Are biological age tests accurate?
On this page
Not accurate enough to trust a single result. Run the same blood sample twice and six popular DNA clocks gave ages up to 9 years apart, and different clocks give different ages for the same person. Track the fitness tests and blood markers you can change instead.
What we’d do
If you’re curious, buy one, but buy the same test every time and only believe a change bigger than the test’s own noise. Spend your money first on a blood panel with ApoB and HbA1c, and your time on the floor test and your fitness. Those are cheaper, steadier and easier to act on.
Why a precise-looking number feels true
A report that says you’re 43.2 years old biologically looks scientific. It comes from a lab, on DNA, with a decimal point. But a decimal point says nothing about how close the number is to the truth. Before trusting one, it helps to know whether it gives the same answer twice, and whether there’s a true answer to check it against.

The same sample, twice
In 2022, researchers ran the same DNA samples through the lab more than once and compared the results on six well-known clocks. Technical noise alone, from the chemistry and the machine, moved the answers by up to 9 years between runs of the same sample. They built a fix, recalculating each clock in a way that cut most of the noise, so that most repeat runs agreed within 1.5 years. It’s fair to ask a company whether it uses a method like that.
Some newer measures were designed with this in mind. DunedinPACE was built only from spots on the DNA that gave steady results on retest, and its creators reported high test-retest reliability.
The simple version
A noisy clock is like a bathroom scale that gives a different number each time you step on it. Averaged over months it can still show a trend. One weigh-in on one morning tells you very little.
There’s no answer key
You can check a thermometer against boiling water. There’s no equivalent for biological age. The first DNA clocks were trained to guess calendar age, and newer ones to predict health and death in big groups. In Levine’s version, each extra year of DNA age went with 4.5% higher risk of dying across a population, which is useful for research and says much less about you alone.
The clocks also disagree with each other about what changes you. In the CALERIE trial, 220 healthy adults were assigned to eat 25% fewer calories or eat as usual for two years. The diet slowed one measure, DunedinPACE, by a small amount, and didn’t change the PhenoAge or GrimAge clocks at all.
What to track instead
The tests that predict survival and respond to what you do are cheaper and steadier: the sitting-rising test, one-leg balance, walking speed, your watch’s VO2 max, blood pressure, ApoB and HbA1c. See which blood tests matter most.
4 sources
- Higgins-Chen AT and others, A computational solution for bolstering reliability of epigenetic clocks: Implications for clinical trials and longitudinal tracking, Nature Aging, 2022pubmed.ncbi.nlm.nih.gov
- Waziry R and others, Effect of long-term caloric restriction on DNA methylation measures of biological aging in healthy adults from the CALERIE trial, Nature Aging, 2023pubmed.ncbi.nlm.nih.gov
- Levine ME and others, An epigenetic biomarker of aging for lifespan and healthspan, Aging, 2018pmc.ncbi.nlm.nih.gov
- Belsky DW and others, DunedinPACE, a DNA methylation biomarker of the pace of aging, eLife, 2022pubmed.ncbi.nlm.nih.gov