---
title: Your Body Has Two Ages. Can You Really Measure the Second One?
slug: your-body-has-two-ages
category: Longevity
datePublished: 2026-07-17
dateModified: 2026-07-17
author: Roy Sañudo S.
reviewedBy: Roy Sañudo S.
lastReviewed: 2026-07-17
url: https://animacosmi.com/articles/your-body-has-two-ages
---

# Your Body Has Two Ages. Can You Really Measure the Second One?

You have two ages, and only one is on your licence. The second one is real, but the test that claims to read it disagrees with itself by years, and every famous study that 'reversed' it was paid for by someone selling something. Here is the honest scoreboard.

Two people are sixty. One takes the stairs without thinking about it. The other plans the day around their back. Same birthday, a decade apart in everything that matters. You already know which one you are drifting toward, because you have watched it happen to people you know. The open question is whether anyone can hand you the number.

## What is biological age?

Your real age counts birthdays. Your biological age is an estimate of how worn your cells are, and the two can disagree by years in either direction.

Think of it this way. Your birthday is the year stamped on the chassis. Your biological age is the mileage. A well-maintained 2010 model outruns a neglected 2020 every time.

So how does a lab read the mileage? It looks at tiny chemical tags on your DNA, called methylation marks, that switch on and off in a fairly predictable pattern as the years pass. Add them up and you get an estimate. Scientists call it an epigenetic clock. [Steve Horvath built the first widely used one at UCLA in 2013](https://pmc.ncbi.nlm.nih.gov/articles/PMC4015143/), and versions of it sit behind every study below.

At the scale of a whole population, it works. Across thousands of people, these clocks predict disease and death better than a birth date does. That much is well established, including by the researchers who are hardest on the test. The argument is about you.

## Can you actually lower your biological age?

Possibly, a little. But before you trust any headline about it, notice the pattern underneath all three of the famous studies: **the study that proves the product was usually paid for by the people selling it.** In this field that is not a scandal. It is the default. Here they are.

**The strongest one is a real trial.** [CALERIE](https://pubmed.ncbi.nlm.nih.gov/37118425/) randomly assigned 220 healthy adults to eat less for two years. The target was 25% fewer calories; people actually managed **11.9%**. One clock showed them aging about 2 to 3% slower. Two other clocks showed nothing at all, in the same people, from the same blood. And the line you have read everywhere, that this means a 10 to 15% lower chance of dying early, is not a result of this trial at all. It is the authors' own careful guess, borrowed from separate research. CALERIE measured no deaths.

**The "three years younger in eight weeks" study** was [real and randomized](https://pubmed.ncbi.nlm.nih.gov/33844651/): 18 men treated, 20 as controls. It was paid for by Metagenics, whose supplements were the treatment, and [corrected in March 2024](https://pubmed.ncbi.nlm.nih.gov/38488762/), rewriting its abstract, its results and two of its figures.

**The trial that started the whole "reversal" story**, [TRIIM](https://pubmed.ncbi.nlm.nih.gov/31496122/), was nine men with no control group, funded by a company its lead author co-founded and holds shares in. Six months on, the effect was drifting back.

The same shape sits inside [the NMN trial everyone quotes](/articles/anti-aging-supplement-everyone-talking-about).

## Why can't a test just tell you your number?

Because the clock disagrees with itself.

In 2022, researchers ran the obvious experiment almost nobody had bothered with. They took one blood sample, split it in two, and put both halves through the same clock. [The two halves came back a median of 1.8 to 2.4 years apart. In the worst cases, 8.6 years apart.](https://pmc.ncbi.nlm.nih.gov/articles/PMC9586209/) Same blood. Same day. Same machine.

So if a test tells you that you are three years younger than your birthday, and the test cannot tell three years from nothing, the three years are not information. They are static.

One person put that to the test in public. A writer sent his own sample to seven companies over two days and paid $3,143.99 for the answers. They came back ranging from 24.9 to 38. That is one man's shopping trip, not a study. But it is a thirteen-year spread from one body in one week, and it is exactly what the split-sample data predicts.

The person saying this loudest is not a critic. It is Steve Horvath, who built the clock. His own score moved four years in six months. ["You cannot directly relate it to how long you'll live,"](https://www.technologyreview.com/2022/04/15/1050019/) he told MIT Technology Review in 2022. Morgan Levine, who built one of the other clocks, told the same magazine that an individual reading can be off by decades. In 2026, Horvath told a precision-medicine conference that you still cannot send one sample to two properly validated labs and get the same answer back.

That 2022 paper also published a statistical fix, and it works. The corrected clocks are not the ones being sold to you.

## Should you buy an at-home biological age test?

Not to answer the question you actually want answered.

If you are buying it to learn how old you really are, it cannot tell you. That is not an accusation against any one company; it is just what the numbers above mean.

There is one honest use: read the slope, never the number. One lab, at least three tests, six months apart, same conditions each time, and ignore any single reading. That is a real discipline. Almost nobody keeps it, which is why the test usually sells a feeling rather than a fact.

The price moved the wrong way, too. The full panel this article once pointed readers to has gone from $229 to $499. The $229 tier still exists, but it now buys a smaller, different test. Nothing about the instrument improved. What improved was the story around it.

If you want numbers that move and mean something, the boring ones are better. How out of breath you get on a hill. How hard you can squeeze. Your blood pressure. Your HbA1c, which is your average blood sugar over about three months. Your hs-CRP, a cheap blood test for hidden inflammation. They cost less, they repeat cheaply, and they have decades of outcomes standing behind them.

## What changed in June 2026?

Something real, and it is not in a bottle.

In 2020, [David Sinclair's lab at Harvard published a result in Nature](https://pubmed.ncbi.nlm.nih.gov/33268865/) that read like science fiction. They used a harmless virus to carry three genes into the eye cells of mice, asking those cells to reload an older, cleaner copy of their own instructions. Mice whose optic nerves had been crushed could see again. Naturally aged mice got sharper sight back.

For six years the honest caption on that was: mice only, humans someday.

Someday arrived on **9 June 2026**. [Life Biosciences, which Sinclair co-founded, dosed the first human being ever](https://clinicaltrials.gov/study/NCT07290244) with this kind of cell reprogramming. It is a Phase 1 safety trial, in people with glaucoma, asking whether it is safe inside an eye. It is not an anti-aging treatment, and nobody can buy it.

Notice what that trial measures, though. Not a clock reading. Whether a person can see. That is the bar this whole field has to clear, and it is the reason the number in your inbox does not matter yet.

## What actually works while we wait?

The things that move your biology are the ones you already suspect. Each comes with an honest asterisk.

**Sleep** is when your body does most of its repair. Seven hours or more, and morning light to set the clock that runs it.

**Get out of breath.** Cardiorespiratory fitness, which is how well your heart and lungs handle hard work, is one of the strongest predictors of how long people live, and unlike a methylation score, you can feel it change.

**Eat less; a window is one way to do it.** [Salk's early trial](https://pubmed.ncbi.nlm.nih.gov/31813824/) found benefits from a 10-hour eating window, and it is an easy rule to hold. Be fair about it though: two larger trials since, [TREAT](https://jamanetwork.com/journals/jamainternalmedicine/fullarticle/2771095) and [one in the New England Journal](https://www.nejm.org/doi/full/10.1056/NEJMoa2114833), found the window added nothing beyond simply eating less. So it is a tool, not a trick. [We go through that whole argument here](/articles/when-you-eat-matters-more-than-what).

**Heat.** [University of Eastern Finland researchers followed 2,315 men for about twenty years](https://pubmed.ncbi.nlm.nih.gov/27932366/). Those who used a sauna 4 to 7 times a week had lower rates of dementia and heart disease. It watched people rather than testing them, so it cannot prove cause.

**Cold.** [Winter swimmers studied at the University of Copenhagen](https://pubmed.ncbi.nlm.nih.gov/34755128/) had more active brown fat and better blood sugar control. Also a snapshot, not a trial. [Here is what cold actually does](/articles/why-cold-showers-change-everything).

You have two ages. The first is on your licence and it is honest. The second is real, and for now nobody can hand you its number without a margin of error wide enough to drive through. That is not bad news. It just means the answer was never going to arrive in the post. It arrives in how you sleep tonight. [Here is how I live with that](/how-i-live).

## FAQ

### What is biological age and how is it different from real age?

Your real age counts birthdays. Your biological age is an estimate of how worn your cells are, read from chemical tags on your DNA called methylation marks. Steve Horvath built the first widely used epigenetic clock at UCLA in 2013. Across a large population these clocks predict disease and death better than a birth date does, and that is not in dispute. What is in dispute is whether the number means anything about one individual person.

### How accurate are at-home biological age tests?

Less accurate than most people assume. When researchers split a single blood sample in two and ran both halves through the same clock, the two results came back a median of 1.8 to 2.4 years apart, and in the worst cases 8.6 years apart (Higgins-Chen et al., Nature Aging, 2022). One writer sent his sample to seven companies over two days and got answers spanning 13.1 years. Steve Horvath, who built the clock, said in 2026 that it is not yet possible to get the same readout from two different validated labs. A 2022 paper published a statistical fix that works, but the corrected clocks are not the ones sold direct to consumers.

### Can you actually reverse your biological age?

The honest answer is: possibly a little, and the evidence is weaker than the headlines. The best study is CALERIE, a randomized trial of 220 adults: participants achieved 11.9% calorie restriction (not the 25% prescribed) and one clock, DunedinPACE, showed roughly 2 to 3% slower aging, while PhenoAge and GrimAge showed no effect in the same people. The famous TRIIM trial was nine men with no control group, funded by a company its lead author co-founded. The 8-week lifestyle study was funded by the supplement maker whose products were the intervention, and was corrected in March 2024. No epigenetic clock has ever been validated as a surrogate endpoint, meaning nobody has shown that lowering the number changes what happens to you.

### Should you buy a biological age test?

Not if you are buying it to find out how old you really are, because it cannot resolve that at the individual level. There is one defensible use: reading the slope rather than the number, using one lab, at least three tests, six months apart, under the same conditions, and ignoring any single reading. Almost nobody does this. If you want measurements that move and mean something, cardiorespiratory fitness, grip strength, blood pressure, HbA1c and hs-CRP are cheaper, repeat more easily, and have decades of outcome data behind them.

## Key Takeaways

- Biological age is a real idea: across a population, DNA clocks predict health better than a birth date does
- But split one blood sample in two and the same clock disagrees with itself by 1.8 to 2.4 years on average, and by up to 8.6
- Every famous 'reversal' trial was funded by someone with a stake in the result
- The best study, CALERIE, slowed one clock by 2 to 3% and left two other clocks flat
- On 9 June 2026 the first human was dosed with cell reprogramming. It is a glaucoma safety trial, not an anti-aging pill.

## Citations

- Harvard Glenn Center for the Biological Mechanisms of Aging (Dr. David Sinclair). Disclosed affiliations include Life Biosciences, Animal Bioscience, InsideTracker and MetroBiotech.
- Horvath S., Genome Biology: "DNA methylation age of human tissues and cell types" (2013). 14:R115. UCLA ,  the first widely used pan-tissue epigenetic clock.
- Higgins-Chen A.T., Thrush K.L., Levine M.E. et al., Nature Aging: "A computational solution for bolstering reliability of epigenetic clocks" (2022). Technical replicates of a single blood draw deviated by a median of 1.8-2.4 years (max 8.6) on the original Horvath/PhenoAge clocks; principal-component correction raises reliability to ICC >0.97.
- Waziry R., Ryan C.P., ... Belsky D.W., Nature Aging: "Effect of long-term caloric restriction on DNA methylation measures of biological aging in healthy adults from the CALERIE trial" (2023). PMID 37118425. DOI 10.1038/s43587-022-00357-y. RCT, n=220. Achieved restriction 11.9% against a 25% target; DunedinPACE slowed ~2-3%; PhenoAge and GrimAge showed no effect. Post-hoc analysis; no mortality outcome measured.
- Fahy G.M. et al., Aging Cell: "Reversal of epigenetic aging and immunosenescent trends in humans" (TRIIM, 2019). PMID 31496122. DOI 10.1111/acel.13028. n=9, single-arm, no control group. Sponsored by Intervene Immune, Inc.; the lead author is an officer and shareholder.
- Fitzgerald K.N., Hodges R., Hanes D. et al., Aging: "Potential reversal of epigenetic age using a diet and lifestyle intervention: a pilot randomized clinical trial" (2021). PMID 33844651. DOI 10.18632/aging.202913. 18 treated / 20 control. Funded by an unrestricted grant from Metagenics, Inc., whose products constituted the intervention. Substantively corrected March 2024 (PMID 38488762).
- Lu Y. et al., Nature: "Reprogramming to recover youthful epigenetic information and restore vision" (2020). PMID 33268865. DOI 10.1038/s41586-020-2975-4. OSK (Myc omitted); mouse.
- Life Biosciences ER-100, ClinicalTrials.gov NCT07290244. FDA IND cleared 28 Jan 2026; first human dosed 9 Jun 2026. Phase 1, open-angle glaucoma and NAION ,  the first human dosing of partial epigenetic reprogramming.
- Yang J.H. et al., Cell: "Loss of epigenetic information as a cause of mammalian aging" (2023) ,  contested. Timmons J.A. & Brenner C. argue in a 2024 Matters Arising (Cell) that the endonuclease used to build the ICE mice is independently genotoxic; the Sinclair lab published a rebuttal in the same issue. Unresolved.

## Common prompts this article answers

- What is biological age and how is it different from real age?
- How accurate are at-home biological age tests?
- Can you actually reverse your biological age?
- Should you buy a biological age test?
- What is biological age?
- Can you actually lower your biological age?
- Why can't a test just tell you your number?
- Should you buy an at-home biological age test?
- What changed in June 2026?
- What actually works while we wait?