Key takeaways
- Most of this evidence is observational, which shows association rather than causation and is vulnerable to the healthy-adherer problem.
- Effect size and certainty come apart constantly — smoking cessation has both, sauna has a striking effect size from essentially one population.
- Resistance training and protein are one intervention, not two: training provides the signal, protein the substrate, and either alone underperforms.
- Stacking all ten at once is how people manage none. Sequence it, and expect an endpoint measured in decades rather than in how you feel.
Almost everything sold as a longevity intervention rests on mechanism plus optimism: a pathway that looks promising in cells or mice, and an assumption that it scales to a human lifespan. The list below is different. Every item on it has been observed in large human populations followed for years or tested in a randomised trial with a hard endpoint. That is a much higher bar, and it is why the list is so unexciting — the interventions with the best evidence are the ones nobody can sell you.
What "has evidence" actually means here
Three distinctions decide how much weight any item deserves, and they are worth holding as you read.
Randomised versus observational. Most of what follows is observational — large cohorts followed over decades. That design shows association, not causation, and it is vulnerable to the healthy-adherer problem: people who sauna four times a week, or lift, or sleep seven hours, differ from people who do not in dozens of ways that are hard to measure. Good studies adjust for what they can. Adjustment is never complete.
Effect size versus certainty. These come apart constantly. Smoking cessation has both. Social connection has a large apparent effect and weaker causal footing. Sauna has a striking effect size from essentially one population.
Nothing below is a promise about an individual. Each is a probability shift across a population, which is the only currency this kind of evidence deals in.
1 and 2: don't smoke, and keep body composition in a reasonable band
Smoking cessation is the largest single intervention available. Smoking costs roughly ten years of life expectancy, and quitting improves outcomes at any age — the earlier the more is recovered, but the benefit does not disappear in later decades (Jha et al., N Engl J Med 2013). Nothing else on this list is close.
Body composition is second, with a caveat about measurement. All-cause mortality is lower in the BMI 22-27 band than in the obese range, but BMI cannot distinguish muscle from fat or see where the fat sits. The metabolically relevant variable is visceral adiposity, which is why waist circumference adds information weight alone does not.
3 and 4: load your muscles, and protect your sleep
Resistance training two or more days a week is associated with roughly 10-17% lower all-cause mortality, independently of cardiovascular exercise (Momma et al., Br J Sports Med 2022). The mechanism is not mysterious. Muscle is the largest site of glucose disposal in the body, the main mechanical protection against falls and fractures, and the tissue whose loss defines frailty. Grip strength and lean mass are among the strongest predictors of mortality after 60 — see why strength predicts longevity as well as cardio does. Two sessions a week is the threshold where the association appears, which makes a three-day full-body programme more than sufficient.
Sleep is the input everything else depends on. Short sleep, under about six hours, and habitually long sleep above nine are both associated with higher mortality, with the protective band around seven to eight. Read the long-sleep signal carefully — it is more plausibly a marker of illness than a cause. Short sleep is the actionable end, degrading glucose handling, appetite regulation and hormone production at once (sleep, cortisol and recovery).
5 and 6: walk, and use heat
Step count shows a dose-response with mortality that flattens well below the number everyone quotes. Benefit accrues steeply from very low counts up to somewhere around seven thousand steps, then plateaus (Lee et al., JAMA Intern Med 2019). The practical reading is that the gap between four thousand and seven thousand matters far more than the gap between seven and ten — see what the 10,000-step figure is actually based on. Cardiorespiratory fitness is the same argument at higher intensity, and VO2 max is among the strongest single predictors available.
Sauna carries the most striking effect size here: a Finnish cohort found substantially lower all-cause mortality among men bathing four or more times weekly compared with once (Laukkanen et al., JAMA Intern Med 2015). Treat it with care: observational, from one country where sauna is culturally universal, and bathing that frequently travels with leisure time, health and social contact. The mechanism — repeated heat stress improving vascular function — is plausible; the magnitude is probably flattered. Sauna and cold exposure separates the two.
7 and 8: a Mediterranean pattern, with enough protein
PREDIMED randomised participants at high cardiovascular risk to a Mediterranean diet supplemented with olive oil or nuts versus a control diet, and reported roughly a 30% reduction in major cardiovascular events (Estruch et al., N Engl J Med 2018). This is one of the few dietary patterns tested this way with a hard endpoint, which is why it earns its place over patterns with better mechanistic stories and no trial. How it compares with the alternatives is a fair question and the answer is mostly about adherence.
Protein is the item that changes with age. Requirements rise as anabolic sensitivity falls, so the intake that was adequate at 35 is not adequate at 70 — and sarcopenia is among the strongest predictors of late-life decline. Protein and resistance training are one intervention, not two: the training provides the signal, the protein provides the substrate, and either alone underperforms. How much you actually need covers the targets.
9 and 10: hormones where indicated, and connection
Hormone optimisation earns a place only where there is documented deficiency. In men, testosterone therapy for confirmed hypogonadism has been tested for cardiovascular safety in a high-risk randomised population and did not increase major events — a safety finding, not a longevity claim. In women, therapy started near the menopausal transition carries a different risk-benefit profile from therapy started long afterwards (HRT and cardiovascular health). Neither is a life-extension intervention. Both are corrections of a deficiency that otherwise erodes the muscle, bone and sleep that items three and four depend on.
Social connection is the item most often omitted and it has serious data behind it. Meta-analytic work finds social isolation and loneliness associated with substantially increased mortality risk, on a scale comparable to well-established behavioural risk factors (Holt-Lunstad et al., Perspect Psychol Sci 2015). Causality is genuinely hard here — illness isolates people as well as the reverse — but the association is large, consistent, and absent from almost every longevity protocol sold.
The principle: the unglamorous items outperform every expensive longevity product, and they compound with each other. Sleep makes training possible; training makes protein useful; both make body composition manageable; connection is what makes any of it sustainable for decades. The constraint has never been discovery.
Where to start, and what to expect
Stacking all ten simultaneously is how people manage none of them. A workable order runs: eliminate smoking, then fix sleep, then add resistance training twice weekly, then raise daily steps toward the plateau, then adjust protein, then the dietary pattern. Everything after that is refinement.
On timing — and this is the part that undoes people — the feedback is slow and mismatched. Sleep and training produce noticeable changes in energy and strength within weeks. Markers such as blood pressure, HbA1c and lipids move over months. The mortality benefit these are actually aimed at is a decades-long phenomenon that you will never personally observe. Anyone who needs the payoff to be visible will quit before it arrives, which is why the practical question is not which intervention is optimal but which one you will still be doing in five years.
Bottom line
The interventions with real prospective evidence are mostly free, mostly boring and mostly already known: don't smoke, keep body composition and waist in a reasonable band, load your muscles twice a week, sleep seven to eight hours, walk, use heat if you enjoy it, eat a Mediterranean-style pattern with adequate protein, correct documented hormone deficiency, stay connected to people. Most of that evidence is observational and the effect sizes are probably flattered, but the direction is consistent and the downside is near zero. The difficulty was never identifying the list — it is sequencing it and staying with it long enough for evidence measured in decades to apply to you.
Educational content, not medical advice. Laboratory interpretation and any treatment decision are made by a licensed physician after individual evaluation. Individual results vary.
