Somewhere in the last three years, "What's your biological age?" replaced "What's your resting heart rate?" as the biohacker's opening question. Epigenetic clock companies are selling direct-to-consumer tests for €200–€500. Longevity influencers post their results like before-and-after photos. The Nordic biohacking community, centered around events like Helsinki's HOLOLIFE (formerly the Biohacker Summit), has made biological age tracking a mainstream practice.
There's real science behind this. There are also real problems with how the science is being applied, marketed, and misunderstood at the individual level. This article is about both.
Chronological age is how many years you've been alive. Biological age is an attempt to quantify how much functional deterioration has accumulated in your body, regardless of the calendar. Two 45-year-olds can have dramatically different cardiovascular systems, metabolic profiles, immune function, and organ reserve. Biological age tries to capture that difference in a single number.
The problem is that "biological age" is not one thing. It's a family of different computational approaches, each measuring different biological signals, trained on different datasets, and — critically — producing different answers for the same person at the same time. Understanding which approach does what is essential before you spend money on any test.
Epigenetic clocks measure biological age through DNA methylation, chemical modifications (methyl groups) attached to specific sites on your DNA that change predictably with age. These marks don't alter your genetic code, but they regulate which genes are expressed and which are silenced. By analyzing patterns across hundreds or thousands of methylation sites, algorithms can estimate how old your biology "looks."
Steve Horvath's original clock, published in 2013, identified 353 CpG sites (specific methylation locations) that correlated strongly with chronological age across multiple tissue types. [1] Gregory Hannum published a similar clock the same year, focused on blood-derived methylation data. [1]
These first-generation clocks were trained directly on chronological age. They're excellent at predicting how old someone is, but that's not the same as measuring how fast they're aging or how likely they are to get sick. A clock trained to match the calendar is, by design, anchored to the calendar. It tells you less about the deviation from normal aging that actually matters for health.
Morgan Levine and Steve Horvath developed PhenoAge in 2018 with a fundamentally different approach. Instead of training on chronological age, they trained on mortality risk, using clinical biomarkers that predict how likely someone is to die in the next ten years. [2] The result was a clock that captures aging-related disease risk, not just the passage of time.
GrimAge, published in 2019, went further by incorporating DNA methylation surrogates for seven plasma proteins plus smoking pack-years, all associated with mortality and morbidity. [1] It has shown stronger associations with socioeconomic status, lifestyle factors, and disease outcomes than first-generation clocks. [1]
DunedinPACE (Pace of Aging Computed from the Epigenome) represents a conceptual leap. Rather than estimating a static "age", how old your biology looks at one point in time, it estimates how fast you're aging right now. [3]
The distinction matters enormously. Think of it this way: first- and second-generation clocks are like an odometer (total mileage accumulated). DunedinPACE is like a speedometer (current rate of decline). [4]
DunedinPACE was built from the Dunedin Longitudinal Study, a birth cohort of 1,000 individuals born in 1972–73 in New Zealand, followed with repeated biological measurements across two decades. Researchers tracked 19 biomarkers of organ-system integrity from ages 26 to 45, derived a composite "pace of aging" score, and then used machine learning to identify DNA methylation patterns that predict that pace from a single blood draw. [3]
A DunedinPACE value of 1.0 means you're aging at the expected rate: one biological year per calendar year. A value of 1.2 means you're aging 20% faster. A value of 0.85 means 15% slower. In the CALERIE randomized controlled trial, the first human RCT of long-term caloric restriction in healthy adults, 25% calorie reduction over two years slowed DunedinPACE by 2–3%, equivalent to a 10–15% reduction in mortality risk. [5] Notably, the same intervention produced no significant changes in PhenoAge or GrimAge. [5]
This suggests that pace-of-aging measures like DunedinPACE may be more sensitive to recent lifestyle interventions than static biological age clocks, which reflect cumulative damage over a lifetime. [5]
While epigenetic clocks require specialized DNA methylation assays (typically €200–€500 per test), there's a parallel approach to biological age that uses something much more accessible: standard blood biomarkers.
PhenoAge's clinical foundation — the step before it was mapped to DNA methylation — was built entirely from routine lab tests. Morgan Levine's team analyzed data from NHANES III (a massive US population health survey) and identified 9 blood biomarkers that, combined with chronological age, most strongly predicted 10-year mortality risk. [2] [6]
Those 9 biomarkers are:
Albumin — The most abundant protein in blood plasma. Reflects liver function and nutritional status. Declining albumin is one of the strongest predictors of frailty and mortality in aging research. Aniva's liver & kidney panel includes albumin.
Creatinine — A waste product of muscle metabolism filtered by the kidneys. Elevated creatinine signals impaired kidney function. Kidney decline accelerates with age and compounds other organ aging. Part of Aniva's metabolic panel.
Fasting glucose — Reflects glycemic control. Even mildly elevated fasting glucose (within the "normal" range) is associated with accelerated aging and increased cardiovascular risk. Aniva measures fasting glucose and HbA1c.
C-reactive protein (CRP) — An acute-phase protein produced by the liver in response to inflammation. High-sensitivity CRP (hs-CRP) is a marker of chronic low-grade inflammation — the "inflammaging" that drives virtually every age-related disease. This is the only PhenoAge input that isn't typically found on a standard metabolic panel. Aniva includes hs-CRP.
Lymphocyte percentage — The proportion of white blood cells that are lymphocytes. Declining lymphocyte percentage reflects immune system aging (immunosenescence) — the gradual deterioration of adaptive immunity that makes older adults more susceptible to infections and cancers.
Mean cell volume (MCV) — The average size of red blood cells. Abnormal MCV (either high or low) can indicate nutritional deficiencies, bone marrow dysfunction, or chronic disease.
Red cell distribution width (RDW) — Measures variability in red blood cell size. Higher RDW indicates less uniform red cell production and has emerged as an independent predictor of mortality across multiple disease states.
Alkaline phosphatase (ALP) — An enzyme found in liver, bone, and intestine. Elevated ALP can indicate liver or bone pathology and is associated with increased mortality risk.
White blood cell count — Total immune cell count. Both very high and very low counts are associated with increased mortality, reflecting either chronic infection/inflammation or immune depletion.
In the original validation, each one-year increase in PhenoAge above chronological age was associated with a 9% increase in all-cause mortality risk. [2] The clinical PhenoAge measure predicted survival with approximately 90% accuracy over 10 years. [2]
Aniva tests nearly every PhenoAge input biomarker. Albumin, creatinine, glucose, hs-CRP, lymphocyte percentage, MCV, RDW, ALP, white blood cell count, plus HbA1c, fasting insulin, and dozens of other markers that feed into biological age calculations. You don't need a €500 epigenetic test to start tracking the biomarkers that actually predict mortality. Aniva's 140+ biomarker panel gives you the clinical foundation. Waitlist is free. Full membership: €199/year.
Here's the part of the story the direct-to-consumer testing companies don't emphasize.
In 2022, Albert Higgins-Chen, Morgan Levine, and colleagues published a landmark paper in Nature Aging showing that technical noise produces deviations of up to 9 years between replicate tests of six prominent epigenetic clocks, using the same biological sample. [7] Not different blood draws from different days. The same sample, processed twice.
As Higgins-Chen put it: epigenetic clocks could report you as biologically 50 years old on one test and 59 on the next. [8]
The sources of this noise include variability in DNA methylation array processing, differences in probe reliability across the Illumina platforms used for measurement, and the fact that many individual CpG sites are inherently noisy, their methylation levels fluctuate for reasons unrelated to aging. [7]
The team developed a computational fix using principal component analysis (PC clocks), which reduces replicate agreement to within 1.5 years for most tests. [7] DunedinPACE was explicitly designed with high test-retest reliability in mind. [3] But most commercial biological age tests don't tell you which version of which clock they're running, whether they use PC-corrected versions, or what the expected measurement error is.
If you take a biological age test, get a result of "4 years younger," make some lifestyle changes, retest three months later, and get "2 years younger" — the most likely explanation isn't that you aged. It's that you measured noise. [9]
A recent review of epigenetic clocks in longitudinal and intervention studies made this explicit: different clocks frequently disagree on whether an intervention had any effect, and changes in some clocks may simply be the result of technical noise creating false positive results. [9]
This doesn't mean biological age testing is useless. It means the field is still maturing, and individual results need to be interpreted with appropriate skepticism — especially when measured over short intervals.
This is the biggest misapplication of biological age testing in the consumer longevity space. Quarterly retesting of epigenetic clocks is mostly measuring noise, not signal. The biological processes these clocks capture — cumulative organ decline, immune aging, metabolic deterioration — change slowly. True aging-related methylation changes require months to years to manifest reliably above the measurement noise floor.
Even in the CALERIE trial, where participants followed a rigorous 25% calorie restriction protocol for two full years, the DunedinPACE effect was only 2–3%. [5] For most lifestyle interventions, expecting detectable changes in under 6–12 months is unrealistic.
Standard blood biomarkers, by contrast, can change meaningfully over shorter timeframes. Your hs-CRP can drop within weeks of reducing systemic inflammation. Your fasting glucose and HbA1c respond to dietary changes within 2–3 months. Your albumin, lymphocyte percentage, and kidney function markers shift in response to sustained lifestyle changes. These are the PhenoAge inputs and they're the fastest-moving levers you can pull.
Aging is not uniform. Your cardiovascular system, immune system, liver, kidneys, brain, and musculoskeletal system can all age at different rates. A single biological age number collapses this heterogeneity into one figure that may obscure organ-specific problems.
A person with excellent metabolic function but deteriorating kidney filtration might score a "normal" biological age, because the kidney signal is averaged away by strong performance in other systems. This is why comprehensive blood testing that measures organ-specific biomarkers: liver enzymes, kidney markers, inflammatory markers, metabolic markers, hormone panels — provides more actionable information than a single composite score.
PhenoAge was developed using statistical relationships derived from large population studies (NHANES, with thousands of participants). [2] It predicts mortality risk at the population level with impressive accuracy. But at the individual level, a single PhenoAge calculation is a snapshot influenced by recent illness, sleep quality, acute stress, time of day, and dozens of transient factors.
This doesn't invalidate the test — it means you need longitudinal data, not isolated measurements. Track the trajectory over multiple draws, across seasons, over years. That's where the signal emerges from the noise.
The epigenetic clock conversation has created an odd dynamic where people obsess over a methylation-derived number they can barely move while ignoring the input biomarkers they can directly influence through evidence-based interventions.
Consider the PhenoAge inputs: reducing hs-CRP through anti-inflammatory nutrition and exercise, improving fasting glucose and HbA1c through metabolic optimization, maintaining albumin through adequate protein intake, preserving lymphocyte function through sleep, stress management, and micronutrient sufficiency. These are concrete, measurable, and modifiable. And they feed directly into the mortality-prediction algorithms that biological age scores are built on.
Not all clocks measure the same thing. First-generation clocks (Horvath, Hannum) are poorly correlated with lifestyle factors and intervention effects. Second-generation clocks (PhenoAge, GrimAge) better capture disease risk but may not respond to short-term changes. Third-generation measures (DunedinPACE) are more sensitive to recent interventions but measure pace rather than cumulative age. [5]
In one large study of a representative US adult population, different generations of clocks showed entirely different patterns by race, socioeconomic status, and education level. [1] When the clocks disagree with each other, which one should you believe? The honest answer is: it depends on what question you're asking.
You don't need an epigenetic test to start optimizing your biological age.Every PhenoAge input — albumin, creatinine, glucose, hs-CRP, lymphocyte percentage, MCV, RDW, ALP, white blood cells — is a standard blood biomarker. Aniva's 140+ biomarker panel measures them all, plus HbA1c, fasting insulin, homocysteine, and the full cofactor network. Track the inputs. Watch the trajectory. That's how you know if your protocol is working.
Despite the limitations, biological age research has produced genuinely important findings:
Aging is modifiable. The CALERIE trial demonstrated in a randomized controlled setting that a lifestyle intervention (caloric restriction) can measurably slow the pace of biological aging in healthy humans. [5] The effect was small (2–3%) but real, and in population-level modeling, even small reductions in the pace of aging translate to large reductions in disease burden and mortality.
The biomarkers that predict biological age are the same ones that respond to intervention. This is the practical takeaway. The clinical markers that constitute PhenoAge — inflammation, glucose regulation, liver function, kidney function, immune status, red blood cell health — are the same markers that change when you improve your diet, exercise, sleep, and stress management. You don't need a €500 methylation assay to track them.
Pace of aging matters more than static age. DunedinPACE's sensitivity to the CALERIE intervention, where static clocks showed no effect, suggests that how fast you're currently declining matters more than how much total decline has occurred. [5] This is good news: it means recent changes in behavior can show up in the right measurements, even if they haven't yet reversed years of accumulated damage.
Faster epigenetic aging predicts real outcomes. Across multiple cohorts, faster DunedinPACE has been associated with lower brain volume, accelerated cognitive decline, increased risk of dementia, heart disease, stroke, and disability. [10] [11] These aren't theoretical associations — they've been validated in the Dunedin Study, the Framingham Heart Study, and the Alzheimer's Disease Neuroimaging Initiative, spanning ages 45 to 75+.
The most actionable biological age strategy isn't an epigenetic test. It's comprehensive blood work.Measure the inputs. Track them over time. Watch for trajectories, not single numbers. Aniva gives you 140+ biomarkers per draw — every PhenoAge input, plus hormones, minerals, lipids, and inflammation markers that epigenetic clocks can't see. Start with data. The rest follows.
Waitlist is free. Full membership: €199/year.
Biological age testing is real science with real limitations. Epigenetic clocks have evolved through three generations:
Each generation captures different aspects of aging, and they frequently disagree with one another.
The measurement noise problem is substantial: technical variability can produce deviations of up to 9 years between tests of the same sample. [7] Newer computational approaches and purpose-built clocks like DunedinPACE have improved reliability, but the field is still maturing. Quarterly retesting is mostly measuring noise. Annual or biannual testing is more appropriate for detecting meaningful change.
The most underappreciated insight from biological age research is that the clinical biomarkers underpinning PhenoAge: albumin; creatinine; glucose; hs-CRP; lymphocyte percentage; MCV; RDW; ALP; white blood cell count; are standard blood tests available from any lab. [2] [6] They respond to intervention faster than methylation marks, they cost a fraction of epigenetic testing, and they tell you exactly which organ systems need attention.
You don't need an epigenetic clock to start slowing down your aging. You need a baseline. You need the right biomarkers. And you need to track the trajectory.
Medical disclaimer: This content is for informational purposes only and is not medical advice. Always discuss results with a qualified healthcare professional.
