Should You Buy an Epigenetic Age Test?

A short answer, and the reasons behind it.

What this is

Consumer epigenetic age tests read DNA methylation from saliva or blood and return an age. The underlying research is legitimate and important. The consumer product inherits problems the research can tolerate and a customer cannot.

There are dozens of clocks, built for different purposes, and they don't always agree on the same person. Results shift with sample collection and processing, tissue type, and transient factors like illness or time of day — so the same person can score differently depending on when they spat in the tube.

What the evidence says

  • Different clocks disagree on the same individual

    Gold· human data

    Documented in the methodological literature and the central practical problem: without agreement, an individual result has no clear referent.

  • Results are sensitive to sampling and processing

    Gold· human data

    Collection method, tissue, preprocessing and computational implementation all move the output — technical variance that a population study averages out and a single customer cannot.

  • Suitable for tracking whether your habits are working

    Not supported

    The measurement noise is comparable to the effect sizes people hope to detect. You cannot read a lifestyle change off it.

What this means in practice

  • If curiosity is the motive, treat it as entertainment and price it accordingly.
  • If health is the motive, spend the money on a standard blood panel via your GP and a fitness test you can repeat.
  • Watch this field. It's improving, and the reason to wait is technical rather than permanent.

The honest summary

Not yet. The science is genuine, the product is premature, and the free alternative — track your fitness and your standard blood markers over time — is more informative today.

Work it out

Related

Frequently asked questions

Why did two tests give me different biological ages?
Because different clocks are built on different training data for different purposes, and because sampling and processing introduce real variance. It's the expected result, not a fault in your sample.

Sources

Written and reviewed by Mathew Beale, MSc Biotechnology, University of Reading.

Last reviewed: