Quick Answer: The strongest evidence sits with unglamorous things: not smoking, cardiorespiratory fitness, maintaining muscle mass, sleep of adequate duration, social connection, and blood pressure control. Moderate evidence: dietary pattern quality, body composition, alcohol reduction. Weak or speculative: most supplements, fasting protocols for lifespan specifically, and nearly everything marketed as a longevity product. The ranking is stable and boring, which is exactly why it gets so little attention.

If you sorted longevity interventions by evidence strength rather than by how interesting they are, the list would look very different from your feed.

What has the strongest evidence?

Cardiorespiratory fitness, and the effect size is larger than most people expect.

VO2 max — your body’s capacity to use oxygen — is among the strongest predictors of all-cause mortality in the literature. Observational data consistently shows large differences in mortality risk across fitness levels, and the gap between the lowest fitness category and merely average is bigger than the gap between average and elite.

That last point is the practical one. The biggest return is at the bottom of the range. Going from sedentary to moderately active is where most of the benefit lives — the returns to going from good to excellent are real but much smaller.

The dose that shows up repeatedly: roughly 150 minutes of moderate activity per week, or about half that at vigorous intensity. This is achievable for most people and it is genuinely the highest-value intervention available.

Not smoking belongs in the same tier and needs no elaboration. It remains the single largest modifiable factor by a wide margin.

Blood pressure control has strong causal evidence from randomized trials, which is rarer in this field than you’d hope. Most longevity research is observational; blood pressure is one of the areas where interventional data exists.

These four — don’t smoke, be fit, control blood pressure, and keep moving — account for most of what’s reliably known.

Where does muscle mass fit?

Strong evidence, and it becomes more important with age rather than less.

Muscle mass and strength are independently associated with mortality, separate from cardiovascular fitness. Grip strength — a simple proxy — is a surprisingly robust predictor across large studies.

The mechanism is partly about function and partly metabolic. Muscle is the primary site of glucose disposal, so more of it improves metabolic health. And functionally, maintaining strength is what preserves independence and prevents falls in later decades.

Sarcopenia — age-related muscle loss — begins earlier than most people think and accelerates after 60. The decline is substantially slowed by resistance training, at any age of starting.

Studies on resistance training in people in their seventies and eighties consistently show meaningful strength gains. There is no age at which it stops working, which is one of the more encouraging findings in the field.

Practical dose: two sessions per week of resistance training covering major movement patterns. That’s the recommendation in most public health guidelines and it’s supported.

Protein intake supports this — older adults appear to need more protein per meal to trigger the same muscle-building response, which argues for adequate intake spread across the day rather than concentrated in one meal.

How strong is the evidence on sleep?

Strong for the association, weaker for the intervention — but strong enough to act on.

Both short and long sleep duration associate with higher mortality, producing the well-known U-shaped curve. The lowest-risk range in most large studies is around 7 to 8 hours.

The long-sleep side of the curve is likely partly confounded — illness causes long sleep as often as the reverse. The short-sleep side has more mechanistic support, including effects on glucose regulation, blood pressure, immune function, and appetite hormones.

Sleep consistency may matter as much as duration. Irregular sleep timing shows independent associations with cardiometabolic risk, which suggests going to bed at wildly different times is its own risk factor.

Sleep apnea deserves specific mention because it’s common, underdiagnosed, and treatable. If you snore heavily, wake unrefreshed, or have been told you stop breathing, that’s worth pursuing — the cardiovascular consequences are significant and treatment works.

What’s less supported: sleep tracking devices improving outcomes. They measure imperfectly and can generate anxiety about sleep, which worsens it. Useful for spotting gross patterns; not useful for nightly optimization.

This is general information, not medical advice.

Does social connection really belong on this list?

Yes, and its effect size is larger than most people assume.

Meta-analyses of social connection and mortality have found effects comparable in magnitude to well-established risk factors like smoking and physical inactivity. This is consistently surprising to people, and it’s one of the more robust findings in the field.

Both objective isolation and subjective loneliness show associations, and they’re not the same thing — people can be objectively connected and still lonely, and both matter.

Mechanisms likely include chronic stress physiology, inflammatory pathways, health behaviors, and practical support during illness. The causal picture is complicated, as it always is with observational data.

Why it’s usually left out of longevity discussions is worth noting: it isn’t purchasable. There’s no supplement, device, or protocol, so it doesn’t generate content or products.

Practically: regular in-person contact, at least one relationship where you can be genuinely honest, and some form of community involvement. These aren’t optimizable in the way training is, which is precisely why they get neglected by people who like optimizing things.

What about diet — how strong is the evidence really?

Moderate, and considerably weaker than the confidence with which it’s discussed.

Nutrition epidemiology is genuinely difficult. Self-reported intake is unreliable, dietary patterns correlate heavily with income, education, and other health behaviors, and long-term randomized trials are nearly impossible to run.

What survives across methods: dietary patterns emphasizing vegetables, legumes, whole grains, nuts, and fish, with lower intake of ultra-processed foods, associate with better outcomes across many populations. The Mediterranean pattern has the most supporting data, including some interventional evidence.

What’s much weaker than commonly claimed: specific superfoods, precise macronutrient ratios, and most individual nutrients studied in isolation. The pattern is better supported than any component of it.

Ultra-processed food intake is an area of growing evidence, with fairly consistent associations to cardiometabolic outcomes. Whether that’s the processing itself or the accompanying nutritional profile isn’t fully resolved.

Fasting protocols have strong animal data and much thinner human longevity data. Time-restricted eating shows metabolic effects in some trials, often attributable to reduced overall intake. Claiming lifespan effects in humans goes well beyond what exists.

Alcohol: the picture has shifted. Earlier findings suggesting benefit from moderate intake have been substantially challenged by studies accounting for sick-quitter bias and using genetic methods. The current position in much of the literature is that less is better, with no clearly protective amount.

Which popular interventions have weak evidence?

Most of what’s actively marketed, which is not a coincidence.

Most supplements lack human longevity evidence. Correcting a documented deficiency has clear benefit. Supplementing beyond sufficiency generally doesn’t, and several large trials of antioxidant supplementation found no benefit or harm. Vitamin D supplementation has repeatedly underperformed expectations in randomized trials despite strong observational associations.

NAD precursors, resveratrol, and similar compounds have interesting mechanistic and animal data and no human lifespan evidence. This isn’t a claim they don’t work — it’s a statement about what’s currently known.

Cold plunges and saunas: sauna use has some intriguing observational data, particularly from Finnish cohorts. Cold exposure has much thinner evidence for longevity specifically, whatever its other merits.

Biological age tests are not validated as outcomes. Different tests give different answers on the same person, and a change in a test score has not been shown to correspond to a change in lifespan.

Extensive supplement stacks, expensive testing panels, and personalized protocols mostly sell certainty, which is the thing this field has least of.

A reasonable heuristic: the more a longevity intervention costs, the weaker its evidence tends to be. The strongly supported items are free or cheap — walking, sleeping, resistance training, not smoking, having friends. That’s not a coincidence either.

How should I actually prioritize? Plus FAQ

In descending order of evidence and effect size, which is roughly the reverse of how attention is distributed.

Tier 1 — do these first: don’t smoke; get cardiorespiratory fitness above the bottom quintile; resistance train twice a week; sleep 7 to 8 hours consistently; know and manage your blood pressure; maintain real social connection.

Tier 2 — worth doing: eat a mostly whole-food dietary pattern; keep body composition in a reasonable range; minimize alcohol; manage chronic stress; keep up with age-appropriate medical screening.

Tier 3 — optional, low expected value: supplements beyond correcting deficiencies, fasting protocols, sauna, cold exposure, tracking devices.

The uncomfortable part: nearly everyone who asks about longevity is already asking about tier 3. Tier 1 is unglamorous, requires sustained ordinary effort, and can’t be purchased.

Is it too late to start?

No. Studies on smoking cessation, resistance training, and fitness improvement consistently show benefit at every age examined, including in people starting in their seventies and eighties. Benefits are typically larger for those starting from the lowest baseline.

Should I get a biological age test?

Probably not for decision-making. They’re not validated as outcomes, different tests disagree substantially on the same individual, and no test has been shown to track actual lifespan changes. Standard clinical markers — blood pressure, lipids, glucose, and a fitness measure — are more actionable.

How much does genetics determine?

Estimates of the heritability of lifespan generally land in the range of 20 to 30%, which leaves substantial room for everything else. Extreme longevity — living past 100 — appears more heritable than ordinary variation.

What’s the single highest-value thing?

If you smoke, stopping. If you don’t, moving from sedentary to regularly active. Those two account for more than everything else on this page combined.

This is general information, not medical advice. Discuss significant health changes with your physician.

TL;DR: Sort by evidence and the list gets boring fast — which is the whole point.

• Strongest: don’t smoke, build cardiorespiratory fitness, resistance train twice weekly, sleep 7 to 8 hours consistently, control blood pressure, maintain social connection.
• The biggest fitness return is at the bottom of the range. Sedentary to moderately active beats good to excellent.
• Muscle mass matters independently and responds to training at any starting age — including the eighties.
• Social connection has an effect size comparable to major risk factors. It’s left out because it can’t be sold.
• Moderate: dietary pattern quality, body composition, less alcohol. Weaker than the confidence around them suggests.
• Weak: most supplements, human fasting-for-lifespan claims, biological age tests, nearly everything marketed.

The more an intervention costs, the thinner its evidence tends to be. The well-supported ones are free.