Biological age: what each test actually measures

Telomeres: a real measurement, but not a stopwatch

Telomeres are the ends of your chromosomes and they shorten as cells divide: across a population they are, on average, shorter in older people. The test measures their average length in the white cells of a blood sample.

Turning that average into your years is another matter. The mean shifts with whatever your immune system has been dealing with lately, and the lab techniques don't produce comparable values: the same test in two places can tell two different stories with neither getting it wrong. And it won't tell you which system is struggling, or what to do on Monday.

Epigenetic clocks: the most interesting models, and the easiest to misread

An epigenetic clock measures DNA methylation at specific sites and feeds those values into a formula. The first generation was trained to predict chronological age, so it is good at telling you the age you already know. Second generation clocks, PhenoAge and its relatives, were trained against clinical markers and health outcomes instead, which is why they add something: when they diverge from your real age, the gap points at something health related.

The limits are serious. There is technical noise, so the same sample processed twice won't return the same number. Passing things weigh on it, like a recent infection or the mix of cell types in the tube. Saliva and blood aren't interchangeable, and every brand runs its own clock, so your result can't be compared with another company's. Four questions before paying:

• Which clock exactly, and what it was trained against.
• Whether the result stays comparable when you repeat it.
• What its technical variability is, stated by the lab.
• What you get besides the number: the underlying markers, or only the headline.

Blood biomarker panels: less impressive, far more actionable

There is no new technology here, and that is the advantage. Glucose and HbA1c, a lipid profile including apoB, high sensitivity CRP, kidney and liver function, full blood count, ferritin, vitamin D and a hormone profile when the history calls for it. Standardised assays with known ranges, which repeat well and any doctor can read.

The composite indices that hand you 'an age' from a blood draw are built from these same variables: the number you're buying was already in your own data. Compressing ten markers into a year gained or lost throws away the part that tells you what to do. And the risk here isn't reliability, it's excess: a hundred markers in a healthy person return values outside range by chance alone.

Functional measures: what your body can do today, in front of you

VO2 max or an estimate of it, grip strength, sit to stand, walking speed, single leg balance. These aren't a model of your capacity, they are your capacity, measured. They repeat well when the protocol is respected, they cost little, and they respond to what you do within months.

They also have something no PDF has: you feel them, on the stairs and getting up off the floor. Their weak point is human, since they depend on how you slept and how much you feel like trying that day. Hence the same protocol, the same person, a similar time, compared against you.

How we use all this, and why the number comes last

The order matters: first the real reason you came, the history and the examination; then which measurements answer a question we have actually asked. A test that won't change anything we do doesn't get ordered. And if it does make sense, it comes with two conditions: the underlying markers are handed over, not just the headline, and a window is set to repeat it by the same method.

In between, few things change and they change in order: change five at once and the repeat test can't tell you which one worked. We're on Méndez Núñez, between Plaza Nueva and Sierpes, and that matters more than it sounds — a plan like this only holds if coming back fits into your week. And on the word most people type when they start looking, anti-ageing: we don't chase youth here. We work on what can be measured, reviewed and sustained.