Telomere length vs epigenetic age: the clocks predict death better

Telomere length vs epigenetic age has been tested head to head in the same people several times, and the epigenetic clocks win: they predict who dies sooner, and telomere length is the weaker marker [4][8]. Telomere length is the length of the protective caps on the ends of your chromosomes, which get shorter each time a cell divides, while an epigenetic clock estimates your age from chemical tags called methyl groups that attach to your DNA and shift in a predictable pattern as you get older [7].

One of those comparisons pooled three American cohorts. Across about 2,500 adults in a national health survey, 1,000 in a long-running study of older adults, and a few hundred in a Baltimore ageing cohort, the epigenetic clocks were the stronger predictors of death, and telomere length was the weaker one [4].

The strongest single predictor there was a clock called GrimAge, one of what the authors call the second-generation clocks [4]. Two others beat telomere length as well: the Hannum clock, and DunedinPoAM, which belongs to a newer group of measures built to capture how fast someone is ageing rather than how old they are [4][11].

A second head-to-head found the same ordering on a different outcome. Among 1,413 adults who had survived cancer in childhood, those whose clocks read older than their birthdays did worse on tests of attention, processing speed and planning, while telomere length showed no relationship with any of those scores [3].

The two measures also disagree with each other more than most people expect. In 635 healthy adults aged 19 to 61, shorter telomeres did tend to come with older-looking clocks, but the link was loose enough that plenty of people had short telomeres and a young clock, or the reverse [5]. Once age, sex and ethnicity were accounted for, only three of the five clocks tested kept any link to telomere length at all [5]. Two dropped out entirely.

The strangest result here is that a methylation-based guess at telomere length beats measuring telomeres in a lab. A model built from 140 methylation sites tracked chronological age far more closely than the laboratory telomere measurement did, and predicted time to death, coronary heart disease and heart failure better - a pattern that held across roughly 10,000 blood samples from seven long-running cohorts including the Framingham Heart Study and Twins UK [8].

The two tests fail to line up because they are not reading the same process. An epigenetic clock is a statistical model: researchers find the DNA sites whose methylation changes most predictably with age, then weight them so the total lands close to a person's chronological age, meaning the number of birthdays they have had [7]. The gap between epigenetic age and chronological age is what the field calls age acceleration, and it is that gap, not the raw age estimate, that the mortality studies use [4].

Telomere shortening is a different signal. In cultured cells, even the methylation-based estimate of telomere length turned out to track how many times a cell had divided rather than the length of the telomere itself [8]. That makes it a record of replicative history, meaning how much dividing a population of cells has already done, which overlaps with ageing without being the same thing.

That difference shows up in which clocks telomere length relates to. In the 635-adult comparison, telomere length was linked to one of the two versions of age acceleration, the version the authors attribute to ageing of the immune system, but not to the other [5]. The overlap between the two tests is largely immune ageing rather than ageing in general, which is what you would expect from a measurement taken in white blood cells.

On whether telomere tests are accurate, the honest answer is that the method decides. The 635-adult comparison measured telomere length two different ways, and the two methods produced noticeably different links to the same clocks, which means two labs can disagree about the same blood [5]. Telomere length's record outside mortality is patchy too: studies of telomere length and cancer have come back inconsistent for years, which is part of what pushed researchers towards methylation in the first place [12].

Neither test has what it would need to be used clinically. There are no agreed protocols for measuring them and no reference values, so there is no number at which a result becomes a finding, and reviews of the field say so plainly [2]. The epigenetic clock is the most promising biological age marker on offer, but its real value still needs confirmation from studies that follow the same people over time rather than photographing them once [1][11].

These are associations, not causes. A clock reading older than your birthdays flags higher risk across a group; it does not show that methylation is doing the damage, and nothing here tested whether changing either number changes what happens to a person.

Who the evidence does not cover:

• children and adults over 61, since the agreement study ran only in 19 to 61 year olds [5]
• the general population, as far as the thinking-and-memory result goes: those 1,413 people had all been treated for cancer as children, many with therapy aimed at the brain [3]
• people outside the United States, for the mortality comparison, which used three American cohorts [4]
• tissues other than blood, because every marker here was read from blood cells [5][8]

One finding in the mortality analysis needs reading carefully: women looked biologically younger than men on GrimAge, but that clock uses sex in its own definition, so part of the difference is built into the tool [4].

Anyone asking which biological age test is best is usually choosing between two products: a biological age telomere test and an epigenetic clock test, often sold by the same company. A telomere kit reports the average cap length in a blood sample, and the measurement method changes the answer enough that results from two labs are not comparable [5]. An epigenetic clock test reports an age estimate from methylation patterns, and which clock the lab picked changes both the number and its meaning, since the older clocks were built to guess a person's birthdays [7] and the newer ones to track death and disease [4]. Longevity clinics sell both, usually inside a wider panel of blood work. What no study here tested is whether watching either number move in one person over months or years tells you anything, and that is the confirmation [1] says is still missing.