How much of ageing is actually modifiable?

Having a sibling who lived past 100 sharply lifts your own odds of extreme old age. Yet most of the variation in ordinary lifespan is not inherited at all. Both statements hold up, and the tension between them is where honest talk about ageing starts.

By the end of this article you will know how large the modifiable share honestly looks, why healthspan is the better question, and why acting on it needs no stronger evidence than we already have.

What does heritability actually tell you?

Studies of twins and family trees put the heritability of human lifespan at roughly a quarter. That means a quarter of the variation in how long people live tracks with inherited differences. The figure gets quoted as a fact about a person, and it is not one. Heritability describes a population at a moment in time. Where half the adults smoke, genes look less important. Where nobody smokes, the same genes explain more of what is left.

The figure is also contested downward. Very large family-tree analyses argue that much of the family resemblance in lifespan comes from who pairs with whom. People choose partners who resemble them in schooling, habits and circumstance, which makes relatives look more alike than genes explain. Corrected for that, some estimates fall to single digits.

Now look at the far tail, where the picture flips. Ordinary lifespan is modestly heritable, while exceptional longevity looks considerably more inherited. Reaching 100 may be substantially a genetic lottery. Dying at 62 of something modifiable is not.

Which hints that lifespan is the wrong quantity to argue about.

Why is healthspan the better question?

Most people carry some version of this claim: ageing is eighty per cent lifestyle and twenty per cent genes, so most of it is in your hands. The arithmetic comes from a heritability estimate, and it breaks in one move. What is not genetic is not therefore modifiable.

The non-inherited share holds behaviour, but it also holds circumstance and plain chance. Which infections you met. Which cells picked up which mutations. Which small blood vessel closed on which morning. Twin studies keep finding that the largest single slice is unshared and unexplained.

So separate two quantities. Lifespan is how long someone lives. Healthspan is how long they live without major chronic disease, or the loss of function that ends independence. The modifiable share of healthspan is larger than that of lifespan. That is both the honest claim and the more useful one.

The reason is structural. The outer boundary of human life is stubborn and has defeated everything tried so far. The age at which a first chronic condition arrives, by contrast, moves with decades of exposure, activity and metabolic load.

Compressing the sick years is the more workable target. How strong is that evidence?

What can the big studies prove, and what can they not?

Large long-term studies, following hundreds of thousands of people for decades, converge on one picture. A small cluster of ordinary behaviours travels with a substantial extension of disease-free years. Not smoking. Regular activity. Body composition away from extremes. Decent diet quality. Alcohol kept low. The reported gap runs from a few years to more than a decade. The direction is consistent; the exact size is not, and quoting one figure as fact would overstate it.

Three things routinely make findings like these look bigger than they are. The first, and the one that hurts the case for activity most, is reverse causation. Illness reduces movement long before diagnosis, sometimes by years. So inactivity before a diagnosis can be an early symptom wearing the costume of a cause. Researchers discard the first years of follow-up, which helps but does not settle it.

The other two are constant company. Confounding: these behaviours cluster with education, income and access to care, and no statistical adjustment separates them fully. And the healthy-user effect: people who adopt one health behaviour adopt others. The sharpest demonstration is that people who take their placebo faithfully do better than those who do not. The cohort evidence supports several extra years of healthy life from a handful of ordinary behaviours, and cannot on its own prove those behaviours produced them.

Which is why the trial evidence matters, and why its current state deserves stating plainly.

Has any drug been shown to extend a human life?

No. That sentence carries no asterisk. Several candidates are in genuine clinical study, which belongs in the same breath. None has yet cleared the bar.

The animal-to-human gap is wider than headlines admit. Extending lifespan in yeast, worms and flies is close to routine, because short-lived organisms are unusually easy to shift. In mice the effects shrink, and depend on strain, sex and timing. The two long monkey studies of calorie restriction disagreed, largely over what the comparison animals were fed. No drug has been shown to extend human lifespan, and the distance between an animal result and a human one widens at every step up the ladder.

The honest position is unglamorous. The best-supported levers are behavioural, and the drug questions remain open, actively studied frontiers. Which raises the obvious objection. If the evidence is this incomplete, why not wait?

Why is acting under uncertainty still sensible?

Because a decision under uncertainty turns on what happens if the claim proves false. Take a compound purely for its longevity promise. If the promise collapses, you hold the cost, the risk, and nothing else. Every point of doubt counts against it, correctly, which is why the bar for drugs stays high.

The behavioural levers sit in a different category. Strength work, sleep, aerobic activity, food quality and not smoking carry strong evidence for outcomes that owe nothing to longevity. Heart health, glucose handling, bone density, mood, and rising from the floor at eighty. The behavioural levers are justified by evidence that does not depend on the longevity claim at all. Acting under uncertainty becomes ordinary good judgement rather than faith.

Which gives you a test for every longevity claim. Strip the promise out and ask what is left. If something solid remains, the uncertainty is survivable. If nothing remains, the promise is carrying the whole weight, and it cannot.

Knowing the size of the modifiable share keeps this field clear of both fatalism and easy promises. The Longevity knowledge check will show how steadily you hold that line.