Where the term comes from

Researchers have searched for reliable markers of biological age for more than three decades, from telomere length to blood proteins. In 2013 two teams published DNA methylation clocks: Steve Horvath’s, built from about 8,000 samples of 51 tissues and cell types using 353 sites on the genome, and a blood-based clock from 656 people. Later clocks such as PhenoAge and GrimAge were trained to predict illness and death rather than birthdays, and DunedinPACE estimates how fast a person is aging.

What the research shows

Across four cohorts of older people, a methylation age five years above calendar age went with a risk of death 21 percent higher, or 16 percent higher after adjusting for health and other factors. For one person, though, readings are noisy: technical variation alone shifted results by up to nine years between repeat measurements of the same sample on six well-known clocks. In the CALERIE trial, 220 healthy adults were randomized to two years of 25 percent calorie restriction or their usual diet; the diet slowed DunedinPACE by 2 to 3 percent but left PhenoAge and GrimAge unchanged. Claims of “reversing” biological age usually rest on small studies, such as an eight-week pilot in 43 men.

What a result can and cannot tell you

Across populations, faster epigenetic aging predicts worse health. For an individual, a single number depends on which clock was used and on laboratory noise, and experts have not yet agreed how aging biomarkers should be validated before clinical use. A result is best read as one data point to repeat and track, not as a verdict. Our longevity page explains how our program uses biological age, as a measure taken before and after the six months.

Related terms

Heart Rate Variability: another body measure that is noisy in a single reading.

Languishing: low well-being that no blood test captures.