#112 How To Slow Biological Aging With a Multivitamin, Vegetables, & Omega-3 | Dr. Steve Horvath

0:00Welcome back to the podcast. Today I'm joined by Dr. Steve Horvath, one of the most influential scientists in the biology of aging and a true legend in the field of longevity science. Steve is best known for pioneering the Horvath epigenetic clock, which is a breakthrough that really helped make biological aging measurable through DNA methylation. Before this work, aging was something mostly described through disease, frailty, organ decline, or simply just the passage of time. Steve's work really helped transform aging into something we could

0:33begin to quantify at the molecular level across different tissues, across different disease states, and across interventions. That contribution is hard to overstate. Epigenetic clocks are now central to some of the biggest questions in aging science, whether we can measure the rate at which someone is aging, whether lifestyle or medical interventions can slow that rate, and whether aspects of cellular age can actually be reversed. In this episode, Steve and I get into all of that.

1:04We talk about what biological aging clocks can tell us and what they cannot. This is important because biological age is not just one number. Some epigenetic clocks are more sensitive to inflammation, immune function, metabolic health, smoking history, or long-term stress exposure. Others are better at estimating our mortality risk, disease risk, or the current pace at which someone is aging. And once we establish that foundation, we move into the questions that people actually want answered.

1:35We discuss whether lifestyle changes can reverse age acceleration, whether these clocks can predict when someone will die, and why a younger biological age does not necessarily mean you have added years to your life. We also get into some of those more practical and provocative areas of longevity science, including whether caloric restriction can slow the pace of biological aging, whether omega-3s, vitamin D, or a daily multivitamin can shift aging clocks, what type of exercise appears most effective for slowing epigenetic aging, whether vegetables matter more than exercise in some

2:10methylation data sets, whether red meat shows up as an aging signal, how sleep disruption and social connection appear on biological aging clocks, whether GLP-1 drugs like semaglutide may reverse epigenetic aging signals, how to interpret consumer biological age tests without overreaching, why two epigenetic age tests may even give different answers, whether AI will build better aging clocks, and one of the most fascinating frontiers in the field, partial reprogramming, the possibility that cells may be made

2:44biologically younger without losing their identity. We also talk about what aging clocks miss because even if a clock moves in a favorable direction, that does not mean every single hallmark of aging has been repaired. DNA mutations, telomere attrition, senescent cells, protein damage, and tissue level decline may somewhat still remain, and that distinction matters. A younger biological clock is not the same thing as complete rejuvenation. So this episode is really about scientific precision. Aging clocks are powerful

3:20tools, but they are not crystal balls. They are not death date calculators, and they are not proof that one supplement, diet, or intervention has quote-unquote reversed aging. But if used carefully, they may help us understand which aspects of aging are measurable, which are modifiable, and which interventions are most likely to move the biology in a meaningful direction. Steve Horvath is one of those rare scientists who did not just contribute to the aging field. He helped define it, and I'm very excited to

3:51have him back on the podcast today. Before we begin, I want to point out just a couple of things. First, we have show notes for this episode, which you can find at foundmyfitness.com forward slash episodes. We've put together detailed notes for my conversation with Steve, including the major aging clocks we discuss, what each one measures, and how to interpret them without overreaching. At the bottom of the show notes, you'll also find a brief consumer guide to biological age testing. This will include what to look for in a

4:23test, why it matters which clock is being used, and which test to use if you are interested in mortality prediction, disease prediction, metabolic health, or your rate of aging. Again, you can find all that, including the consumer guide, at foundmyfitness.com forward slash episodes, E-P-I-S-O-D-E-S. You can click on the Steve Horvath episode to find all that info. And lastly, you may have noticed that Found My

4:53Fitness is ad-free. We don't run sponsorships or interrupt these episodes with ads because our goal is to keep the science as objective and independent as possible. That is made possible by listeners like you who directly support the show. If you find value in the evidence-based conversations that you listen to on this podcast, please consider becoming a Found My Fitness premium member. Premium membership directly supports our work, gives you access to exclusive benefits like the aliquot. This is our members-only podcast. Also monthly live and recorded Q&As with me and our curated science digest that we

5:30send out to you twice a month. You can learn more about supporting the show and becoming a Found My Fitness premium member at foundmyfitness.com forward slash premium. Again, that's foundmyfitness.com forward slash P-R-E-M-I-U-M premium. Thank you so much for your support. Now onto the podcast with Dr. Steve Horvath. Just a quick heads up, the first 20 minutes may be a little technical for a few of you because we have to explain and define these epigenetic aging clocks since they measure different

6:03things. But if you stick around, after that, we get into all the practical questions that everyone once answered. Hope you enjoy. Welcome back to the podcast. I am sitting here with Dr. Steve Horvath. Steve, good to see you again. This is the second time you've been on this podcast. You have been incredibly influential in the longevity field. You are the developer of the original Horvath epigenetic aging clock, which has really revolutionized the way the aging field has been able to measure

6:37biological aging. So thanks for coming back on the show. Yeah, thank you. I'm very excited to be here. The science has evolved quite a bit from the last time we spoke. So it's a wonderful opportunity for me to talk to you and your audience. I'm so excited. I mean, the last time we spoke was in 2019. So I would hope that there's been a lot of new, exciting data to discuss. But maybe the way we

7:07could start this is for people who might be new to the field, this idea of biological aging and explaining what biological aging means. It's a good question. Everyone talks about biologic age, but it has so many different definitions. So for many people, biologic age refers to fertility issues, as an example. Broadly, it relates to this phenomenon that people of the same age have

7:40different mortality risks, morbidity risks, or people who you know from your high school, they look older or younger than you. All of that is in that concept of biologic age. However, longevity researchers or geroscientists who study aging really conceptualize biological age using measurement technologies. How do you get a number for measuring biologic age?

8:13And the field has really exploded over the last 13, 14 years. People have developed biologic age measures based on wearables, step counts, gait speed, which is very exciting. Many imaging data, you can measure your brain age based on imaging, for example. My field is in the realm of molecular markers of aging. So I work on epigenetic marks and we can talk

8:48about it later, but I just want to give an overview of the field. There are so many so-called genomic technologies for measuring anything from gene expression, proteome, metabolome, glycom, really any ohm. And for any readout, people have developed clocks, aging measures. I started with DNA methylation back in 2011. We published

9:22our very first epigenetic clocks. And why methylation? Because the signal for aging and even mortality is very strong in methylation. But when you want to measure biologic age, you really need to look at many levels of readouts, molecular, then biochemical readouts, blood biochemistry, various measures of organ function, fibrosis, as an example. And then, of course, above all, functioning measures,

9:56VO2 max, gait speed, and daily living activities. Frailty. Frailty, all of that. It's so important for people to understand that as we have this chronological age, everyone knows their age, right? This is how long you've been alive since the day you were born. And the interesting thing, you talked about biological aging. You know, you have these processes that are happening that affect your daily function. They affect your disease risk. And not everyone has the same disease risk at

10:31the same age. And so there could be this disconnect where some people perhaps, you know, genetic and also lifestyle factors contribute to them not aging quite as good. And so they may get cardiovascular disease earlier or cancer earlier, right? And the opposite is true. And that's what people are really interested in. Well, let's say I'm 50 years old, but I want, you know, the organs in my body and the cells in my body to seem like they're 30 years old, right? Yes. To be younger. And so that's why it's exciting to have these tools that do measure function,

11:01like you mentioned, I think, you know, cardiorespiratory fitness and VO2 max, you know, frailty. And there's a lot of different ways that people are measuring function. But then on the molecular level, that's very exciting because it's quantifying this process of aging. So I'd... I mean, there is one key word that has to be mentioned in that context. It's all about prevention, you know. So what motivated my work was to understand aging in people who do everything

11:31right. For example, in you, why do you age? Given that you and me, we really take care of ourselves and we try to optimize lifestyle prevention, all of that. But something still changes deep inside of us in our cells. And what is it? What drives aging? And these methylation clocks that I've developed, they really track damage accumulation on one way or another, you know, because that is

12:04something that just happens, you know, and that drives then organ dysfunction many years and decades later, you know, but it's almost unavoidable to age. And what I wanted to accomplish with these methylation clocks is to have a precise tool to allow researchers to actually identify novel interventions. How do we truly reverse the age, the ages of individual cells of organs and the whole

12:38organism, you know? You mentioned something that caught my attention. You said these methylation clocks or patterns are able to track the damage that occurs. And that's, to me, always been a question, is it tracking the damage that occurs? And, or if you're changing these patterns, does that change the damage? So because you're globally affecting, you know, the way genes are activated or not activated, gene expression as we call it, then you would imagine, is it like a two-way street possibly where

13:12you're able to increase, you know, genes that we have that help take care of damage, repair, you know, all these, you know, stress response genes better as well. But I guess we'll get into that. So I think one of the points of confusion I've heard repeatedly, you know, from my audience and just from, you know, in general, like out there, is that people think these biological aging clocks are just sort of like one thing. You hear biological age, it's like this one thing. You reverse biological

13:43age, it's just this one thing that's happening. But we actually have very, you know, different clocks that seem to be perhaps tracking or have different, you know, strengths and weaknesses and what they're tracking and what they're sensitive to in the context of the aging process. So maybe we can kind of just, can you walk us through some of these clocks and, you know, what their core strength is, like what they are tracking and perhaps even what a common misconception is in terms of what it's tracking. Yes. So maybe we start with the big picture. Aging is, of course, associated with

14:21the accumulation of damage on all levels, the proteomic level, metabolomic, intercellular communication, but also damage accumulation surrounding the DNA molecule, these chemical changes to the DNA. There is an accumulation of damage and that impairs then the cell function. For example, certain cell identity genes need to be active in a liver cell and different genes need to be acting

14:53in a brain cell and so on. And so the damage impairs the function of cells and then tissues and then organs. And interestingly, methylation changes can be observed at really millions of locations on the DNA molecule. And many of these changes have actually no consequences. And by the way, when I talk about changes on the DNA, I talk about gain of methylation at the wrong places, but also loss of methylation at the

15:29wrong places. So what is happening with aging is that the methylation landscape really flattens out. And conversely, in a young cell, you want to have really peaks of methylation at regions that need to be shut down and conversely low methylation at regions that need to be accessible, you know, on the DNA. So anyways, these methylation clocks look typically at hundreds of locations on the DNA that are carefully

16:03chosen. However, one can really look at tens of millions of locations and people have developed different clocks based on tracking changes at different locations. And one of the great misconceptions is to expect that all clocks disagree, agree with each other, that all clocks give you this one readout. That wouldn't be reasonable, right? Because you have millions of locations. So methylation clocks

16:38really capture, again, different properties of aging. Some clocks are very good at tracking inflammation. Other clocks are very good at metabolic syndrome. Then the so-called, these are now second generation clocks, you know, they really relate to inflammation and various stressors, smoking. But then the earlier generation of clocks, so-called first generation clocks, had a totally different

17:08goal. They just want to measure calendar age, you know. And yeah, so the misconception is that people get disappointed that two different clocks lead to slightly different readouts. But the metaphor I want to use is, think of the word of proteomics. If I told you protein one measures the same as protein two, you would just not believe it. And the same happens in the case of methylation. If you target certain parts of the DNA, they give you a different readout from other parts.

17:41Okay. Yeah, that's really good to kind of clear up. And, you know, I guess if you understand that concept, you wouldn't want all these clocks to be giving you the same readout, because then that would be kind of a problem, I think. It would be overly simplistic. Right. And most clocks were tailor-made for blood, for the sake of convenience. But arguably, you would want to develop special clocks for the brain, for the liver, for the kidney, you know. And the field is moving

18:15in that direction. So people develop actually single-cell clocks and organ-specific clocks. Oh, that's cool. Yeah. So let's talk about some of the main ones that are used. And so we have the one that was, you know, the original Horvath epigenetic aging clock, first generation for chronological age. And then we kind of get into these other clocks, which were first generation, and then they have second generation versions as well. But so the DNA pheno age, does that clock lean more towards inflammatory and metabolic

18:46function than pure chronological age? Yeah, for sure. You know, so the so-called pheno age clock was a giant step forward when it came to mortality risk prediction. This clock was very much constructed to track really biochemical markers and also changes in blood cell composition. These markers measure organ dysfunction one way or another. And the idea was to actually develop a methylation surrogate of these clinical parameters.

19:24And we should discuss the pros and cons of that idea, you know. But yes, so this clock was then impressive mortality risk predictor for humans, you know. However, it was then superseded by the grim age clock, which was named after the grim reaper. It was published also many years ago, 2019. But it continues to be very impressive for mortality risk. And from a mathematical perspective, these clocks were

20:00constructed in very different ways. But overall, they very often agree, you know, when you have an intervention that appears to slow grim age progression, it also slows pheno age progression, you know. So, often there is actually agreement to some extent, which is impressive, given that these clocks were constructed in very different ways. Can you talk a little bit about the different ways they were, I mean, constructed? So, with the grim age,

20:34you know, I think it was my understanding that there's smoking is somehow embedded in that calculation or, you know, whatever you want to call it, stress-related proteins, inflammation. But the pheno age also has some inflammation as well in there. Yes, yes. Yeah, maybe I'll start with a much simpler example. C-reactive protein is a marker of chronic inflammation. It's a very important biochemical readout. And the doctor will measure it when you have

21:09certain conditions. But interestingly, you can actually estimate C-reactive protein levels based on methylation. And I should say the estimate is not very tight. For the experts, I will say correlation may be 0.3 or lower, you know. So, it's not a tight correlation. But I want to mention it as an example for this idea of using methylation to estimate a famous marker. But now, imagine you actually get these two readouts from the lab. Let's say you go to a

21:42longevity clinic. They can easily measure both. But which one is more informative for you? Now, a medical doctor will always focus on the plasma-based readout. They're trained to look at thresholds and then they diagnose maybe an acute infection. But the surprising finding is that the methylation estimate is actually a better predictor of your mortality risk, far better predictor of your mortality risk than the plasma measure. And that's one famous example that has

22:19been validated. I give you now another example, smoking. So, you can ask someone, how many cigarettes do you smoke per week? For how many years have you smoked? And this is known as the smoking pack year estimate of smoking exposure. Interestingly, you can use methylation to estimate smoking exposure as well. And you can ask the same question. Well, which measure is more predictive of how long you end up

22:51living? Is it the self-reported measure or the blood measure? And again, we know the answer from many studies by now. Again, the methylation estimator is actually superior to self-reported. So, I mention it because that then gives rise to an idea. Well, why don't I build a clock that uses these methylation estimators of C-reactive protein, of many other famous proteins, and also the methylation estimator

23:24of smoking history? Why don't I use these methylation biomarkers in a linear combination? I combine it in optimal way to build a mortality risk predictor. And this idea is underlying the grim age clock, the grim reaper clock. And it just worked beautifully in many validation studies. We now know this idea worked. But that's really the idea of the grim age clock.

23:55And this was developed in your lab, the grim age clock? Yes, by Ake Liu in my lab. Pheno-H was developed by Morgan Levine when she was a postdoc in my lab. Morgan was on our podcast a few years back as well. So, why should methylation patterns be able to predict your mortality? I mean, that's... And I mean, how accurate is that? I mean, like, what are we talking about, you know? Yeah. The first question, why does methylation relate to mortality?

24:29It is a good question. Because when I published the very first clocks, I remember I was extremely nervous about what I had published. I thought maybe these clocks have no use. Maybe they just measure your calendar age, you know. And now I was so relieved that basically six weeks after I had published it, somebody came to me at UCLA and said, you know, we just applied your clock and it predicts mortality. But anyways, but then by now we know that these methylation clocks very much

25:05predict mortality risk to the point that certain startups pursue the idea of using methylation clocks for pricing, life insurance policies or financial products, you know. And of course, methylation clocks are used in serious randomized control trials, you know. So, the evidence is very strong and without any debate. But why is that, you know? And it could be that these clocks really track

25:42long-term exposures of a stressor. So, for example, smoking again, you know. So, maybe if you just smoke a bit, it doesn't really show up in other biomarkers. But if you have really this prolonged stressor, it really alters the epigenome. Why? Because the epigenome creates a memory, really. Think of the epigenome as a memory of stressors and it primes the cell to

26:18respond. And you can imagine if the cell senses this onslaught of various stressors, it tries to remodel its regulatory system so that it prepares for future stress. That's maybe one way to conceptualize it. Yeah. And I mean, we even, I think this concept of, you know, the stressor affecting the epigenome, you know, we even know it can affect the epigenome in gonads, right, in sperm and eggs. And that's why certain, we have those studies out of Sweden where they went through these periods of starvation,

26:53like famine, and then there was feasting depending on, you know, the, what food was available at the time. And I know that some researchers had looked at how the epigenome had changed and that also seemed to affect life expectancy of the offspring as well. So, and then smoking. Yes, I think this is a study from the Netherlands, the so-called Dutch hunger, perhaps. Yes. I'm not sure whether that's what you meant. Yes, the Netherlands, yes. Yeah, that's right. Yeah, very exciting work, you know, that maybe a couple of years of starvation could already change the gonad methylome and that could then

27:28lead to changes in the offspring. So, I think it's very exciting, but I need to tell you, I haven't worked in that space. Yeah. Well, as far as, it's been years since I've looked at those studies, but I think there were like prepubescent boys too, where it's like if they have gone through these periods of like hunger where they were calorically restricted, it obviously changed their, you know, their gonads, their epigenome and their sperm in a way that was, you know, more permanent. And so, they had offspring that were like more resilient against type 2 diabetes and,

27:59you know, other age-related diseases as well. So, I think they also lived longer, like their grandkids or something, like it affected their life expectancy as well. But smoking is another one that would also go, it gets, I mean, that's something that goes, you know, deeper, right, and affects the gonads, if I'm correct. I mean, or not. Yeah, I need to tell you regarding these findings, I'm hugely interested and excited about them, but I want to emphasize to your audience, they are controversial, fundamentally. Okay. Very smart people disagree with these findings.

28:31Personally, I'm completely neutral, but I just want you to know that, you know. They disagree that the epigenome changes? Yeah, so that, no, just to be very precise, that an exposure from your parents, for example, has an effect on you. Again, this brilliant people publish on it, so, and these studies go through rigorous peer review, but I just want you to be aware that there are strong counter-arguments, you know, so it remains to be seen, I want to say. Because

29:05I want to think that if your parents or grandparents went through severe stressors, I want to think that you still are born with a clean slate, you know, so that you are, in certain ways, not predisposed or doomed in one way or another, you know, so. Well, sure, that would be a nice thing to think. Yes. But, I mean, on the bright side of things, even if there is, and I've seen, you know,

29:38I've seen evidence that convinced me that there's an epigenetic change that does happen. Yes. And in, you know, sperm DNA, for example, like, if you have an obese male, and then they lose weight, like, you can look at their sperm DNA, and it changes from being obese to lean. And epigenetic changes, gene expression changes are happening. So, but even if it's on the bad side, you know, the good news is that once you're born, you can do things in your life to change things in a positive way too, right? So, it's not like, you know, even if you don't have the cleanest slate.

30:12Yes. Yeah. What I want to tell you briefly about sperm, that is, yes, there are methylation changes and also changes with aging. So, the sperm methylome of a 50-year-old is different from that of a 20-year-old. However, the changes that occur are at completely different location from the changes that we use in any of our other clocks, you know. Another way to say it is, if I take grim age or pheno age or any

30:43of my clocks and apply it to sperm, it completely fails. So, for example, what is known as the Horvath pan-tissue clock? You apply it to sperm, you get one number, 37 or so. But everybody has the same number in essence, you know, uninformative. It's just very different locations. The same statement also holds, by the way, for the placenta. So, people have developed clocks applied to placenta to

31:13estimate the age of the newborn, meaning gestational age, or also various stressors from the mother. But again, these changes are very different from what we observe in blood or adult tissues. Have there been clocks for sperm? Yes. That have been developed that are more precise? Yes. Okay. I have a question about grim age, but it sort of leads us into the next clock that I want to have you discuss, which is the duoden pace, as I say, right?

31:45I think it's called duoden pace. Duoden pace. That's right. Duoden pace. So, the question I have is, you know, with the DNA grim age, we're talking about this methylation pattern being able to predict mortality and very, you know, pretty accurately mortality risk. And it's able to measure, you know, this accumulation of damage that's changed the epigenome in a way that's obviously, you know, quantifiable. So, what if you're 45 years old, you get your DNA grim age test done, it gives you your mortality

32:20because you've had all this, you know, lifetime exposures up until the age of 45 of, you know, let's say air pollution, maybe you smoked a little bit, whatever, alcohol, poor diet, stress, chronic stress, all those sorts of things. But, you know, you change your lifestyle and it gets better. Does that, does that grim age change? If you had asked me that question two years ago, I would have humped and hard, you know, and I was always very cautious about that. In certain ways, how reversible are these changes,

32:56you know, but the science has really advanced. And now I'm confident in saying that you can reverse grim age to some extent. The key word is to some extent, because these changes appear to be very minor. We can talk about it later, but there have been very rigorous randomized control trials with supplements and medications. So, there's a hopeful message. You can reverse it. So, that's what we're going to get into, folks. That's going to be the exciting stuff. So, this, this, this other clock that's able to measure the pace of aging. Can you talk a little

33:32bit about the, can you say? Yeah. Yeah. So, there is another widely used clock, which is known as a Dunedin pace clock. It was developed by Dr. Moffitt and Dan Belsky. And it was constructed in a very different logic from other clocks. And the metaphor is, it's supposed to be an odometer. It's supposed to measure the speed of aging or what they call the pace of aging, whereas previous clocks really

34:06measured in certain ways the accumulation of damage. So, the idea is very compelling. Maybe I'll just review how it was constructed. So, the team really looked at rate of change in established physiologic markers and biochemical markers, including also importantly, and we should discuss that, change in body mass index. So, but also measures of waist to hip ratio. So, also measures of glucose

34:44impairment, markers of inflammation, many readouts. And the study leveraged a unique epidemiologic cohort study in New Zealand, in the city of Dunedin. And so, it's a study where they tracked middle-aged people and younger people for many years and assessed these readouts repeatedly. For the experts, it's a longitudinal studies. And by having these longitudinal data, multiple

35:17measurements per person, they could really estimate the pace, each person's individual trajectory. So far, so good. So, you have these pace measures. But then they went to the next step, which is similar to Grimmage. They said, why don't we use methylation to estimate the pace of aging based on these physiological measures? And I think that's a very good idea. Why? Because people care about

35:49what's my current pace of aging? Actually, it's a good question. I'm not quite sure what people care about. Some people want to know. Let me know where I stand right now. Or how an intervention is affecting how they're aging, perhaps. Exactly. Yes. Thank you. Yeah. That's a good point. You know. So, if you have an intervention, you want to see, does it really change the pace? Does it affect the odometer? You know. And so, therefore, people use Dunedin pace along with all the other clocks that I

36:26mentioned in when they study interventions. It's by now part of the standard repertoire of clocks. When people publish a paper on longevity interventions, they hopefully report about five clocks, I want to say, just in order to give the reader a chance to judge the evidence. Because the very best intervention will touch on many clocks, you know. That would be a robust rejuvenation of the methylone. In my experience from reading the literature,

37:00that's pretty much what I've seen. I see the main clocks that are being used are the Pheno-Age, Grim-Age, perhaps Grim-Age II. Yes. And I see the Duodenian pace. Those are at least three of the ones that they seem to be. And then there's a few others that sometimes are in the mix. But those three stand out to me when I'm reading literature maybe because I know them the best. But those are the ones that stand out. With this disagreement, and you kind of touched on this already, you know, if you're looking at an

37:33intervention, and we're going to get into those in a minute, and you see, you know, your DNA Grim-Age doesn't change. So your mortality risk is the same, or, you know, determining when you're going to die is the same. And yet your rate of aging perhaps slows a little bit. Maybe it's not much. Maybe it's 2%. Some people will look at that and go, oh, these are all like if you're changing your pace of aging, why are you not changing the Grim-Age? And then the question in my mind is, well, how long

38:05was the trial? You know, so if you're changing the pace at which you age by 2% and the trial was six months, is that going to be reflected in the Grim-Age? Or, you know, what's the standard deviation here that we're even talking about with Grim-Age, right? Yes, I think you make a very important point. If you have an intervention that has a very strong effect, I would expect that most of these clocks will show it. Why? Because these clocks are correlated with each other. And just to throw

38:40out a number, correlation 0.5 after you regress out age, sex, and various variables. But there's still a fairly good agreement. This is the typical glass half full, half empty. It's a correlation of 0.5 high or low, you know. To me, it's reasonably high if you have a very strong rejuvenating intervention. Now, when it comes to, I need to tell you, I'm obsessed about the question,

39:12which clock is best? And so I... Best for what? For judging longevity interventions. Okay. Yeah, because when it comes to mortality risk prediction, we know the answer right now after several large studies. There was a study in Scotland, Generation Scotland, 18,000 people were evaluated, and Grim-Age was best. And then there was a study from Harvard, I want to say 30,000 people were evaluated, Grim-Age was best. So we know which clock is best for mortality risk prediction.

39:47Can I pause you right there? Yeah. And I just want to make this point, because usually when we have these studies, at least observational studies looking at diet, because you'll never have a randomized controlled trial that's going to last, you know, 30 years or 40 years. So if you have, if you have, if you're looking at observational data and how different lifestyle like effect or diet and lifestyle affect mortality, you're typically looking at, okay, how much seafood did they eat? How many people died from cardiovascular disease? How many people died from cancer, right? So you get this all-cause

40:18mortality, right? Number. And what you're saying is that you can actually now, instead of having to just have observational data looking at that all-cause mortality, you can now have an intervention. We're going to give people, you know, you know, fish or whatever, we're going to do this intervention for a period of time. And you can have the Grim-Age, which is kind of like this surrogate all-cause mortality, but it's very actually a very good estimate of it. Is that, am I thinking of it a little bit correctly or? Yeah, you think of it correctly and that's certainly the ambition,

40:53but I want to be very precise using the language of the FDA because I think we should do that. So the dream of the longevity field is to develop what is known as a surrogate endpoint for a clinical trial. In other words, you have a study where you apply, let's say, multivitamin for two years and then you see a change in any clock. It could be a proteomic clock, it could be a Grim-Age, any other clock. And let's say you see a reversal.

41:24Now you would like to, I want to call it, jump to the conclusion that this actually translates into a lower mortality risk. And this, so we would like to think that is the case, but from a regulatory perspective, that hasn't been proven. And in general, the FDA evaluates biomarkers. Why? They want to give guidance to companies, to biotech, where they say,

42:00if you show us that your treatment reverses that biomarker, therefore we believe that it actually helps patients, you know. And I just need to tell you and the audience, the biomarker field has not yet developed any biomarker that is credible to the FDA when it comes to this ambition of being an official surrogate endpoint of a clinical trial. Having said this, we just can't wait

42:31for this regulatory approval yet. Why? They are urgent questions, right? People have exciting interventions. So we need to make assumptions, you know. And for the longest time, I've been very cautious when it comes to this claim, do methylation clock meet this high standard, you know. And I'm coming around, you know, just because I see increasing evidence, you know, that these changes

43:04track what I call validated interventions, you know, where we know the intervention has a benefit for human mortality risk. And then I see that it also touches a methylation clock in the expected direction. It gives me confidence, you know, that the clock does what it's supposed to do, you know. So that's where I'm at, you know. What have you, what's the most robust intervention, or it doesn't have to be, you don't have to tell me

43:39what the intervention is, or you could, but what's the most robust data that you've seen in terms of, you know, reversing biologic age by some of these clocks, grim age, pheno age? What's like? Yeah, I will start with interventions that are in certain ways boring to you and me. Why boring to you? You and I, we are hopefully healthy people, and we want to optimize our health. But I want to start with people who have a condition, you know, to answer your question. HIV positive people exhibit

44:16epigenetic age acceleration. It's actually a pronounced pro-aging effect, maybe five to seven years in blood. And sure enough, if they stick to their antiretroviral therapy, that will reverse their epigenetic age. And that, and I mention it because that... For how much? Several years, you know. Several years. To give you a number, four or five years. Four or five years. Does it happen pretty immediately after taking the drug? Yes. Probably several weeks or months, you know. But there have been many studies all over the world

44:54that have shown it, you know. So it's very well established. And yeah, so that's one application. I mentioned it. I trust it 100%. But many people are not HIV positive, you know. So therefore, I say do not take antiretroviral therapy, you know. It's just not. So the other intervention that has very strong evidence is anti-TNF-alpha therapy, really anti-inflammatory drugs. For people who have an autoimmune disease, again, that just makes sense, you know. But yes. And metformin is an interesting

45:35intervention to many of us. The problem is... And I'm coming around to believing that metformin affects epigenetic age. There have been a couple of studies that suggest it. But I need to emphasize the effect is way weaker than the above. So these are really medical interventions. And in general, as you can expect, a medical intervention has a much stronger effect than a supplement, you know.

46:05Yeah. When it comes to supplements, we do have some answer. Omega-3 has a beneficial effect. Apparently, vitamin, multivitamins have an effect. The problem is that these supplements have much weaker effects. Suddenly, we talk about a couple of months of rejuvenation, you know. So yeah. And we're going to talk about those more in depth and what that means. But I want to kind of... This gets me into the controversies and hype because you're talking about like these really

46:36robust effects if someone has HIV, which is obviously devastating for your body. And then they take the antivirals. And that's really kind of... It does have a pretty robust effect on obviously their life expectancy. You know, many different features of health as well as epigenetic aging. So that makes sense. But I've heard people out there talk about reversing their biological age by seven years in... Sorry. They reverse their biological age by five years in seven months by doing

47:08lifestyle interventions. Is that something that you think could be a real biologic effect? Do you think that could be noise? Do you think it could be cherry picking the best clock to get whatever outcome that they're wanting? Or I mean, how do you feel about that statement? Yes. So it's a very good question. I think the first thing I would ask, what was their BMI before they started? And many other clinical readouts. So if you start with a

47:44person who was obese, had inflammation, diabetes, many of these stressors in their lives, and they really changed everything. And they take their GLP-1 receptor agonists, they suddenly go to the gym, and they do everything right, then it would perhaps be possible. But there are many pitfalls, and I can discuss them later. But I don't think it's not possible when you start with this baseline.

48:18So you're very unhealthy. You're very unhealthy. All the things. And above all, you actually start with an epigenetic age measure, let's say grim age, that shows you are eight years older than you should be. You see, you're in this highest percentile of risk. So then maybe you can go back to the average. However, now let's talk about the opposite case. A biohacker obsessed about healthy lifestyle. And now they say,

48:50I changed my diet, and now I reversed my age by five years, I would have the hardest time believing it. And by the way, this is something we see over and over again with various rejuvenating interventions. They seem to work in people whose epigenetic age is already accelerated, you know, but not in the people who are very healthy, you know. But yeah, so anyways, I would be very skeptical, but I'm open-minded. I'm strictly data-driven, you know. So I would have a long

49:26conversation with that person, you know. Yeah. Well, you make a really good point. And that is, you know, people that are already accelerating their aging at a faster rate. So they have this age acceleration, right? Their grim age is already, you know, they're going to, it's higher than it's supposed to. Is that correct? Is it? Yes. Higher. Higher. Higher than it's supposed to be. You know, their biological age, their pheno age is higher. Their pace of aging is higher. So they're already age acceleration for whatever reason. They're sedentary, they're obese, they're sedentary, and obese, and they smoke. Or perhaps

50:01they have vitamin deficiencies. That's another one I've seen, like vitamin D deficiency has been shown to be, you know, associated with age acceleration. And if you correct those problems by losing weight, by getting physically active, by quitting smoking, you know, by eating healthy, by getting your micronutrients and filling the gaps so you're not deficient, then you see a more robust effect. And that is also a recurring theme that I've seen from reading the scientific literature, where it's like, okay, if you already have enough vitamin D, and we'll talk about this, like, you know, if you're already sufficient, taking a vitamin D supplement's not going to

50:34slow your aging. The thing you're doing, you're already doing it. You're avoiding deficiency. And that's the key, right? You're trying to stop that accelerate things that cause the acceleration of aging seem to be easy, more responsive. Exactly. Yeah. The other question I wanted to ask you goes back to something that you mentioned earlier, when you were talking about, you know, these insurance companies being able to predict your mortality risk pretty accurately, using the DNA Grimmage. I've also heard people say that you can take this DNA Grimmage test, and predict the day you're going to die,

51:08like, within a month. No, no, that's not true. Okay. Now, why is that not true? Yeah. So, I want to start out by commenting on insurance companies. They are in the business of predicting how long you live. If they make an error, it will cost them a fortune. And they are superb at that. And just to emphasize, they look at so much. So, they will, above all, look at very traditional readouts, such as, what's your blood pressure? What's your medical history? Prior history of

51:43cancer, you know, substance abuse. So, they will look at all of the above, because all of these variables I mentioned are very strong predictors of mortality risk. And the question is, does Grimmage add something, or Grimmage or another methylation? That's really the question for these companies. And scientifically speaking, I can say, yes, it adds something, but not that much, you know. Clearly, the life insurance companies have done very well without having a methylation readout. But the

52:17exciting thing is methylation adds something. But then these companies have to weigh the costs, you know, because these tests are not cheap. They cost several hundred dollars, or is it worth it to measure? And by the way, that's the same question for any consumer. Is it really worth it to you to measure it, you know? Sorry, the other part of the, you had a second part of the question. Yeah, the question is, I guess I can word it a different way. If I were to go out and get a DNA

52:47Grimmage test, and it said that I was going to die when I was age 80, am I actually going to die at age 80? How reliable is that number? How accurate is that number? Or am I going to die at perhaps age 85? Yes, I want to tell you that Grimmage could lead to a prediction of when you die, let's say age 85. It could. And we know, though, that this estimate is accompanied by a large error bar, plus minus six years. I'm just

53:25making it up. So let's say you are a 50-year-old, you measure your Grimmage, and we apply the math, the mathematical algorithm, which, by the way, is very complicated, you know, for estimating your age at death. But the error rate is substantial. And this makes sense, because human beings are so complex, you know? Think how many things can happen, even in the next year. You can go through a

53:57divorce, you get hit by a car, you get depressed, you start smoking, you stop smoking, you know? So it would be unethical to report literally the age at death to a person. Therefore, we have decided to only ever give people an age estimate, right? We will say your Grimmage is 50. And what I want to really explain to anyone who listens is that please do not translate that age estimate in your mind

54:31into an estimate when you will die. In other words, if your Grimmage is 10 years younger than your calendar age, it does not mean you will now live 10 years longer than the average person. You see, you cannot compare this differential into a lifespan differential. Then what does it mean? Yes. So what does Grimmage really measure in a mathematical sense? What does it measure? It really

55:03measures the instantaneous hazard that you drop dead. I always say to people, it's your risk that you will die in the next year. That's how you need to think of it, you know, compared to a person of the same age and the same sex, you know? So let's start with the 50-year-old.

55:25And let's say their Grimmage is 58, eight years older than expected. Then their risk of dropping dead in the next year is more than twice that of the average 50-year-old of the same sex. Does that make sense? So it's really, mathematically speaking, it's a hazard ratio. And the hazard ratio measures instantaneous mortality risk. Now, you can translate that then into an estimate of your lifespan. It's

56:01easy to do. But it's a very complicated formula, certainly highly non-linear. And as I mentioned, associated with a strong arrow bar. Are there companies that have consumer-available tests doing that, where they're measuring the Grimmage and then doing that translation to when you will die? Is that something you've seen? No, I have not seen that. And I'm glad, because I would have a problem with that on two grounds. I find it, on some level, perhaps unethical. But I believe in freedom. So if

56:34people want to do something, I'm okay with it. My concern is it's scientifically unsound. It really is, you know, for the reasons I mentioned. There's a strong arrow bar, you know? Right. If you're talking five or six years, either way, that's a pretty big arrow bar for when you're going to die. But it seems like people are using it more to estimate their biological age, right, in a way, right? And that's typically what people are usually using. We use Grimmage, of course, to understand the effect of various stressors. And I'm a longevity

57:11researcher. I'm very excited about finding interventions that reverse it in humans and, of course, in animal models. So that's how I use it, you know? For these clocks, when we're looking at the, like, aging process as a whole, you know, we were talking about damage. You know, there's the insult that is the initial insult. And then you have perhaps the damage response, maybe the amplification of that damage with inflammation.

57:43Then you start to have tissue breakdown, right? Stem cell exhaustion, like things that are more downstream of the damage and amplification of that damage. Do these aging clocks, where do they sit on that? Yes, we have gained a lot of insights into aging in general, by the way, and also which aging hallmarks really affect epigenetic clocks, you know? So 10 years ago, we barely knew anything

58:19about mechanism. Epigenetic clocks were rightly criticized as black box readouts. But after really 10, 12 years of research by the very best labs in the world, you know, we really have characterized these changes. Maybe for the biologists, they're these hallmarks of aging. And we know that clocks relate to mitochondrial dysfunction, the energetics. They relate also to stem cell changes, very much

58:55so. Stem cell biology. They relate to metabolic changes, nutrient sensing to some extent as well. And also aspects of DNA repair, you know, so that is part of the biology. They clearly relate also to changes in what is known as cell composition. So in blood, we have many different blood cells.

59:26And some cells are aged, so-called stressed memory T-cells, cytotoxic T-cells that are exhausted. This is actually a technical term, exhausted T-cells from aging. And conversely, there are these naive T-cells, you know, so we understand that epigenetic clocks also relate to inflammation and that biology. So epigenetic clocks should be conceptualized really as integrators of many

1:00:02different stressors, but not all. They don't capture everything. And the most striking blind spot I want to highlight, which is frustrating to me, but I want to emphasize it. People in the aging field have heard of senescent cells, senolytics, very exciting intervention. I'm very much following that literature. However, epigenetic clocks really don't capture that well, you know. So let me give you the prime

1:00:38example. You have cells growing in a dish. You radiate them, any radiation. You induce senescence. The cells can no longer proliferate. And by the way, radiation leads to double-strand breaks. It really very much stresses the cells. And wouldn't it be nice if methylation clocks pick that up? But they don't, you know. So radiation damage, at least for cells. They don't pick up double-stranded breaks, even?

1:01:10Yes, at least when you induce it by radiation, you know. So we know radiation is very bad for you, but methylation changes do not result directly, you know. And I give you the converse of that when it comes to senescence. Many people have heard of telomeres. In theory, you want reasonably long telomeres at the ends of your DNA. And for many years, people have thought aging is about telomere

1:01:43attrition. Now we know better. It's not. But anyways, it's a famous hallmark of aging, telomere shortening. However, many of the clocks have only a weak correlation with telomere biology. It's a frustrating aspect. And 20 years ago, people had an exciting idea. Overexpress, a part of telomerase, the TERT, Overexpress TERT. And there were companies that pursued that as a rejuvenating intervention.

1:02:17And at least in our hands, we did not see a beneficial effect, at least in vitro, you know. So although I like epigenetic clocks for many studies, but they don't capture the totality of aging, you know. So you really want to complement epigenetic clocks with other readouts. That's interesting that they're not, because you mentioned that they do track with the DNA repair process, but not to some extent. I know I'm giving conflicting messages, but that's

1:02:54the biology. So there are certain experiments that show that some aspects of DNA repair relate to epigenetic aging, but others don't. It's just not a tight story, you know. So I think the field really needs to nail that down. Yeah. I mean, well, there's a lot of things that lead to aging. You know, it's a very complicated, you know, multifactorial process.

1:03:27When you actually are able to perhaps reverse, you know, biological aging, or I guess there's two ways of thinking about it. You're slowing age acceleration, right, if you're taking away something that's negatively accelerating aging or negatively affecting your health. But then also, let's say, if you can actually somehow slow the aging process, at least on the readout, the clock is showing that you're younger after doing something. Where do you think, do you

1:03:59think that's like inflammation, like these processes that you described that are sort of tracking with these clocks are being affected? So the, you know, mitochondrial function, inflammation, those processes are improving and the clocks are sort of picking that up?

1:04:17Yes and no. I mean, so epigenetic clocks such as Grimmage and Dunedin-Pace and Pheno-H, they do track inflammation to some extent, no question. So yes, if you reverse that, these clocks will pick it up. But it would be a grave error to assume that the clocks only measure that biology. It's really not true, you know. The clocks very much relate also to stem cell functioning, you know, and other aspects, you know. So again, they're integrators. And so there

1:04:54will be interventions that actually don't even touch the inflammasome in one way or another, but they could have a very strong effect on reversing your epigenetic age. And the prime example would be therapies that, for example, completely rejuvenate your hematopoietic stem cells. Just assume you have an intervention where you really replace your bone marrow, you know, or hematopoietic stem cells that produce all of these blood cells. And you just get hematopoietic stem cells with an epigenetic age of

1:05:32zero. That would very much rejuvenate your blood drastically, you know. We know that from mouse studies, but also human studies, you know, the epigenetic age in a bone marrow transplant recipient often reflects the age of the donor, you know. So do you see there are various interventions that could have a very strong effect, but they just don't touch on that biology you mentioned. If it's rejuvenating the blood, is it also perhaps rejuvenating other organs?

1:06:07That's a great hope, you know. So my responses, assume not, because it would, wouldn't it be nice? What have animal studies shown? Have they looked at that? Yes, there have been animal studies, I want to say, in the lab from Vadim Gladyshev at Harvard. And the studies, my reading of the studies is that they have been disappointing. They didn't rejuvenate other organs. If anything, there was a disappointing result that after X number of months, actually,

1:06:42the stem cells had aged. So the body has the memory of the old mouse, and that then aged the blood, really. Does that make sense? So did these mice get a hematopoietic stem cell graft, or what? Yes. They did? Okay. Vadim carried out really an elegant set of experiments, various transplantation experiments. And the scientific question is the following. Okay, if I replace, let's say, the blood by that

1:07:18of a very young mouse, or take other organs, by the way, should we replace the kidney, or the heart, or any other organ? Would the rejuvenation of one organ translate to a body-wide rejuvenation? And my current reading of the literature is that we haven't found any such organ as a target. I thought there was some evidence that if you did some of these transplants, where you take

1:07:49young blood and put it into older mice that rejuvenated the brain, for example, or am I, I mean, that's, I don't know if they were measuring, using clocks, but they were doing cognitive function, and a battery of tests, and the cognitive function improved, and things like that. No, for sure, you know. So maybe to remind the audience this, this idea of heterochronic parabiosis, for example, where you really connect the circulation of an old mouse with a young mouse.

1:08:21And this is really a phenomenal paradigm of rejuvenation, arguably one of the best ones we have, along with caloric restriction. And so, yes, we know that when an old mouse is exposed to the circulation of a young mouse, it has multiple benefits, cognitive benefits, also muscle benefits, so. And also, importantly, epigenetic clocks get rejuvenated, many organs. So we know that again

1:08:55from several studies, including from Vadim Bladeschev's lab, but others have found that too. So yes, young circulation rejuvenates the liver, the kidney, all of that, you know, on the methylation level. But there's a problem. You disconnect these mice, so they are no longer connected. They're no longer exposed to the young circulation. Then the things bounce back. The epigenetic age bounces back to that of the recipient mouse.

1:09:28It's very frustrating to all of us who work in the longevity field, because that is a very common story. You have a powerful intervention. It actually rejuvenates the organ. The problem is transient, you know, but yes. It's kind of like the probiotics flow through. You have to keep taking them to have a benefit in the gut as soon as you stop taking them, because they don't stick there, right? They're not taking residence there. Well, let's talk about caloric restriction, since you just mentioned that as a rejuvenating therapy. I mean, at least many animal studies have

1:09:59shown that. And I don't know that anyone wants to be calorically restricted for the rest of their life, although GLP-1 receptor agonists are kind of doing that in a way. There was a very recent trial, the calorie trial. And I'd love for you to talk about this. It was a two-year randomized control trial where individuals were basically eating 25% fewer calories than they otherwise would, or they were eating their normal, you know, daily food intake as usual. And I wanted to ask you, were the participants overweight in this trial, or were they normal weight? Do you know?

1:10:36I don't remember. I know it's a U.S. population, so assume that they are on the chubby side for sure, you know. Okay. Yeah. Yeah. So in this trial, it was a two-year randomized criminal trial, and it seemed there was many clocks that were measured, and it seemed like they had different readouts. Do you want to talk a little bit about the findings? I mean, I think… Yes. Yeah. I have a lot to say about weight loss. We should discuss it, you know. But the calorie study is a very famous study. U.S. population, very rigorous study, many, many readouts.

1:11:12But I want to acknowledge something, and the experts know it. The adherence was not good, you know. So there was an ambition that these people would lose more weight than they did. But as everyone knows, it's so hard to adhere to a diet. So the age reduction was on some level very weak, I would say. I apologize. I don't know the number, but I remember it was weak. I mention it because later we should talk about GLP-1 receptor agonists, where the weight loss can be

1:11:46unpronounced, and they're discrepant findings, actually. But anyways, back to the calorie study.

1:11:56Again, there were multiple blood draws from these people, and so one could evaluate which methylation clocks pick up a beneficial effect. And I was disappointed that grim age and pheno age did not pick up an effect. But this new clock, a new clock at the time, Dunedin Pace, really picked up an effect, which was reassuring, you know. Reassuring because everything I know about the

1:12:28biology of methylation clocks tells me that they should pick up a reduction in weight, if it's strong enough, you know. How much weight did they lose, do you remember? It was pretty, like, not very much. Yeah, it was not impressive to me, at least. So the clock that did pick up the... It was Dunedin Pace. Dunedin Pace. And in hindsight, let's discuss why it picked up the effect. There's like a 2% to 3% slowing of the rate of aging over the two years. Yes, that's true. Yeah. So it picked it up. And it makes mathematical sense to me,

1:13:03because Dunedin Pace, again, was trained. That's the lingo of machine learning. But it was developed to track changes in BMI. So yes, it picked it up. By contrast, grim age was never trained to look at weight loss. It was trained on mortality. So yes, Dunedin Pace worked. And my reading of Dunedin Pace is that it is good at that biology. People losing weight, it will pick it up.

1:13:36And now the question is, why didn't the other clocks pick it up? And there could be several explanations. But my view is, if there had been a larger sample size, if the people had adhered to the protocol, I'm as sure as you can be that the other clocks would have picked it up. It's a sample size issue. Or conversely, small effect size. And what I can tell you is, there was a very exciting study that involved actually obese people, BMI 30 and higher, who had been put on a GLP-1 receptor

1:14:15agonist treatment, semiglutide. And these people really lost a lot of weight over 33 weeks. And by the way, this study was published in MedArchive. It's a preprint. So take it with caution. It was Michael Corley's group in San Diego. But very beautiful study, very rigorous again, and a large sample size, so credible. And they looked at all methylation clocks. And suddenly,

1:14:49all methylation clocks picked it up, really all, you know. And so that's my thinking, you know, if you have a strong weight loss intervention, you have really a strong reduction in fat, lipolysis, you know. And this inflammatory signal is reduced. I think all methylation clocks will pick it up. Right. And I think it goes back to this concept that we were discussing earlier, where if your baseline is unhealthy, if you are obese, you are accelerating your aging, right? You're in age

1:15:26acceleration mode, right? So you need to slow it down. And with any clinical trial, it's always, you always get a better signal when you're starting with something that's on a population that's either deficient or unhealthy. And then you're giving something to improve that deficiency or negate it or to, you know, improve their health. And you get a more dramatic effect. So we know obesity accelerates aging. We know, you know, that it's associated with, you know, decrease in life expectancy, you know, increase in cardiovascular disease, type 2 diabetes, cancer, right? All these diseases of

1:15:58aging. So it's not surprising that you would give someone a drug that does cause rapid weight loss in a short amount of time. So you're going to get a much more robust signal. Right. And you're obviously picking that up with the aging clocks. With the calorie trial, you know, again, I didn't know what the adherence was. But also, as you mentioned, these clocks are trained with different, there's different specialties of them, so to speak, right? And BMI, being trained on BMI, wow, that's going to make you sensitive to weight loss for sure. And so the, you know, the duodenum pace clock, which is

1:16:35measuring the pace of aging, you would imagine would be more sensitive than something that would. But, you know, my question as a longevity researcher is, which clock should a clinician use, you know? If we could briefly talk about multivitamins, this study. Interestingly, here, Grimm-Age found an effect, Pheno-Age found an effect based on multivitamin use, but duodenum pace failed. It was not significant, you know? So, and you can ask this question now for many interventions, what should be the go-to

1:17:11clock? And, you know, I even, I want to stay clear of this debate because we will never agree, you know? Therefore, I just love it that the field by now simply reports at least five clocks, you know, so the reader can just look at it and be the judge. Let's talk about the multivitamin. So this was the COSMOS trial. I've talked a lot about the COSMOS trial in the context of brain aging. So the larger, there's, you know, the larger trials, and there was three randomized controlled trials where these older adults were given a standard

1:17:44Centrum Silver multivitamin a day, every day. And it was, what is it, about 3.6 years for this trial? And they were looking at, I mean, there's a lot of endpoints of this trial, but one of them was cognitive function and brain aging. And at the end of the trial, the people that were given the multivitamin had slowed their brain aging by 2.1 years. And there was a battery of tests that were done there. And I'm not sure if, in fact, some clocks were used as well. But I know that the global brain aging was slowed by 2.1 years and their episodic brain aging, so episodic memory is a kind

1:18:21of memory where you're remembering experiences, people, right? Like those sorts of things. That was slowed by almost five years compared to the placebo group, which is quite significant. And they did better on a battery of cognition tests. And so that was very, that's very encouraging, you know, and it's something that I do talk about a lot because I feel like it's a very easy, safe intervention that people can take a standard multivitamin. These have a variety of vitamins and minerals, trace elements that people are not getting from their diets. And so they're

1:18:51kind of filling that nutritional gap. And so, you know, who doesn't want their parents and grandparents to have better brain aging? So my parents are on a multivitamin, right? When it comes to looking at these epigenetic aging clocks, the pheno-age and grim-age clocks were the ones that stood out to me. As you mentioned, there was a battery of clocks that were looked at, but it seemed as though they were slowing, or at least, I'm not exactly sure all the calculations that go into this, but 2.7 months to five months, right? Like they were basically,

1:19:24they're slowing the aging by roughly that amount. Yes. Which to me is, if you think about now, this trial that was done with the aging clocks, I think it was like a subset of it, of the larger trial. Was it two years or did they do the 3.6 years for that? Do you? It was, yeah, I think it was two years. Two years, yeah. And so to me, the question is, now this wasn't the, you said the duodenum pace didn't change. No, it changed in the right direction. It just wasn't statistically significant.

1:19:58Oh, I see. It's a, no, in the right direction, you know, maybe a larger sample size would have led to a significant finding. It was definitely in the right direction. Well, the question I have for you is if you're changing it by, you know, three to five months within that two-year range, according to the Grimm-Age and Pheno-Age clocks, and you're to keep doing that, you know, you know, for years. So now we're talking not just two years, we're talking 20, we're talking 30, 40 years. Yes. How do you think that, do you think that you get this accumulative effect?

1:20:32Yes, I think so. I think so. Maybe to step back, if you tell an 80-year-old that a multivitamin will reduce his or her age by three months, they will roll their eyes. They will say, okay, give me something that reverses my age by 30 years, you know. So fair enough. The effect, we just need to acknowledge, the effect is very minor, you know. However, I like the way you

1:21:02conceptualize it. If you really use it for 30 years, right, you're 50-year-old and you use it until you're age 80, my expectation is that suddenly these three-month benefit, they accumulate and suddenly you have a benefit of maybe two and a half years, you know. It's still not great, you know, but there is a benefit, you know. So... But think of the effort you have to put into just taking a multivitamin,

1:21:33right? I think it's pretty great for that amount of effort, you know, if you're just having to take one vitamin supplement. Yes. And it's gonna delay your brain aging, you know, by 2.1 years just after, you know, in that trial, it was 3.6-year trial. But, you know, that's pretty robust. Five years delaying brain, episodic brain aging. Yes. And now we're talking about, like, globally, like, biological aging. If it's slowing it by, let's say, on the high end, five months after two years,

1:22:03I don't know. That seems like a pretty great effect if you're just taking a vitamin supplement. For two years, it's doing that. Well, let's continue on and then combine other things. And we'll get into some of the other trials that do show synergy. But I think it's interesting. The other question is that, and this is where, you know, the COSMOS trials, people, they're looking at everything, right? Cancer, mortality, cardiovascular mortality, all cause mortality. And those didn't really seem to change, at least within the timeframe that was looked at. And so, you know, we see these epigenetic clocks giving us a signal. We see the

1:22:37brain aging effects. And the question is, why are those showing up before? Yeah, I tell you my reading of it. And to me, this whole study was one triumph for epigenetic clocks. And I explained to you why. Assume you knew nothing about multivitamins. You would think that there is a benefit, you know? Clearly, vitamins are important. It's a trivial tautology.

1:23:07Avoiding deficiencies are important. Especially, you know. So you would say, okay, I administer that to the U.S. population. I would hope to see an effect. And that, of course, is the reason why these large-scale studies were even initiated. Think about how difficult it is to raise the funding for such a large-scale study. Clearly, there must be very compelling reasons. Okay, but there is a problem now. And these hard endpoints, mortality, cardiovascular disease, they didn't detect an effect. Deeply frustrating.

1:23:44Oh, is there a trend? It wasn't statistically significant? Yeah, I'll let you summarize it. But to me, you know, I just looked at it from the point of you as a consumer. Five years ago, you know, I wouldn't take a multivitamin. I looked at the literature. No benefit. I won't take it, you know? Now, so a person can now make their own judgment, you know? So what does it mean? To me, I take it as a wonderful triumph of epigenetic clocks that they

1:24:17did pick up the signal. And I call this testing the test. You have an intervention where you really think it's got to move the needle, you know? And then if a readout doesn't show it, one interpretation is, well, maybe the readout is too crude. Maybe all-cause mortality is a real—I mean, I like it as a readout. I used it for grim age. Don't get me wrong. I like that it's hard and

1:24:50definitive. You can't argue with it. However, people die for a hundred different reasons, you know, that may really not relate to the biology of aging, you know? And so now that we have actually biomarkers that did pick up that signal, even though it's very weak, is to me really reassuring, you know? Yeah. And I think it's reassuring in combination with the brain aging signal that it picked up. And just knowing that, you know, so many—so many globally people are not getting

1:25:25enough of these important vitamins and minerals and trace elements and essential fatty acids from their diet, then it's kind of like an insurance. Like, okay, I'm going to fill some of these nutritional gaps. They won't all get filled because you can't stuff everything into one pill. I mean, you can only get a little bit of some things in there, right? Yes. But I do think that it's—again, I agree with you. I think it is a triumph, and it's something that I do think that is safe. I mean, it's really been shown to be safe. And so maybe you pee a little

1:25:57bit more of it out. So what? It seems to be doing something beneficial for the brain, and at least for, you know, looking at these aging clocks, it seems like for the way you're aging as well. Yes. Yeah. So what have you got to lose, you know? Yeah. Let's go back to maybe some of these other vitamins. There's other, I guess, lifestyle interventions as well that I wanted to cover. But since we're on the vitamin train, the big one is omega-3, right? I mean, that I've seen, at least in the literature. And this is something that isn't

1:26:29surprising to me because going back to this theme that we've been talking about, if you're starting out with a deficiency, if you're starting out at an unhealthy point and you improve that, you fill that deficiency gap, or you, you know, improve your health, lose weight, whatever, then you're going to have a stronger signal, right? Yes. 90% of Americans don't get enough omega-3 fatty acids. Nobody's eating seafood in the U.S. It's just, you know, so you're starting with a population that's already, you know, I don't

1:27:03want to say deficient, but they're not getting a sufficient amount of omega-3 fatty acids. And so I think it's probably why it's easy to keep getting this stronger signal because you're, if you start out with someone who's already getting enough omega-3, maybe you go to Japan and do this study. I don't know. It would be interesting to see. Perhaps, perhaps there, it just keeps improving inflammation and then, you know, you'll keep, seeing an effect. But it seems like many, many studies have shown that omega-3 fatty acid, whether it's from food, supplementation, a combination of both, seem to slow epigenetic

1:27:39aging by different clocks. Yes. There has been quite some literature. It started with observational studies that you cannot trust. But last year, we published a study which was very rigorous. This was a study conducted by a Swiss professor, Heike Bischoff-Ferrari, who looked at 780 people and followed, again, the most rigorous design, randomized controlled trial, placebo controlled trial, in a population that I

1:28:18was very interested in. People 71 years or older, really older people, reasonably healthy. Average age, I want to say 75, I think, or 73. So older people. And she evaluated famous interventions. Number one, omega-3, one gram, vitamin D. And we should talk about the intervention about vitamin D was tricky. It was high vitamin D versus low vitamin D. It wasn't vitamin D versus no vitamin D. That's a key

1:28:53distinction. What was the low? I know the high vitamin D was 2,000 IUs. Yes. And the low was 800 IUs. Oh, so it was only double. Yeah, exactly. And that's a limitation. Because the results for vitamin D were disappointing. No effect on epigenetic clocks. But that's why I hasten to add. Yeah, but we have other randomized controlled trials showing the opposite if you start with the deficiency and add it. And we can talk about that. Okay, so vitamin D. That's true. But there was also another disappointing. If you look at the exercise.

1:29:25Yeah, we need to talk about exercise. So this was called a home exercise intervention. Now, to remind you, these are people in their 70s. And think in terms of ethics approval. You cannot stress these people too much. So this home exercise intervention was very modest. Okay. It was resistance training, right? Three times a week. Yes. But it was, I'm telling you, it was mild, mild, mild resistance training. Because no effect.

1:29:59Right. I was very disappointed. Did you read that the starting population, like around 88% of them already identified as being physically active? Exactly. I mean, which is, if you were to get a US population, not a chance. Yes. Like that, there's no way you would have had that many people physically active. But anyway, so that's another... No, these are people in Switzerland. Hopefully, they hike in the mountains. They're walking everywhere. Exactly. No, but that was interesting to me because I'm very interested in that population. People who already do a lot of good things, what can they do to improve, you know, their outcomes, you know?

1:30:32Great framing of it. Yeah. And yeah, so I think we already discussed the result. The most credible result was omega-3 on epigenetic clocks. A couple of epigenetic clocks picked it up. Grim-Age version 2, Pheno-Age, Dunedin-Pace also worked very well in that context. So, nice result for omega-3. The other interventions disappointed. By themselves. By themselves. By themselves. But yes, there was this one treatment arm where people actually

1:31:06did, used all three beneficial interventions, high dosage vitamin D, omega-3 plus exercise. And according to Pheno-Age, that treatment arm did the best, you know? So, that's the finding that we would have liked to see, you know, for all clocks. But it's just the Pheno-Age picked it up. Well, I think there was even a dose dependent where there was the group that just got the omega-3 and vitamin D, and that also improved more than the omega-3 alone. And then all three improved

1:31:39the most. So, you see this nice dose dependent effect with adding in these healthy lifestyle interventions, even in an already presumably healthy population, which is exciting. And I have the numbers here. I think it was 3.8 months, the Pheno-Age delayed, the biological aging was delayed by 3.8 months. Yeah, over three years. Over three years. Three years of that, yeah. And that doesn't sound like a lot again, but they also correlated with some other outcomes, right? So, I think there was, in all three interventions, yes, it was 3.8 months that delayed

1:32:14the biological aging, but also that was associated with outcomes that were important. 61% reduced, chance of getting metastatic cancer. It was like a 20% reduction in pre-frailty, which is also nice to see these outcomes correlated with this as well, right? I agree with that. And I can tell you the same study looked at a new concept in the field called intrinsic capacity, which looks at various domains of functioning, frailty, cognition, psychology.

1:32:47Anyways, also intrinsic capacity was improved in that population. Oh, okay. So, it's not just the molecular readout, yes. I think, for me, the take-home, again, is something that you mentioned when you have this already, you know, healthy. They have to be a healthy population of 88% with them physically active, right? Yes. So, and you take that healthy population, you can still improve, right? You still can improve things. Do you, again, this comes down to the compounding factor, right? So, this is three years, and then let's say, okay, well, they're going to start doing this for

1:33:19the rest of their lives, you know, decades. We're talking, well, in this case, they're a little bit older, but people listening to this podcast, maybe in their 30s, maybe in their 20s, in their 40s. It's like, okay, well, I'm going to start training, getting, making sure I'm not vitamin D deficient, getting my omega-3, and then you have, like, how is that going to compound over time? And I know it's speculation, but it makes sense. That's the way I think about it. I think of it the same way. I wish I could go back in time and tell myself to stop eating

1:33:51chocolates, which really messed me up. So, yeah, good health behaviors, you know, and supplements included, I think, will have benefit, major benefits. When it comes to vitamin D, that's the one that, I mean, this one study was a bit disappointing, but as you mentioned, I mean, comparing 800 IUs to 2,000 IUs, I wouldn't imagine to see a big difference there because you're already feeling the gap. Exactly, and most of the participants had no insufficiency in vitamin D. They really started

1:34:22at normal levels. And we know there have been, in my opinion, so many studies that I've come across and read over the years showing that vitamin D deficiency causes age acceleration, in some cases severe, like three years. And if you correct that deficiency, it'll slow age acceleration, where then you say, you know, reversed aging by, you know, four years or whatever. I mean, so I think my take home, and I know the one that I like, the most recent one was the base

1:34:58to the Berlin study where they took, which was, the thing that was nice about that was they had a deficient population and then a sufficient population and gave them vitamin D. Yes, and this was a study in Berlin and they followed people for seven years, which was also impressive, was a large population. And you can imagine Berlin is, of course, not blessed by sunshine, so they start out deficient, you know, so it all made sense, you know. Yeah, reverse aging, if you're deficient and feel that sufficiency, but the people that were

1:35:33not deficient, actually, there was no effect, which is, again, what you expect. It's not about, this is a magic supplement that's slowing aging. It's not doing that. It's helping people that are deficient correct their deficiency. And that's why there's so much in the, even in the scientific literature with vitamin D, for example, if you're looking at outcomes, it's the same thing. Yes. You know, and it drives me nuts when studies don't measure their baseline levels, or if they only measure 10% of the population and then use that to extrapolate like everyone else. Like, you can't do

1:36:04that. There's so many, you know, and there's gene snips that are affecting vitamin D. There's other micronutrients. Magnesium really affects vitamin D. You need magnesium to convert vitamin D3 into, you know, the steroid hormone. So there's so many different things that are affecting your vitamin D. If you don't measure it before and after, it's hard to really make a statement that it did what it did or didn't do what it was supposed to. Yes. Okay, so I want to... We should talk about exercise. Yeah, we can go into exercise. Oh, no, but follow your script. Sorry.

1:36:36Oh, no, no. We were talking about calorie restriction, and I just wanted to mention dietary patterns in general, you know, because you mentioned weight loss, and we've talked a little bit about it with the GLP one. Obviously, if you lose weight, it's probably a big confounder with all some of these dietary pattern trials, right? Like, if you're getting someone who's overweight, overweight, and these participants are overweight, and you're putting them on a healthy diet or a Mediterranean diet or something like this, and they lose weight on all the diets, then

1:37:08how much of what you're seeing is due to that weight loss, right? Exactly. So do you want to talk about that? It was like the direct trial. Is that what it was called? Yeah, I need to tell you, I don't know too much about it, but I want to explain some properties of grim age that I'm aware of. So grim age very much correlates with what is known as carotenoid levels in the blood. So what are those? So, you know, let's maybe back off and think of

1:37:41nutritional studies. Many people have so-called food questionnaires where they evaluate the diet of participants. And from all I know from analyzing data is that these food questionnaires often don't reflect reality. I don't remember what I ate for breakfast. Well, I didn't eat breakfast today, but... Yeah, I mean, and people always know what they should answer, you know, but so that may bias their memory, they will say, oh yeah, I ate X servings of broccoli, but it just doesn't reflect reality.

1:38:14But fortunately, they're blood tests. You can measure the so-called carotenoid levels in the blood and have an objective readout of fruit, vegetable consumption. And the striking finding in post-menopausal women from the Women's Health Initiative was that this measure of vegetable intake has a strong correlation with grim age and other epigenetic clocks. Strong meaning maybe minus 0.3. So it's, to me, a very strong effect, which really changed my behavior. By now, I really

1:38:50eat a lot of vegetables. Can you translate that to like months? Like what would minus 0.3? Yeah, sorry. I could translate it, but... Or just an estimate. Yeah, no, let me put it this way. Smoking has a correlation of 0.4. So if you smoke a lot, it increases your age. Vegetable consumption minus 0.3. So it's actually... Wow. Yeah, I was very surprised. So in this... Sorry, I add one more statistic. Exercise. The correlation would be 0.1. So do you see...

1:39:23So vegetable intake has a much stronger effect. I mean, orders of magnitude, stronger effect on grim age and these methylation clocks than, for example, exercising. And you think it might come down to even the carotenoids perhaps, or just the vitamins and minerals and everything in the vegetables kind of compounding? Yeah, you know, I never looked into that, but I feel that would be such a worthwhile research study. What I can tell you is this vegetable association is 100% accurate, but now teasing

1:39:59it apart, what is it, you know, to be seen? Probably so many things. I mean, you've got the fiber matrix, you're getting micron... Like vegetables, especially greens. And if you're talking about carotenoids, you know, lutein, zeaxanthin, these are carotenoids that are in greens. And interesting, there's been a lot of studies coming out looking at blood levels of lutein and zeaxanthin. People usually associate them with eye health. They accumulate in the eye. There have been randomized controlled trials showing they can help prevent age-related macular degeneration. They also accumulate in the brain.

1:40:35And they're associated with improved cognitive function, crystallized intelligence, improved brain aging in general. All right. And there's other carotenoids. Beta-carotene is probably what most people are familiar with, lycopene in tomatoes. So there's a variety of these carotenoids, which are very powerful at basically, I would say, buffering oxidative stress and singlet oxygen, for example, if you're talking about in the eye. Yes. But it's interesting that vegetable intake can have such a profound effect.

1:41:09And there was a vegan trial too, I think. Also, there was a trial looking at people that are eating a lot of vegetables versus like a healthy omnivore trial. And I think the vegan trial, they also had slow their epigenetic aging more. But there's always weight loss as a confounding factor because they were eating fewer calories. But that's really interesting that there's a minus 0.3. That is pretty strong. You gave me that reference point of smoking being, you said it was, wait, smoking was 0.4. Okay. Smoking, 0.4, maybe 0.45. So it's increased correlation. Exercise, 0.1.

1:41:46Okay. So, and we can talk later about exercise, but very weak effect. In order to see an effect of exercise, you really need to study many thousands of people. With vegetable intake, the effect is so strong, you probably see an effect when you analyze a couple of hundred people, you know. So, but regarding the question vegan versus carnivores, you know, I honestly have not seen convincing data.

1:42:16Omnivore, omnivore. Oh, omnivore. Yeah. So, or, yeah. Carnivore would be the extreme opposite. That's true. Yeah. Let me rephrase it. So I have not seen any evidence that people who, let's say, eat a lot of red meat age much faster than people who are vegans. And we looked again in the Women's Health Initiative. I mean, there was a hint, I want to say, when we analyzed 3,000 women and then women who ate red meat, it was barely noticeable that red meat

1:42:50was ever so slightly increasing epigenetic age, but it was truly negligible, you know. So what I can tell you is I eat so much meat. Hopefully it's not bad for me. You eat meat and vegetables. I eat meat and vegetables. I try to be easy on the carbs, you know. I eat carbs, but I try to reduce them. Well, I mean, vegetables are carbohydrates. They're just complex carbohydrates, not simple. So you're not eating the simple carbohydrates. Yeah, exactly. Yeah. That vegetable stuff is interesting. There's so much in vegetables with the micronutrients

1:43:24and the phytochemicals, right? That's another thing in them. The fiber. I mean, there's a lot of things going on here. Yeah. Somebody should really tease that apart. What kind of vegetables should be eaten, you know, and dosages, yeah. Lots of exciting PhD dissertations could be written on that topic. Um, exercise. Yes. That, so let's talk about that. So you're, you know, what, there's a trial that you sent me that was pretty convincing and it was kind of, it was a new one in 2025

1:43:55showing that six months of cycling, it seemed to slow epigenetic aging or grim age, right? Grim age by 7.4 months. Yes. Yes. I, maybe I'll frame it like that. So there have been very nice studies on the effect of exercise on grim age and pheno age and other clocks. And, um, so why, what do I mean by nice? Um, studies where they use one way or another wearable to really measure your step count and

1:44:26activity. So it's a very rigorous readout of your physical activity. And the studies were also convincing because they were large scale studies, many thousands of people. And in different countries, Japan, Germany, US. And the finding is the following. Yes. If you move more, yes,