Not smoking, and it is not close
Nothing else on this list comes near it. Stopping is worth doing at any age and the gains begin within weeks; stopping before middle age recovers most of the lost life expectancy, and stopping in the sixties still recovers a meaningful part of it. Anyone looking for a single thing to change has already found it, and anyone who does not smoke can skip to the next section having banked the largest available win.
The same applies with less force to vaping, where the long-term picture is genuinely not known yet, and where the honest position is that it is clearly less harmful than smoking and not established as harmless.
Moving, in the amounts that have been studied
The UK guidance for adults is at least 150 minutes a week of moderate activity — brisk walking, cycling, anything that raises the breathing rate — or 75 minutes of vigorous activity, or a mixture, plus muscle-strengthening work on at least two days.
Three things about that are worth knowing:
- The strength component is the one people skip, and it is the one that protects independence later. Muscle and bone are lost steadily from middle age and the loss is slowed by loading them, not by walking.
- Over 65, balance and flexibility join the list, on two days a week, because a fracture from a fall is one of the sharpest turning points in an older person's health and the risk responds to training.
- The largest benefit is at the bottom of the range. Going from nothing to a little is worth far more than going from a lot to slightly more, which is the opposite of how exercise advice is usually pitched.
Long uninterrupted sitting carries its own risk, separate from how much exercise is done elsewhere in the day, so breaking it up with standing or a short walk is worth doing in its own right.
The two numbers with no symptoms
Raised blood pressure and raised cholesterol do not announce themselves. They are among the largest modifiable contributors to heart attack, stroke and vascular dementia, they are measurable in minutes, and they are treatable — which together make them the best return available from any single appointment.
Older writing on longevity often recommends an annual physical examination. That is not how it works here: the NHS Health Check is offered every five years to adults aged 40 to 74, covers blood pressure, cholesterol, weight and diabetes risk, and is free. Taking it up when the letter arrives is the whole action required. Private annual full-body screening packages, by contrast, generate a large number of findings of no consequence, and the resulting investigations carry their own risks.
Alcohol
UK guidance is no more than 14 units a week for both men and women, spread across three days or more rather than concentrated, with some alcohol-free days. There is no identified level below which it is clearly protective; the older claim that moderate drinking extends life has weakened considerably as studies have better accounted for the fact that some non-drinkers have stopped because they were already unwell.
Food, without the mysticism
The pattern that survives in the evidence is unglamorous: plenty of vegetables, fruit, pulses and whole grains, fish, less red and processed meat, less salt, and not much in the way of sugary drinks. Variety across a week does more than perfecting any one item.
Two things commonly said in this context are not right. Food does not deposit "toxins" that a diet then removes — the liver and kidneys do that continuously, and no eating pattern improves on them. And industrial trans fats, which older advice warns about at length, have been all but removed from the British food supply, so the warning is now about a problem that was largely solved rather than one to watch for.
On weight: activity and diet quality shift health outcomes more reliably than the number on the scales does, and for someone inactive and eating poorly at any weight, changing those two is the useful move. Where weight is high enough to be driving blood pressure, blood sugar or joint damage, it is worth addressing with a GP rather than alone, because the approaches that work are not the ones marketed as rapid.
Sleep
Persistently short sleep is associated with worse cardiovascular and metabolic health, and so is persistently long sleep, which is usually a signal of something else. What matters most is treating the fixable causes rather than chasing a number, and the commonest fixable cause is undiagnosed sleep apnoea, which is set out under sleep apnoea.
Hearing, sight and the dementia list
This is the least known item here and one of the more useful. The 2024 Lancet Commission on dementia identified fourteen modifiable risk factors associated with up to around half of cases worldwide, and two of them are untreated hearing loss and untreated vision loss.
The full list: hearing loss, untreated vision loss, high LDL cholesterol, raised blood pressure, obesity, diabetes, physical inactivity, smoking, excessive alcohol, depression, social isolation, head injury, air pollution, and low educational attainment in early life.
Most of that list is the same list as everything above, which is the point — the things that protect the heart largely protect the brain. The additions worth acting on are getting hearing tested rather than adapting to it over years, keeping sight corrected and attending eye appointments, and treating head protection in cycling and similar activities as a long-term matter rather than a short-term one.
Other people
Social isolation appears on that list, and on the mortality evidence more broadly, at an effect size that surprises people who assume it is a soft factor. It is also the hardest item to act on by instruction, since company cannot be prescribed. What can be said usefully is that it is worth treating contact as a thing to maintain deliberately rather than something that will take care of itself, particularly after a retirement, a bereavement or a move, which are the points at which a social circle quietly contracts.
The dull interventions that work
Vaccination and screening are the least interesting items in the whole subject and among the most effective. Flu each autumn, and the pneumococcal, shingles and RSV vaccines at the ages they are offered, prevent illnesses that are genuinely dangerous in later life. The national screening programmes — bowel, breast, cervical, and the abdominal aortic aneurysm scan for men at 65 — exist because they find things early enough to treat. Uptake is far from complete, and the invitation arriving is not the same as it being used.
Things that get recommended and do not help
Drinking extra water for the skin. Water is necessary and thirst manages it adequately for most people. Drinking more than needed does not plump or moisturise skin, which is affected by the barrier on the outside rather than by fluid volume on the inside, and there is no "natural curative" effect to be had.
Brain-training and "neurobics". Puzzles, apps and tricks such as writing with the other hand produce improvement at the specific task practised and do not transfer to general cognition or reduce dementia risk. Education, hearing, blood pressure, activity and company are the levers that appear in the evidence; crosswords are a pleasant way to spend an hour and should be enjoyed as one.
Detoxes, cleanses and antioxidant supplements. No benefit on mortality, and in some cases harm — high-dose beta-carotene increased lung cancer in trials of heavy smokers. More on this under vitamins and minerals.
Most things sold on the promise of longevity. The field attracts an unusual volume of supplements, tests and protocols with no outcome data behind them. The interventions with decades of mortality evidence are the ones above, and all of them are either free or cheap.
What centenarians actually tell us
Interviews with very old people make entertaining reading and poor evidence. Whatever a hundred-year-old credits — a daily whisky, hard work, never marrying — is a single case recalled late, and the people who did the same things and died at seventy are not available for interview. Reaching extreme old age involves a substantial element of luck, including genetic luck, and that part is not actionable.
What is actionable is the difference between an unhealthy sixty-year-old and a healthy one, and that difference is made almost entirely by the unspectacular list above. The realistic goal is not an extra decade of life so much as more of the existing decades spent in good health, which is both more achievable and rather more to the point.