Guide

Longevity biomarkers: which tests serve, and how to read them

Marker by marker: what it predicts, what it does not say, and the moment to hand over.

12 analyses · updated 2026-10-30

Four elements carry most of the decision in prevention: HbA1c, lipids read against global risk, kidney function and blood pressure. Trendy markers arrive far behind, and several of them vary more than the difference one seeks to measure. A test is worth only what it predicts, and the decision it changes.

This guide gathers the analyses the School devotes to biomarkers and prevention. Each answers a real consultation scene: a person arrives with a result, and one must know what can be said about it.

The foundation fits into four elements, and it is remeasured

A prevention panel starts with a standardized cardiometabolic foundation, and its value comes less from the first number than from remeasurement under comparable conditions. Same laboratory, same time of day, same fasting state. Without that, the difference between two panels mixes the person's evolution with the variability of the method.

The rest of the work consists in classifying each value and setting a remeasurement date. That is where the difference lies between a panel that serves and a list of numbers commented once and then forgotten.

the useful markers of a prevention panel and their limits

Fasting glucose is the last marker to move

Normal fasting glucose does not say that a metabolism is well. Insulin resistance sets in years before sugar rises, and this window is what matters in coaching. The same number does not carry the same weight depending on the context in which it is read.

Orders of magnitude are known: prediabetes multiplies risk, and lifestyle does better than metformin in the trial that compared the two. Exercise acts on the marker without going through the scale, which changes the conversation to have with a person whose weight does not move.

what the number announces, and what makes it recede

High ferritin is almost never iron overload

The scene repeats every week: a ferritin above the reference range, someone has said the word hemochromatosis, sometimes the word bloodletting. In the vast majority of cases, iron is not the cause. Ferritin also rises with inflammation, alcohol and fatty liver.

One other test settles the direction of the workup, transferrin saturation. The most frequent case in practice is metabolic, not genetic. Hemochromatosis exists and is inherited, and its investigation belongs to the physician.

why it is almost never iron overload

A slightly high TSH does not explain fatigue

TSH is the most requested test in the face of persistent fatigue, and the box most often ticked to explain what is wrong. Two large trials tested the idea directly: correcting a slightly elevated TSH improves neither fatigue nor symptoms. Twenty-one trials reach the same result.

There is something more troubling for the shortcut: in about a third of cases, the number comes back down on its own with no intervention. Fatigue therefore remains a poor entry point to the thyroid, and interpretation of the test is the physician's call.

what a slightly high TSH changes, and what it does not change

Salivary cortisol does not support adrenal fatigue

A systematic review of 58 studies looked at whether cortisol measurements supported the concept of adrenal fatigue. The answer is no, and the reason lies as much in the measurement method as in the hypothesis. Variability from one sample to another often exceeds the difference being sought.

Capillary cortisol, presented as the solution to that problem, does not solve it. Cortisol excess exists in pathology, and it has another face than chronic fatigue. What remains to offer an exhausted person is documented, and does not go through a salivary kit.

what cortisol measurements really measure

Vitamin D: the error runs both ways

Twenty-five thousand participants followed for five years, and no reduction in fractures in the general population. The result surprised, and it does not allow the opposite conclusion: the effect exists elsewhere, at a moderate daily dose and in people who are deficient. On falls, the dose and the rhythm of intake change the result.

The 2024 recommendation moves the question from dose to age, which modifies management in routine practice. The frequent error runs both ways: supplementing without deficiency, or leaving an established deficiency unaddressed.

what the large trials allow us to assert

Supplements are classified by the outcome measured in trials

A supplement is not judged on its mechanism but on what trials measured. Vitamin D shows a documented benefit only in deficiency. Antioxidants provide no benefit in mortality prevention. NAD+ related molecules are still evaluated on surrogate markers.

Three constraints bound any supplementation: overdose, interactions with ongoing treatment, and opportunity cost. Targeted supplementation is documented, dosed and reassessed at three months, or it serves no purpose.

what has evidence and what does not

A biological age is worth only its predictive value

The term covers methods with no relation to each other. The most useful comparison involved 9,389 people from the NHANES III survey followed for eighteen years, with 1,843 deaths: one method came out ahead, and when it is placed in the same model as chronological age, chronological age ceases to be associated with mortality. That is the only argument that counts.

Measurement noise, for its part, often exceeds the difference being sought. A paper published in 2022 ran six clocks on duplicated samples: the same sample, analyzed twice, produced differences of up to nine years. Two free measures already predict a great deal, walking speed and grip strength.

what the measures predict, and what they do not say

Epigenetic clocks remain a research tool

Clocks read DNA methylation at thousands of sites and translate it into a number. DunedinPACE does not give an age but a rate of aging, which is another question. The result reads as a reflection of trajectory, not as an individual verdict.

Two technical limits are enough to explain the caution: methylation varies with the tissue sampled and with the analysis platform. Consumer offers remain outside the frame of a prescription, and the useful consultation is the one where the person arrives already with their result.

their real usefulness and their limits in practice

Before any test, validated and free instruments

Most first consultations begin with a home made questionnaire. Yet validated, free instruments exist, whose sensitivity, specificity and gap with objective measurement are known. One still needs to know which rules out a disorder and which quantifies nothing.

Two gaps are systematic and useful to know: declared activity is not activity done, and five questions are enough to spot muscle frailty. A questionnaire serves to rule out, not to quantify, and the interview adds what the form does not capture.

which validated tools, and what they really measure

Knowing when to hand over, and when not to

Every non-physician practitioner knows the instruction: refer when there is a warning signal. Nobody ever says which ones, or what they are worth. The data exist and surprise in both directions. Several red flags taught everywhere have no validation, and unintentional weight loss deserves the detour.

Over-referral also costs something, in trust and in access to care. Before intense effort, the main risk is even of referring too much. Referring is a technical decision, not a confession, and it holds in three steps.

which signals require handing over

What each marker predicts and the decision it allows, or does not allow.
MarkerWhat it predictsDecision it changes
HbA1cMetabolic riskYes, after confirmation
LipidsGlobal cardiovascular riskYes, read with the rest of the profile
Blood pressureMild cognitive impairment, eventsYes, best demonstrated lever
Fasting glucose aloneLittle, and lateNo, it moves last
FerritinInflammation more often than ironNo without transferrin saturation
Slightly high TSHNot fatigueNo, a third normalize on their own
Salivary cortisolNothing validated on fatigueNo
25-OH vitamin DDeficiency, in those likely to have itYes in case of deficiency
Epigenetic clocksPopulation trajectoryNo individually, up to 9 years of noise
Walking speed, grip strengthMortality, independenceYes, and free

What each marker predicts and the decision it allows, or does not allow.

The analyses in this guide

Questions

Which tests should be requested for a longevity panel?

Prescribing is the physician's domain. The foundation that carries most of the decision is known: HbA1c, a lipid panel read with global risk, kidney function and blood pressure. What gives follow-up its value is not the length of the list, it is remeasurement under comparable conditions.

Are biological age tests sold online reliable?

Not at the individual level. A 2022 paper ran six epigenetic clocks on duplicated samples: the same sample analyzed twice produced differences of up to nine years, more than the difference a practitioner hopes to observe after a year of coaching.

Can a non-physician practitioner interpret a blood panel?

No. Interpretation of a panel, diagnosis and prescription are the physician's domain. A practitioner can explain what a marker measures, recall what trials show, work the lifestyle levers and refer when a result falls outside the frame.

Does adrenal fatigue exist?

A systematic review of 58 studies concluded that cortisol measurements do not support this concept. Measurement variability often exceeds the difference being sought, and capillary cortisol does not solve that problem. Pathological cortisol excess exists, with a different presentation.

Which free marker is worth following?

Two: walking speed and grip strength. They predict mortality and independence, they cost nothing, and they are remeasured under the same conditions without depending on a laboratory or an analysis platform.