The Biological Age Optimization Protocol: Test, Intervene, Retest

Key Takeaways

Every biological-age programme on the market follows the same script: buy the test, get a frightening number, buy the programme, retest, celebrate. The script works commercially because the second number is usually lower — and it is usually lower for reasons that have nothing to do with the programme. Regression to the mean, a different lab run, a cold you had the first time, a clock with a noise floor of several years.

The Biological Age Optimization Protocol inverts the script. It starts with twelve months of the behaviours that have the largest clock signals in the cohort data, adds the one supplement combination that slowed a clock in a large randomised trial, and only then addresses testing — with rules derived from the reliability studies rather than the sales pages. This article is the reasoning behind each layer and a month-by-month plan. The two companion pieces explain what the tests measure and what the trials have shown.

Why the Protocol Is Graded C

Grades on this site go to the weakest load-bearing claim. The behaviours in Layer 0 are graded A for the outcomes that matter — cardiovascular death, cancer, all-cause mortality — and B for their association with clock acceleration in large cohorts. The supplement step in Layer 1 rests on one randomised trial with a small effect. The testing rules are sound, but what a change in a clock means for an individual is, in the words of the field's own consortium, not yet established.

So: C. Not because the protocol is weak, but because the thing it claims to measure is. Run it for Layer 0 and you are following the best-evidenced health advice there is. Run it for the number and you are chasing something that cannot yet be reliably seen.

Layer 0: The Behaviours With the Largest Clock Signals

Tobacco: none, in any form. In 2,978 US adults over 50, GrimAge rose with pack-years, with smoking in youth, and even with having two smoking parents; GrimAge then mediated the path from smoking to cancer, hypertension, heart disease and death. This is the largest effect any exposure has on any clock, and it is also, of course, the largest preventable cause of death. Cannabis smoking accelerated blood epigenetic age in one cohort, and stopping normalised it.

Alcohol: at most seven units a week, no binges. Each additional standard drink per day raised PhenoAge acceleration by 0.71 years in middle-aged adults and 0.60 in older ones in the Framingham cohort; liquor was worse than wine or beer. In young adults, any recent binge added 1.38 years to GrimAge acceleration. In 2,316 women, lifetime consumption raised GrimAge independently of recent intake. Three cohorts, three clocks, dose-response.

Aerobic fitness: a measurable VO2max, improving. A 2026 meta-analysis across 145,465 people found each standard deviation of activity associated with 0.09 SD less GrimAge acceleration — small, and almost all cross-sectional. High-fit adults read 1.5 to 2 years younger on a fitness-trained clock. The clock signal is modest; the mortality benefit is the largest of any modifiable factor. Three sessions a week of zone-2 work plus one of intervals, and the home strength programme from Layer 1.

Eating: early, modest, plant-rich. CALERIE's 25% restriction for two years slowed DunedinPACE and nothing else, with small effects. DIRECT PLUS found no between-diet differences but slower ageing with adherence to a polyphenol-rich Mediterranean pattern. The TwiNS vegan arm reduced acceleration, with fewer calories as a confounder. The common thread is less, and from plants — not a named diet, and not a supplement powder. The circadian side of eating — an early, consistent window — is in the Circadian Optimization Protocol.

Sleep: treat apnoea, avoid chronic night work. Ten or more years of night shifts added 3.16 years of epigenetic age in 2,574 women. Insomnia symptoms added a year in older women; sleep duration did not. In working-age Finns, only sleep-apnoea symptoms were consistently linked to GrimAge and DunedinPACE after adjustment. If you snore and wake unrefreshed, a sleep study belongs in month one.

Layer 1: Omega-3, Vitamin D and Strength Training

The DO-HEALTH trial is the reason this layer exists. Over three years in 2,157 adults aged 70 and older, omega-3 at 1 g/day slowed PhenoAge, GrimAge2 and DunedinPACE in the 777 participants with methylation data; vitamin D3 at 2,000 IU/day alone did not, but the combination of both with a simple home strength programme was additive on PhenoAge. Standardised effects 0.16 to 0.32; 2.9 to 3.8 months over three years.

What to do with that: take the omega-3, for this and for its cardiovascular evidence. Take the vitamin D if your serum level is below 30 ng/mL or you are not training; skip it if you are replete and already strength-training, because that is the combination the trial showed. Do the strength programme regardless — it is thirty minutes, three times a week, of squats, push-ups, rows and a hinge, and it overlaps completely with Layer 0.

What not to add, and why:

• Metformin. Unrelated to epigenetic ageing in 4,575 people observationally; a cell-type-specific effect in one small HIV trial. Not a general human result.
• Rapamycin. No human clock data. The profile covers what it does have.
• A "methylation-support" stack. A 12-week multi-nutrient supplement in 80 older adults changed no clock overall; the apparent benefit in the baseline-high subgroup is regression to the mean. TMG and B-vitamins have a mechanistic case and no independent clock trial.
• Growth hormone and DHEA. The TRIIM reversal was uncontrolled, in nine men. The compounds carry real risks.
• NMN, senolytics, peptides. No controlled human clock data.

Layer 2: Testing Rules That Survive the Reliability Studies

This layer is optional. The protocol works without any clock test, because every element of Layers 0 and 1 is justified by outcomes rather than by the readout. If you test, do it in a way that can produce information.

The clinical panel is the part most people skip and the part that pays. ApoB, HbA1c, fasting insulin, hs-CRP, GGT, creatinine and blood pressure are the inputs to a Klemera-Doubal or clinical PhenoAge calculation, available free, and each is a lever your physician recognises. The Biohacker's Blood Panel covers the full list.

The Twelve-Month Plan

Month 0 — baseline. Clinical panel. VO2max estimate or a fixed time trial. Alcohol and tobacco log. If testing a clock: second-generation, blood, when well. Record the vendor, platform and clock name so you can match them in a year.

Months 1–3 — Layer 0 only. Stop tobacco. Alcohol to seven units a week. Three aerobic sessions and three strength sessions weekly. Eating window closed three hours before bed, plant-forward. Sleep study if indicated. Start omega-3 1 g and, if indicated, vitamin D 2,000 IU.

Month 6 — clinical check. Repeat the panel and the fitness test. Serum vitamin D. This is where you will see change: ApoB and HbA1c respond in months, and a falling resting heart rate is a clock you can read without a laboratory.

Months 7–12 — hold. The clock trials ran for two to three years. The behaviours need to be habits, not a programme, by the time the second test comes.

Month 12 — retest. Clinical panel, fitness test, and — if you tested at baseline — the same clock from the same lab on the same tissue, six weeks clear of any illness. Compare the trend. If the clock moved less than its noise floor, that is the expected result and not a failure. If the clinical panel moved, that is the result that matters.

Who Should Not Run This Alone

• Anyone on anticoagulants — discuss the omega-3 dose with the prescriber.
• Anyone with hypercalcaemia, sarcoidosis or kidney stones — vitamin D under supervision.
• Anyone who would change medication, diet or mood on the basis of a clock number. The consortium that built the clocks says they are not yet valid for individual decisions.
• Anyone recently ill or recovering from surgery — delay baseline testing by six weeks.
• Anyone under 30. The clocks are least validated and least responsive in young adults; run Layer 0 and skip Layer 2.

Frequently Asked Questions

Will I see my biological age drop? Probably not in a way you can distinguish from noise, and if you do, be suspicious of a result that beats the best randomised trials by years. Expect to see your clinical panel move; that is the signal.

Which test should I buy? If any: one that reports DunedinPACE or a principal-component GrimAge from venous blood. If the vendor reports only a Horvath or Hannum age from a cheek swab, the number cannot be tracked.

Why omega-3 and not NMN or rapamycin? Because omega-3 slowed three second-generation clocks in a 777-person randomised trial and the others have never been tested against a clock in people. The protocol follows the evidence, not the catalogue.

Does caloric restriction belong in the protocol? Modest intake, yes — it is in Layer 0. Formal 25% restriction for two years moved one clock by a small amount and is hard to sustain; the protocol takes the direction without the dose.

My first result says I am ten years older than my age. What now? Confirm it was blood, that you were well, and which clock it was. Then run the clinical panel, which will show you where the actionable risk is. Do not start a drug on a clock result.

Where do I start? With the protocol page for the layers, doses and monitoring table — and with the month-zero clinical panel.

Related Research

• The Biological Age Optimization Protocol
• 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 Biohacker's Blood Panel: 40 Biomarkers Worth Tracking
• The Circadian Optimization Protocol