The Biological Age Optimization Protocol

Read This Before the Protocol

A biological-age test is a research instrument you are buying as a consumer product. The Biomarkers of Aging Consortium — the field's own standards body — wrote in 2024 that robustness and responsiveness at the level of individuals are prerequisites for clinical use that have not yet been met. First-generation clocks differ by up to nine years between technical replicates of the same sample. Oral and blood samples from the same person can disagree by decades. A single reading, from a cheek swab, interpreted to the year, is not a measurement; it is a number.

That is not a reason to ignore the tests. It is a reason to use them the way the trials do: a reliable clock, a consistent tissue and laboratory, a long interval, and interventions chosen for their own evidence rather than for what they do to the readout. This protocol does exactly that, and it is graded C because the one randomised supplement effect on a clock is a few months over three years.

Who should not run this. Anyone who would change a medication, a diet or their mood on the strength of a clock number. Anyone recently ill or recovering from surgery — biological age rises transiently under stress and recovers, so your reading will be wrong. Anyone under 30, in whom the clocks are least validated and least responsive.

How This Differs From the Other Longevity Protocols

Foundational Longevity and the Ultimate Longevity Stack are built around compounds with mechanistic and animal evidence for lifespan. This protocol is built backwards from the human measurement: what has actually moved a clock in a randomised trial, and how to measure that without fooling yourself.

They are compatible. Run this as the measurement layer under whichever stack you follow — with the warning that no epigenetic clock has been shown to move in response to NMN, rapamycin or senolytics in humans.

The Layer Logic

Layer 0 — The Behaviours With the Largest Clock Signals

Tobacco. In 2,978 US adults over 50, GrimAge rose with pack-years, youth smoking and even parental smoking, and GrimAge mediated the path from smoking to cancer, heart disease and death. GrimAge is partly built on a methylation signature of smoking, so this is expected — and it is also the single largest clock effect in the literature. Nothing on this page approaches the size of not smoking.

Alcohol. Each additional daily drink raised PhenoAge acceleration by about 0.7 years in middle-aged adults in the Framingham cohort; in young adults, any recent binge added 1.4 years to GrimAge acceleration; in 2,316 women, lifetime drinking raised GrimAge independent of recent intake. The associations are consistent across three cohorts and three clocks.

Exercise. A 2026 systematic review and meta-analysis of 44 studies in 145,465 people found each standard deviation of physical activity associated with lower GrimAge acceleration — a small effect, almost entirely from cross-sectional data. A randomised two-year diet-and-activity trial in 219 postmenopausal women found the diet arm slowed GrimAge; the activity arm reduced stochastic methylation errors but was not reported to change a clock. Exercise is in Layer 0 because it does more for mortality than anything else here, not because it reliably moves a readout.

Eating pattern. The CALERIE trial — two years of 25% caloric restriction in 220 non-obese adults — slowed DunedinPACE and changed PhenoAge, GrimAge and the other clocks not at all, with effect sizes the authors called small. In the eight-week twin study, a vegan diet reduced epigenetic age acceleration, but the vegan twins also ate fewer calories. An 18-month trial of three diets in 256 people found no between-arm clock differences; only adherence to the polyphenol-rich green Mediterranean pattern tracked slower ageing. The common thread is modest intake and plant-rich composition, not a named diet.

Sleep and shift work. Ten or more years of night-shift work added about three years of epigenetic age in 2,574 women. Insomnia symptoms added one year in 2,078 older women, while sleep duration did not matter. In working-age Finns, only sleep-apnoea symptoms were consistently linked to GrimAge and DunedinPACE. Fix apnoea; the circadian layer is in the Circadian Optimization Protocol.

Layer 1 — The One Supplement Result Worth Building On

The DO-HEALTH trial randomised 2,157 adults over 70 to vitamin D 2,000 IU, marine omega-3 1 g, a simple home exercise programme, or their combinations, for three years. A 2025 analysis of 777 participants with methylation data found that omega-3 alone slowed PhenoAge, GrimAge2 and DunedinPACE, and that all three interventions together had an additive effect on PhenoAge. Standardised effect sizes were 0.16 to 0.32 — the authors translate that to 2.9 to 3.8 months of biological ageing over three years.

That is the entire randomised supplement evidence for slowing a clock in a large trial, and it is small. It is also real, in a population old enough for the clocks to be meaningful, and the components are cheap and safe. Omega-3 at 1 g and vitamin D3 at 2,000 IU are the steps; the exercise is 30 minutes of simple strength work three times a week, which Layer 0 already requires.

What is not here, and why:

• Metformin. Observationally unrelated to epigenetic age in 4,575 people. In a small HIV trial it reduced monocyte clock age but not CD8 T-cell clock age. No general human clock effect.
• Rapamycin. The PEARL trial reported no clock endpoint. No verified human clock data exist.
• Growth hormone + DHEA + metformin. The TRIIM study reported 2.5 years of reversal, in a one-year, uncontrolled study of nine men. Hypothesis-generating, not protocol material.
• "Methylation-support" supplement stacks. A 12-week trial of a combination nutritional supplement in 80 older adults changed no clock overall; the only "effect" was in the subgroup that started with a raised epigenetic age, the pattern that regression to the mean predicts.
• NMN, senolytics, spermidine. No human clock trial.

Layer 2 — How to Test Without Fooling Yourself

If you test at all, follow these rules. Each one comes from a reliability study, not from a vendor.

1. Use a second-generation or principal-component clock. DunedinPACE excluded unreliable probes by design and showed high test-retest reliability. Principal-component versions of GrimAge and PhenoAge agree within about 1.5 years between replicates, where the original clocks drift by up to 9. If the vendor reports only a first-generation Horvath or Hannum age, the number is too noisy to track.
2. Blood, not saliva or cheek swab. Oral and blood estimates from the same person can differ by almost 30 years for some clocks, with low correlation even after cell-type correction.
3. Same laboratory, same platform, same clock, every time. Pre-processing choice alone changes which samples agree, and two labs can return different ages from one sample.
4. Twelve months apart, not three. The DO-HEALTH effect was three months over three years. A quarterly test cannot resolve that against noise.
5. Not within six weeks of illness, surgery or major stress. Biological age rises during these and recovers afterwards; severe infection accelerated clocks in several cohorts and the effect partly reversed in survivors.
6. Pair it with a clinical blood panel. Blood-chemistry biological age — the Klemera-Doubal or PhenoAge calculation from routine labs — predicted mortality at least as well as methylation clocks in head-to-head analyses and costs a fraction. The Biohacker's Blood Panel covers the inputs.
7. Act on the trend, never on a single reading. A one-year change inside the clock's known noise band is not information.

When to Stop and See a Clinician

• A clinical biomarker is out of range — ApoB, HbA1c, blood pressure, liver enzymes. That is actionable today; a clock number is not.
• You are considering metformin, rapamycin, growth hormone or any prescription compound on the basis of a biological-age result. No clinician should prescribe on that basis, and the data do not support it.
• Your clock result is dramatically older than your age. Before concluding anything, repeat it in blood, at the same lab, after confirming you were well at the time of sampling.
• You notice the number is changing how you feel about yourself. Twelve-month intervals exist partly for this reason.

Monitoring

Safety and Scope

Omega-3 at 1 g/day and vitamin D at 2,000 IU/day are within standard safety limits for adults; people on anticoagulants should discuss omega-3 dose with their prescriber, and anyone with hypercalcaemia or sarcoidosis should not take vitamin D without supervision. Nothing on this page diagnoses, prevents or treats any disease. Epigenetic clocks are not approved diagnostic devices anywhere, and this protocol does not treat them as one.

Related Research

• Biological Age Tests Explained: GrimAge, DunedinPACE and What They Can't Tell You
• How to Actually Lower Your Biological Age: What the Trials Show
• The Biological Age Optimization Protocol: Test, Intervene, Retest
• The Epigenetic Clock: How Peptides and Interventions Change DNA Methylation Age
• The Biohacker's Blood Panel: 40 Biomarkers Worth Tracking