Check · Protocols · Blueprint In review

Has the Blueprint protocol slowed Bryan Johnson's aging and made him biologically younger?

Claim attributed to Bryan Johnson (Project Blueprint / "Don't Die") , Johnson states he has "the world's slowest speed of aging: 0.48," that "my birthday now happens every 2 years," and that he is "#1 out of 5,677 global Rejuvenation Olympics competitors" (a three-test average of 0.54). He frames Blueprint as proof a person can "meaningfully slow down their speed of aging." He also sells the supplements, foods, and the DNA-methylation "Speed of Aging" test that generate these metrics, the measurement and the intervention share an owner.

Verdict Unproven

Evidence grade D Very low certainty

A real clock reading, an n=1 design, and a seller's own metrics. Johnson's DunedinPACE scores appear genuinely low, but a low clock reading is not proof that Blueprint slowed his aging, there is no control, no randomization, an invalid "years" extrapolation, and no published trial of the protocol. The biology is plausible; the causal claim is untested.

A genuinely low clock reading is not the same as a slowed life: the metric is a population research tool, the design is n=1, and the man reporting the result sells both the test and the protocol.

The theory

What it’s supposed to target

Multi-biomarker optimization
Bundle of dozens of interventions
Self-experiment (n=1)
Biological-age readouts

Blueprint is not one mechanism but a bundle: dozens of simultaneous interventions (a fixed diet, heavy supplementation, hard exercise, strict sleep and constant testing) aimed at pushing every measurable biomarker toward youthful ranges. The theory is that if you optimize enough organ-level markers at once, aggregate biological age, read off various clocks and tests, should fall.

Many individual pieces have real rationale, exercise, sleep and not smoking are genuinely pro-longevity. The problem is the bundle and the n=1: with everything changed at once in a single highly resourced person, you cannot tell which parts help, which do nothing, and which carry risk, and the biological-age readouts are themselves unvalidated for tracking one individual. A motivated experiment, not evidence that the protocol reverses aging.

Mechanism is theory, not proof. A plausible pathway explains why something might work, not whether it does. The verdict rests on the evidence below, not the elegance of the theory.

The claim

What would have to be true

DunedinPACE would have to be a validated, accepted measure of one individual's biological aging, not just a population-research instrument. It is not: peer-reviewed work finds individual-level use can be uninformative and potentially harmful.

A single person's clock reading would have to be attributable to the Blueprint bundle specifically, rather than to diet, exercise, weight loss, regression to the mean, or assay batch. With no control arm, this cannot be established.

A momentary pace-of-aging ratio would have to be extrapolatable into 'years younger' or a birthday every two years. Clock methodologists say it cannot be.

Moving the clock would have to translate into real outcomes, less disease and longer life. Even the best controlled trial has not shown that yet, and Blueprint has no outcome data at all.

The lifestyle core of Blueprint (a careful diet, heavy exercise, good sleep, not smoking) genuinely improves many health markers, but that supports 'healthy living helps,' not the claim that the bundle slowed aging by about 50%.

The evidence

What the evidence actually shows

The headline rests on one clock that was never built to judge one person

Johnson's claim leans almost entirely on DunedinPACE, a DNA-methylation estimate of the pace of aging. The clock is real and, as clocks go, unusually steady: its developers report test-retest reliability of ICC ~0.96 in the Dunedin birth cohort (N=1,037, ages 26 to 45). That reliability is genuine, and it is the strongest thing the claim has going for it.

But the same paper draws the line Johnson crosses. Establishing DunedinPACE as a surrogate for aging, the authors write, 'will require evidence that it is modifiable by intervention and that intervention-induced changes result in changes in healthy-lifespan phenotypes.' In other words, the tool's own creators say it has not yet been shown to certify that an intervention slowed anyone's aging, let alone one person's.

A 2025 peer-reviewed review in *Epigenomics* puts the individual-use question to rest more bluntly. Examining the leap 'from population science to the clinic,' Apsley, Etzel, Ye and Shalev conclude that epigenetic clocks 'fail to meet common standards for clinical utility' and that, used on a single person, they 'can be uninformative and potentially harmful.' The same group, writing for a general audience, calls them 'not appropriate tools for individual health decisions,' noting that results shift with diet, illness, and time of day, and that the effects show up at the level of groups, not individuals.

The best controlled evidence moved the clock by 2-3 percent, not by half

There is exactly one randomized trial that comes close to testing this kind of claim, and it did not test Blueprint. In the CALERIE trial (Waziry, Ryan, Belsky et al., *Nature Aging*, 2023), healthy adults were randomized to 25% calorie restriction or to eat freely for two years (197 with methylation data, of 220 randomized). Calorie restriction slowed DunedinPACE by a 'reduction in the pace of aging of 2-3%.' The two other leading clocks, PhenoAge and GrimAge, did not differ between groups.

Two to three percent is not nothing, the Columbia and NIA investigators noted it is in the range associated with meaningfully lower mortality risk in observational data, so it is a real, encouraging signal. But it is a group-average effect from a sustained, measured 25% calorie cut, and it sits roughly an order of magnitude below what Johnson's numbers imply. A 'speed of aging' of 0.48, or a 'birthday every 2 years,' implies a slowing of around 50%, many times the best controlled result, achieved without a control group.

CALERIE's authors are careful about what even their result means: 'It is therefore unclear if the changes in DunedinPACE observed during the 2-yr intervention will translate into reduced morbidity and mortality.' A conclusive test, they write, 'will require trials with long-term follow-up' on disease and death. The clock moved; whether the people live longer is still unmeasured.

An n=1 self-experiment, and noisy at that

Blueprint has never been tested in a controlled trial. As an independent critique by the American Council on Science and Health puts it, Johnson's results are 'anecdotal evidence or, at best, a case study', the lowest rung on the evidence ladder, and there is 'no way to guarantee' that the protocol, rather than the diet, the exercise, or chance, produced any change. The protocol itself is also a moving target: Johnson has abandoned plasma transfusions, growth hormone, and rapamycin after side effects or lack of benefit, which makes 'the Blueprint protocol' a hard thing to credit with a fixed result.

Even the measurement carries more wobble than one reading suggests. Higgins-Chen, Levine and colleagues (*Nature Aging*, 2022) showed that standard epigenetic clocks can disagree by up to about 9 years between two reads of the same blood sample; their fix, a principal-component retraining, was needed to bring replicate disagreement under 1.5 years. A single favorable reading in one person, absent rigorous noise control, is weak evidence on its own.

Finally, the '31 years younger' framing that surrounds these scores is not a measurement but a multiplication. Morgan Levine, who helped build advanced clocks, explains that aging rate is 'not measured in units of years,' that a momentary pace cannot be projected across a lifespan, and that a clock is 'a single imperfect estimate of aging, not a direct measure.' Taking an age of 47 and a pace of 0.66 to manufacture '31 years' is, as she describes it, not a valid calculation.

The man with the result also owns the scoreboard

The favorable story is produced and sold by the same person. Blueprint operates as a direct-to-consumer store for 100-plus supplements and 'longevity' foods, and Johnson's own site sells a third-generation DNA-methylation 'Speed of Aging' test for $325, measuring biological age across 11 organ systems, the same class of metric used to market his results. Independent reviewers describe the presentation as effectively 'a poorly disguised advertisement for a commercial product.'

The leaderboard is not independent either, and its credibility took a notable hit from inside. The Rejuvenation Olympics, where Johnson ranks #1, runs on tests from TruDiagnostic, a company that sells them. In a development that matters more than any outside skeptic's objection, Dr. Oliver Zolman, Johnson's own former chief scientist and a co-creator of the Rejuvenation Olympics, publicly stated that 'Bryan and TruDiagnostic made the leaderboards unscientific and clinically meaningless.' The eligibility rules were changed (the field was uncapped and the ranking shifted to absolute pace of aging), and the person who helped build the contest no longer vouches for it.

None of this proves the readings are wrong. It means the headline metric, the test that generates it, and the leaderboard that ranks it all sit inside one commercial ecosystem with a financial interest in a flattering number, exactly the configuration that calls for independent confirmation, and exactly what is missing.

Evidence quality

Studies, graded, and who paid

Johnson's DunedinPACE scores are low and repeated B Moderate certainty

The underlying readings appear real; DunedinPACE has high reported reliability (ICC ~0.96). This is a measurement, not a fabrication.

A low clock reading proves Blueprint slowed his aging D Very low certainty

An n=1, uncontrolled self-experiment cannot separate a protocol from diet, exercise, regression to the mean, or assay noise. Lowest tier of evidence.

Epigenetic clocks can certify one person's biological age D Very low certainty

Peer-reviewed work concludes clocks 'fail to meet common standards for clinical utility' at the individual level; they are population tools.

The protocol made him '31 years younger' / a 'birthday every 2 years' D Very low certainty

A momentary pace ratio cannot be multiplied across a lifespan. A clock architect calls the arithmetic invalid.

Cited studies with type, size, funding/conflicts, and limitations.
#	Study	Type	Size	Funding / COI	Key limitations
1	Belsky et al., DunedinPACE (eLife)	Biomarker development/validation cohort	N=1,037 birth cohort, ages 26-45, 4 timepoints	Independent Academic (Duke/Columbia/King's College London); NIH and related public funding. Authors developed the algorithm.	Population-level reliability (ICC ~0.96); authors state surrogate-endpoint status is not yet established. Validates the clock for research, not for certifying one person's response.
2	Waziry/Ryan/Belsky, CALERIE caloric restriction RCT (Nature Aging)	Randomized controlled trial (secondary methylation analysis)	220 randomized; 197 with methylation data; 2 years; 25% calorie restriction vs ad libitum	Independent NIH/NIA-funded CALERIE consortium; not industry-funded.	Only ~2-3% group-level DunedinPACE slowing; PhenoAge/GrimAge unchanged; no disease/mortality outcomes yet. Tests calorie restriction, not Blueprint.
3	Higgins-Chen & Levine, clock reliability (Nature Aging)	Methods/reliability study	Multiple methylation datasets and technical replicates (six clocks)	Independent Academic (Yale); NIH/NIA + NIMH public funding. Authors are clock methodologists, not product sellers.	Shows up to ~9-year disagreement between replicate reads of the same sample; reliable tracking needed a PC retraining fix. A single n=1 reading is weak without noise control.
4	Apsley, Etzel, Ye & Shalev, limits of clocks as personal biomarkers (Epigenomics)	Peer-reviewed review/analysis	n/a (review)	Independent NIH/NSF-funded academics (Penn State; Univ. of Illinois Urbana-Champaign); no product interest.	Concludes clocks 'fail to meet common standards for clinical utility' and individual use 'can be uninformative and potentially harmful.' Argument, not a Blueprint-specific experiment.
5	Shalev & Apsley, plain-language explainer (The Conversation)	Expert commentary (NIH/NSF-funded academics)	n/a	Independent Authors are NIH/NSF-funded academics (Penn State; Univ. of Illinois); no product interest.	Popular-press piece; states clocks are 'not appropriate tools for individual health decisions' and best used on populations. Notes fluctuation with diet/illness/time of day. Does not itself discuss seller conflicts of interest.
6	Borrus et al., 'When to Trust Epigenetic Clocks' (bioRxiv preprint)	Retrospective re-analysis (preprint, not peer-reviewed)	10 longitudinal datasets (6 interventions, 2 positive controls, 2 diet arms)	Independent Academic preprint (Yale Dept. of Psychiatry; NIA grants + Yale fellowships); no product-seller funding.	Preprint, used for context only. Shows first-generation clocks throw false positives that vanish after correction; reliable signals need multiple clocks to agree.
7	Levine critique of the '31 years' math (commentary, via NMN.com)	Expert critique / secondary reporting	n/a	Funding unknown Published on a site in the longevity-supplement space (NMN.com); used only as a conduit for Morgan Levine's methodological argument, which aligns with sources 1 and 3.	Not peer-reviewed and hosted on a commercial site; cited only for the methodological point that pace cannot be converted into 'years.'
8	ACSH critique of the Blueprint protocol	Independent expert critique	n/a	Independent American Council on Science and Health, an independent 501(c)(3) nonprofit; no stake in Blueprint.	Opinion/analysis, not original data. Characterizes results as 'a case study' with 'no way to guarantee' causation; notes abandoned interventions.
9	Blueprint 'Speed of Aging' test product page (money trail)	Primary claimant source / commercial listing	n=1 (self)	Industry-funded Seller's own site; Johnson profits from the test and products underlying the claim.	Confirms the $325 DNA-methylation test across 11 organ systems (conflict-of-interest evidence). Does not itself carry the headline quotes.
10	Johnson, 'I have the world's slowest speed of aging' (Blueprint blog)	Primary claimant source	n=1 (self)	Industry-funded Johnson's own commercial site; he sells the products and test behind the claim.	Source of the verbatim claims (0.48; 'birthday every 2 years'; three-test average 0.54; #1 of 5,677). Self-reported, uncontrolled, marketing context.

The pattern is ownership. The metric (DunedinPACE), the test that sells it ($325 on Johnson's own site), and the leaderboard that ranks it (Rejuvenation Olympics, run on TruDiagnostic's tests) all sit inside one commercial ecosystem with an interest in a flattering result. The independent science points the other way with unusual consistency: the clock's own developers say it is not yet a validated surrogate; a peer-reviewed review says individual use 'can be uninformative and potentially harmful'; the one rigorous RCT (independently, government-funded) moved the clock only 2-3% by testing calorie restriction, not Blueprint. The most damaging line does not come from an outside critic at all, it comes from Johnson's former chief scientist and the leaderboard's co-creator, Oliver Zolman, who calls the rankings 'unscientific and clinically meaningless.'

See who funded the studies behind every Caveat check

Stay neutral

Unproven ≠ disproven

Unproven is not disproven. Johnson is not making this up: he genuinely has repeated, low DunedinPACE readings on a clock with real reliability, and several of his health markers may have improved from a careful diet, heavy training, and weight loss. The failure here is inferential, not fabricated, a low clock number is simply not the same thing as 'Blueprint slowed my aging.' Nothing in the evidence shows the claim is false; it shows the claim has never been tested in a way that could be true or false.

And it is structurally hard to test. There is no validated clinical endpoint for 'biological age', the clocks are surrogates with population-level, not individual-level, validity, and the FDA has accepted none as a validated endpoint. An n=1 design has no control and no randomization, so it cannot separate a protocol from diet, placebo, regression to the mean, or assay noise. The outcomes that would actually settle it, disease, death, healthspan, take decades to appear, which is why even CALERIE has only a surrogate result so far. On top of that, the claimant owns both the intervention and the measurement and has not published the protocol as a controlled trial. These obstacles explain why good evidence is missing; they do not substitute for it.

The gap

Where claim and evidence diverge

The claim and the evidence diverge at the word 'slowed.' What is shown is a low, repeatable reading on one population-level clock. What is claimed is that a specific 100-plus-component protocol caused that reading and made a single man biologically younger, a causal, individual-level statement the instrument was never validated to support.

The numbers diverge by an order of magnitude. The best controlled result for any lifestyle intervention is a 2-3% group slowing of DunedinPACE from sustained calorie restriction; Johnson's figures imply a slowing near 50%, from an uncontrolled design. And the most vivid framing, '31 years younger,' a 'birthday every 2 years', is not a measurement at all but an arithmetic projection that clock methodologists say is not valid.

Follow the funding

The money trail

Blueprint runs as a direct-to-consumer store selling 100-plus supplements and 'longevity' foods (olive oil, cocoa, protein, the 'Longevity Mix') under the Bryan Johnson / Blueprint brand.

Johnson's own site sells a third-generation DNA-methylation 'Speed of Aging' test for $325, measuring biological age across 11 organ systems, the same class of metric used to market his results.

Johnson is founder and owner of Project Blueprint and the public face whose personal 'results' function as the primary marketing for the products; an independent reviewer characterizes the content as effectively a poorly disguised advertisement.

The Rejuvenation Olympics leaderboard he tops runs on TruDiagnostic's commercial tests, and the contest's co-creator and Johnson's own former chief scientist, Oliver Zolman, has publicly disowned the rankings as 'unscientific and clinically meaningless.'

Bottom line

The honest read

The honest read is narrow and specific. Bryan Johnson appears to have genuinely low, repeatable readings on a respectable epigenetic clock, and the lifestyle backbone of Blueprint, good diet, hard exercise, sleep, not smoking, plausibly improved real health markers. To that extent, 'healthy living helps' is fair. What is not established is the headline: that the Blueprint protocol slowed his rate of aging or made him biologically younger. That is an n=1, uncontrolled, self-measured case study using a metric validated for populations, not individuals, marketed by the man who profits from both the test and the supplements.

The size of the claim makes the gap starker. The best controlled trial moved the same clock by 2-3% with sustained calorie restriction; Johnson's numbers imply roughly ten times that, with no control group, and the most quotable version ('31 years younger,' 'a birthday every 2 years') rests on arithmetic that clock experts call invalid. None of this means he has been disproven, it means the claim sits at the lowest tier of evidence, downgraded further by an invalid extrapolation and a measurement system he owns. Plausible, sponsor-controlled, and unproven.

Falsifiable

What would change this verdict

A randomized controlled trial of the Blueprint protocol (or a close approximation) versus a credible comparator, run and analyzed independently of Johnson's commercial ecosystem, showing a clock effect of the magnitude he claims.

Validation of an epigenetic clock as an accepted individual-level endpoint, independent, standardized, noise-controlled, such that a single person's reading could legitimately certify that an intervention slowed their aging.

Hard-outcome data (reduced disease incidence or mortality, or robust functional/healthspan measures) tied to the protocol, since clock movement is a surrogate and the real test is whether people actually live longer or healthier.

Receipts

Sources

Common questions