DunedinPACE (Pace of Aging)
A DNA test that estimates how fast your body is aging. A value of 1.0 means one biological year of aging for each calendar year.
AI-verified against cited sources. Awaiting expert reviewAlso called DunedinPACE
Check your result
The usual range for DunedinPACE depends on age. Add your age above to see where you stand.
Nothing you type here leaves your browser.
What it is
DunedinPACE is a research score. It is read from chemical tags, called methylation, at 173 spots on the DNA in your blood cells. Researchers built it by following about a thousand people born in Dunedin, New Zealand, from age 26 to 45 and tracking 19 measures of body health. The score estimates how fast those measures were declining. It is scaled so that 1.0 is the typical pace in that group: one year of biological aging per calendar year. A value of 1.1 means a pace about 10% faster, and 0.9 means about 10% slower. It is a speed, not an age.
Why it matters
In research studies, people with higher DunedinPACE scores were more likely to develop chronic illness and disability and to die earlier, on average. For example, each standard-deviation increase went with about 26% higher risk of death in one study of older men (average age 77) and about 65% higher in another study of adults averaging 66 years old. These are patterns across large groups. The score can't tell you what is driving your number or what will happen to you, and no medical guideline sets a normal range or a target for it.
If it’s low
A value below 1.0 means the score estimates a slower pace of decline than in the group it was built on. Younger adults tend to score lower. There is no medical target for this score, and a low value doesn't guarantee good health.
Common causes
- Younger age, since scores are lower on average in younger adults
- A blood sample rather than a cheek swab, since blood-based results are lower
- Normal differences between people, plus some test-to-test variation
What can help
- Keep up regular checkups and the screening your clinician recommends, since a low score doesn't replace them
If it’s high
A value above 1.0 means the score estimates a faster pace of decline than in the group it was built on. In studies, higher scores went with more chronic illness, more disability and earlier death on average. That is a pattern across large groups. Many things can raise a score, and it doesn't tell you which one applies to you.
Common causes
- Older age, since scores are higher on average in older adults
- A cheek-swab or saliva sample instead of blood, which scored much higher in one study (the algorithm was built on blood)
- Smoking history, which affects DNA methylation (the effect is smaller for DunedinPACE than for the older DunedinPoAm score)
- Long-term health conditions, which were more common in people with higher scores
- Hard early-life circumstances, such as childhood poverty or victimization, which were linked to faster scores in young adults
- Differences between labs, array types or data processing
What can help
- Check what kind of sample the number came from (blood or cheek swab) before comparing it with anything else
- Ask your clinician how it fits with measures that have clear targets, such as blood pressure, blood sugar and cholesterol
- Not smoking, regular physical activity and a balanced diet support health in general
What can affect your result
- Age: scores rise with age on average (correlation 0.32 in 1,175 UK adults aged 28 to 95)
- Sample type: in one study of 83 people, cheek swabs averaged 1.61 and saliva 1.50, compared with 0.89 to 0.96 for blood-based samples
- Lab and array type: the same samples agreed less closely across two array types (reliability 0.87) than across two runs on the same type (0.96 to 0.97)
- Smoking history, long-term health conditions and early-life hardship
- Who the scale was built on: the 1.0 anchor comes from a New Zealand birth cohort that is mostly white (93%)
- Random lab variation: reliability is high but not perfect, so a small difference between two tests may not mean anything
- Sample type changes the number. The algorithm was built and tested on DNA from blood. In one 2025 study of 83 people, cheek swabs averaged 1.61 and saliva 1.50, compared with 0.89 to 0.96 for blood-based samples.
- Array type and lab processing can shift results. Reliability was 0.96 to 0.97 when samples were run twice on the same type of array, but 0.87 when the same samples were run on two different array types.
- Smoking history and the mix of white blood cell types can influence methylation scores. In the original study, DunedinPACE's links to illness and death largely held after adjusting for both.
- The older DunedinPoAm score is a different algorithm. Its values are not interchangeable with DunedinPACE.
Not well established. In the one randomized trial that measured it (CALERIE), the group assigned to cut calories by 25% (they averaged about 12%) scored about 2 to 3% lower on average at 12 months, and the difference was still there at 24 months.
When to talk to a clinician
No score level calls for urgent care. If you have symptoms, or other results that your lab marked as urgent or outside its range, those matter more than this number. Talk with your clinician about them.
Questions you could ask
- Is it worth retesting, and if so, should I use the same company and sample type?
- How does this score fit with the rest of my results?
- Which of my routine results, such as blood pressure, blood sugar and cholesterol, should we look at first?
- Does the type of sample used for this test change how to read the number?
- Does this score change anything about my usual checkups?
Questions people ask
What does a DunedinPACE of 1.0 mean?
It is the reference pace: one year of biological aging for every calendar year, as measured in the group the score was built on. Below 1.0 means a slower pace and above 1.0 a faster one, so 1.1 is about 10% faster. Scores tend to be lower in younger people and higher in older people, so 1.0 is not a pass mark at every age.
Can I compare a score from a cheek swab, or from a different company?
Not safely. The algorithm was built and tested on blood. In one study it gave much higher values for cheek and saliva samples (about 1.5 to 1.6) than for blood (about 0.9 to 1.0), and results for the same person differed between sample types. Array type and lab can also shift the number. Compare results only from the same company and sample type.
How is this different from a biological age?
A biological-age clock estimates an age. DunedinPACE estimates a speed, and its developers say it is not a clock and does not measure an age value. Neither kind is a diagnostic test, and researchers still lack a gold-standard measure of biological aging.
Show the science and sources
For clinicians and researchers
DunedinPACE (Pace of Aging Calculated from the Epigenome; Belsky 2022) is an elastic-net regression on 173 CpG sites. It was trained on blood DNA methylation (Illumina EPIC) from 817 Dunedin Study members at age 45 to predict a longitudinal Pace of Aging phenotype: a composite of each person's rates of change in 19 organ-system biomarkers measured at ages 26, 32, 38 and 45, scaled to a cohort mean of 1 (SD 0.29). The in-sample correlation with measured Pace of Aging was 0.78, and effect sizes in the Dunedin cohort itself may be overestimated. Scoring quantile-normalizes a sample against a 20,000-probe reference panel (the 173 scoring probes plus 19,827 normalization probes) so that values are comparable to the Dunedin cohort. The developers' R package accepts Illumina 450k, EPIC and EPICv2 array data and describes DunedinPACE as a blood DNA methylation biomarker. Test-retest ICC was 0.96 in 36 replicate samples on 450k arrays, 0.97 in 28 replicates on EPIC arrays and 0.87 in 350 replicates run on both array types, compared with 0.89 for the earlier DunedinPoAm, which used 12 years of follow-up instead of 20. Per 1 SD, DunedinPACE was associated with all-cause mortality in the Normative Aging Study (HR 1.26, 95% CI 1.14 to 1.40, n = 771 older men) and the Framingham Offspring cohort (HR 1.65, 1.51 to 1.79, n = 2,471), and with incident cardiovascular disease (HR 1.39) and stroke or TIA (HR 1.37) in Framingham. Within-person change was small: +0.021 per 5 years in the Normative Aging Study. In the CALERIE randomized trial, 25% calorie restriction lowered DunedinPACE by about 2 to 3% at 12 and 24 months (d = −0.29 and −0.25), with substantial overlap between the groups, while the DNA methylation versions of PhenoAge and GrimAge did not change. Tissue matters: in 83 people aged 9 to 70, mean DunedinPACE was 1.61 in buccal cells and 1.50 in saliva versus 0.89 to 0.96 in blood-based samples, and within-person correlations between tissues ranged from 0.15 to 0.72. The developers state that commercial use requires a license from TruDiagnostic. No clinical reference interval, decision threshold or consensus standard exists for DunedinPACE, and researchers note there is no gold-standard measure of biological aging.
Sources
- DunedinPACE, a DNA methylation biomarker of the pace of aging, eLife (2022)
- DunedinPACE R package and README (what values to expect, inputs, licensing), Belsky lab, Columbia University (GitHub)
- Effect of long-term caloric restriction on DNA methylation measures of biological aging in healthy adults from the CALERIE trial, Nature Aging (2023)
- Cross-tissue comparison of epigenetic aging clocks in humans, Aging Cell (2025)
AI-verified against cited sources. Awaiting expert review. Last reviewed 2026-10-03. How we build and check these pages.