Pace-of-Ageing Measures vs Biological Age Clocks

Biomarkers described as biological age clocks and biomarkers described as pace-of-ageing measures can use similar laboratory data while representing different statistical targets. A clock generally asks how old or risk-burdened a biological profile appears at one time; a pace measure asks how rapidly age-related change appears to be occurring. [1] [4]

Who This Is Useful For

This page is useful for readers comparing epigenetic test results, cohort studies, or trial endpoints. It clarifies why a result expressed in years above or below chronological age is not equivalent to a result expressed as biological years of change per chronological year.

Biological Age Clocks: Estimating State

First-generation DNA methylation clocks were trained to predict chronological age from methylation patterns measured across people of different ages. Their output is age-like, and the difference between predicted and chronological age is often called age acceleration. Later clocks such as PhenoAge and GrimAge were trained partly on clinical or mortality-related targets, so the shared word “clock” does not mean that all models estimate the same construct. [2] [3] [4]

Pace-of-Ageing Measures: Estimating Rate

The original Dunedin Pace of Aging measure combined within-person change in multiple organ-system biomarkers measured repeatedly from ages 26 to 38. DunedinPoAm and the later DunedinPACE algorithm then used blood DNA methylation at one assessment to approximate a longitudinally derived rate. Their scale is centred near one biological year per chronological year, rather than an age in years. [5] [6] [7]

State and Rate at a Glance

Why Similar Assays Can Produce Different Measures

DNA methylation is the input for many clocks and for DunedinPACE, but the training outcome determines what the algorithm is designed to reproduce. A methylation model trained on chronological age learns a different statistical mapping from one trained on mortality proxies or on repeated physiological decline. Comparisons in the same cohort have found only limited agreement among several proposed biological ageing measures, supporting the view that they capture overlapping but non-identical information. [8]

How to Interpret Disagreement

A person or group can have an older clock estimate without a correspondingly faster pace score, or the reverse. This need not be a contradiction: accumulated state reflects prior development, exposures, illness, and ageing, whereas a rate-oriented model targets change over a defined part of adulthood. Model-specific training data and measurement error can also contribute. [7] [8]

Use in Longitudinal and Intervention Research

Rate measures are conceptually attractive when a study asks whether ageing-related change differs between groups or across time. However, a single blood sample remains a model-based estimate of a rate, not repeated observation of the underlying organ systems. Clocks can also serve as longitudinal outcomes, but the reliability of change scores is crucial because technical noise can be large relative to short-term biological change. [6] [7] [9]

Evidence Quality and Interpretation

Confidence is strong that state-oriented clocks and longitudinally trained pace measures differ in construction and intended interpretation. Their foundational studies specify distinct training targets and output scales. [2] [5] [7]

Confidence is moderate that these measures capture useful differences in health and ageing at the population level. Several measures associate with morbidity, mortality, or functional decline, but the strength and pattern of association vary by model and cohort. [3] [4] [7] [8]

Confidence is weaker for interpreting a small change in one individual as a verified change in their ageing rate. Repeatability, cell composition, batch effects, regression modelling, and the interval between samples all affect longitudinal interpretation. [9]

What This Does Not Mean

• It does not mean every measure labelled a clock is purely a chronological-age predictor.
• It does not mean a pace score directly observes every tissue changing in real time.
• It does not mean an age-acceleration value can be converted into an annual pace by simple arithmetic.
• It does not mean one model is universally superior; suitability depends on the research question and validation setting.

Practical Interpretation Examples

• A clock reports “five years older”: This is a model-relative age or risk estimate, not proof that five literal years of lifespan have been lost.
• A pace score is above one: This indicates a faster modelled rate relative to its reference scale, not a forecast that the same rate will persist indefinitely.
• Two measures move differently: First check their training targets, scales, tissues, and repeatability before treating the difference as biological conflict.

Summary

Biological age clocks mainly summarize accumulated age-related state or risk, while pace-of-ageing measures aim to summarize the rate of ongoing change. They can be complementary, but their outputs are not interchangeable. Interpretation should follow the model's training target, validation evidence, population, and technical reliability. [7] [8]

References

1. Jylhava, J., et al. (2017). Biological age predictors. EBioMedicine. https://doi.org/10.1016/j.ebiom.2017.03.046
2. Horvath, S. (2013). DNA methylation age of human tissues and cell types. Genome Biology. https://doi.org/10.1186/gb-2013-14-10-r115
3. Levine, M. E., et al. (2018). An epigenetic biomarker of aging for lifespan and healthspan. Aging. https://doi.org/10.18632/aging.101414
4. Lu, A. T., et al. (2019). DNA methylation GrimAge strongly predicts lifespan and healthspan. Aging. https://doi.org/10.18632/aging.101684
5. Belsky, D. W., et al. (2015). Quantification of biological aging in young adults. Proceedings of the National Academy of Sciences. https://doi.org/10.1073/pnas.1506264112
6. Belsky, D. W., et al. (2020). Quantification of the pace of biological aging in humans through a blood test, the DunedinPoAm DNA methylation algorithm. eLife. https://doi.org/10.7554/eLife.54870
7. Belsky, D. W., et al. (2022). DunedinPACE, a DNA methylation biomarker of the pace of aging. eLife. https://doi.org/10.7554/eLife.73420
8. Belsky, D. W., et al. (2018). Eleven telomere, epigenetic clock, and biomarker-composite quantifications of biological aging: Do they measure the same thing? American Journal of Epidemiology. https://doi.org/10.1093/aje/kwx346
9. Higgins-Chen, A. T., et al. (2022). A computational solution for bolstering reliability of epigenetic clocks: implications for clinical trials and longitudinal tracking. Nature Aging. https://doi.org/10.1038/s43587-022-00248-2