Longevity isn't about living forever

Public health keeps two counts. The first is how long people live. The second is healthspan: how long they live in good health, not just how long they last. For most of the last century the two rose together, and the first was a fair stand-in for the second.

The shortcut no longer works. Life expectancy kept climbing. Healthy life expectancy climbed more slowly. The years between them are years lived after the body has stopped fully cooperating. Worldwide that stretch now runs to roughly a decade. In many places it's getting wider, not closing.1

The gap is itself the finding. Medicine has been very good at pushing death back and much less good at pushing decline. Most of the years we added landed at the end.

That puts the target in question. A treatment that adds years without adding function hasn't delivered what people thought they were buying. It lengthens the dependent decade, not the capable ones. Nearly the whole longevity market is priced and measured on how long you last. The mistake runs the length of the product description.

The claim here's that lifespan is the wrong thing to measure. The right measure is whether you can keep taking part. Enough strength to carry your own shopping and get up off the floor unaided. Enough thinking to follow an argument, hold a view and change your mind. Enough slack that a bad week doesn't become a bad year. People who want you around. Something you're in the middle of. Judged that way, the variables with the most leverage are almost none of the ones for sale.

A phenotype, not a panel

In 2001, Linda Fried and colleagues did something plain and durable. They used data from more than 5,000 adults over 65, living at home rather than in care. Frailty, in older adults, is a loss of physical reserve. Fried's team argued it wasn't a vague clinical impression but a phenotype: a pattern you can see and score from outside the body. Theirs had five parts. Weight loss the person hadn't intended. Exhaustion they reported themselves. Weakness, measured as grip strength. Slow walking speed. Low physical activity. Three or more marked a person as frail, one or two as pre-frail.2

What makes the phenotype interesting is its crudeness. There's no blood panel and no lab test. There's a grip meter, a stopwatch, a set of scales and two questions. That unimpressive kit predicts falls, worsening movement, trouble with everyday tasks, hospital stays and death.2

Gerontology drew a lesson the wider longevity debate hasn't. Our best guide to how the last decades of a life will go measures what a body can do, not what it holds. Panels and scans have their uses, but they sit upstream of function and are often wrong. A person can produce a clean set of results and still be unable to rise from a chair unaided. The results are accurate. They answer a question about contents, not about what someone can still do.

The trouble with successful ageing

Four years before Fried's paper, John Rowe and Robert Kahn proposed a definition of successful ageing. It had three parts: a low chance of disease and the disability it brings, high physical and mental function, and active engagement with life.3 It was a correction. The field had been describing usual ageing as though decline were even and certain. Rowe and Kahn pointed instead at the huge differences between people born in the same year. Much of what was blamed on age, they argued, came from how people had lived.

The correction was right, and the framing has aged badly. Define successful ageing by the absence of disease and disability, and you place many people outside success by definition. Everyone with a long-term condition, an inherited risk, a workplace injury, or decades of physical labour. Critics have made this point for a quarter of a century. The model is individual in a way the evidence doesn't support. It gives too little weight to income, schooling, work, neighbourhood and care. And it quietly turns an outcome shaped by circumstance into a report card on personal conduct.

That move is worth naming. It's the default grammar of most longevity writing. Decline gets described as a thing a person allowed to happen. That story is comfortable for whoever tells it, unkind to whoever lives it, and poorly supported. A body asked to absorb more than it had to spend, for longer than it could manage, behaves as adaptive systems do under load. That's not a failure of character, and treating it as one has never improved anyone's path.

None of this means throwing the model out. It means taking active engagement with life more seriously than the other two parts. Almost nobody measures it.

Range goes before baseline

Something else runs underneath the frailty criteria. The reading that follows goes further than the studies it rests on.

The usual picture of the body is a set of values held near fixed points: temperature, acidity, blood sugar, pressure. That's accurate as far as it goes, and incomplete. Living systems hold steady through change, not in spite of it. They keep adjusting to meet expected demand, and the adjusting has a cost that adds up over years. McEwen called that build-up allostatic load.4 The contrast is between defending a set point and moving well across a wide range and coming back.

Read that way, ageing narrows the corridor and slows the return to it. Resting numbers stay normal longest, which is why they reassure and tell you so little. What goes first is the top end, the recovery, and the room for a knock once absorbed without comment. After a poor night's sleep the numbers may match in a younger body and an older one. The time taken to get back doesn't.

This is what reserve is, and reserve is what frailty measures. Each of Fried's five criteria measures margin rather than state. Force on hand above what's needed. Energy left at the end of a day. Frailty is a corridor narrowed to the width of ordinary life, with nothing spare. That's why frail people tend to be undone not by disasters but by minor infections, small falls and short hospital stays.

That reserve narrows with age isn't in dispute. Calling reserve the right unit of a long life, the thing to protect ahead of any blood marker, is our reading. It rests on work never built to test it.

The variables with the most leverage

If the target is taking part, the evidence points somewhere the market has almost no product line for.

One review pooled 148 studies covering more than 300,000 people, followed for 7.5 years on average. People with stronger ties to others had a 50 per cent greater chance of surviving the follow-up.5 A later review took in 70 studies and more than 3 million people. It treated loneliness, being cut off, and living alone separately. Each was tied to a rise in the chance of death of between a quarter and a third.6 Those effects are about the size of the death risks that carry their own public health campaigns. The authors draw that comparison themselves.

Purpose behaves similarly, and in one respect more strangely. Hill and Turiano followed more than 6,000 adults for 14 years. A stronger sense of purpose in life predicted a lower risk of death. The finding worth citing is that the link didn't vary by age. It predicted as much in people's twenties as in their seventies.7 Whatever is happening here, it doesn't belong to late life alone.

Steptoe, Deaton and Stone pull three of these apart. One is how a person judges their life from a distance. One is the moment-to-moment feel of a day. The third is the sense that a life has purpose and is worth its cost. All three relate to health and survival, and they don't move together.8 The third shows up most steadily in the death data, and wellbeing culture speaks to it least. Carol Ryff's model sets purpose beside personal growth, good relations and accepting yourself. Her point was that flourishing isn't the same as feeling good, and can't be measured as if it were.9

Frankl made the same claim from a vantage point no study can copy. The power to endure, he argued, is bound to having something worth enduring for. When the second goes, the first follows.10 That's testimony and argument rather than evidence. It lines up with the numbers in a way that's hard to ignore.

What strength is actually for

The case for strength is the most settled part of this, and it's usually made for the wrong reason.

Booth, Roberts and Laye argued that not moving is a primary cause of long-term disease rather than a side effect. They listed dozens of conditions where that case can be made.11 Warburton and Bredin reviewed the reviews. They confirm the breadth of the benefit and, more usefully, its shape. More activity brings more benefit, but the curve is steepest at the bottom. The largest returns come from moving from almost nothing to something modest.12

This belongs in a case about taking part, not one about disease. Physical capacity is the ground everything else runs on. Thinking narrows when you can't leave the house. Friendships thin out when getting to people becomes an event. Someone who can no longer rise unaided from a chair has lost a category of taking part, with the roles and habits inside it. That's how physical decline becomes social and personal decline. It runs the opposite way to the usual wellness framing.

What's actually being sold

Set that beside what the market sells and the mismatch is hard to miss.

Take fasting windows, one of the better-evidenced ideas in the category. De Cabo and Mattson's review lays out a real case for how it might work. It covers fuel switching, how cells answer stress, and changes that may bear on ageing.13 This is serious work. Its most dramatic lifespan results come from animals, and mice aren't small humans. The human trials are shorter and smaller, aimed at markers on a test rather than at how well anyone is doing at 80. The gap worth watching sits between what the research shows and what gets sold on it.

Then there's the most influential longevity story ever told. The Blue Zones made popular the idea that a few regions produce remarkable numbers of people over 100. Their food, movement and shared life were said to explain it.14 Some of the behavioural observations may well be sound. The arithmetic underneath isn't established. The extreme-age claims rest on birth and death records whose completeness and reliability have been seriously challenged. Checking ages in the very oldest groups is a known problem, unsolved and often abused. We raise the book because its grip on the public debate is unavoidable, not because it settles anything. Nothing here rests on it, and no argument about how to live should.

The honest accounting

Almost every finding in the middle of this piece comes from observational studies. Researchers watch what people already do and can't assign anyone to anything. That limits what the results can show, and the tangle is large.

People with strong ties and a clear sense of direction tend to have more money, more schooling and better work. They also tend to have better care and better health to begin with. Purpose is easier to hold when you're well. Being cut off is often a result of illness as much as a cause of it. Reverse causation is live throughout: the arrow may run the other way. Being healthier may be what makes a life feel worth the effort. Someone who has withdrawn and lost their direction may have a reason: a disease already under way that nobody will find for three years. Studies here adjust for what they can measure. What's left over, from everything they can't measure, pushes the numbers the way they already point. The wellbeing researchers say as much themselves.8

The argument survives this, for a structural reason rather than a rhetorical one. It doesn't depend on purpose causing survival. Suppose purpose and strong ties were purely markers of a life going well, not causes of it. The case for changing what we measure would be unchanged. It rests on the frailty and capacity work, which isn't fragile in the same way. A person whose friendships have thinned and whose direction has gone is in a state the data tie to a worse path. That holds whichever way the arrow runs.

What would settle it's a study nobody can run. You can't assign people to meaning at random.

Where the current metric leads

There's a version of the coming decades in which the gap keeps widening. The technology of postponement keeps improving, because it's well funded, easy to measure and simple to sell. The technology of taking part doesn't. Strength, friendship, purpose and the room to adapt are slow, unglamorous and largely unsellable. They're handed out by circumstances no one person arranges. The added years pile up at the dependent end. Everyone lives longer, and fewer people live well.

That outcome wouldn't register as failure by the measure now in use. Year after year it would register as progress. Which is the strongest argument for changing the measure. And it's an argument about what gets counted long before it's one about what anyone should do.