BioHackBuyer
Written by

David Westbrooks

Published October 17, 2026

Not medical advice. Talk to a qualified clinician before starting any new health protocol, especially if you take medication or have a medical condition.

Bryan Johnson Blueprint protocol: what is actually evidence-based

The most-copied protocol in longevity, graded intervention by intervention. The two strongest things in it are free, the biomarker headline rests on an instrument the field says is not validated, and three of the advanced therapies have no data in healthy adults at all.

The short version

Bryan Johnson's published protocol is the most-copied self-experiment in the longevity genre, and almost every page about it does one of two things: repeats it as a plan, or dismisses the man and skips the evidence. This page does neither. I graded each published intervention against the primary literature for that intervention, and I am treating his own pages as the source for what he claims, never as the source for whether it works.

Two findings carry the page. First, the strongest effects in the whole protocol are free: tested cardiorespiratory fitness and sleep duration around seven hours both have Strong human evidence, and neither is sold in his store. Second, his headline biomarker claims are measured with instruments the field itself says are not validated, and those instruments can move by years for technical reasons. A clock moving and a life lengthening are two different claims.

If you are deciding whether to spend $98 a month or $4,499 once, the graded sections below are the decision aid. Everything asserted here is sourced, and every number is either a live first-party page or a resolvable PMID.

Biological age and living longer are two claims

Johnson publishes biological age figures, and his own pages report things like "biological age" milestones next to blood pressure and vascular measurements. Those numbers come from aging clocks, mostly DNA methylation clocks, and they mean something narrower than the way they get repeated.

The field does not yet agree on what makes a reliable biomarker of aging. The consensus framework published in Cell in 2023 states that standards are still lacking and that validation is incomplete, which is a polite way of saying these markers are not established surrogate endpoints for living longer (PMID 37657418). A surrogate endpoint is a stand-in that has been shown to track the outcome you actually care about. Biomarkers of aging have not cleared that bar.

There is a second problem underneath that one, and it is technical rather than philosophical. When the same sample is measured twice, six prominent epigenetic clocks produced deviations of up to nine years between replicates, which is why the authors proposed a retrained version to make the measurements agree (PMID 36277076). If a test can move nine years on the same blood sample, then a single person tracking their own clock over months is reading signal and noise through the same window.

None of that means the protocol does nothing. It means the measurement cannot carry the claim. Movement on a clock is an interesting data point about a biomarker. It is not evidence of a longer life, a lower disease rate, or a slower death.

Evidence: Strong that no aging biomarker is yet a validated surrogate endpoint PMID 37657418, and strong that the same sample can read years apart PMID 36277076.

What this does not mean: that the interventions are useless. It means the clock cannot prove they work.

The free tier beats the stack

This is the part of the audit that matters most, and it is the part you will not see in a supplement advertisement. The largest mortality effect anywhere in this dossier belongs to cardiorespiratory fitness. In 122,007 patients referred for treadmill testing and followed for a median of 8.4 years, risk-adjusted all-cause mortality fell steadily with fitness, and the elite group had a hazard ratio of 0.20 against the lowest group (PMID 30646252). Grade: Strong.

Sleep is the second. A dose-response meta-analysis found a U-shaped curve, with the lowest risk at roughly seven hours a night and rising risk on both sides of it (PMID 28889101). A separate pooled analysis of 16 prospective studies, covering 1.38 million participants and 112,566 deaths, found short sleep associated with a 12 percent higher risk of death (PMID 20469800). Grade: Strong for the association. These cohorts mostly rely on self-reported sleep, so read it as a strong pattern rather than a proven lever.

The eating window is where the popular version of the protocol and the trials part ways. In a 12-month randomized trial, patients with obesity on calorie restriction plus an eight-hour eating window lost no significantly more weight than patients on calorie restriction alone, a net difference of -1.8 kg with a confidence interval crossing zero (PMID 35443107). The TREAT trial likewise found 16:8 did not beat three structured meals for weight loss over 12 weeks (PMID 32986097). Grade: Strong, as a null.

Put those three together and you have the most effective part of the protocol, available for nothing, with no product to sell. Nothing in the Blueprint storefront addresses fitness or sleep because there is nothing to sell.

Evidence: Strong for fitness PMID 30646252 and sleep duration PMID 28889101. Strong nulls for the eating window PMID 35443107.

Who this is for: anyone deciding whether to spend money at all. These three cost nothing and carry the largest measured effects on the page.

The supplement tier, graded

The supplements are his own brand, so the protocol and the storefront are the same business. That is worth stating once, plainly, before any price: the person publishing the protocol is the person selling the version of it you can buy.

Prices below are the variant prices on his own store, read from the store's product data on October 17, 2026, for the exact size named. The Easy Stack is $98.00, the Medium Stack is $270.00, and the Big Stack is $449.00. Individual products run from Ashwagandha + Rhodiola at $24.00 for a 30-day supply to Creatine at $40.00 for a single pack, with Longevity Mix at $49.00 and Omega-3 at $39.00 per 30 days.

Then the grades, which are less exciting than the packaging. Vitamin D at 2000 IU a day did not reduce cancer or cardiovascular events in the VITAL trial, which randomized 25,871 people for a median of 5.3 years (PMID 30415629). Marine omega-3 at 1 g a day missed its primary cardiovascular endpoint in the same trial (PMID 30415637). Grade: Strong, both as nulls. If a stack sells you fish oil on the strength of cardiovascular prevention, that is the trial to read.

A daily multivitamin had the same shape in COSMOS, which randomized 21,442 adults: no reduction in invasive cancer (PMID 35294969). The cognition substudies did find a small signal, a 0.06 standard deviation difference in global cognition and 0.12 in episodic memory, and the authors were explicit that it is not enough for a clinician to recommend multivitamins for prevention (PMID 38244989). Grade: Moderate for the small effect, Strong for the cancer null.

Cocoa flavanols are the clearest example of why you report both numbers. In COSMOS the primary endpoint, total cardiovascular events, missed, with a hazard ratio of 0.90 and a p value of 0.11 (PMID 35294962). Cardiovascular death, a secondary endpoint, came in at 0.73. Grade: Moderate. The headline you would want to quote is the secondary one, and it is not the finding the trial was powered for.

NAD precursors are the current favourite. A 60-day randomized trial in 80 middle-aged adults found blood NAD rose significantly at every dose from 300 mg to 900 mg, with no safety issues (PMID 36482258). A smaller 10-week trial in postmenopausal women with prediabetes found improved insulin-stimulated glucose disposal by clamp (PMID 33888596). Grade: Moderate. Both are honest results about narrow outcomes, and neither is a mortality endpoint. His own protocol page names nicotinamide riboside in its pregnancy-avoid list, which is the clearest public statement of what is actually in the stack.

Spermidine has cohort data: 829 participants followed up to 20 years, with a hazard ratio of 0.74 per standard deviation of higher intake, and the association held after adjustment for lifestyle factors (PMID 29955838). Grade: Moderate, association only, no randomized trial with a clinical endpoint.

Taurine is where the field got ahead of itself and then corrected. The 2023 Science paper found circulating taurine declines with age and that supplementation extended healthspan in mice and monkeys (PMID 37289866). Two years later, a study of 137 men aged 20 to 93 found no association between circulating taurine and age, muscle mass, strength, physical performance, or mitochondrial function, and said so in the title (PMID 41061678). Grade: Limited for the human claim, however good the animal data looks.

Two products on the store have no graded row in this audit, and I would rather say that than dress them up. Creatine at $40.00 has a large evidence base for muscle and strength in the sports literature, and the graded ledger for this guide contains no creatine trial with a mortality or disease endpoint. Ashwagandha + Rhodiola at $24.00 is not graded here either. A curated bundle is a shopping convenience, not an evidence claim, and the free tier above is still the part with the biggest measured effect.

Prices, checked October 17, 2026

What you are paying for: a curated bundle of components whose individual evidence is graded above, most of it Moderate or null, assembled by the same business that publishes the protocol. If you want the cheaper version of the same category, our under $2,000 home setup guide rebuilds it from parts.

The prescription layer is not a consumer product

This layer is not a product and cannot be bought. As published on his protocol page on October 17, 2026, the prescription list runs: Armour Thyroid 60 mg, Levothyroxine 100 mcg, Minoxidil 3.75 mg, Metformin 500 mg, Tadalafil 5 mg, Jardiance 10 mg, Acarbose 200 mg, and Repatha 140 mg every other week. Two of those, metformin and acarbose, are being used off-label against a longevity hypothesis rather than a diagnosis, which is a different thing from being prescribed for diabetes.

The evidence for that hypothesis is where it gets thin. Metformin monotherapy was associated with better survival than matched non-diabetic controls in the UK Clinical Practice Research Datalink, covering 78,241 treated patients and 90,463 matched controls, with sulfonylurea doing worse (PMID 25041462). Grade: Moderate, observational, and first-line treatment choice is not random. Metformin did reduce progression to type 2 diabetes by 31 percent in a high-risk cohort, which is a real result about a different question (PMID 11832527). A 2026 review of the pharmacology puts the honest summary on the record: for repurposed drugs like metformin and rapamycin, clinical evidence for broadly applicable geroprotection remains limited and effects on aging remain unproven (PMID 42737424).

Rapamycin deserves its own paragraph because it is the drug people ask about. In genetically heterogeneous mice, rapamycin started late in life increased lifespan at the 90th percentile by 14 percent in females and 9 percent in males (PMID 19587680). In humans, the mTOR inhibitor RAD001 improved influenza vaccine response by about 20 percent in elderly volunteers (PMID 25540326), and the 48-week PEARL trial of weekly low-dose rapamycin in healthy adults found adverse events similar across arms while its primary endpoint, visceral adiposity, did not change (PMID 40188830). That is a safety trial with a null primary result, not a longevity result. The mechanism also cuts both ways: rapamycin impairs glucose tolerance and insulin action in mice by disrupting mTORC2 (PMID 22461615). One fact about the page as it stands: rapamycin does not appear anywhere on the protocol page as fetched on October 17, 2026. I am not going to narrate a start or a stop that the page does not contain.

His own protocol page also contradicts itself on acarbose dosing. The stack line says 200 mg. The prescription section says 400 mg daily. I record that as a fact about the page, not a claim about what he swallows. Prescription drugs need a physician and bloodwork, and that is the end of what this page has to say about acquiring them.

Evidence: Moderate for metformin, from observational data only PMID 25041462. Animal-only for rapamycin lifespan PMID 19587680, with a human safety trial whose primary endpoint did not move PMID 40188830.

Who should skip this section entirely: anyone looking for a product to buy. These require a physician, monitoring, and a diagnosis that makes the prescription appropriate.

The advanced therapies have no healthy-adult data

This is the layer where the evidence genuinely runs out, and it is advertised with the most confidence.

Mesenchymal stem cell injections: his page describes his own use as a trial, which is the accurate word. A critical review of the field concludes that therapeutic effects are still not spectacular and that long-term safety in specific tissue sites needs follow-up (PMID 31093291). Grade: Limited.

Cerebrolysin, taken for cognition: the randomized evidence is in post-stroke recovery, where a meta-analysis of nine trials found superiority over placebo on neurological scales (PMID 29248999), plus old trials in dementia of the Alzheimer type (PMID 8159781). There is no trial of cerebrolysin for cognition in healthy adults. Grade: Limited for the use case.

Follistatin gene therapy: the supporting work is preclinical, showing durable muscle growth in cynomolgus macaques from an AAV-delivered follistatin isoform (PMID 20368179). The human trial on record is a phase 1/2a in six patients with Becker muscular dystrophy, with mixed individual results (PMID 25322757). No healthy-adult data, and no long-term human safety data. Grade: Limited.

None of these have data in healthy adults. When a protocol lists three interventions at this level next to a multivitamin, the list implies a similarity that the evidence does not support.

What the measurement layer can and cannot prove

The measurement layer is where the protocol is most confident and the evidence is weakest. Two facts from the research above do most of the work here: aging clocks have no consensus validation as surrogate endpoints (PMID 37657418), and the same sample can read years apart on repeat measurement (PMID 36277076).

The screening half has its own null. The Cochrane review of general health checks in asymptomatic adults found no reduction in all-cause, cardiovascular, or cancer mortality, with more diagnoses and more drug treatment along the way (PMID 30699470). Grade: Strong. The positive whole-body MRI literature exists in defined high-risk groups, such as carriers of a TP53 mutation in the UK SIGNIFY study, where a one-off scan found cancers in a population already known to be at high risk (PMID 28091804). That is a different claim from scanning an asymptomatic adult at average risk.

The self-reported biomarker figures, including the percentile claims about muscle and fat, have no third-party audit. I grade those Anecdotal and use the word self-reported, which is not an insult. It is the correct grade for a number produced by the subject of the experiment.

The costs are real, and they are the biggest numbers on this page. Speed of Aging, the measurement product on his own store, is $325.00. A Prenuvo Longevity Scan runs from $1,199.00 for the focused scan to $4,499.00 for the enhanced one. Spend that money knowing the two findings above: the instrument is not validated as an endpoint, and broad screening in asymptomatic adults has not reduced mortality in the trials that tested it.

How to test the cheap parts yourself

If you read this far wanting to act, act on the part with the strongest evidence and the lowest price, which is zero. This is the honest version of a self-experiment, and it takes one season rather than a lifetime.

Get a fitness baseline you can repeat. The dossier evidence about cardiorespiratory fitness is about tested fitness, not about one workout plan, so choose an effort you can measure the same way twice: a timed route, a treadmill test, a step count at a fixed pace. Repeat it in three months. A supplement stack cannot tell you whether that number moved, and a wearable readiness score cannot either, though our wearables guide explains what those numbers are worth.

Protect seven hours of sleep and keep the wake time steady. Sleep duration is a Strong association and the one intervention here that costs nothing, needs no product, and shows up in the largest cohorts on record.

Then spend deliberately rather than enthusiastically. Longevity Mix at $49.00 per 30 days is the cheapest entry product and its evidence is the same modest evidence as the rest of the supplement market. The $98.00 Easy Stack, $270.00 Medium Stack, and $449.00 Big Stack are bundles of the same components, priced for convenience. If you want a cheaper route through the same category, our under $2,000 home setup guide builds the whole thing from scratch.

Do not buy measurement expecting proof. Buy it if you want the data for its own sake, and read the grade before you read the number.

For the sleep hardware question, our sleep optimization guide grades the devices, and the wearables guide covers what a readiness score is actually measuring. If you want the device-level version of everything on this page, the evidence-backed gear hub is where it lives, and the morning routine guide owns the caffeine timing and morning light questions that come up whenever this protocol does.

Frequently asked questions

What does Bryan Johnson actually take every day?

His protocol page, as read on October 17, 2026, lists a prescription stack: Armour Thyroid 60 mg, Levothyroxine 100 mcg, Minoxidil 3.75 mg, Metformin 500 mg, Tadalafil 5 mg, Jardiance 10 mg, Acarbose 200 mg, and Repatha 140 mg every other week. The daily routine on the same page adds a pre workout drink of Longevity Mix with creatine 5 g and 11 g of collagen peptides at breakfast, and its pregnancy avoid list names nicotinamide riboside among the stack ingredients. I did not find NMN, taurine, or cocoa extract on the protocol page. Those are products in his store, which is a different list.

Is the Blueprint protocol scientifically proven?

No. Nobody has randomized this protocol, or any comparable multi intervention protocol, against a control group. Each component has been studied on its own, in its own population, and the graded sections on this page say what that evidence is worth. The largest effects belong to the free parts: fitness (PMID 30646252) and sleep duration (PMID 28889101, 20469800).

Does Bryan Johnson protocol work? Has it been tested?

It is an n of 1 with no control arm and no blinding, and the endpoints are largely self reported by the team running the experiment. That does not make the results false. It makes them unverified, which is why the biomarker claims on this page are graded Anecdotal rather than treated as findings.

How much does the Blueprint protocol cost per month?

On his store, checked October 17, 2026: the Easy Stack is $98.00, the Medium Stack $270.00, the Big Stack $449.00, Longevity Mix $49.00 per 30 days, Omega-3 $39.00 per 30 days, Creatine $40.00 for one pack, and Ashwagandha + Rhodiola $24.00 per 30 days. The measurement products are separate: Speed of Aging is $325.00 and a Prenuvo Longevity Scan runs from $1,199.00 to $4,499.00.

Is Bryan Johnson biological age real, or is it a marketing number?

It is a measurement from an aging clock, and the clock is the weak link. The field has no agreed standard for a reliable biomarker of aging, and these markers are not established surrogate endpoints for living longer (PMID 37657418). Technical noise alone has produced deviations of up to nine years between replicate measurements of the same sample on six prominent epigenetic clocks (PMID 36277076).

What is the difference between pace of aging and life expectancy?

Pace of aging is a biomarker reading. Life expectancy and disease incidence are outcomes. Moving the first does not demonstrate the second, and no intervention that changes a clock has yet been shown to change mortality in a randomized trial. Treat a clock change as a data point, not a result.

Which parts of the Blueprint protocol have strong evidence?

Two, and both are free. Cardiorespiratory fitness has the largest effect on the page, with an elite versus low hazard ratio of 0.20 for all cause mortality across 122,007 patients (PMID 30646252). Sleep duration around seven hours carries the lowest mortality risk in a dose response meta analysis (PMID 28889101, 20469800). The eight hour eating window, by contrast, added nothing to calorie restriction at twelve months (PMID 35443107).

Can you get the benefits of Blueprint for free?

The strongest parts, yes. Fitness and sleep duration are the two interventions with the best human evidence in this audit and neither is for sale. Everything you can buy in the store sits in the Moderate, Limited, or null column of the graded evidence on this page.

Is rapamycin safe for healthy people who are not sick?

The human data is early and narrow. Lifespan extension is established in mice, not people (PMID 19587680). A small human trial found an mTOR inhibitor improved influenza vaccine response in elderly volunteers (PMID 25540326). The 48 week PEARL trial in healthy adults found adverse events similar across arms while its primary endpoint, visceral adiposity, did not change, so it is a safety result rather than a longevity result (PMID 40188830), and in mice rapamycin impairs glucose tolerance by disrupting mTORC2 (PMID 22461615). Rapamycin does not appear on his protocol page as fetched on October 17, 2026.

Should I buy the Blueprint supplements or use cheaper alternatives?

Read the grades first. Omega-3 missed its primary cardiovascular endpoint in a 25,871 person trial (PMID 30415637) and a daily multivitamin did not reduce invasive cancer in COSMOS (PMID 35294969). Taurine animal data was undercut by a 2025 human study finding no association with age or performance (PMID 41061678). If you want the bundle for convenience, buy it knowing that the evidence lives mostly in the Moderate column, and that the free tier still carries the biggest measured effect.

Sources and methodology

Every PMID below was resolved through Europe PMC on October 17, 2026 and read as an abstract at minimum before it was cited here. Every price, protocol item, and self-reported figure came from a live first-party page fetched on the same date: the protocol page at protocol.bryanjohnson.com/protocol, the storefront at blueprint.bryanjohnson.com, and the store product data at blueprint.bryanjohnson.com/products.json. The protocol items and biomarker claims are the subject of the audit, not evidence for it.

Grades used on this page: Strong, Moderate, Limited, Anecdotal. A Strong null is still a null, and it is graded Strong because the trial was well powered for the question it answered.

Gaps to keep in mind: nobody has randomized this protocol against a control, a stack can interact in ways single-intervention trials never test, animal results are assumed to translate, aging clocks carry technical noise that a single person cannot separate from signal, and self-reported endpoints have no independent audit. I am not a clinician, this is not medical advice, and the prescription layer above requires a physician and bloodwork.