DNA-based biological age tests cost several hundred euros and sometimes give different results from one laboratory to another. PhenoAge starts with a standard blood test.
Can nine blood markers reveal your body’s age? They can produce a useful estimate of physiological risk, not read the actual age of every cell.
PhenoAge converts your chronological age and nine laboratory results into an age expressed in years. Its advantage is tangible: you can calculate it from a blood test available in France, without DNA analysis. Its main pitfall is just as tangible: an incorrect unit, a recent infection or an opaque calculator can shift the result by several years.
If you already have the nine values, use the PhenoAge calculator with French unit conversion. The calculation runs locally in your browser and displays every value actually fed into the formula.
There is strong evidence that PhenoAge is associated with morbidity and mortality in large cohorts. The evidence is much weaker for converting your personal score into life expectancy or claiming that a protocol has “rejuvenated” you. So use this number as an imperfect summary of nine markers, then return to the results that make it up.
What is PhenoAge?
PhenoAge, or Phenotypic Age, is a clinical clock developed using the US NHANES III survey. Researchers first selected, from 42 variables, nine biomarkers and chronological age that together predicted mortality. They then converted this statistical risk into a more intuitive unit: years.
The model validated in another NHANES cohort therefore answers this question: what age, on average in the reference population, has a mortality risk comparable to that associated with this blood profile? It measures neither an instantaneous rate of aging nor the wear and tear on each organ.
This distinction changes the interpretation. A PhenoAge of 52 years in a 45-year-old does not mean that their liver, brain and arteries are all 52 years old. The score aggregates signals related to inflammation, metabolism, kidneys, liver, red blood cells and immunity. A single temporary abnormality can influence the whole result.
The original model nevertheless has genuine prognostic value. In its US validation, PhenoAge remained associated with mortality after accounting for chronological age, including in several subgroups. This is a cohort-level association, not a certain prediction for an individual.
The blood markers you need
You need an age and nine results from samples collected on the same date. Here are the units required by the original implementation of the BioAge scientific package :
| Data | Required unit | Common presentation in France | Check or conversion |
|---|---|---|---|
| Chronological age | years | years | age at the time of blood collection |
| Albumin | g/L | g/L | no conversion if the report states g/L |
| Creatinine | µmol/L | µmol/L or mg/L | mg/L × 8.84 |
| Glucose | mmol/L | mmol/L or g/L | g/L × 5.55 |
| CRP | mg/dL, then natural logarithm | mg/L | divide mg/L by 10 before ln(CRP) |
| Lymphocytes | % | % and sometimes G/L | use the percentage, not the absolute count |
| VGM or MCV | fL | fL | no conversion |
| IDR-CV or RDW-CV | % | % | RDW-SD expressed in fL is not suitable |
| Alkaline phosphatase | U/L | U/L | no conversion |
| White blood cells or WBC | 10⁹/L | G/L or 10⁹/L | the numerical value is identical to 10³/µL |
The hematology results generally come from a complete blood count with differential: white blood cells, lymphocyte percentage, MCV and RDW-CV. Albumin, creatinine, glucose, CRP and alkaline phosphatase are blood chemistry tests. They are not all automatically included in every panel.
So yes, the calculation is possible in France. You do not need a laboratory to sell a “PhenoAge test.” However, ordering and interpreting the tests must be appropriate to your clinical situation. This exploratory score alone does not justify repeated blood tests.
CRP deserves special attention. The original publication expresses it in mg/dL, whereas French laboratories mainly use mg/L. A CRP of 2 mg/L becomes 0.2 mg/dL before taking the natural logarithm. If the result is reported as “less than” a detection limit, you do not know the exact value. Never enter zero, because ln(0) does not exist.
How to perform and check the calculation
The safest option is to use a local spreadsheet or a script with a visible formula. Avoid sharing identifiable health data with a website that does not disclose its code, units or privacy policy.
Follow this order:
- Copy the nine values and your age into a column separate from the laboratory report
- Convert only the units that differ from the table above
- Convert CRP from mg/L to mg/dL, then calculate its natural logarithm
- Calculate the predictor
xb, the intermediate riskM, then PhenoAge - Keep the source values, conversions, formula and date with the result
The constants below follow the option orig = TRUE in the BioAge package. A correction published in 2019 added the intermediate step missing from the validation paper. One tricky detail: this correction shows 0,09165 in the final denominator, whereas the BioAge code uses 0,090165 and more precise coefficients. This guide follows the BioAge code to make the example reproducible. Do not compare two calculators that use different variants.
xb = -19,90667 - 0,03359355 × albumine + 0,009506491 × créatinine + 0,1953192 × glucose + 0,09536762 × ln(CRP) - 0,01199984 × lymphocytes + 0,02676401 × VGM + 0,3306156 × IDR-CV + 0,001868778 × phosphatases alcalines + 0,05542406 × leucocytes + 0,08035356 × âge
M = 1 - exp[(-1,51714 × exp(xb)) / 0,007692696]
PhenoAge = 141,50225 + ln[-0,0055305 × ln(1 - M)] / 0,090165
Fictional example
Take a 45-year-old with these results, all fictional:
| Variable | Input value |
|---|---|
| Albumin | 42 g/L |
| Creatinine | 90 µmol/L |
| Glucose | 5.4 mmol/L |
| CRP | 2 mg/L, converted to 0.2 mg/dL |
| Lymphocytes | 30 % |
| MCV | 92 fL |
| RDW-CV | 13.3 % |
| Alkaline phosphatase | 75 U/L |
| White blood cells | 6.5 × 10⁹/L |
With the full coefficients from the BioAge package, the result is approximately 43.3 years. The simple difference is therefore 43.3 - 45 = -1.7 years.
Before interpreting this result, run these checks:
- check that the CRP used is indeed 0.2 mg/dL, not 2
- check that creatinine is in µmol/L and glucose in mmol/L
- check that RDW is RDW-CV expressed as a percentage
- calculate it a second time from the original values, without rounding intermediate steps
A biologically implausible result should first raise suspicion of a unit or transcription error. It is neither something to celebrate nor something to treat.
Interpreting the difference from chronological age
The calculation you can do at home is the raw delta :
delta = PhenoAge - âge chronologique
A positive delta means the score exceeds your actual age. A negative delta means it is lower. This interpretation is practical for personal tracking, but it is not exactly the PhenoAgeAccel used in scientific publications.
In the validation study, PhenoAgeAccel is the residual from a regression of PhenoAge on chronological age in a reference population. This adjustment is necessary because the raw delta tends to overestimate age in younger people and underestimate it in older people. Calculating a true residual therefore requires a cohort’s reference regression line, not just your ten numbers.
Positive acceleration then means that the profile is older than expected for someone of the same age in that reference population. Recent studies also find an association with mortality in people with diabetes or prediabetes, as well as in certain stages of cardiovascular-kidney-metabolic syndrome. These findings from specific clinical populations do not establish that five years of acceleration takes five years off an individual’s life.
Nor does the score tell you what to correct. Go back to the variables. High blood glucose, abnormal creatinine and persistently elevated CRP do not call for the same investigation. Compare each result with the laboratory’s reference range and your medical history before looking at its contribution to the score.
Interpretation mistakes
“+6 years = six years of life lost”: false. PhenoAge expresses a statistical risk in years, not a date of death. “The score is falling, so my protocol slows aging”: not demonstrated. A decrease may come from a temporary change in a marker. “All my results are within range, so the score is useless”: not so fast. The model combines continuous gradients, but it never replaces clinical thresholds. “I need to optimize every coefficient”: dangerous. The formula is linear and can reward extreme values that are not medically desirable.
PhenoAge, epigenetic age and biological age
The term PhenoAge refers to two related but different tools. The PhenoAge in this article is calculated directly from blood chemistry. DNAm PhenoAge is a DNA methylation clock trained to reproduce this phenotypic target using 513 CpG sites.
The useful differences fit into three lines:
| Measure | Input | What it best summarizes |
|---|---|---|
| Clinical PhenoAge | nine blood markers and age | physiological profile associated with risk in the original cohort |
| DNAm PhenoAge | blood DNA methylation | epigenetic signature trained on clinical PhenoAge |
| Clocks such as GrimAge or DunedinPACE | methylation and their own models | risk linked to certain exposures or pace of aging, depending on the algorithm |
An epigenetic clock is therefore not automatically a “truer” version. It measures a different layer, costs more and depends on the tissue, platform and computational processing. Clinical PhenoAge is more accessible, but more sensitive to your current state.
To compare these categories without diluting this tutorial, see the guide to methods for measuring biological age · FR. For methylation, the guide to biological age tests and epigenetic clocks · FR details the distinctions between age and pace.
Limitations and a reasonable measurement frequency
PhenoAge was developed in a US population using specific laboratory procedures. There is no universal French clinical threshold that turns, for example, +4 years into a disease or an indication for treatment. The score does not directly incorporate sex, blood pressure, physical fitness, medication or known diagnoses.
A systematic review published in 2026 evaluated 81 clocks based on routine clinical indicators. Most had a risk of bias or limitations in applicability, and many lacked external validation. The message is not to discard PhenoAge, which has substantial validation behind it. It is to refuse to treat a research formula as a clinical gold standard.
The field also continues to evolve. A study dated July 12, 2026 proposed an 11-marker biochemical clock, BBAC, then compared it with PhenoAge in UK Biobank and NHANES. Performance varied by cohort and statistical model. Several authors are employed by the company behind the tool, and a patent application has been filed. This finding supports the value of blood tests for estimating risk, not the definitive superiority of any one clock.
After an infection or acute stress
Do not calculate a baseline during an infection, just after surgery, during an inflammatory flare or while dehydrated. The formula specifically includes variables that respond to these states: CRP, white blood cells, lymphocytes, albumin and sometimes creatinine or glucose.
A rise after an infection can therefore be real for that blood sample, without representing a lasting acceleration of aging. Wait for clinical recovery and for results to normalize as agreed with the healthcare professional overseeing the tests. Elevated CRP or a kidney, liver or blood abnormality must never be ignored on the grounds that it “distorts” PhenoAge.
A cautious tracking protocol
To make two measurements more comparable, standardize the protocol:
- Have your blood drawn when your condition is stable, following your doctor’s and laboratory’s instructions
- Use the same laboratory if possible, with a morning blood draw and a comparable fasting period, as the validation study required at least eight hours
- Record recent infections, medication changes, unusual training sessions and significant weight changes
- Keep the same formula and conversions for every calculation
- Repeat after three to six months if a lasting change warrants follow-up, rather than every week
The three-to-six-month interval is a conservative practical choice, not a universally validated frequency. In the CALERIE trial, the BioAge package enabled comparisons of group trajectories at 12 and 24 months. This suggests possible sensitivity to change in research, but does not establish the smallest reliable change in an individual.
If the score changes, compare each marker first. A repeated trend under the same conditions is more informative than an isolated result. And if a result falls outside the reference range, symptoms appear or a treatment has changed, the priority is medical interpretation of the results, not waiting for the next PhenoAge calculation.
The Verdict
PhenoAge is probably one of the most accessible entry points into biological age, provided you treat it for what it is: a risk model built from a blood test, not a youth meter.
You can calculate it in France. Accuracy mainly comes down to CRP in mg/dL, the type of RDW, blood collection conditions and keeping the formula unchanged. The raw delta helps track a trend. The PhenoAgeAccel used in studies, however, requires a reference population.
I would use it to ask better questions about blood test results and observe a trajectory at spaced intervals. I would not use it to diagnose accelerated aging, buy an anti-aging protocol or convert a three-year decrease into an established clinical benefit.
