Is Longevity Science Portable?

Some months ago I gave a health talk in Tamil to the Japan Tamil community in Tokyo, part of a series named for a phrase every Tamil household knows: வருமுன் காப்போம். Prevent before it arrives. I opened with a single plot. Functional capacity on one axis, age on the other. For the first two or three decades of life the curve climbs. Somewhere around thirty it flattens. After that, it declines. Nearly everything we argue about in longevity, I told the room, is an argument about the shape of that decline.

Steeper is worse. A body that loses capacity quickly spends fewer years above the line that matters: the line below which a person can no longer live independently. So the goal is not to abolish decline; nothing known does that. The goal is to buy years above it, by making the slope gentler and by starting the descent from a higher peak.

Two refinements keep the picture honest. First, a steep drop at the very end of a long, high plateau is not a failure. It is close to the ideal. Fries named it the compression of morbidity in 1980: extend vigor deep into the lifespan, and the sick years get squeezed into a short window at the end. What we are trying to avoid is the other shape, the early, shallow slide that spends decades below the line. Second, and more important for everything that follows, the slope is not a single parameter waiting to be optimized. The body is a portfolio of systems. Muscle strength, aerobic capacity, kidney function, hearing, vision, immune competence and processing speed each decline on their own schedule, and each bends to different interventions. “Slowing aging” is shorthand for bending a bundle of slopes at once, and any specific claim to bend one of them deserves its own audit.

I closed that talk by telling the room that most of them could see a hundred if risk was managed early and well. Then, recently, I spent two days at RISE for Healthy Aging at the Indian Institute of Science in Bengaluru: nearly 800 participants, close to 60 speakers, and sessions running from the cell biology of aging to Indian cohort studies, diagnostics and dementia. The field is taking shape in India, and quickly. I went in interested in the science. I came away holding a question the curve alone cannot answer.

India is beginning to build its own science of aging. Suppose we succeed. Suppose we build biological-age models that actually work in Indians, understand our metabolic differences, and generate excellent Indian multiomic datasets. What exactly are we then entitled to import from the global longevity movement?

My own work sits around computational biology and AI, so I tend to see this as a translation problem across a system. A biological mechanism can be universal while the measurement used to capture it is population-dependent. A measurement can be valid while the intervention derived from it is not. An intervention can work biologically while being impossible to sustain inside a particular household. And a biomarker can improve without anyone yet knowing whether the person will stay healthier for longer.

Epidemiology has a formal name for this family of problems: transportability, the question of whether an effect demonstrated in one population can be expected in another whose biology, baseline risk or exposures differ. I find it useful to break any longevity claim into five steps.

Each arrow is an assumption. And India is an unusually good place to test where the assumptions hold.

Start with what probably does travel

Indians do not age according to a different set of cellular laws. The processes geroscience studies are not American phenomena: cellular senescence, altered nutrient sensing, mitochondrial dysfunction, chronic inflammation, the loss of protein homeostasis.

Nor should an Indian discussion of longevity become an excuse to rediscover what medicine already knows. Tobacco is harmful. Hypertension matters. Diabetes matters. Physical inactivity matters.

So the argument here is not that Western longevity science fails in India. That would be both scientifically weak and demonstrably wrong. The better question is narrower: when a result moves from one population into another, how much of it survives the journey?

India is already solving the measurement problem

This part is not new. The BHARAT study at IISc starts from exactly this observation: most biological-age models were built in Western populations and may have limited applicability to Indian demographics. It is integrating multiomic, biochemical, clinical and lifestyle data to develop aging signatures for Indian populations; its framework paper appeared in April 2026.

Genomics tells the same story. GenomeIndia has reported whole genomes from 10,000 healthy, unrelated Indians across 83 population groups: a reminder of how much diversity disappears when more than a billion people are filed under a single word, Indian. And we already know why representation matters: polygenic risk scores trained largely on European-ancestry data have repeatedly lost accuracy when transferred to other ancestries.

So building Indian datasets, reference ranges and clocks is clearly necessary. The question I find more interesting comes after.

Suppose we build an excellent Indian biological-age clock. What intervention does the number justify?

A correct measurement does not automatically produce a correct prescription. Even the Biomarkers of Aging Consortium, the group working hardest to push these tools toward the clinic, reports that clinical translation remains limited, and that the field still needs to connect its biomarkers to actionable decisions and validate that they respond reliably in individuals. A clock can be scientifically useful long before it is medicine. That distinction matters.

The same BMI can describe different biology

Consider something far simpler than an aging clock: body weight. South Asians tend to develop type 2 diabetes younger, and at lower BMI, than white populations. Reviews of the underlying phenotype describe lower lean muscle mass alongside greater abdominal and ectopic fat, including liver fat, at the same BMI.

This does not make BMI useless. It makes BMI incomplete.

Now apply the standard intervention: lose weight. For someone carrying substantial excess adipose tissue, that can be excellent advice. For someone already light, with little muscle reserve, making the weighing scale go down may be the wrong objective entirely. A better target may be to reduce excess fat while maintaining or even building muscle and strength.

Same mechanism. Same measurement. Different starting phenotype, so a different intervention. That is transportability failing quietly, two steps downstream of the biology.

Muscle may be India’s most underappreciated longevity variable

The longevity world is drawn to sophisticated measurements: epigenetic clocks, multiomic panels, continuous monitoring, new molecules targeting aging pathways. Muscle is harder to market. It may be far more important to preserve.

South Asian researchers have already concluded that imported definitions of sarcopenia do not completely fit the region, because body composition and function differ, and that excess fat and inadequate muscle often coexist in the same person. Muscle is not only what lets you lift things. It is a major site of glucose disposal, and it underwrites balance, mobility and independence.

So perhaps the highest-return question for an Indian in middle age is not what is my biological age? but:

How much useful muscle will I still have at 70?

That changes what we measure. Can you rise from a chair without using your hands? Climb stairs comfortably? Carry something heavy? How fast do you walk, and are you getting stronger or weaker over time? These are unsophisticated measurements. That does not make them unimportant. A programme that improves a panel of molecular markers while the person grows physically weaker has optimized the wrong endpoint.

Longevity may begin long before middle age

India complicates one more assumption baked into longevity culture: that the story begins when a 40-year-old decides to optimize himself.

In the classic Pune Maternal Nutrition Study, 631 term babies born in rural Maharashtra were compared with 338 babies born in Southampton. The Indian babies were smaller overall, particularly in measures reflecting muscle and abdominal viscera, while skinfolds were relatively preserved. The researchers called this the thin-fat Indian baby and proposed that it could contribute to later insulin resistance.

This should not be turned into an ethnic law; later Indian newborn studies using more direct body-composition methods have not reproduced the phenotype uniformly. But the broader lesson survives. Adult metabolic health is partly a life-course phenomenon:

So an Indian longevity strategy cannot begin and end with today’s affluent 50-year-old. Part of the next generation’s healthspan is being built now.

Then biology runs into the household

Suppose we get the biology exactly right. Someone needs more muscle, less visceral fat, a better diet. Now a different question: can they actually do it?

Most dietary advice assumes a person constructs every meal individually: more protein, fewer refined carbohydrates, more vegetables, different timing. But in many Indian homes, food is a household system. What someone eats depends on what is bought, what is cooked, who cooks it, what parents and children will accept, whether the family is vegetarian, religious practice, festivals, hospitality, and habits that predate everyone at the table.

Meal timing makes the collision concrete. Controlled studies of early time-restricted feeding suggest that when we eat can influence metabolic physiology, independent of weight loss. In one small crossover trial, men with prediabetes who restricted eating to an early six-hour window (last meal before 3 pm) improved insulin sensitivity and blood pressure within five weeks. That is scientifically interesting. It is also an intervention designed around a schedule many working Indian households cannot reproduce: the office closes late, the commute is long, the family eats when the last person is home. Dinner at 9:30, bed by 11. Whether smaller, livable shifts (an earlier or lighter dinner, a longer overnight gap) deliver benefit on that schedule has not been shown by this trial; it is exactly the kind of transportability question Indian studies should answer. India even has culturally embedded early-dinner traditions, like the Jain practice of finishing the last meal before sunset.

Portion size creates a similar implementation problem. In much of South India, polished white rice is the staple around which the meal is organized, and refilled. Meta-analysis associates higher white-rice consumption with higher type 2 diabetes risk, particularly in Asian populations, though observational data cannot establish causation. Reducing the glycaemic load of that plate is biologically plausible. Sustaining the change inside a shared household diet is the harder problem.

The household framing has trial evidence behind it. The iHealth-T2D trial enrolled 3,684 South Asians at 120 sites across India, Pakistan, Sri Lanka and the UK. Instead of treating lifestyle change as an individual problem, it delivered an intensive family-based programme through community health workers. After one year the investigators reported modest improvements in weight and waist circumference against usual care, no overall difference in HbA1c or blood pressure, and better results among those who attended more sessions.

I like this result partly because it was modest. Real prevention usually is. But the design carries the insight: sometimes the household is the unit of intervention. If everyone eats the same food, changing one person’s diet means changing the default meal. If the family treats exertion in the old as dangerous, an exercise prescription is not enough. Implementation is not what happens after the science. Implementation determines how much of the science reaches the person.

Why a pill travels more easily than a lifestyle

There is a corollary worth stating plainly. Some pharmacological interventions transport well partly because they compress the implementation problem. A fixed-dose pill is not context-free: access, adherence, adverse effects and health systems still matter. But the active intervention itself is standardized.

Compare two trials. HOPE-3 randomized 12,705 people across 21 countries to rosuvastatin or placebo, followed real cardiovascular events for a median of 5.6 years, and found consistent benefit across racial and ethnic subgroups; nearly half of the participants were Asian. iHealth-T2D asked thousands of South Asian families to change how they live, and moved the needle modestly.

This does not mean pills beat lifestyle. It means a lifestyle intervention must cross an additional biological–social interface before it reaches the outcome, and that interface is one of the places where India differs most.

Social capital has two arrows

That interface is not necessarily a disadvantage. India retains what many wealthy societies are trying to rebuild: family, friendship, neighbourhood and religious networks that supply daily contact, practical help, caregiving and belonging. In a 2026 analysis of 31,063 older Indians in the Longitudinal Ageing Study in India, higher social participation was associated with more physical activity, yoga, adult vaccination and preventive check-ups. Observational, so no causal claim, but the connection between social structure and health behaviour is hard to miss.

Notice the difference in starting points. The US Surgeon General declared an epidemic of loneliness in 2023, and much of the rich world is now trying to prescribe community the way one might prescribe a drug. India often begins from a different social baseline: many people still live inside dense family and community networks, though migration, nuclear households and widowhood are loosening them, and Indian loneliness is real. So the design problem here is not always building connection from scratch. It is understanding, and sometimes redirecting, what the existing network carries.

Families transmit food norms. Communities can normalize tobacco. Older adults are lovingly told to take rest when what they need is to keep their strength. Feeding is affection, which makes restraint socially expensive. So I would not file social connection as one more checkbox beside sleep and exercise. The network is part of the intervention environment.

Religious and traditional practice belongs inside the same frame. Rather than asking vaguely whether religion or Ayurveda “improves longevity,” decompose them into testable exposures: participation, fasting, dietary rules, movement, meditation, routine, specific compounds and formulations. An old practice is not correct because it is old. It is not wrong because it is old, either. Test the specific claim.

Some exposures cannot be optimized away

A person can exercise, eat well, stop tobacco and control their blood pressure. They cannot individually remove particulate matter from their city’s air, cool a heatwave, or make an unsafe road walkable. As we move from molecular mechanism toward lived healthspan, the environment increasingly enters the system. Some determinants belong to the person, some to the household, and some to the city and the state. A serious longevity programme needs all three.

The strongest objection is the most useful one

A reasonable reader might now say: why complicate this? Statins work in Indians. Blood-pressure treatment works. Quitting tobacco works. Just practise good medicine.

I largely agree, and that is the control case. The interventions that transport well show us what strong transportability looks like. HOPE-3 did not merely show that a statin moved a lipid number. It counted heart attacks, strokes and cardiovascular deaths. Now compare that evidentiary chain with the one behind much of the longevity frontier:

Perhaps every arrow will hold. But they have not all been demonstrated, and that difference, not novelty or age, is what separates the two kinds of claim.

Longevity has a strange prestige hierarchy

Technological sophistication and evidentiary strength are not the same thing. The field sometimes assigns the most prestige to what sits furthest from a demonstrated human outcome.

I work in multiomics. I am interested in these technologies precisely because they may let us see biology long before clinical disease appears, and a biological-age score may yet become a powerful surrogate endpoint. But the number of biomarkers measured should not be confused with the amount of clinical evidence generated. A 400-analyte panel contains more information than a blood-pressure cuff. It does not automatically contain more useful information today. Until changing the surrogate is reliably connected to changing outcomes people care about, confidence should reflect the distance.

What should someone in India actually do?

The uncertainty lives mostly at the frontier; the basics are not mysterious. Start with what has earned confidence: no tobacco in any form, know and treat your blood pressure, understand your cardiovascular and diabetes risk, stay physically active, fix real nutritional deficiencies, use established preventive medicine.

Then pay particular attention to muscle and body composition. Do not assume a normal BMI means low metabolic risk. Do not pursue weight loss while ignoring strength. Make some of your exercise progressive resistance work, not only walking. With age, treat function itself as a biomarker: chair, stairs, grip, gait, balance. Improving or declining?

Make interventions livable. If the household eats together, change the household default rather than sustaining a private “longevity diet” forever. Use family and community as infrastructure when they help; notice when the same structures block change.

And at the frontier, be curious but demanding. If someone hands you a biological-age result, a multiomic score or a new longevity intervention, ask:

• Was this measurement validated in people like me?
• Does it predict something I care about?
• If the number changes, does the outcome change?
• Which intervention changes it, and was that tested in humans, for how long?
• What happened to disease, function, cognition, mortality?

These questions are not anti-innovation. They are how interesting science becomes medicine.

India can do more than localize

The opportunity is bigger than making Indian versions of American longevity products. India is scientifically valuable because the variables longevity research tends to study separately all coexist here, at enormous scale: genetic diversity, distinct metabolic phenotypes, sharp rural–urban contrasts, strong but changing family structures, living traditional practice, heavy environmental exposure, and over a billion people moving through all of it at once.

BHARAT and GenomeIndia are laying the foundation. The next step is to connect measurement to intervention to behaviour to function to disease, and to follow people through time. Longitudinal science, not expensive snapshots: blood pressure beside proteomics, grip strength beside metabolomics, household structure beside genetics, air beside inflammation.

That is a direct ask of Indian funding agencies, institutes, biotech and philanthropy: put serious money not into another panel to sell, but into cohorts held together for decades: the unglamorous infrastructure every future longevity product will quietly stand on. Whoever builds that will not be importing longevity science. They will be exporting it.

Import the science, not the assumptions

The American longevity movement has done something genuinely important: it turned aging from a fact we accept into a process we can study and, perhaps increasingly, modify. India should learn from it aggressively. But copying is not learning.

The mechanism may travel. The measurement needs validation. The intervention may depend on the starting phenotype. Implementation depends on the household and the environment. And the outcome still has to be demonstrated. Transportability need not decay at every step; some interventions survive the whole journey intact. But each step is another place where an assumption can fail.

The farther a longevity claim sits from a demonstrated human outcome, the more questions we should ask before acting on it.

Because the endpoint was never the number. At 70, 80 or 90: can I think clearly, move independently, get up off the floor, eat normally, see and hear well enough to take part, manage my own affairs, keep the relationships that make the effort worthwhile? That is healthspan: the years above the line, on the curve this essay opened with. Everything upstream matters only insofar as it protects those years.

India has already begun asking whether the clocks need localization. The harder question comes next.

Do the prescriptions?