aen

How Aeon calculates your assessment

Everything we show you is either calculated mathematically or supported by a specific peer-reviewed study. Here’s exactly how.

Your biological age

Aeon uses PhenoAge (Levine et al., 2018) to calculate biological age. PhenoAge is a validated mortality-risk algorithm published in the journal Aging and trained on NHANES III data (n=9,926, validated on NHANES IV, n=6,209).

What it uses: 9 blood biomarkers + chronological age

  1. Albumin (g/dL)
  2. Creatinine (mg/dL)
  3. Glucose (mg/dL)
  4. C-reactive protein (ln mg/dL)
  5. Lymphocyte percent (%)
  6. Mean cell volume (fL)
  7. Red cell distribution width (%)
  8. Alkaline phosphatase (U/L)
  9. White blood cell count (1000 cells/uL)

How it works

These 9 values are combined using a parametric Gompertz model that estimates all-cause mortality hazard. The result is converted to a “phenotypic age” that corresponds to the average age at which someone would have your specific mortality risk profile. If your PhenoAge is lower than your birthday age, your blood markers suggest lower-than-average mortality risk for your age.

What it does NOT do

  • It does not predict lifespan.
  • It does not diagnose disease.
  • It is not a “longevity score” — it’s a mortality-risk proxy.
  • It cannot be improved by manipulating a single biomarker (the model is multivariate).

The math is deterministic. No AI is involved in the biological age calculation. The same inputs always produce the same output. The formula is public and verifiable.

Levine ME, Lu AT, Quach A, et al. An epigenetic biomarker of aging for lifespan and healthspan. Aging. 2018;10(4):573-591. PMID: 29676998.

Optimal vs. clinical reference ranges

Standard lab reports show you whether you’re in the “normal” range — meaning the range that includes 95% of the population tested at that lab. This tells you whether you’re sick, not whether you’re optimized.

Aeon uses a second set of ranges: longevity-optimal. These are derived from studies that correlate biomarker levels with the lowest all-cause mortality, lowest cardiovascular event rates, or best functional outcomes in long-term cohort studies.

Fasting Glucose

Clinical: 70–99 mg/dL

Optimal: 72–90 mg/dL

Source: NHANES mortality data

LDL-C

Clinical: <130 mg/dL

Optimal: <100 mg/dL (ideally <70 if high ASCVD risk)

Source: ACC/AHA guidelines

hsCRP

Clinical: <3.0 mg/L

Optimal: <1.0 mg/L (ideally <0.5)

Source: JUPITER trial (PMID: 18997196)

Vitamin D

Clinical: 30–100 ng/mL

Optimal: 40–60 ng/mL

Source: Endocrine Society + meta-analyses

When Aeon shows a biomarker bar on your dashboard, the colored zones represent these longevity-optimal ranges, not standard lab ranges.

How we rate evidence strength

Every recommendation includes an evidence rating. Here’s what each means:

Strong

Supported by at least one large randomized controlled trial (RCT) with clinically significant effect size, replicated in independent studies, and consistent with meta-analyses. The intervention has a well-characterized safety profile.

Moderate

Supported by smaller RCTs, prospective cohort studies, or mechanistic trials with plausible biological pathways. Evidence is consistent but not yet definitive at the “change clinical guidelines” level.

Emerging

Supported by preclinical data, early human trials, or observational studies with biological plausibility. Promising but not yet proven at scale. Aeon labels these clearly so you know the difference.

Every recommendation cites a specific PubMed ID (PMID). You can look up any study yourself at pubmed.ncbi.nlm.nih.gov.

The role of AI in your assessment

Aeon uses AI (a large language model) for two specific tasks:

1. Parsing your documents

When you upload a lab PDF, AI extracts the biomarker names and values from the document. You then review and confirm every extracted value before anything else happens.

2. Synthesizing recommendations

After your values are confirmed, AI reads your specific biomarker profile against our curated evidence base and generates personalized recommendations with appropriate context.

What AI does NOT do

  • Calculate your biological age (that’s pure math)
  • Decide which values are critical (that’s a deterministic threshold check)
  • Choose the evidence (that’s a curated, human-reviewed database of 36 verified studies)
  • Store or remember anything about you (stateless — each session is independent)

The AI operates within strict constraints: it cannot diagnose, prescribe, or recommend prescription medications. It can only cite studies that exist in our pre-verified evidence base. Its output is validated against a schema before being shown to you.

Where your information goes

Nowhere.

  • Your lab values exist only in your browser’s memory (RAM) during your session.
  • When you click “End Session,” everything is cleared from memory.
  • We do not have a database of user results. We cannot look up your assessment. We cannot be breached because there is nothing to breach.
  • The PDF emailed to you is your only copy. If you delete that email, the assessment is gone.
  • We do not use your values to train AI models.
  • We do not sell, share, or aggregate anonymized data.

This architecture is not a policy choice (policies can change). It is a technical constraint (there is no database to write to).

What Aeon cannot do

  • Diagnose any medical condition
  • Replace a physician’s clinical judgment
  • Account for your full medical history, family history, genetic risk, or medication interactions beyond those you disclose
  • Guarantee the accuracy of values you enter manually or that are extracted from your documents
  • Predict future health outcomes
  • Provide emergency medical advice

Aeon is a decision-support tool that helps you understand your lab results in the context of longevity research. It is not a medical device, not FDA-cleared, and not a substitute for professional medical advice.

© 2026 5M Ventures LLC. All rights reserved.