal government, to try to get people to eat more

01:35:12.660 organ meats because they wanted to send the steak and all to the soldiers overseas. Didn't work.

01:35:18.560 Even Margaret Mead couldn't get people to eat a lot more organ meats. I happen to love organ meats

01:35:22.620 and some of my go-to protein sources are chicken gizzards, fantastic protein to calorie ratio and

01:35:28.980 I love them. A lot of people look at that and go, no, we need to figure out the way to get people

01:35:34.260 to do this, but it's the food that matters, not this. Now, if we go back to those idea of wanting

01:35:41.540 to understand the food, the effects of the food, then I think working with ultra processed food is 0.85

01:35:47.300 just silly. I think it's not a meaningful category. I think any attempts to say, let's find the right

01:35:53.040 definition. I wouldn't even bother. I don't even think it's worth even discussing. I would say is,

01:35:58.200 let's talk about the substances in the food. What is the effect of eating dried apples? Or what is

01:36:04.260 the effect of eating things with this composition? Or what is the effect of eating bottled wine or

01:36:09.820 dried dates or grilled chicken gizzards? That's, I think, where the knowledge is.

01:36:15.380 I want to continue on this point a little bit more from a public health perspective. I think we could

01:36:20.480 sit here and convince ourselves, I think quite easily, and everybody else, that at the individual

01:36:26.640 level, the heuristic of don't eat ultra processed foods is simply not helpful. It's too coarse a tool

01:36:35.680 to parse out too nuanced of a subject is the bottom line, right?

01:36:41.100 I'll go 10%, some percent of the way with you on that. I would say it's a very suboptimal tool.

01:36:47.040 There are probably better tools. There certainly could be better tools. I don't think it would be

01:36:52.220 very effective, and I don't think it would be effective for the long term.

01:36:55.780 You're talking at the level of the individual or at society? Because I want to distinguish these.

01:36:59.600 Either one. I think at the heuristic level. And what I mean by that, again, is nothing to do with

01:37:04.700 causation or really understanding the foods. If you said, give me a rule, and it's different for you.

01:37:10.180 You're very sophisticated. You have a lot of knowledge. If you were to say to my dad when he was alive,

01:37:15.020 who was a smart man, but probably intentionally kept himself not too smart about nutrition so he

01:37:19.860 could feign ignorance and then make the choices he wanted, for him, something like that might work

01:37:25.120 a little bit. It would only work a little until he got sick of it and bored with it, and until good

01:37:30.540 marketers were smart and said, we'll market you things that skirted the definition of ultra processed,

01:37:37.440 so we could say not an ultra processed food, but that still had the same degree of fat,

01:37:43.340 sugar, calories, and whatever. David Kessler has said this nicely in some of his work.

01:37:48.540 He makes the distinction between, I think he says, ultra processed and ultra formulated.

01:37:54.620 And he says, ultra processed, waste of time.

01:37:57.740 Waste of time as a terminology?

01:37:59.920 As a thing we should focus our attention on for helping people. He says, ultra formulated is better

01:38:05.760 because then it talks about what's in the food. Process talks about the process. I don't care how you

01:38:09.900 got there. I want to care where you got to. And so that's one way of making the thinking a little

01:38:14.520 better. Do you think that there are public health solutions to the metabolic situation we're in as

01:38:22.800 a country? If you define public health solution as things that exclude things we really consider

01:38:29.100 clinical, usually like pharmaceuticals and surgery, and you say things where there's palpable

01:38:35.380 and demonstrable success at present, I would say no. It's painful to say that, but I would say that

01:38:43.400 after, depending on your point of view, after roughly 50 years of looking at this, no. I just

01:38:49.340 put up a LinkedIn post about one of the latest papers that came out of this group. I think it was

01:38:54.900 an Australian group that just said, we looked at all the parent training type stuff with kids

01:39:00.400 and obesity and a big meta-analysis of study after study after study of year after year after year.

01:39:05.720 There's nothing there. You could argue that the only area in which public health has changed the

01:39:12.520 course of societal health in this country, smoking cessation seems to be a success. There have been

01:39:18.920 public health measures- Oh, I thought you were only in obesity.

01:39:21.420 I am. I am. I am. I'm now trying to talk about other areas. So I'm saying smoking cessation efforts

01:39:26.260 at the policy level. So excise taxes, advertising laws, rules for where you can smoke, those have

01:39:33.300 appeared to have a significant reduction in the number of people. Seatbelts. Yeah. Great one.

01:39:37.640 So we've got some wins in public health. Why do you think public health has been unsuccessful in this

01:39:46.460 arena? And it can't be for a lack of trying. It can't be for lack of trying at all. It may be for

01:39:54.520 lack of trying in a little more intelligent and unbiased manner. I think there's two really big

01:40:01.380 reasons. One is intrinsic to the problem. One can never start smoking. For some people, as difficult

01:40:07.920 it is, one can abandon it entirely. One can't never start eating and one can't for practical purposes

01:40:14.520 abandon it entirely. So it's a different problem. You've got a very strong intrinsic push there. I mean,

01:40:23.060 it's so linked to our survival. It's a hard goal to manipulate. I think it's intrinsically just very,

01:40:29.120 very difficult. We like our freedom. We like our variety. I don't want to give up my freedom and

01:40:34.260 variety. If somebody were to say, if we eliminated all of these choices or you're only allowed to shop

01:40:40.740 on Monday or you're only allowed to buy X calories, it's like rationing in a war because some people are

01:40:47.740 obese, even though I myself have struggled with my weight, I would say, I don't want to live in that 1.00

01:40:52.680 world. I don't want to give up my freedom of choice, even if it means I and some other people

01:40:57.900 are going to struggle with our weight. So I think there's some real intrinsic difficulties.

01:41:02.720 Everything we do in the public health realm almost seems like either we can't get it to stick

01:41:07.720 or we didn't think through enough, if it stuck, it would work. Or it's like whack-a-mole. That is,

01:41:13.620 you get me to consume fewer calories here or expend more energy there. And then I expend less here or

01:41:20.860 consume more there. So I think that's the intrinsic problem. And then I think the other problem is that

01:41:26.400 we started off with the obvious stuff. We looked for the keys under the lamppost because that's where

01:41:32.440 the light was best. We said, ah, school-based approaches, farmer's markets, walking trails,

01:41:37.860 calories on the menu, so on. Good ideas, all good ideas. Nudge, we tried them. None of them really

01:41:46.240 seemed to work meaningfully. At the beginning, that was fine. What I think we have is a failure

01:41:51.840 of courage, honesty, and creativity to say, we've shown that none of those things have large, demonstrable,

01:42:00.540 meaningful effects. We've not shown that none of them could ever have any effect under any circumstance.

01:42:06.020 But I think for practical purposes, we've shown that those things don't have large, demonstrable,

01:42:11.120 meaningful effects. And I think it's time to stop pretending that they might and proposing

01:42:16.140 and funding the next study that's only a trivial variant on the many, many such studies that went

01:42:22.720 before and were shown to not be successful. I think we need to start taking radically different

01:42:28.220 views and say, the next time someone comes and says, I've got this parent-based or school-based

01:42:32.840 or community-based idea. Tell us how it's radically different than what's gone before,

01:42:38.700 not trivially different. Do you have any radical ideas that if you ignore the challenges or issues

01:42:47.640 around implementation, you would want to implement? I think you asked me this once before and I'm going

01:42:52.620 to change my answer slightly, but not that much. I think before I said two things. If you really want

01:43:00.060 to decrease suffering now for some subset of the population for which we could afford it,

01:43:05.480 make bariatric surgery freely available. And then for investing in research, invest in research on

01:43:12.000 the effects of general quality of upbringing and general education, especially, but not only for

01:43:17.180 women and girls. I'm going to stick with the second one, say we still need that. We need to look at what 1.00

01:43:22.720 are the effects of general education, not nutrition education, general education, general parent training,

01:43:30.500 general security, financial and other kinds of security growing up on obesity levels. That would be

01:43:38.820 my research public health piece. Sorry, what's the hypothesis there?

01:43:43.020 The hypothesis that general education, security so that you're not constantly worried about,

01:43:48.760 can I pay the bills? Will I be able to get food at all? General good parenting, which delivers

01:43:55.380 whatever psychological benefit it delivers. There's, I would say, at minimum circumstantial

01:44:00.900 evidence that all of those things lead to reductions in obesity and diabetes decades later. The two best

01:44:07.760 studies I know of this are what's called the Moving to Opportunity Study, funded by the Department of

01:44:13.100 Housing and Urban Development, and the Abecadarian Study done by the Rameys, Craig and Sharon Ramey.

01:44:20.200 But of those three things, isn't education and financial well-being higher today than it was 50

01:44:27.080 years ago? The third one, parental support is probably less today. There are probably more

01:44:31.900 broken families today, but two of those three are better today than 50 years ago, aren't they?

01:44:37.060 I'm not sure. I don't want to say they're better, but I think there's tremendous disparity.

01:44:40.860 So the quality of the education and security and so on that some people get is very different than

01:44:46.920 others. And I think Confucius famously said, we are not concerned about poverty. We are concerned

01:44:52.920 about differentials of wealth. It's possible that the public health solution to obesity is becoming

01:44:58.200 a Marxist. I'm not sure we all want to volunteer, but I think some of these things are worth looking

01:45:02.940 at. We do see that there seem to be these long-term benefits. In studies, as I said, I mentioned when I

01:45:08.480 think of the two best, but they're not the only ones, where they weren't meant to be primarily

01:45:12.380 nutrition, obesity studies.

01:45:14.780 But how does that explain what you see in the Middle East, for example, where you have countries

01:45:19.340 that are remarkably wealthy, there is no poverty. Obviously, there is disparity in wealth between

01:45:25.040 the very affluent, which is effectively everybody, and then the ultra affluent. Families aren't broken,

01:45:32.140 everybody is educated, and yet obesity and diabetes rates are greater than even in the US.

01:45:36.840 Yeah. I think there's at least two ways to look at that. One way is the idea of interaction.

01:45:42.600 I'm talking about Gulf nations, obviously not.

01:45:44.500 Right. The same exposures may not have the same effects in one place or another. For example,

01:45:49.420 we consistently see that in less developed countries, poorer countries, less industrialized

01:45:55.260 countries, greater wealth is associated with greater obesity. We consistently see in more industrialized,

01:46:02.420 wealthy countries, greater wealth and education, especially among adult women, tends to be 1.00

01:46:08.260 associated with lesser obesity. So you have the interaction. The other is, as you and I had said

01:46:13.480 earlier, the strength of the evidence. One of the things people frequently like to say is when you say,

01:46:17.700 well, we think temperature has this effect. And they say, what about those people in Iceland?

01:46:21.800 They don't have. Well, alcohol is this. Yes, but those people over there, they drink a lot and

01:46:25.640 they don't have. They're different. They're different in many ways. We don't know that any

01:46:30.300 one factor has to explain why Qatar has this versus Samoa or something. But I think it's a good

01:46:36.400 hypothesis. So that's what I would look at at the public health level in what we think of as

01:46:40.220 traditional public health. Now with the GLP-1 agonist-related drugs and some other drugs

01:46:45.280 being as profoundly beneficial as they appear to be, I think we're going to get to the point where

01:46:52.000 it's going to be hard. I think we're already asking the question and I think it's going to be

01:46:55.760 hard not to take the question seriously of should it almost be the default? That is just like we've

01:47:03.260 asked the question and people give different answers, but we asked the question, should it be

01:47:07.420 that the default is you get this vaccine when you're a kid, that you get your teeth fluoridated

01:47:12.700 at the dentist, that it's not, hmm, maybe some people should get that. I think we are at the point

01:47:18.300 where we're saying, as these drugs continue to be tested and experienced, if we continue to see the

01:47:23.880 effects we're getting, are we getting to the point where we should start to say, you know what,

01:47:28.420 that idea of the poly pill that came up decades ago where young adults, even if you don't have

01:47:33.480 diabetes, obesity, hypertension, et cetera, you'd get a low-dose diuretic, low-dose metformin,

01:47:40.000 et cetera, low-dose statin. Are we at the point where we probably still should do that

01:47:44.200 and then say, anybody get a low-dose GLP-1 agonist-related drug? And we'll roll it out

01:47:50.440 and we'll pay for it. And anybody in the country who wants it can get it or almost anybody. I think

01:47:56.060 that may be the future. I'm too hungry having this discussion. So David, thank you for making

01:48:01.340 the trip down. As always, it's a pleasure. Truly my pleasure, Peter. Let's keep doing it.

01:48:06.200 Thank you for listening to this week's episode of The Drive. Head over to

01:48:10.500 peteratiamd.com forward slash show notes. If you want to dig deeper into this episode,

01:48:17.300 you can also find me on YouTube, Instagram, and Twitter, all with the handle peteratiamd.

01:48:22.880 You can also leave us review on Apple podcasts or whatever podcast player you use. This podcast

01:48:29.200 is for general informational purposes only and does not constitute the practice of medicine,

01:48:33.720 nursing, or other professional healthcare services, including the giving of medical advice,

01:48:38.440 no doctor-patient relationship is formed. The use of this information and the materials linked to

01:48:44.580 this podcast is at the user's own risk. The content on this podcast is not intended to be a substitute

01:48:50.420 for professional medical advice, diagnosis, or treatment. Users should not disregard or delay in

01:48:56.000 obtaining medical advice from any medical condition they have, and they should seek the assistance of

01:49:01.120 their healthcare professionals for any such conditions. Finally, I take all conflicts of interest

01:49:06.320 very seriously. For all of my disclosures and the companies I invest in or advise, please visit

01:49:12.220 peteratiamd.com forward slash about where I keep an up-to-date and active list of all disclosures.

The Peter Attia Drive - October 13, 2025

#368 ‒ The protein debate： optimal intake, limitations of the RDA, whether high-protein intake is harmful, and how to think about processed foods ｜ David Allison, Ph.D.

Episode Stats

Summary

Transcript