Peer-Reviewed Science

How Eumaia Measures Biological Age

Every number in Eumaia is grounded in published research — not wellness trends. Here's exactly how we estimate biological age, which studies we rely on, and what our limitations are.

Plain-English Summary

Eumaia uses a blood-biomarker phenotypic age model — not an epigenetic DNA clock. It does not sequence your genome or measure DNA methylation. It uses 9 routine blood markers to estimate biological age based on a validated population-level mortality risk model. Results are estimates for educational and self-tracking purposes, not clinical diagnoses.

The PhenoAge Algorithm

Eumaia's biological age estimate is based on the PhenoAge algorithm, developed by Dr. Morgan Levine and colleagues at Yale School of Medicine and published in Aging (2018). The algorithm was trained and validated on over 11,000 participants from the National Health and Nutrition Examination Survey (NHANES III).

PhenoAge outperforms chronological age at predicting all-cause mortality, cancer incidence, disability, and multi-morbidity — making it one of the most clinically validated biological age clocks available from standard blood tests.

Primary Citation

Levine, M.E., Lu, A.T., Quach, A., et al. (2018). "An epigenetic biomarker of aging for lifespan and healthspan." Aging, 10(4), 573–591. doi:10.18632/aging.101414

The 9 Biomarkers

PhenoAge uses nine routine blood markers — all available from a standard metabolic panel and CBC — weighted by their contribution to biological aging:

Albumin (g/dL)

Liver function, nutritional status, longevity predictor. Declines with inflammation and poor nutrition.

Optimal range

> 4.4

Creatinine (mg/dL)

Kidney filtration efficiency. Elevated levels indicate declining renal function.

Optimal range

0.7–1.0

Fasting Glucose (mg/dL)

Metabolic health and insulin sensitivity. Chronic elevation accelerates cellular aging.

Optimal range

< 85

C-Reactive Protein (CRP) (mg/L)

Systemic inflammation — the #1 driver of accelerated biological aging across all systems.

Optimal range

< 0.5

Lymphocyte % (%)

Immune competence. Low levels predict infection risk and immune senescence.

Optimal range

25–40%

MCV (fL)

Red blood cell size. Elevated MCV indicates B12/folate deficiency or liver stress.

Optimal range

80–95

RDW (%)

Red cell variation width — a powerful, underappreciated marker of oxidative stress and inflammation.

Optimal range

< 12.5

Alkaline Phosphatase (ALP) (U/L)

Liver and bone health marker. Elevations linked to liver inflammation and accelerated aging.

Optimal range

< 80

White Blood Cell Count (K/μL)

Total immune activation. Chronically elevated WBC reflects systemic immune stress.

Optimal range

4–7

How the Calculation Works

The PhenoAge calculation involves two steps:

Step 1 — Phenotypic Mortality Score

The nine biomarkers are combined using a linear equation with coefficients derived from Cox proportional hazards regression on the NHANES dataset. This produces a "mortality score" that represents 10-year mortality risk better than chronological age alone.

lin_combo = −19.907 − 0.0336×albumin + 0.0095×creatinine + 0.1953×(glucose/18) + 0.0954×ln(CRP) − 0.012×lymphocyte + 0.0268×MCV + 0.3306×RDW + 0.00188×ALP + 0.0554×WBC

Step 2 — Convert to Biological Age

The mortality score is converted to an age equivalent using a Gompertz survival model — mapping the predicted mortality risk to the chronological age at which the average person would carry that same risk.

mortality = 1 − exp(−exp(lin_combo) × (exp(0.0076927 × 120) − 1) / 0.0076927)
PhenoAge = 141.5 + ln(−0.00553 × ln(1 − mortality)) / 0.09165

Results are clamped between 18 and 100 to avoid extrapolation artifacts. All coefficients are taken directly from the published paper and have not been modified.

Validation & Evidence

11,000+

NHANES participants used in training and validation

3.2 yrs

Average biological age improvement seen in lifestyle interventions (Fitzgerald et al., 2021)

p < 0.001

Statistical significance of PhenoAge vs chronological age in predicting all-cause mortality

6 systems

Body systems tracked: metabolic, immune, inflammatory, hepatic, renal, cardiovascular

Additional Supporting Research

Fitzgerald, K.N., et al. (2021). "Potential reversal of epigenetic age using diet and lifestyle interventions." Aging. doi:10.18632/aging.202913
Liu, Z., et al. (2020). "Phenotypic age: a novel signature of mortality and morbidity risk." EBioMedicine. doi:10.1016/j.ebiom.2020.102981
Belsky, D.W., et al. (2015). "Quantification of biological aging in young adults." PNAS. doi:10.1073/pnas.1506264112

Epigenetic Clocks — The Next Frontier

PhenoAge is a blood-biomarker clock — it estimates biological age from routine lab values. A more advanced class of biological age tests uses epigenetic methylation patterns — directly reading which genes are being silenced or expressed as you age, at the DNA level.

Horvath Clock (2013)

Steve Horvath, UCLA

The first generation epigenetic clock. Trained on 8,000+ samples across 51 tissue types. Highly accurate but slow to respond to interventions.

GrimAge (2019)

Lu et al., Harvard

Second-generation clock trained on time-to-death rather than chronological age. Strongest predictor of all-cause mortality and healthspan of any current clock.

DunedinPACE (2022)

Belsky et al., Duke

Measures the pace of aging — how fast you are aging right now, not just your biological age. Responds to lifestyle interventions within months. Sinclair's preferred current metric.

Why Eumaia uses PhenoAge today

Epigenetic clocks require DNA methylation testing — a specialised lab process currently costing $300–500 per test (TruDiagnostic, Elysium Index, Biological Age Test by Chronomics). PhenoAge requires only a standard blood panel available at any lab for under $100. We chose accessibility without sacrificing validation quality. Epigenetic clock integration is on our roadmap — our goal is to be the platform that synthesises both blood-biomarker and methylation data into a unified biological age picture.

References: Horvath (2013) Genome Biology · Lu et al. (2019) Nature Aging · Belsky et al. (2022) eLife · Sinclair & LaPlante, Lifespan (2019)

Research Foundations

Eumaia draws from two distinct traditions — peer-reviewed longevity science that forms the methodological core, and the practitioner-experimenters who have shaped how millions of people actually engage with longevity. We are precise about which is which.

Scientific Foundation

Published researchers whose peer-reviewed work directly informs Eumaia's biological age model, protocols, and recommendations.

Dr. Morgan Levine

PhenoAge, Biological Clocks

Yale → Altos Labs

Dr. Steve Horvath

Epigenetic Clocks, DNA Methylation

UCLA / Altos Labs

Dr. Valter Longo

Fasting Mimicking Diet, Longevity

USC Longevity Institute

Dr. Matt Kaeberlein

Aging Biology, Rapamycin, mTOR

University of Washington

Dr. Rhonda Patrick

Micronutrients, Sauna, Inflammation

FoundMyFitness

Dr. Iñigo San Millán

Zone 2, VO2 Max, Mitochondria

CU Anschutz

Dr. Matthew Walker

Sleep Science, REM, Circadian

UC Berkeley

Dr. Gabrielle Lyon

Muscle-Centric Medicine, Protein

Institute for Muscle-Centric Medicine

Dr. Jason Fung

Therapeutic Fasting, Insulin, Autophagy

IDM Program

Dr. Ben Bikman

Insulin Resistance, Metabolism

Brigham Young University

Dr. Robert Lustig

Metabolic Health, Sugar, Liver

UCSF

Dr. Tim Spector

Microbiome, Twin Research, Nutrition

King's College London

Voices Shaping Longevity

The popularisers and practitioner-experimenters who built the consumer longevity conversation — and whose protocols Eumaia's users actively follow and want to evaluate.

David Sinclair

NAD+, Sirtuins, Epigenetics

Harvard Medical School

Andrew Huberman

Neuroscience, Protocols, Performance

Stanford

Dave Asprey

Biohacking, Mitochondrial Health

Bulletproof / unlimited.life

Peter Attia

Longevity Medicine, Zone 2, VO2 Max

Early Medical

All expert names and research are referenced for educational attribution only. No endorsement implied. Eumaia is not affiliated with, or partnered with, any individual listed above.

Honest Limitations

We believe in transparency. Here's what Eumaia cannot do:

  • PhenoAge is a population-level predictor — individual results carry inherent uncertainty (±3–5 years typical range).
  • We do not account for genetics. Two people with identical biomarkers may have different genetic aging trajectories.
  • Lab values fluctuate with hydration, recent meals, illness, stress, and timing of blood draw. A single reading is not definitive.
  • Eumaia is not a diagnostic tool and cannot diagnose, treat, or predict any specific disease.
  • Our intervention recommendations are based on population research — not clinical trials of your specific health profile.
  • VO2 Max, HbA1c, ApoB, and epigenetic clocks (DunedinPACE, GrimAge) would improve accuracy but require additional testing.

Medical Disclaimer

Eumaia is a longevity education and tracking platform. Content is for informational purposes only and does not constitute medical advice, diagnosis, or treatment. Always consult a qualified physician or licensed healthcare provider before making decisions about your health. The PhenoAge algorithm and all referenced research are used for educational purposes — Eumaia is not affiliated with Yale University, the NHANES program, or any referenced researcher.

Ready to Measure Your Biological Age?

Use the free calculator — no account required.