What is biological age?
It’s an estimate of how old your body acts, set against your birthday. There are three kinds, built from fitness tests, from a blood panel or from your DNA. They measure different things, they don’t average into one true age, and the free kind is the most useful.
What we’d do
Start with the free kind: the floor test, one-leg balance, a dead hang, your walking pace and your watch’s VO2 max. Add blood markers you can change, like ApoB and HbA1c. Leave the DNA tests until those are in place, and never average the three into one number, because the average hides which part of you is lagging.
Why it’s confusing
Ads promise “your real age” from a cheek swab or a blood draw, as if there were one hidden number waiting to be found, and there isn’t. Each method is built a different way, trained to predict a different thing, and they often disagree.
The three kinds
| Kind | Built from | What it tracks | Cost |
|---|---|---|---|
| Fitness age | Tests like getting up off the floor, one-leg balance, grip or hang, walking speed, VO2 max | How well your body works now | Free |
| Blood age | Routine blood markers plus your age | How your organs and metabolism are doing | A standard panel |
| DNA age | Chemical tags on your DNA, called methylation | Patterns that shift with age and illness | A specialty test |
Fitness age. Each of these tests predicts survival in big studies. The least fit quarter on a treadmill had about 5 times the death rate of the fittest 2% (Mandsager 2018), and people scoring 0 to 4 on the floor test had about 4 times the risk of a perfect 10 (Araujo 2025). They change with training, which makes them easy to track.
Blood age. The best-known version, Phenotypic Age, comes from Levine and colleagues, who built it from nine routine blood markers plus age, using a big US health survey: albumin, creatinine, glucose, C-reactive protein, lymphocyte percent, red cell size, red cell width, alkaline phosphatase and white cell count.
DNA age. A clock published by Horvath in 2013 used 353 spots on the DNA and was trained on 8,000 samples from 51 tissues and cell types. It gave embryonic stem cells an age close to zero, and cancer tissue came out 36 years older than its owner on average. Later clocks were trained to predict health and death rather than birthdays. In Levine’s DNA version of Phenotypic Age, each year older went with 4.5% higher risk of dying. DunedinPACE, from the long-running Dunedin study in New Zealand, gives a pace of aging rather than an age.
The simple version
Think of a used car. Its mileage, its rust and its service record are all real measures of wear, and each answers a different question. Averaging them into one score would tell you less than looking at each.
How to use it
Track each kind on its own, against itself, over months. A change you can trust is one that’s bigger than the test’s own wobble from one time to the next, which for DNA clocks can be large. See are biological age tests accurate.
5 sources
- Levine ME and others, An epigenetic biomarker of aging for lifespan and healthspan, Aging, 2018pmc.ncbi.nlm.nih.gov
- Horvath S, DNA methylation age of human tissues and cell types, Genome Biology, 2013pubmed.ncbi.nlm.nih.gov
- Belsky DW and others, DunedinPACE, a DNA methylation biomarker of the pace of aging, eLife, 2022pubmed.ncbi.nlm.nih.gov
- Mandsager K and others, Association of Cardiorespiratory Fitness With Long-term Mortality Among Adults Undergoing Exercise Treadmill Testing, JAMA Network Open, 2018jamanetwork.com
- Araujo CG and others, Sitting-rising test scores predict natural and cardiovascular causes of deaths in middle-aged and older men and women, European Journal of Preventive Cardiology, 2025pubmed.ncbi.nlm.nih.gov
