Peter Attia on Longevity: 12 Episodes on What to Measure and When to Act

Peter Attia on longevity, ranked across 12 Drive episodes: VO2 max as the strongest mortality predictor, cumulative apoB exposure, Alzheimer’s as a midlife disease, and the screening that catches cancer early. Free to read in full.

Peter Attia’s longevity argument reduces to four claims he returns to again and again. Cardiorespiratory fitness is the strongest modifiable predictor of all-cause mortality anyone has measured, beating blood pressure, cholesterol, BMI, smoking status and age. Strength and power decide how much of later life you spend using your body. Atherosclerosis tracks apoB particle concentration multiplied by time, which makes lipid exposure a lifetime problem. And the diseases that kill most people, including Alzheimer’s, begin in midlife and surface decades later. We ranked 12 episodes from The Peter Attia Drive that carry those claims with the most specificity.

Each entry explains why the episode earns its rank, the takeaways worth acting on, and a verbatim quote from the conversation. Every one links to our full summary of the full episode, free to read, and most include the official video so you can go straight to the source.

A word on scope. These come from the Drive episodes in our summary library, which skews to the recent run (December 2025 through September 2026), so treat this as the current state of Attia’s thinking and the guests he brought in to argue it, as opposed to a career retrospective. Three recurring collaborators do a lot of the heavy lifting here: lipidologist Tom Dayspring, cardiologist Michael Davidson, and NMR pioneer Jim Otvos.

1. A Guide to Cardiorespiratory Training at Any Fitness Level to Improve Longevity (AMA 79)

The Peter Attia Drive · Peter Attia · 38m · January 2026

The single most direct answer to the question this page exists for, and at 38 minutes it is also the cheapest. Attia states the hierarchy outright: VO2 max outperforms every other measurable variable for predicting all-cause mortality, the bottom 20-25% of the population carries a 4-5x higher annual mortality risk than the top 2-3%, and moving up a single quartile buys a 50-75% risk reduction. He then gives the training model that follows from it (a base of zone 2 volume under a peak of maximal aerobic output) and, unusually for this genre, tells people with only 150 minutes a week to spend all of it at high intensity.

Key takeaways

  • VO2 max beats blood pressure, cholesterol, BMI, smoking status and age as a predictor of all-cause mortality. Bottom quartile carries 4-5x the annual risk of the top 2-3%.
  • Think of aerobic fitness as a triangle: a base of submaximal efficiency (mitochondrial, fat-fueled) under a peak of maximum oxygen delivery (cardiac output).
  • Zone 2 earns its place once you have training time to spend. On a 150-minute week, high-intensity work is the efficient use of all of it.
  • Three lactate levels matter: local shuttling, the first threshold near 2 mmol (zone 2), and the second near 4-5 mmol where clearance is overwhelmed.
  • Recoverability becomes the binding constraint in your 40s and 50s, which is exactly when high-volume zone 2 earns its keep.
Cardiorespiratory fitness outperforms every other variable we can measure. This includes blood pressure, this includes cholesterol, this includes BMI, smoking, it even includes age, which just blows my mind. — Peter Attia

Read the full episode summary

2. Michael Davidson on Lowering LDL Early: Preventing Cardiovascular and Alzheimer’s Disease

The Peter Attia Drive · Peter Attia · Michael Davidson · 2h 19m · September 2026

The most recent episode here and the one that reframes the whole lipid conversation as a timing problem. Davidson argues for primordial prevention: lowering LDL before plaque is measurable, because atherosclerosis is concentration multiplied by duration and a 10-year risk calculator systematically clears young adults who are accumulating exposure. The analogies land hard. We treat hypertension before the stroke and diabetes before the blindness, so the same logic should apply to apoB. The episode also pushes into APOE4 and whether CETP inhibition reaches the brain, which is where cardiovascular and dementia prevention stop being separate projects.

Key takeaways

  • Primordial prevention: act on LDL before plaque exists, since short-horizon risk calculators make younger adults look safe while exposure compounds.
  • Cumulative LDL exposure (concentration multiplied by years) is the mental model, so decades of modest lowering beats a late aggressive correction.
  • A high HDL-C gives no license to tolerate high LDL. Genetic studies and failed HDL-raising trials both point the same way.
  • Combination therapy can reach an LDL target while keeping statin doses lower, which matters because high-dose statins carry a small diabetes signal.
  • Pull your own numbers this week: LDL-C, non-HDL-C, apoB, and Lp(a) if your lab offers it, and read them against family history.
We have the knowledge to prevent heart disease now. It's applying it earlier in life that really makes the difference. — Michael Davidson

Read the full episode summary

3. Building Strength and Muscle Mass: Optimize Training and Nutrition for Longevity (AMA 71)

The Peter Attia Drive · Peter Attia · Nick Stenson · 1h 57m · July 2026

The second half of Attia’s exercise case, and the one people underweight. He separates strength, hypertrophy and power, then makes the claim that power fades first as type 2A fibers atrophy, which is why explosive work belongs in a program long before you feel old enough to need it. What makes this a top-three listen is how low the prescribed heroics are: three one-hour full-body sessions, most sets stopped with one to two clean reps in reserve, progression through reps or sets or a slower eccentric when added load gets risky, and roughly 0.8 to 1 gram of protein per pound of body weight to turn the work into adaptation.

Key takeaways

  • Strength tracks mortality and disease risk more tightly than muscle size alone, so train for capacity you will want in your eighth decade.
  • Progressive overload has several levers: reps, sets, shorter rest, and a slower eccentric all raise demand when more weight raises injury risk.
  • End most sets with 1-2 good-form reps in reserve. Attia argues this captures nearly the benefit of training to failure at a fraction of the cost.
  • Power declines before strength and size. Scale explosive concentric work, warm up properly, and keep the ability to move fast.
  • Protein target is roughly 1.6-2.4 g/kg per day, with sleep, hydration and stress management treated as conditions for adaptation.

Read the full episode summary

4. Gayatri Devi on the Evolution of Alzheimer’s Disease and Dementia Care

The Peter Attia Drive · Peter Attia · Gayatri Devi, M.D. · 1h 58m · July 2026

The most hopeful episode in this set and the one carrying the most provocative single observation. Devi describes two patients whose siblings have advanced Alzheimer’s pathology while they, after decades of aggressive treatment for neuroimmunological disease, have none, which puts neuroinflammation upstream of amyloid as a target. Her clinical argument is just as useful: Alzheimer’s behaves as a spectrum across functional domains, cognitive reserve lets high performers ace a MoCA ten years into their condition, and the interventions worth doing in your 30s and 40s include suppressive treatment of herpes viruses, the shingles vaccine, and serious dental hygiene.

Key takeaways

  • Neuroinflammation may precede amyloid, which in turn precedes tau and synaptic loss. Early anti-inflammatory action could interrupt the whole cascade.
  • Stage the person across functional domains (memory, language, visual-spatial) and retire the single global number.
  • Cognitive reserve hides early disease. A patient at the 99th percentile overall but the 10th in one domain is already declining.
  • Forgetting proper names is normal; trouble retrieving common nouns like chair, book or teal warrants evaluation.
  • Every Alzheimer’s risk factor is also a stroke risk factor, which is why vascular health and brain health converge.
I have two patients in my practice where the sibling is riddled with Alzheimer's, but these two patients have severe neuroimmunological disease that they've been treated with aggressively for decades, and neither of them have a lick of Alzheimer's pathology. It's quite impressive. — Dr. Gayatri Devi

Read the full episode summary

5. Jim Otvos on NMR Blood Analysis and Reading Mortality Risk From a Single Blood Sample

The Peter Attia Drive · Peter Attia · Jim Otvos · 2h 27m · August 2026

Two and a half hours with the man who invented the NMR lipid test, and the payoff is a belief correction most listeners are still carrying. Small dense LDL looked dangerous for twenty years because at any given LDL-C, smaller particles mean more particles; match people on particle number and size adds nothing to risk. Otvos closes the argument with familial hypercholesterolemia, where particles are large and plentiful and homozygous patients die in their thirties. The practical half is the discordance finding: in roughly 20% of patients LDL-C and particle number disagree, and particle number is the one that predicts events.

Key takeaways

  • LDL particle number (LDLP) drives cardiovascular risk. Once LDLP is matched between patients, particle size carries no additional information.
  • In about 20% of people LDL-C and LDLP are discordant, and in those cases LDLP is the better predictor of outcomes.
  • NMR measures the containers, so results come back as nmol/L of particles where a standard panel reports mg/dL of cholesterol mass.
  • Familial hypercholesterolemia settles the size debate: large particles in high numbers still kill homozygous patients by age 30-35.
  • Trials that enlarged particles (niacin, CETP inhibitors) failed. Drive particle number down and let size fall where it will.
People with FH have very high LDL cholesterol levels. They also have very high LDL particle levels. But those particles are large. They're not small. And there's, you know, these are people that die when they're 30 or 35 years old when they're homozygous FH. So this idea that somehow fluffy large LDL particles are not to worry about or benign is completely facious. — Dr. Jim Otvos

Read the full episode summary

6. Tom Dayspring on Brain Lipidology: APOE and Cholesterol Homeostasis

The Peter Attia Drive · Peter Attia · Tom Dayspring · 1h 44m · June 2026

The episode that dismantles the most common objection to aggressive lipid lowering, which is the fear that a low LDL starves the brain. Dayspring’s case is anatomical: the brain holds 20-25 grams of cholesterol, roughly twenty times the liver’s store, synthesizes all of it behind the blood-brain barrier from the second trimester onward, and moves it between astrocytes and neurons on APOE particles that have no contact with plasma lipoproteins. His clincher is the two-year-old with an LDL-C of 30 mg/dL whose brain is growing faster than it ever will again. He also inverts what LDL is for: returning cholesterol to the liver.

Key takeaways

  • The brain makes 100% of its own cholesterol and holds 20-25 grams, about twenty times the liver’s content, fully separated from plasma.
  • Lowering plasma LDL leaves brain cholesterol untouched. Toddlers with LDL-C near 30 mg/dL build adult-sized brains.
  • The brain runs its own lipoprotein system on APOE: astrocytes synthesize, package, and ship cholesterol to neurons.
  • Around age 10 neurons stop making cholesterol to conserve ATP, since one molecule costs over 30 ATP, and outsource it to astrocytes.
  • Atherosclerosis has one requirement: cholesterol in the artery wall. Every apoB particle the liver clears is one that cannot enter.
The brain holds on to cholesterol like the bank holds on to its gold in the vault — Tom Dayspring

Read the full episode summary

7. Lisa Mosconi on Alzheimer’s Disease in Women, Hormonal Transitions, and New Therapies

The Peter Attia Drive · Peter Attia · Dr. Lisa Mosconi · 2h 14m · January 2026

The best available treatment of the 2:1 female-to-male Alzheimer’s ratio, and Mosconi spends the episode closing the obvious escape hatches. Longer life expectancy fails as an explanation: women outlive men by two to three years, and the gender skew is absent from vascular dementia, reversed in Parkinson’s and frontotemporal dementia, and flat in Lewy body disease. What remains is the menopausal transition, which her 2017 imaging work showed to be the most neurologically active phase of a woman’s midlife. Her reframe (a midlife disease with late-life symptoms) is the single most useful sentence on this page for anyone deciding when to start paying attention.

Key takeaways

  • Alzheimer’s pathology accumulates through midlife and only exceeds the brain’s compensation decades later, which moves prevention into the 40s and 50s.
  • Imaging of at-risk adults aged 45-65 shows women carrying more pathology and progressing faster than age-matched men.
  • Life expectancy cannot explain a 2:1 prevalence gap, and other dementias show no comparable female skew.
  • Menopause is the candidate mechanism: estrogen is neuroprotective and co-localizes with memory centers like the hippocampus.
  • Women’s verbal-memory advantage inflates scores on standard diagnostics, delaying diagnosis while pathology advances.
Alzheimer's is not a disease of old age. It's a disease of midlife with symptoms that start in old age. Alzheimer's starts in midlife with negative changes in the brain and that later on lead to the symptoms and the clinical diagnosis of dementia. — Dr. Lisa Mosconi

Read the full episode summary

8. Breast Cancer Screening: Understanding Risk and Deciding When to Start

The Peter Attia Drive · Peter Attia · 50m · June 2026

Fifty minutes that turn a guideline argument into a personal decision, and the clearest example on this page of Attia optimizing for one patient over population averages. The numbers do the work: CISNet modeling puts annual mammography at a 42% mortality reduction against 30% for biennial, with interval cancers at 11% versus 38%. At least 9% of women qualify for supplemental MRI and 0.4% receive it, which he calls an execution failure as opposed to an evidence gap. His scheduling fix is the actionable part: formal risk assessment by 25, because learning your Tyrer-Cuzick score at 42 arrives too late to change the plan.

Key takeaways

  • Annual mammography models out to a 42% mortality reduction versus 30% biennial, with 76% of cancers found at stage 1 against 56%.
  • Run a formal risk assessment (Tyrer-Cuzick or similar) by age 25 so high-risk women can start MRI years before their first mammogram.
  • About 9% of women meet guideline thresholds for breast MRI and 0.4% get one. The tool exists and the referral pathway fails.
  • Dense tissue raises risk and hides tumors, since both read white. Half of screening-age women are dense; 3D mammography is the floor.
  • Abbreviated MRI keeps nearly all the sensitivity of a full exam in 10-15 minutes, and halves interval cancers from 5 to 2.5 per thousand.
When breast cancer is caught at stage 1, the 10-year survival is over 96%. By stage 4, 5-year survival is only around 30%. — Peter Attia

Read the full episode summary

9. Dom D’Agostino on the Ketogenic Diet, Ketosis, and Hyperbaric Oxygen

The Peter Attia Drive · Peter Attia · Dom D’Agostino · 2h 25m · December 2025

The most honest keto episode in the library, largely because D’Agostino spends part of it describing his own failure. He cut protein from 250 g to 70-80 g daily, lost 10-15 pounds, and watched his bench press collapse, which is the detail that makes the rest of his protein guidance credible. His clinical ground is firmer than the wellness conversation implies: two-thirds of drug-resistant pediatric epilepsy patients respond to a ketogenic diet and a third achieve full seizure control. The practical through-line is measurement, with blood and breath ketones telling different stories during a calorie deficit.

Key takeaways

  • Protein is the variable that decides whether keto works. D’Agostino lost 10-15 lb of mass and real strength after cutting to 70-80 g daily.
  • Log macros and total calories, meter blood ketones daily, and pull lipid panels regularly. Correlate the three before changing anything.
  • In a calorie deficit, blood ketones can read low while breath acetone reads high, because tissues clear ketones from blood quickly.
  • The strongest clinical case stays therapeutic: two-thirds of drug-resistant pediatric epilepsy responds, one-third reaches full seizure control.
The ketogenic diet is indeed a magical diet in the way that it remarkably changes our physiology and there's no other diet that exists that can for example manage drug resistant seizures — Dom D'Agostino

Read the full episode summary

10. Layne Norton on the Seed Oil Debate and How to Read Nutrition Evidence

The Peter Attia Drive · Peter Attia · Layne Norton · 2h 25m · January 2026

Booked as a debate and better for having become a seminar. Norton makes the evidence-based case that seed oils carry no unique harm while Attia steelmans the opposition, and the forensic detail is what earns the slot: the Minnesota Coronary Experiment and the Sydney Heart Study both fed margarine that was 25-40% trans fats, a confounder potent enough to explain their mortality signals on its own. Norton’s second argument is the one that generalizes beyond fats, since cardiovascular disease compounds over decades and a 1-to-7-year trial has no power to see it. His compounding analogy (8.5% against 9% annual returns) is the clearest framing of cumulative exposure on this page.

Key takeaways

  • The landmark anti-PUFA trials used margarine that was 25-40% trans fats, so their mortality findings cannot be attributed to seed oils.
  • Nutrition RCTs run 1-7 years while atherosclerosis compounds over decades, which is why short trials miss real differences.
  • Chase converging lines of evidence across study types, and name your own priors out loud. Norton names his low-carb training and his funding.
  • Fat structure matters mechanically: cis double bonds kink and fluidify membranes, while saturated and trans chains pack rigid.
  • Reverse causality distorts cholesterol-mortality data in sick and elderly cohorts, where a low number can mark wasting disease.
The scientific method is perfect it is a perfect method but it is done by people who are not and that is why it is so important to look at the overall consensus of the evidence. — Layne Norton

Read the full episode summary

11. Peptides: Separating Scientific Promise From Marketing Hype

The Peter Attia Drive · Peter Attia · 50m · August 2026

The most useful 50 minutes on this page for anyone being sold a longevity compound. Attia takes BPC-157 and CJC-1295 as case studies and shows how far biological plausibility, clinician access, third-party testing and a wall of testimonials fall short of evidence. His five-question screen (mechanism, human outcome data, dosing and safety and pharmacokinetics, personal risk-benefit, better-characterized alternative) converts a vague debate about whether peptides work into a falsifiable question about one molecule at one dose for one person. The sharpest point is that a trial studies a product, with its formulation, purity and lot-to-lot consistency, so a gray-market vial inherits none of that evidence.

Key takeaways

  • A peptide is a short amino-acid chain, so the label carries no information about safety or efficacy. Insulin and GLP-1 agonists sit in the same category.
  • Screen any compound on five questions: mechanism, human-outcome evidence, dosing and safety and pharmacokinetics, personal risk-benefit, alternatives.
  • CJC-1295 raises growth hormone and IGF-1, and a moved biomarker still owes you proof of better function, recovery or lifespan.
  • Anecdotes arrive pre-confounded: people start when symptoms peak and simultaneously change sleep, training, diet and physical therapy.
  • Clinical evidence attaches to a specific manufactured product, including purity, sterility, stability and lot consistency.

Read the full episode summary

12. Abbie Smith-Ryan on Women’s Health and Performance Across Life Stages

The Peter Attia Drive · Peter Attia · Abbie Smith-Ryan · 2h 24m · January 2026

The longevity window almost nobody frames as a longevity window. Smith-Ryan points out that women reach their genetic ceiling for bone density around 19 and spend the following sixty years defending it, which makes adolescent resistance training a fracture-prevention intervention with a fifty-year lag. Attia’s line that osteoporosis is a childhood disease is the compressed version. The rest of the conversation is practical physiology for training across the cycle: a luteal phase that burns an extra 200-300 calories with more inflammation and fluid retention, 1.6 g/kg of protein as a floor, and creatine at 5-10 g that also pulls extracellular fluid into muscle.

Key takeaways

  • Peak bone density arrives near age 19, so varied sport and resistance training in adolescence is osteoporosis prevention decades ahead of time.
  • Training is viable in every cycle phase, while the luteal week brings 200-300 extra calories burned, more inflammation, and fluid retention.
  • Target 1.6 g/kg of protein daily regardless of cycle phase. Muscle protein synthesis holds steady when intake is adequate.
  • Creatine at 5-10 g daily after loading supports performance and eases luteal bloating by moving extracellular fluid into muscle cells.
  • Menarche is when many girls leave sport. Teaching cycle physiology and food as fuel keeps them in it, which protects long-term health.
The earlier you start and the better base that you have, the easier it is over time to maintain that fitness. — Abbie Smith-Ryan

Read the full episode summary

What these episodes have in common

Every guest moves the intervention earlier, and they arrive there from different diseases

The strongest agreement across these conversations has little to do with any specific tactic. It is about when the clock starts. Michael Davidson wants LDL lowered before plaque is measurable, because a 10-year risk calculator clears a 35-year-old who is quietly accumulating the exposure that determines his lifetime risk. Lisa Mosconi calls Alzheimer’s a disease of midlife with symptoms in old age, and her imaging of adults aged 45-65 shows the pathology already separating women from men. Gayatri Devi pushes further upstream still, to anti-inflammatory action in the 30s and 40s. Attia’s own breast-cancer episode sets the formal risk assessment at 25.

What earns this the word convergence is that four specialists reached it from four different organ systems. A lipidologist, a neuroscientist, a neurologist and a cancer-screening framework all concluded that the decisive window opens decades before the diagnosis is available. Layne Norton supplies the quantitative reason in the seed-oil episode: cardiovascular disease compounds like an investment return, so 8.5% against 9% looks identical after two years and diverges violently after forty.

The practical consequence is uncomfortable. Most of the high-value actions on this page are ones you take while every test still reads normal, and the feedback loop that would tell you they worked arrives thirty years late.

The dashboard Attia actually trusts counts particles and capacity

Across these episodes a consistent measurement philosophy emerges: prefer the thing being counted over the category it falls into. Jim Otvos built a career on that distinction. NMR reads lipoprotein particles as containers, which is why results come back in nmol/L, and the twenty-year panic about small dense LDL dissolved once particle number was matched between groups. Size turned out to be a proxy that looked causal. In roughly 20% of patients the proxy and the count disagree outright, and the count wins.

The exercise episodes run the same play on fitness. VO2 max and strength are what Attia calls integrators of work done, which is why they predict mortality better than the categorical risk factors stacked next to them and also why they take years to move. Tom Dayspring applies the principle to anatomy: asking whether a low plasma LDL starves the brain is the wrong question, because the brain synthesizes its own supply behind the blood-brain barrier and ships it on APOE particles that plasma lipoproteins cannot touch.

Even the keto episode lands here. Dom D’Agostino’s advice reduces to instrumentation: log macros and calories, meter blood ketones daily, pull lipid panels, and treat the blood-versus-breath discrepancy as information about flux. The pattern across all 12 episodes is that a measurement earns its place by being mechanistically close to the thing you care about.

Where they diverge is evidence standards, and that turns out to be the real skill

These episodes disagree less about biology than about what counts as proof, and watching the disagreement is where the listening pays off. Norton treats the anti-seed-oil literature as structurally unreadable, because the trials everyone cites fed 25-40% trans-fat margarine and ran for a handful of years against a disease that takes decades. Otvos describes an entire therapeutic program, the drugs built to enlarge LDL and HDL particles, that was chasing a correlation. Attia’s peptide episode generalizes the lesson into a five-question screen, and the one question it insists on is whether a human outcome improved.

Devi pulls in the other direction, and that tension is worth sitting with. Her neuroinflammation hypothesis rests on two patients in her own practice, a sample size the peptide episode would reject out of hand. The difference is what each claim is used for: she is generating a hypothesis and recommending interventions that are cheap and independently justified (shingles vaccine, dental hygiene, suppressive antivirals), while a gray-market vial asks for money and an unquantified risk on the same quality of evidence.

That calibration is the transferable skill in the set. Norton makes the point most directly when he says the scientific method is perfect and the people running it are human, which is why he looks for converging evidence across study designs and names his own priors out loud before interpreting anything. Read these 12 episodes for the specific numbers, and keep the filter.

Every episode referenced

Frequently Asked Questions

What does Peter Attia say matters most for longevity?

Cardiorespiratory fitness, measured as VO2 max. In the AMA 79 episode he says it outperforms every other variable he can measure, including blood pressure, cholesterol, BMI, smoking status and age. The bottom 20-25% of the population carries a 4-5x higher annual all-cause mortality risk than the top 2-3%, and climbing one quartile is associated with a 50-75% reduction.

What biomarkers does Peter Attia track for longevity?

The lipid episodes give the clearest list: LDL-C, non-HDL-C, apoB and Lp(a), read against family history as early in life as possible, per Michael Davidson in episode 407. Jim Otvos adds LDL particle number from NMR, which disagrees with LDL-C in about 20% of people and predicts events better when it does. Alongside the blood work sit VO2 max and strength, which Attia treats as integrators of years of training.

Does Peter Attia think Alzheimer’s disease can be prevented?

The guests on these episodes argue the window is open far earlier than most people assume. Lisa Mosconi frames Alzheimer’s as a midlife disease with late-life symptoms, with pathology visible in at-risk adults aged 45-65. Gayatri Devi goes upstream to neuroinflammation, citing two patients treated aggressively for neuroimmunological disease for decades who show no amyloid pathology while their siblings have advanced disease.

What does Peter Attia say about peptides and longevity supplements?

Episode 403 is a 50-minute filter for exactly this question. Using BPC-157 and CJC-1295 as case studies, Attia shows that plausible mechanism, clinician access, third-party testing and testimonials all stop short of evidence, and proposes five questions: mechanism, human-outcome data, dosing and safety and pharmacokinetics, your personal risk-benefit, and whether a better-characterized option exists.

Where can I read Peter Attia Drive summaries for free?

Right here. Each of the ranked entries above links to our full write-up of that conversation, with takeaways, verbatim quotes and the official video where one exists, at no cost and behind no signup. The same treatment covers a 1,500-episode library spanning nutrition, neuroscience, training and business, all browsable from the episode index.

Keep exploring