Organ age tests: what the result really means

You open a report and see something like: “brain, 62; heart, 49; kidney, 57”. It is hard not to read it as a school report for the body. But an organ age test does not look directly at your organs or count how many years they have left. It compares proteins in your blood with a statistical model.

The underlying science is real and promising. In large cohorts, some proteomic age gaps are associated with future disease and mortality. The problematic leap comes next: a population association does not turn an “older heart” into a diagnosis or tell you which supplement, diet or treatment will rejuvenate it.

Short answer in August 2026: these clocks are useful for studying how different systems age and may help enrich clinical trials. Their individual clinical utility has not yet been demonstrated. Before buying one, ask what it adds to your history, function and established tests, and which decision will change.

How an age comes out of a blood sample

Proteomics measures hundreds or thousands of plasma proteins at once. Researchers identify proteins enriched in or informative about a tissue, then train an algorithm to predict chronological age. They compare the predicted age with what is expected for someone of the same age.

If you are 50 and the model predicts 58 for the cardiac pattern, the gap is positive. That means “more similar to the average at an older age within this assay”, not “your heart has lived eight extra years”. The result depends on four components: the proteins measured, the rule assigning them to an organ, the reference population and the algorithm.

The 2023 Nature study measured 4,979 proteins in 5,676 people across five cohorts and built models for 11 organs. It called a protein “organ enriched” when the corresponding gene was expressed at least four times more strongly in that tissue than elsewhere. This reasonably improves interpretability, but it does not prove that every circulating protein came exclusively from that organ.

Three evidence levels that should not be blurred

Level	Question	Current position
1. Recognises age	Does the pattern predict chronological age in another sample?	Yes, several models do this accurately
2. Associates with risk	Does an unfavourable gap precede more disease or mortality?	Yes, prospective associations exist in large cohorts
3. Improves decisions	Does it add enough information, change treatment and improve health?	Not yet demonstrated for routine individual use

A commercial test can show evidence from levels one and two and let the reader imagine level three. That is where we need to pause. Predicting age and associating with risk are scientific achievements. Clinical utility means showing that knowing the score leads to a better decision than not knowing it.

What large cohorts have found

In 2025, a Nature Medicine study analysed 2,916 proteins in 44,498 UK Biobank participants. It estimated age for 11 organs and found associations with future heart failure, COPD, type 2 diabetes, Alzheimer's disease and mortality. Extreme brain and immune profiles produced some of the strongest signals.

For example, an especially aged brain was associated with a hazard ratio of 3.1 for Alzheimer's disease, while an especially youthful brain was associated with 0.26. These figures describe extreme groups inside a model during follow-up. They are not personal probabilities, do not diagnose Alzheimer's and do not imply that changing one protein will change the outcome.

Another study developed clocks in 43,616 UK participants and validated them in 3,977 people in China and 800 US women. That transportability is a strength, although an all-female US cohort does not represent the whole population and every platform still needs calibration.

In 2026, EPIC added 17,473 European participants and up to 28 years of follow-up. Age gaps were associated with lifestyle, mortality and several diseases; predictive performance was comparable to classical lifestyle factors. “Comparable” does not mean the clocks replace those factors or improve a decision once those factors are already measured.

The score can change without an organ becoming younger

Proteins vary with disease, inflammation, kidney clearance, medication, menopause, smoking, activity and other processes. Laboratory variation also exists. The model sees the final pattern but cannot automatically separate cause, consequence and noise.

The UK Biobank study repeated samples in 1,176 people across several visits. Gaps had relative stability, yet 68% of those in an extreme category at baseline no longer met it later. That does not make the measure useless. It means an extreme label is a photograph with uncertainty, not a permanent biological identity.

There is therefore no validated retesting schedule or universal target to remove a certain number of years. If a post-intervention change has not been shown to predict less disease, it may be an interesting research signal but not a substitute for actual health.

What to do if an organ looks “older”

First, breathe. Then translate the signal into clinical questions with established tests and decisions.

Report signal	Useful confirmation	What to avoid
Heart or arteries	Symptoms, blood pressure, ApoB/Lp(a), glucose, smoking and targeted testing if indicated	Starting medicine or imaging from organ age alone
Brain	Noticed change, function, sleep, hearing and vascular risk	Reading the gap as an Alzheimer's diagnosis
Kidney	Creatinine/eGFR, urine albumin, blood pressure and medication	Concluding kidney damage without clinical testing
Liver	Alcohol, medicines, enzymes, platelets and ultrasound or elastography when appropriate	Buying a detox to lower the score
Muscle	Strength, chair rise, gait and mass in context	Confusing one protein with a diagnosis of sarcopenia

The foundation remains a proportionate set of clinical biomarkers that change decisions, together with symptoms and function. A proteomic clock may add a hypothesis, but it should not displace a confirmable abnormality.

Proteomics, epigenetics, imaging and function do not compete

An epigenetic clock measures DNA methylation; proteomics, circulating proteins; imaging, structure; and functional tests, what you can do. Each layer answers a different question. The guide to choosing a biological age test compares the broader map.

The mistake is looking for one universal winner. A useful tool answers a defined question, has known error, is validated in people like you and leads to a proportionate decision. Sometimes that will be a simple biomarker or strength test, not the newest panel.

Before buying: six questions

1. Which platform and version does it use? SomaScan, Olink and other assays are not interchangeable.
2. Who was it trained and validated in? Ask about age, sex, ancestry, health state and external validation.
3. What is the error? Reporting whole years may suggest precision the method does not have.
4. What does it add to my clinical data? Novelty is not enough if blood pressure, ApoB, kidney function or strength already resolves the decision.
5. How will an unfavourable result be confirmed? Name the owner, standard test and timing before measuring.
6. What would we do if it changes on retesting? Without an answer in advance, retesting may become a subscription to anxiety.

Risks and red flags

The blood draw usually causes no more than pain, bruising or dizziness. More likely harm comes from interpretation: anxiety, false reassurance, testing cascades, unindicated supplements, repeat expense and exposure of omics data. Ask how the sample, result and algorithmic data are stored.

Chest pain or pressure, severe breathlessness, fainting, one-sided weakness, trouble speaking or sudden confusion requires urgent care. Jaundice, visible blood in urine or stool, markedly reduced urination, rapidly increasing swelling, unintended weight loss or progressive cognitive decline needs prompt assessment. An organ dashboard cannot rule out these causes.

Frequently asked questions

Can a blood test measure the age of each organ?

It can estimate relative age from plasma proteins and a model trained on a reference population. The result is a statistical score, not the organ's literal anatomical age or a direct measure of how long it will keep functioning.

What does it mean if my heart is ten years older?

It means the protein pattern resembles the model average for an older age. It does not prove heart disease or ten years of damage. Translate it into symptoms, blood pressure, lipids and indicated clinical tests.

Which organs can these clocks estimate?

Studies have modelled the brain, heart, arteries, lung, liver, kidney, muscle, pancreas, intestine, adipose tissue and immune system. The set changes with the platform, available proteins and algorithm.

Is a proteomic organ age test better than an epigenetic clock?

There is no universal winner. Proteomics, DNA methylation, clinical tests, function and imaging capture different layers. Proteomics may be easier to interpret by organ, but no clock replaces clinical risks that already change decisions.

Can an organ age test diagnose disease?

Not on its own. An unfavourable gap is associated with higher risk in cohorts, but it does not identify a specific disease, its cause or whether it will occur. Any signal needs confirmation through history, examination and standard tests.

How often should I repeat the test?

No clinical interval has been validated. If repeated, use the same assay in a comparable clinical state and define the decision in advance. Change may reflect biology, medication, illness or technical variation.

Can I rejuvenate one organ?

Lowering a proteomic age score has not been shown to cause less disease or longer life. Real risks such as hypertension, high ApoB, diabetes, smoking, inactivity or kidney disease can be treated with interventions of established benefit.

When might an organ age test be useful?

Today it fits best in research, trials or carefully explained exploratory assessment. Routine care should add it only if it is validated for your population, has a confirmation pathway and could change a concrete decision.

Sources and certainty

1. Oh et al., Nature, 2023: 5,676 adults, 11 organs, SomaScan method, disease and mortality.
2. Oh et al., Nature Medicine, 2025: 44,498 participants, Olink, prospective risk and longitudinal stability.
3. Wang et al., Nature Aging, 2025: UK development and external validation in China and the United States.
4. Robinson et al., Nature Aging, 2026: 17,473 Europeans, lifestyle, 24 diseases and mortality.
5. Li et al., Ageing Research Reviews, 2026: translational framework, calibration, transportability and incremental utility.

Certainty is moderate to high that these clocks capture population differences associated with risk, and still insufficient for diagnosing an individual, prescribing from the score or treating a change as proof of rejuvenation.

A good test does not hand you an age to chase. It helps you ask a better question and accepts that, for now, many useful answers still come from much simpler measurements.