Biomarkers and longevity. How to read results without getting lost in the numbers

The more biomarkers we can measure today, the easier it is to get lost interpreting them. The science of healthy longevity is not about collecting as many numbers as possible, but about understanding which biomarkers are well validated, what they actually measure, and how to tell chronological age from biological age — a key concept for assessing real longevity.

1. How does biological age differ from chronological age?

Chronological age is simply the number of years since birth — independent of lifestyle or health status. Biological age is meant to describe something else: the actual functional state of the body, which correlates with disease risk and healthspan better than the calendar alone. Two people of the same chronological age can have markedly different biological age — and that difference is what longevity science is interested in, even though the way we measure it is still evolving and far from free of limitations. See also: healthspan.

2. Three levels of aging biomarkers: clinical, omic and epigenetic

A large consensus review by an aging research consortium (Moqri et al., Cell, 2023) organizes aging biomarkers into three levels of maturity and validation. Clinical biomarkers — glucose, lipid panel, hs-CRP, eGFR — are the best validated, available in any laboratory, and backed by decades of population data linking them to disease risk. Omic biomarkers (proteomics, metabolomics) analyse hundreds or thousands of molecules at once, offering a potentially richer picture but with greater interpretive complexity and limited standardisation. The third level — epigenetic clocks based on DNA methylation patterns — is the youngest and most intensively studied category. The reviewers state it plainly: no single biomarker at any of these levels currently meets all the criteria of an ideal, fully validated marker of aging. Examples of well-validated clinical markers: HbA1c and hs-CRP.

3. PhenoAge and GrimAge — epigenetic clocks that predict more than age

PhenoAge (Levine et al., Aging, 2018) was among the first epigenetic clocks designed not to predict chronological age but a clinically defined "phenotypic age" — and it showed a stronger association with all-cause mortality, cancer and declining physical function than earlier clocks focused purely on the calendar. Its successor, GrimAge (Lu et al., Aging, 2019), combines seven methylation surrogates of plasma proteins with an estimator of smoking pack-years and showed an even stronger association with time to death, coronary heart disease and cancer in large cohorts. Importantly, both tools were developed and validated on population data — applying them to precise individual risk assessment carries more limitations than the marketing around commercial "biological age" tests suggests.

4. DunedinPACE — measuring the pace of aging, not just its state

DunedinPACE (Belsky et al., eLife, 2022) takes a different approach: instead of estimating biological age at a single moment, it measures the pace of aging — whether a person is aging faster or slower than average over a period. The tool was built on the unique Dunedin cohort, tracking 19 indicators of organ function in the same people at four time points across two decades. That is a conceptually important difference from PhenoAge or GrimAge: pace of aging is in theory easier to link to lifestyle change over time than a one-off "biological age", although direct evidence that lifestyle interventions measurably slow DunedinPACE in randomised clinical trials remains early-stage.

5. Why clinical biomarkers still win in practice

Despite the enthusiasm around epigenetic clocks, clinical biomarkers remain the foundation of practical longevity prevention — for a simple reason: they are inexpensive, widely available, standardised across laboratories, and backed by decades of data linking them directly to specific, well-studied interventions. Epigenetic clocks, promising as research tools and for evaluating population-level effects of anti-aging interventions, do not yet have comparable clinical standardisation or clear pathways for "what to do if the result is unfavourable" — and that actionability is what determines a test's practical value. More: preventive medicine and longevity and heart health and longevity.

6. How to read a biomarker result without drawing the wrong conclusion

Three practical rules. First, check the validation level of the biomarker — clinical, omic or epigenetic — and treat it with the certainty that level deserves. Second, ask not only "what is my biological age" but "what will I do with this result" — if the answer is "nothing specific", the test's practical value is limited. Third, no single result — not even from the most advanced epigenetic clock — replaces a trend over time and the full clinical picture, including a conversation with a doctor, especially when the result suggests a meaningful deviation. See also: brain health and longevity.

Study limitations

Epigenetic clocks (PhenoAge, GrimAge, DunedinPACE) were developed and validated mainly in Western, largely Caucasian research cohorts — their accuracy in other ethnic groups is the subject of ongoing validation work, not settled fact. The link between these clocks and mortality or disease comes from observational studies — this does not prove that an intervention lowering a clock score automatically extends life; randomised trials with hard endpoints are still missing. Moqri and colleagues explicitly note the lack of consensus on standardisation and comparability across laboratory platforms offering commercial "biological age" tests — the same sample can yield different results across providers. DunedinPACE, despite its innovative methodology, rests on a single source cohort (Dunedin, New Zealand) with a specific demographic profile.

One small step

Before paying for a commercial "biological age" test, check which epigenetic clock or method the provider uses and what validation data it has — and if you plan to interpret the result, ask a clinician what concrete action would follow from it.