Peter Attia MD222 ‒ How nutrition impacts longevity | Matt Kaeberlein, Ph.D.
CHAPTERS
- 0:02 – 5:25
Why Nutrition and Aging Research Is So Hard to Interpret
Peter Attia and Matt Kaeberlein explain why they trade emails on nutrition and longevity. They note that diet is harder to study than a drug and that animal studies leave out environmental complexity. They also point out that long-running epidemiology reflects environments that have since changed.
- •Geroprotective molecules are cleaner to test than food, yet even rapamycin dosing questions remain unresolved
- •Lab mice live in controlled, relatively pathogen-free environments unlike humans
- •Epidemiological nutrition studies are decades old and the human environment has changed
- •Looking back over decades makes it risky to assume causation from correlation
- 5:25 – 14:48
Evolving Convictions and a First-Order Framework for Nutrition
Peter describes loosening his nutritional convictions over time and his simple clinical sort by nutrition, muscle and metabolic health. Matt agrees there is no one-size-fits-all diet, explains his distrust of diet gurus, and says being near optimal intake and body composition gets most of the benefit. Both argue that effort is better spent on exercise than on nutritional minutiae.
- •Peter sorts patients as over- or undernourished, under- or adequately muscled, and metabolically healthy or not
- •Matt says the optimal strategy differs between people and beliefs should change as data arrive
- •Matt recalls the early-'90s low-fat, high-carbohydrate advice as exactly wrong
- •Peter prefers the 80/20 approach for nutrition and puts extra effort into exercise
- 14:48 – 21:35
Calorie Restriction in Animal Models and the Cancer Question
Matt traces calorie restriction research back to the 1930s rat studies and its lifespan effects across species. He and Peter discuss that lab mice die mostly of cancer rather than atherosclerosis. Matt argues calorie restriction also delays functional declines in many tissues.
- •Early rat studies found 40-50% longer lifespans with apparent healthspan benefit
- •Calorie restriction has slowed aging in many organisms across wide evolutionary distances
- •Most lab mice die of cancer, unlike humans where atherosclerosis leads
- •Matt favors functional measures of aging over clocks or biomarkers
- 21:35 – 30:26
Epigenetic Clocks: What They Can and Cannot Tell Us
Peter asks Matt to justify the utility of aging clocks. Matt says epigenetic clocks work as chronological measures but remain unproven as measures of biological aging, since no one has shown they predict an individual's future outcomes. He also cautions that the claim that reversing the epigenome reverses aging has no supporting data.
- •Epigenetic clocks are useful for chronological age, such as in forensics or rescued dogs
- •Matt is not convinced clocks measure biological aging and wants individual-level prediction
- •Longitudinal studies show some correlation with mortality, but whether they beat other markers is unclear
- •Epigenetic change is one of several consensus molecular processes of aging
- 30:26 – 37:07
Yamanaka Factors and Partial Reprogramming
Matt explains how the four Yamanaka factors wipe epigenetic changes and can restore cells to a pluripotent state. He describes the progeroid mouse work and David Sinclair's optic nerve results, and sets out what he would need to see in an old mouse to be excited. Peter pushes on why epigenetic changes matter and how they relate to mutations and wound healing.
- •Yamanaka factors reset the epigenome in cultured cells, and partial reprogramming is attempted in animals
- •Progeroid mice gained lifespan, but no one has turned an old mouse into a young one
- •Matt wants reprogramming to match rapamycin's roughly 25% lifespan extension
- •Epigenetic changes affect gene expression and loss of homeostasis but will not fix mutations
- •Peter raises his son's fast healing versus his own as an example of age-related differences
- 37:07 – 47:31
Reprogramming Limits, Hype, and Hedging Your Bets
The two discuss whether reprogramming can work in tissues like heart and brain, and Matt raises the problem of preserving memories and identity. They worry that hype leads people to neglect their health, and Peter frames hedging as risk management. Matt expects long timelines and regulatory hurdles, and both see narrow first applications such as eye disease, osteoarthritis and senolytics.
- •Matt has yet to hear how reprogramming could work in the brain without altering memories and experiences
- •Reprogramming too far can be fatal, and certain cancers appeared in partial reprogramming in mice
- •Peter likens counting on future therapies to relying on a lottery ticket
- •Matt advises not expecting major lifespan-extending treatments in the next 20 years
- •Targeted indications like Unity's eye program may come first
- 47:31 – 49:41
Exercise Gets Most of the Benefit with Modest Effort
Peter and Matt note that most people do not need to do everything perfectly. Peter describes the exercise dose-response curve, saying about 50% of the benefit comes from going from nothing to about 15 MET hours per week. Matt adds that most people do not know how much benefit a little movement brings.
- •Matt says people don't need to optimize every nutrition or activity detail
- •Peter says 30-50% of exercise benefit is captured by going from sedentary to 15 MET hours per week
- •That equals about three hours a week at five METs, such as brisk walking or slow jogging
- •Matt says the public is unaware of how much benefit modest activity gives
- 49:41 – 59:01
Calorie Restriction, Immunity, Muscle, and Body Weight Normalization
Matt discusses mixed data on immune function in calorie-restricted animals, including the sepsis findings, and concerns about CR without optimal nutrition in people. They cover how muscle strength is normalized to body weight in mouse studies and what that means for human sarcopenia. The conversation then turns to bone density, the unknown optimal body weight, and the neglect of adverse events in nutrition research.
- •CR animals responded better to some pathogen challenges but did worse in sepsis experiments
- •CRON or CRAN diets are controllable in mice but hard to replicate in real-world human CR
- •Mouse muscle function is usually normalized to body weight, with CR mice weighing about 30-35% less
- •Peter cites a study where diet-driven weight loss reduced bone mineral density but exercise-driven loss did not
- •Matt says nutritional interventions are wrongly assumed safe and adverse events are rarely tracked
- 59:01 – 1:09:19
The Wisconsin and NIA Rhesus Monkey Studies
Peter and Matt walk through the two famous primate studies. The Wisconsin study found benefits from CR on a high-sugar diet, while the NIA study found no lifespan extension, and they discuss why the results differed. They note that molecular signatures were shared with rodents and that it is unclear if benefits come from CR or from avoiding obesity.
- •Wisconsin monkeys on a processed, high-sucrose diet lived longer with CR and had less cancer, heart and metabolic disease
- •The NIA study used a higher-quality diet and mixed ages, and found no lifespan benefit
- •NIA controls ate less than Wisconsin controls, which could have narrowed the gap
- •Molecular changes in mTOR, mitochondria and metabolism were shared between mice and monkeys
- •Neither study answers whether eating less of a good diet extends life
- 1:09:19 – 1:19:10
Human Evolution, Metabolic Flexibility, and the Grandmother Test
The two discuss Herman Pontzer's work on human energy expenditure and why monkeys are an imperfect comparison. They then challenge the idea that an ancestral diet is optimal and note that humans can be healthy on very different diets. Peter proposes the grandmother test and both urge humility as food environments change.
- •Peter describes humans as having high energy needs and a unique capacity to store energy
- •Matt sees no compelling reason to think the ancestral diet was optimal for longevity
- •Peter says people can be healthy on diets as different as vegan and ketogenic
- •Peter proposes shopping the grocery perimeter and the great-grandmother test
- •Matt says the food environment is changing quickly, so epidemiology from 20 years ago is hard to apply
- 1:19:10 – 1:32:47
Matt's Science Review of Anti-Aging Diets and Time-Restricted Feeding
Matt explains why he wrote the 2021 Science review with early-career co-authors and how the group defined the dietary interventions. He and Peter separate time-restricted feeding from intermittent fasting. They discuss mouse work showing that calories account for most of the benefit while the feeding window adds a smaller portion.
- •The review covered caloric restriction, intermittent fasting, time-restricted feeding, fasting-mimicking diets, ketogenic diets and protein restriction
- •Most mouse studies of these diets are really flavors of calorie restriction
- •Time-restricted feeding limits daily eating hours, while intermittent fasting means fasts of 24 hours or more
- •Recent mouse studies suggest about two-thirds of the CR benefit comes from calories and one-third from the fasting window
- •Matt finds mouse diet studies powerful for mechanisms but hesitates to recommend them to people
- 1:32:47 – 1:41:47
Early vs Late Feeding, Ketogenic Diets, and Mouse CR Intensity
They discuss early versus late feeding windows and glucose, with Peter citing CGM observations. Matt then covers ketogenic diet lifespan studies in mice and notes that calorie restriction and protein restriction are the two interventions with consistently large effects. He describes the Weindruch and Walford graded CR study and his concerns about the psychological effects of CR in people.
- •Peter sees lower average glucose with early feeding in CGM users
- •Matt says what counts as better depends on the endpoint studied
- •An intermittent ketogenic diet extended mouse lifespan about 15%, but the mice were calorically restricted
- •Weindruch and Walford saw a roughly linear lifespan response up to 65% restriction
- •Matt worries about the psychological and social effects of true caloric restriction in people
- 1:41:47 – 1:55:38
Late-Life Calorie Restriction and the Rapamycin Story
Peter reviews a 2019 mouse study with crossover CR designs, and Matt interprets it as showing smaller benefits from late-onset CR. Matt contrasts this with rapamycin, which gives nearly the same benefit when started late, and tells the story of the Interventions Testing Program's accidental late-start rapamycin experiment. They also discuss bioavailability of compounded rapamycin.
- •Late-onset CR gave a significant but smaller lifespan benefit than early CR
- •Mice on early CR then switched to ad lib lived longer than those starting CR late
- •Rapamycin started at 20 months in the ITP gave robust lifespan extension, changing the field's paradigm
- •Rapamycin needed encapsulation to survive gastric pH
- •Matt has heard that compounded rapamycin may have lower bioavailability than Rapamune
- 1:55:38 – 2:12:16
Protein Restriction, mTOR, and Protein Needs in Humans
Peter explains why protein is the nutrition topic that interests him most and contrasts his intake with the RDA. Matt reviews protein and amino acid restriction in mice, the role of mTOR and FGF21, and why this may not extrapolate to people. They discuss mTOR activation time course, why rapamycin has not caused muscle loss, and the Simpson data showing a high-protein diet with the longest mouse lifespan.
- •Peter aims for about a gram of protein per pound spread over four meals, while the RDA is 0.8 g/kg
- •Matt says the RDA is a minimum for sedentary people, not an optimum
- •Protein restriction in mice works through mTOR inhibition and FGF21, though mechanisms vary
- •Matt says mTOR effects depend on location as well as amount, which helps explain no muscle loss with rapamycin
- •In the Simpson nutritional geometry study, the longest-lived ad lib mice ate roughly 40% protein, suggesting many routes to longevity
- 2:12:16 – 2:22:16
Low vs High Protein: Muscle, Mortality, and the Levine Study
Peter lays out the consequences of being under-muscled in humans, and the two debate how much dietary protein matters for muscle in older people. Matt discusses the epidemiological study in which low protein was linked to lower mortality before about 65 and higher mortality afterward. Matt's own modeling put the crossover nearer 50, and they agree that high protein in the context of obesity and high calories may be problematic.
- •Peter says low muscle mass drives metabolic, sarcopenia and osteoporosis problems that dominate late life
- •The Levine study found low protein beneficial for all-cause mortality below 65, with the pattern reversing above 65
- •Matt's own modeling placed the crossover at around age 50
- •Matt is confident that above-RDA protein is beneficial after 50, especially with exercise, but is wary for overweight, sedentary or diabetic people
- •Peter thinks much of the epidemiology conflates high protein with high calories
- 2:22:16 – 2:41:29
IGF-1, Growth Hormone, and Cancer Risk
Matt explains IGF-1 and the growth hormone pathway, from C. elegans DAF-2 mutants to long-lived GH-signaling-deficient mice and a late-life IGF-1 receptor antibody study. Peter discusses his reluctance to prescribe growth hormone given sparse human data. They cover the Laron syndrome study in Ecuador and why IGF-1 is a muddy biomarker, and Matt describes how immune function may mediate cancer risk.
- •Mutations reducing GH/IGF-1 signaling extend lifespan in worms and mice, mostly when deficient from development
- •A late-life IGF-1 receptor antibody gave modest lifespan extension in mice
- •Peter does not prescribe growth hormone because he sees too little human safety data, though he also sees no clear evidence of harm
- •People with low GH signaling in Ecuador had a dramatic reduction in cancer but no lifespan difference
- •Total IGF-1 is limited as a biomarker because of binding proteins, insulin interplay and population differences
- •Matt believes higher GH/IGF-1 signaling raises cancer risk all else equal, though immune function may offset it
- 2:41:29 – 2:43:49
Closing Takeaways on Nutrition, Longevity, and Not Overthinking
Peter and Matt acknowledge they raised more questions than answers. Matt notes genotype affects how animals respond to calorie restriction and that animal nutrition studies are valuable for mechanisms and led to rapamycin. His practical take-home is to eat a relatively healthy diet, avoid overeating and exercise, without obsessing over details.
- •Different mouse genotypes respond differently to the same caloric restriction
- •Nutritional studies in lab animals have revealed key pathways and led to candidates like rapamycin
- •Matt says most people can get much of the benefit with a healthy diet, no overeating and exercise
- •He cautions that overly detailed recommendations can create anxiety about protein, ketosis or fasting windows
