This chapter is about how longevity claims get sorted before they get judged: two taxonomies (Huberman's four categories, Attia's three mutually exclusive ones) plus a six-question screen for any exogenous molecule a person takes. It then applies that machinery to the only two interventions with cross-species lifespan evidence — caloric restriction and rapamycin — and to the NAD/sirtuin story those same standards falsified, including why resveratrol and nicotinamide riboside failed Interventions Testing Program testing outright.
The organising move of this chapter is that you do not evaluate a longevity intervention until you have classified it. Two complementary taxonomies do that work, and Longevity Intervention Classification Frameworks holds both.
Huberman's version has four categories: behavioral dos-and-don'ts; management of calories, glucose, insulin and mTOR; targeting specific cellular pathways such as NAD; and the 'kitchen sink' approach of stacking many interventions at once, with Brian Johnson as the exemplar. Attia's version is tighter — three buckets, explicitly mutually exclusive and collectively exhaustive. First, essential behavioral factors: eating, sleeping, moving, which are mandatory rather than optional. Second, exogenous molecules targeting specific disease processes — metformin, SGLT2 inhibitors, GLP-1 agonists, PCSK9 inhibitors, statins, bempedoic acid. Third, geroprotective molecules that target aging hallmarks directly, which per this framing contains rapamycin and, on the strength of Interventions Testing Program evidence, caloric restriction.
The payoff is that the bucket determines the evidence bar. A candidate that addresses baseline survival behavior is judged differently from one that lowers a specific disease risk, which is judged differently again from one claiming to touch the aging process itself. The methodology for that judgment — the ITP standard, statistical significance, what counts as a real result — belongs to Evaluating Evidence & Supplement Quality and is assumed here rather than re-derived.
Two framing claims sit alongside the taxonomies. The Titanic analogy: emotional and mental health is the ship's heading, the dominant directional factor; sleep, exercise and nutrition are the other essential pillars; supplement choices such as NAD precursors are deck details — 'lobster vs. steak' — that matter far less than where the ship is pointed. And the arithmetic behind it: in modern society, living longer is mathematically equivalent to delaying the onset of chronic disease — cardiovascular, cerebrovascular, cancer, neurodegenerative, metabolic — because infection and childbirth complications are no longer primary mortality drivers. That equivalence is why foundational behaviors dwarf any single supplement's contribution.
Attia extends the classification schemes into a screen for any exogenous molecule — medication or supplement alike — to be run before deciding to keep taking it. The six questions, in order, as recorded in Longevity Intervention Classification Frameworks:
The walkthrough is capped by a risk-reward assessment: side-effect profile, effect size, confidence in the safety data, and supplement quality and sourcing. That final step sorts the molecule into a bucket — and it is where creatine lands as low-risk, modest reward. That is creatine's only appearance in this chapter's material; nothing here tests creatine against lifespan endpoints, and the substance of its mechanism, dosing and safety evidence lives in Creatine Fundamentals. What this chapter contributes to the creatine question is the yardstick, not a verdict.
Two mental models travel with the framework. The bucket-of-exogenous-molecules audit treats medications, supplements and hormones as one undifferentiated inventory, refusing the default assumption that supplements are automatically lower-stakes than drugs. The kidney-filtration or zero-base review says to periodically dump the whole bucket out and ask, item by item, what should be added back — rather than keeping something merely because it is already on the list. Note that questions 2 and 3 reuse the lifespan-versus-healthspan and disease-specific-versus-geroprotective axes from the taxonomies above; the framework is those axes generalized from longevity interventions to everything a person swallows or injects.
The bar that separates the serious longevity interventions from the rest is cross-species replication. Caloric restriction is one of only two interventions — the other being rapamycin — that has ever extended lifespan across all four model-organism categories tested by the Interventions Testing Program: yeast, worms, flies, and mammals. That is the whole claim to distinction, and it is a high one; the ITP standard itself is the subject of Evaluating Evidence & Supplement Quality.
This is also where the taxonomies get interesting rather than tidy. Caloric Restriction & the Sirtuin Hypothesis points to CR fitting Attia's essential behavioral category — the eating, sleeping, moving bucket — while Longevity Intervention Classification Frameworks lists it in the geroprotective molecules bucket on the strength of the ITP evidence. The two placements are recorded side by side in this chapter's material without being reconciled, which is worth noticing given that Attia's three categories are specified as mutually exclusive.
What CR does not come with, in this material, is a human protocol: no dose, no percentage restriction, no duration is given here. The cross-species lifespan result is the claim; the translation to a person's plate is not covered.
For years the accepted mechanism for CR's longevity effect was the sirtuin pathway. That explanation was falsified, and the falsification is the most consequential single event in this chapter.
The evidence came out of yeast, across multiple strains. Some strains respond to caloric restriction but not to sirtuin overexpression; others show the reverse. A 2004 study by Kaplan and Kennedy found the two effects were independent and additive rather than synergistic — which is the signature of two separate mechanisms, not one pathway mediating another. The verdict quoted in Caloric Restriction & the Sirtuin Hypothesis is blunt: 'there's no evidence whatsoever that sirtuins have anything to do with caloric restriction and vice versa.'
One residual signal survives. A single transgenic mouse model, SIRT6 overexpression, has shown a survival benefit — 10–15% longer life, and only in males; females showed no extension. That is the entire positive column.
The commercial casualty is resveratrol, which was marketed on exactly the sirtuin/CR-mimetic story and its 'French Paradox' packaging. In ITP testing it showed categorically nothing. And the dose-response arithmetic makes the marketing claim self-defeating: achieving the doses studied would require drinking one's body weight in wine daily. The falsification matters beyond resveratrol, because the same sirtuin link was load-bearing for the NAD hypothesis taken up later in this chapter.
Rapamycin is the second of the two interventions to have extended lifespan across all four ITP model-organism categories, and per Rapamycin as a Geroprotective Molecule it shows uniform life extension unmatched by any other molecule tested. It acts on mTOR — a pathway robustly expressed from infancy through puberty, one of the most rapid aging phases of life, and tapering off afterward. That expression profile is what underwrites the framing that 'development is aging.'
On dosing, the material is specific and personal rather than protocolised: Peter Attia takes 8 mg once weekly, cycling roughly two months on and one month off because of side effects — around 10% of his rapamycin patients develop mouth sores. The drug is off-patent and cheap, roughly $40 a week at a therapeutic geroprotective dose, and the source explicitly reads that price as the reason rapamycin attracts so little marketing noise compared with heavily promoted NAD precursor products.
The most striking finding is that rapamycin's immune effect flips direction with the dosing schedule, not the total dose. Taken daily, it appears immunosuppressive — consistent with its origin in transplant medicine. Taken in pulsed weekly doses, it appears to enhance immune function instead: the reference point is Mannick and Klickstein's 2014 study of elderly subjects' response to flu vaccination. That reframes rapamycin as an immune modulator rather than a pure suppressant, and makes schedule the operative variable. In the three-bucket taxonomy of Longevity Intervention Classification Frameworks, rapamycin is the cleanest example of a geroprotective molecule — one aimed at aging hallmarks directly, not at a named disease.
Caloric restriction and rapamycin are the only two known non-genetic interventions shown to extend lifespan across yeast, worms, flies and mammals, and both act by downregulating mTOR. That shared mechanism is exactly what creates the chapter's sharpest tension, because mTOR activation in skeletal muscle is desirable. The resolution offered in Rapamycin as a Geroprotective Molecule is a distinction between local and systemic: chronic systemic mTOR activation may be harmful, but the goal is to activate mTOR locally through exercise plus leucine and protein rather than to suppress or activate it body-wide. The mechanics of that local activation are the business of Protein & Muscle-Building Nutrition and Resistance Training & Performance Supplements.
Two translatability caveats then cut against reading mouse lifespan numbers as human promises. The first is the 'longevity quotient' framework attributed to Eric Verdin — body-size-predicted lifespan compared against actual lifespan across species. Mice live far below their predicted ceiling; humans are already close to their roughly 80-year predicted ceiling. There is simply less theoretical headroom for an effect of comparable magnitude in a species that has already largely closed its gap.
The second is direct and empirical: human resistance-training studies found that rapamycin blunted muscle protein synthesis. This is not a theoretical worry but a measured conflict with the muscle-preservation goals that run through the rest of this subject — the same protein and training-to-failure targets covered in Protein & Muscle-Building Nutrition and Resistance Training & Performance Supplements. The material attributes this discussion to Drive Podcast ep. 369, Attia and Galpin. What it does not resolve is whether pulsed weekly dosing, which flips rapamycin's immune profile, also spares muscle protein synthesis.
NAD — nicotinamide adenine dinucleotide — is a coenzyme and electron shuttle active in 500–600 cellular pathways, though 99% of its role is the simple business of shuttling electrons as the NAD/NADH pair. The critical fact in NAD Precursor Supplementation (NR, NMN, IV NAD) is that the total NAD pool is tightly regulated regardless of age. What actually declines with age is redox potential, the electron-shuttling capacity — which reflects declining mitochondrial function rather than NAD depletion. The supplement category's founding image, a draining NAD tank that needs refilling, is not what the biology shows.
The second pillar was the sirtuin link: NAD is consumed as a substrate by sirtuins during DNA repair, so boosting NAD was supposed to boost sirtuin activity and thereby replicate caloric restriction's longevity benefit. That inference depended on the sirtuin/CR connection falsified above and described in Caloric Restriction & the Sirtuin Hypothesis. Remove it and the rationale is missing a middle term.
Because NAD itself is not orally bioavailable, precursors are used. NR (nicotinamide riboside) crosses cell membranes directly and is typically dosed at 300–600 mg. NMN (nicotinamide mononucleotide) carries an extra phosphate group that must be cleaved before uptake, giving it weaker bioavailability evidence than NR — though no head-to-head blood-NAD trial has ever directly compared the two. NMN is typically dosed around 1500 mg. IV NAD infusion bypasses oral bioavailability entirely: 30 minutes to 3 hours, 500–1000 mg, roughly $300–1000 per session, with evidence of meaningful benefit relative to that cost lacking.
Then the hard test. NR failed ITP testing entirely — no lifespan extension, no healthspan improvement, and no change in blood NAD — while rapamycin, acarbose, SGLT2 inhibitors and 17-alpha estradiol all extended lifespan in the same program. The NMN-versus-NR debate is characterised here as commercial posturing rather than a scientific one; the FDA reclassified NMN as an investigational drug, blocking supplement sales, after a company began drug trials with it, paralleling what happened with NAC. The lesson carried forward is that GRAS status, which NR retains, confers the legal ability to sell, not proof of efficacy — a distinction that belongs to Evaluating Evidence & Supplement Quality.
One plausible positive human signal survives: NAD precursors showed a 60–80% reduction in basal and squamous cell carcinomas — not melanoma. That aligns with skin showing the largest age-related NAD decline of any tissue, which makes it a coherent finding rather than a stray one. It is a specific dermatological result, not a longevity claim, and it sits alongside an otherwise failed lifespan record. Against rapamycin and caloric restriction, the two interventions with robust cross-species evidence, NAD precursors currently lack comparable support.