Short telomeres and mortality: what the cohorts actually measured

A meta-analysis of 25 studies, 121,749 individuals and 21,763 deaths found a hazard ratio of 1.09 per standard deviation shorter telomere length, and 26 per cent higher hazard in the shortest quartile. Two large prospective cohorts give 1.23 for the shortest quintile. Every one of these is observational: none of them shows that lengthening telomeres extends life.

goodWell supported, with the usual caveats about effect size.Revised 05 Nov 2026

Two coils of fine thread resting on a pale grey surface beside a closed notebook

The number everyone quotes, and the design that produced it.

Telomere length is the most quoted number in the longevity conversation, and it is quoted in a way that quietly changes its meaning. The sentence that usually appears is “short telomeres mean shorter life”. That part is well supported. The sentence that follows it — the one that sells things — is a different claim, and it has not been tested.

The largest synthesis

The 2018 meta-analysis pooled 25 studies, including unpublished data from the Swedish Twin Registry: four cohorts of 12,083 individuals with 2,517 deaths plus 21 published studies. Across all eligible studies: 121,749 individuals with 21,763 deaths.

The results, exactly:

One standard deviation decrement in leukocyte telomere length corresponded to a hazard ratio of 1.09, 95 per cent confidence interval 1.06 to 1.13.
Individuals in the shortest quartile had a 26 per cent higher hazard (95 per cent CI 15 to 38) than those in the longest quartile.
In the Swedish Twin Registry alone, one standard deviation decrement corresponded to 13 per cent increased all-cause mortality risk (CI 7 to 19), and the shortest quartile had a 44 per cent higher hazard (CI 27 to 63).
Stratified by age, the hazard ratio was smaller in individuals over 80.

The heterogeneity between studies is real, and the authors name its sources: telomere measurement technique and age at measurement both contribute.

Two cohorts, checked separately

The 2017 individual-participant analysis used 12,199 adults from two population-based cohorts in Europe (ESTHER) and the United States (Nurses’ Health Study), with blood samples from 1989 to 1990 and 2000 to 2002, measured by quantitative polymerase chain reaction. 2,882 deaths during follow-up.

Age-adjusted hazard ratios, shortest quintile versus longest:

all-cause mortality: 1.23 (95 per cent CI 1.04 to 1.46)
cardiovascular mortality: 1.29 (95 per cent CI 0.83 to 2.00)
cancer mortality: 1.10 (95 per cent CI 0.88 to 1.37)

Only the all-cause figure reaches significance. The cardiovascular and cancer intervals both cross 1. The authors conclude that leukocyte telomere length predicts all-cause mortality beyond its association with age — the word “predicts” here means prognostic association, not causation.

And in a diseased population

The 2022 analysis in Heart looked at patients with heart failure, measuring telomere length in two cohorts: 2,260 and 1,413 patients. Shorter telomere length was independently associated with all-cause mortality, meta-analysis hazard ratio 1.16 per standard deviation decrease (95 per cent CI 1.08 to 1.24), remaining significant after adjustment for a clinical risk score (1.12, CI 1.05 to 1.20). There was no association with heart failure hospitalisation (hazard ratio 0.99, CI 0.92 to 1.07).

A useful counterweight: the association was real for mortality and absent for hospitalisation. A biomarker that predicts one outcome in a population does not automatically predict the others.

Why none of this supports “lengthen them”

Every dataset above is observational. A cohort measures telomere length once (or at a few timepoints) and then follows people for years. That design can show two things travel together. It cannot show which one causes the other, and it cannot show that changing the first changes the second.

The gap between the two claims is the whole space where longevity products live:

The cohorts support	The cohorts do not support
Short telomeres associate with higher all-cause mortality	That lengthening telomeres lowers mortality risk
The association survives adjustment for age	That the association is causal
It holds in general populations and in disease cohorts	That any available intervention changes the outcome

The peptide end of this story has been written about in detail on peptheon.eu, including which human data exist for telomere claims and epithalon.

What this evidence does not show

Nothing causal. No randomised trial exists on telomere length and lifespan. The largest dataset here is 121,749 people followed observationally.
No mortality data in people under 43. The individual-participant cohorts ranged from 43 to 75 years; the meta-analysis reports heterogeneity partly driven by age at measurement.
No effect in the over-80s of the size seen earlier. The meta-analysis explicitly found the hazard ratio smaller above 80, which undercuts any simple “short telomeres, short life” reading at the oldest ages.
No conclusion from a single measurement. Blood was drawn once per cohort, and leukocyte telomere length varies between tissues, methods and laboratories — the authors say so.
No cardiovascular or cancer mortality signal in the two cohorts: both intervals cross 1.
Nothing about anyone who has already been through a health programme. The mortality outcomes here are all-cause, long-horizon, and measured over decades; they are not the endpoints a product trial could plausibly measure.

The line

Short telomeres track higher mortality, by about 9 per cent per standard deviation, across 121,749 people — and not one of those 121,749 people was given anything to lengthen theirs. The number is real. The leap from it is not.