Longevity Interventions: What the Human Evidence Shows

By /

Longevity Interventions: What the Human Evidence Actually Shows

By Anas, Editor at HealthForge · Published 20 September 2026 · 19 min read

A person walking alone along a tree-lined path in soft early morning light
Photo by ActionVance on Unsplash.

No supplement, drug, or protocol has been shown to extend human lifespan in a randomized trial. That single fact should reshape how you read almost everything sold under the “longevity” banner. Most of what circulates online rests on animal studies, cell experiments, or shifts in biomarkers that stand in for aging without proving anyone actually lived longer.

The field isn’t empty. A handful of interventions carry decades of human data for the outcomes that matter — death and disease — while others are biologically fascinating but clinically unproven. Both get sold with the same confident language, and the gap between a mouse result and a human benefit is where most of the hype lives.

This guide sorts longevity interventions by the strength of their human evidence, keeps lifespan claims separate from healthspan and biomarker effects, and closes with a way to prioritize your effort and money.

In this article
  1. The short version
  2. What counts as real evidence in longevity research
  3. Longevity interventions ranked by evidence strength
  4. Lifestyle interventions with the strongest longevity payoff
  5. Metformin: proven in diabetes, unproven for longevity
  6. Rapamycin: the most interesting candidate, still without lifespan data
  7. Popular supplements and why the hype outpaces the data
  8. Interventions with promising early data but no human proof yet
  9. How to read longevity biomarkers
  10. The one place “anti-aging” is genuinely proven — skin
  11. What this means for you
  12. What the evidence does not establish
  13. Common questions
  14. Where this leaves us
  15. Related reading

What counts as real evidence in longevity research

A researcher's gloved hands transferring a liquid sample into a row of small vials in a laboratory
Most longevity claims rest on lab work that has never been tested for outcomes that matter in people. Photo by Julia Koblitz on Unsplash.

Aging research runs on the evidence hierarchy, and where a claim sits on it should govern how much weight you give it. A finding in yeast or worms is a starting hypothesis. Results in mice are more relevant but still often fail to translate. A small human trial measuring a blood marker is a step up. The gold standard — a large randomized trial measuring death, disease, or disability — is rare in longevity research.1

Mechanisms are seductive, and that is the core tension. A compound that switches on a “longevity pathway” in a dish generates headlines long before anyone knows whether it helps a person. A 2026 review of longevity pharmacology put it plainly: mechanistic progress is outpacing clinical translation, which leaves human evidence limited and often indirect.7

The preclinical literature has a quality problem of its own. A 2025 methodological appraisal of the DrugAge database — a large catalog of aging-intervention studies — found substantial heterogeneity and bias, which weakens any confident leap from animal data to human longevity.8

Two distinctions run through everything below:

  • Lifespan is how long you live. Healthspan is how long you stay healthy and functional. An intervention can plausibly improve one without touching the other — a difference worth understanding in depth, since lifespan vs healthspan shapes how you weigh nearly every claim here.
  • A surrogate endpoint (a biomarker, a lab value) is a stand-in. A hard endpoint (mortality, a diagnosed disease, disability) is the outcome that actually matters. Many longevity claims quietly swap the first for the second.

Longevity interventions ranked by evidence strength

A small arrangement of loose capsules and pills scattered on a neutral surface
Popular longevity compounds are sold with the same confidence as interventions with far stronger human data. Photo by laura adai on Unsplash.

Here is the landscape at a glance, graded by the quality of human evidence for meaningful outcomes.

Intervention Human evidence for hard outcomes Verdict
Not smoking, physical activity, Mediterranean diet Consistent observational, some trial support Proven risk reduction
Metformin (in diabetes) Observational mortality benefit Proven in disease, unproven for longevity
Rapamycin Short-term safety data, no mortality data Promising, unproven
NAD+ boosters (NR, NMN) Reliable biomarker change, null functional effects Biomarker effect, no proven benefit
Senolytics (dasatinib + quercetin) Early feasibility trials only Early-stage, no human proof
Resveratrol No convincing human target engagement Largely hype
Topical retinoids (skin aging) Repeated randomized trials Proven — for skin, not lifespan

The rest of the article works through these rows.

Lifestyle interventions with the strongest longevity payoff

A wooden table with fresh vegetables, greens, and a bottle of olive oil in daylight
A Mediterranean-style diet is among the least glamorous but best-supported longevity habits. Photo by Margit Knobloch on Unsplash.

The least glamorous interventions have the best human track record, and they map closely onto the longevity pillars of sleep, exercise, and nutrition.

A 2004 prospective cohort of aging women in the Archives of Internal Medicine linked a Mediterranean diet, physical activity, and nonsmoking each independently to lower all-cause mortality.2 The pattern holds in higher-risk groups: a prospective cohort of people with coronary artery disease found nonsmoking, regular activity, and Mediterranean diet adherence each associated with lower mortality.3 A 2025 nutrition evidence synthesis went further, reporting that Mediterranean-style diet programs — usually combined with physical activity and behavioral support — reduced all-cause and cardiovascular mortality over five years in higher-risk adults.4

Two honest caveats. Most of this evidence is observational, so it establishes strong, consistent association rather than airtight causation — a good reason to understand RCTs vs cohort studies. And these studies test bundles of behavior, so the effect of any single component — sleep, for instance — is hard to isolate. Even with those limits, this is the most robust human evidence in the entire longevity field, and it costs nothing.

Metformin: proven in diabetes, unproven for longevity

Metformin is the most credible “longevity drug” candidate with real human data — but the data are narrower than the enthusiasm suggests.

A 2017 systematic review and meta-analysis in Ageing Research Reviews found metformin associated with lower all-cause mortality and reduced age-related disease risk in people with type 2 diabetes, apparently beyond its glucose-lowering effect.5 Plausible geroprotective mechanisms sit behind that finding,6 which is why researchers keep testing it.

Translation is the problem. A 2022 meta-analysis in Aging Cell found heterogeneous effects on model-organism lifespans and no significant overall lifespan benefit in mice, and it noted that no randomized placebo-controlled trial has confirmed pro-longevity effects in people with normal blood sugar.10 Long-term observational data are cautionary too: a 2023 Aging Cell comparison suggested that early survival advantages in diabetics may be outweighed over roughly 20 years — not the profile of a proven longevity drug.9

Safety in older adults is bounded by kidney-function thresholds, avoidance in decompensated heart failure, and gastrointestinal side effects.11 This is a legitimate diabetes medication with intriguing aging biology, not a validated intervention for healthy people who want to live longer.

Rapamycin: the most interesting candidate, still without lifespan data

Rapamycin (and its analog sirolimus) inhibits mTOR, one of the most reproducible longevity pathways in animals. Human data are emerging, and they read more soberly than the online conversation.

The PEARL trial, a 48-week randomized placebo-controlled study in healthy older adults, found low-dose intermittent rapamycin relatively safe, with adverse events similar to placebo — but it did not significantly change visceral fat, one of its prespecified targets.12 A systematic review in The Lancet Healthy Longevity found no serious adverse events attributable to rapamycin-class drugs in healthy participants, though in people with aging-related disease the drugs caused more infections and raised cholesterol and triglycerides.13 The RAPA trial reported that weekly sirolimus did not improve exercise outcomes in older adults while adding minor adverse events.14

A 2025 review captures the bottom line: human evidence is insufficient to establish rapamycin as a proven therapy to delay aging or extend lifespan in healthy adults.15 The drug is genuinely promising and, at low intermittent doses, appears reasonably tolerated in the short term — but there is no human mortality data behind it.

NAD+ boosters (NR and NMN)

This is a clean case study in the surrogate-versus-outcome trap. Oral nicotinamide riboside (NR) and nicotinamide mononucleotide (NMN) clearly and reliably raise NAD+-related biomarkers in humans — that part is well established.16 A randomized crossover trial in Nature Communications confirmed NR is well tolerated and elevates blood NAD+ in middle-aged and older adults,17 and a small trial found NMN raised NAD+ and shifted some muscle-function measures in older men.18

Moving the biomarker, however, has not translated into reliable benefit. A 2023 meta-analysis in Advances in Nutrition found effects on functional, metabolic, and aging outcomes were heterogeneous and often null.16 A 2025 expert review reached the same conclusion: clear biochemical target engagement, inconclusive evidence for anti-aging, longevity, or broad healthspan benefits.19 You are paying to move a number, not for a demonstrated result.

Resveratrol

Resveratrol’s story is weaker still. Popularized as a sirtuin (SIRT1) activator, it showed no significant overall effect on human SIRT1 gene expression, protein expression, or serum levels in a 2025 systematic review and meta-analysis of 11 randomized trials.20 Its life-extending effects are strongest in yeast and nematodes and fade in higher organisms — a 2012 cross-species analysis in Aging Cell explicitly questioned marketing it as a human life-extension supplement.21

There is also a pharmacological wall. Resveratrol has poor oral bioavailability, producing low plasma concentrations despite being well tolerated,22 and adding piperine — a common “absorption enhancer” — did not significantly increase exposure in a pilot trial.23 In people, the compound largely fails to reach or engage its supposed target.

Interventions with promising early data but no human proof yet

Senolytics

Senolytics aim to clear “zombie” senescent cells that accumulate with age, and the human evidence is early and honest about its limits. A phase 1 feasibility trial of dasatinib plus quercetin in mild Alzheimer’s disease found the combination well tolerated, with biomarker shifts suggesting biological activity but no established efficacy.24 A separate preliminary study in people with diabetic kidney disease found that a short course reduced the senescent-cell burden in fat tissue.25

These are proof-of-concept signals: the drugs appear to do something biologically, but they do not yet show that anyone lives longer or better. Dasatinib in particular is a chemotherapy agent with real toxicity, which makes casual self-experimentation unwise.

How to read longevity biomarkers

Biomarkers are where the field gets both its rigor and its confusion. They are useful for research and risk stratification, and misleading when treated as outcomes.

Epigenetic “clocks” are the headline example. A 2019 meta-analysis found accelerated DNA methylation age associated with higher all-cause mortality — roughly an 8–15% increase in risk per five-year increase in epigenetic age, with noted publication bias.26 Age acceleration also tracks with cardiovascular disease, cancer, and diabetes.27 Newer clocks predict better: the DNAm PhenoAge clock improved prediction of mortality and morbidity over earlier versions,28 and GrimAge and GrimAge2 show the strongest mortality associations among second-generation clocks.29

Two problems limit their use as personal feedback tools. Clocks show excellent technical reproducibility but only moderate biological reliability across repeated short-interval samples, so a single reading is noisier than it looks.30 Change over time carries more information: longitudinal increases in clock values predict mortality independent of baseline.31

Conventional biomarkers deserve at least as much attention. Glycemic, lipid, and inflammatory markers are associated with both healthspan and lifespan in observational and genetic analyses.32 Higher HbA1c, fasting glucose, triglycerides, and CRP predict worse healthspan and shorter lifespan, while higher HDL-related measures are protective.33 These are cheap, standard, and actionable.

One caveat holds for all of them: these are surrogates. Improving a clock reading or a lab value is a plausible signal, not a guarantee of a longer or healthier life.

The one place “anti-aging” is genuinely proven — skin

Cosmetic aging is the exception where randomized human evidence is strong, precisely because the outcome — visible skin — is easy to measure.

Topical tretinoin significantly improves fine and coarse facial wrinkles compared with vehicle in photodamaged skin, confirmed by both a 2025 meta-analysis and a classic vehicle-controlled trial.34 A network meta-analysis found that retinoids broadly improve fine wrinkles, tazarotene helps coarse wrinkles, and tretinoin and retinol improve hyperpigmentation.35 Among over-the-counter options, retinol and vitamin C earn the strongest recommendation grades, while many other cosmeceutical ingredients remain inconclusive.36

Two qualifiers keep this honest. Evidence for OTC retinol products specifically is thinner than for prescription tretinoin — one review found only nine rigorous vehicle-controlled trials, several of them flawed.37

The individual signals are still real. Topical retinol beat placebo for fine and deep wrinkling over 26 weeks,38 and bakuchiol matched retinol on wrinkles and pigmentation with less irritation in one trial.39 Glycolic acid improved texture and discoloration but not wrinkles to a significant degree.40 The broader OTC skincare market suffers from a genuine shortage of controlled efficacy trials.41 This is anti-aging that works — on your skin, not your lifespan.

What this means for you

The evidence points to an unglamorous but reliable order of priorities. Spend your effort where the human data are strongest before spending money where they are weakest.

  • Do the proven basics first. Not smoking, exercise for healthspan, and a Mediterranean-style dietary pattern have the best human evidence for lower mortality.234 Nothing in a bottle competes with these.
  • Track standard biomarkers, not exotic ones. HbA1c, fasting glucose, triglycerides, HDL, and CRP are cheap, validated risk markers you can act on with a clinician.33
  • Treat epigenetic clocks as research tools, not verdicts. A single reading is noisy, trends matter more, and moving a clock is not proven to move your outcome.3031
  • Be honest about supplements. NAD+ boosters raise a lab number without proven benefit,19 and resveratrol largely fails to reach its target.2022
  • If you’re considering metformin or rapamycin, that’s a clinician conversation — both carry real safety constraints, and neither is proven for healthy longevity.1115

For readers who want to go deeper on the science, evidence-first books tend to be a better investment than most supplement stacks; How Not to Age is one grounded option.

What the evidence does not establish

This article describes associations and short-term trials, not proof of life extension. Most lifestyle evidence is observational, so it cannot fully rule out confounding — healthier people differ in many ways at once.23 The strongest longevity signals also come from combinations of behaviors, which leaves the independent contribution of any single factor, including sleep, uncertain.4

For drugs, the honest ceiling is clear: no randomized trial has confirmed a longevity benefit for metformin in normoglycemic people,10 and human evidence is insufficient to call rapamycin a proven anti-aging therapy.15 Biomarker changes — whether NAD+ levels or epigenetic clocks — are surrogates whose relationship to your personal mortality is probabilistic, not deterministic.1930 And the preclinical literature that fuels most of the excitement carries substantial heterogeneity and bias.8 Where this article says “promising,” read it as worth watching rather than settled.

Common questions

Do any supplements extend human lifespan?

No supplement has been shown to extend human lifespan in a randomized trial. Some, like NAD+ boosters, reliably change biomarkers,16 but functional and aging outcomes are heterogeneous or null,19 and resveratrol largely fails to engage its target in people.20 Claims of lifespan extension go well beyond the human evidence.

Is rapamycin safe and effective for longevity?

Low-dose intermittent rapamycin appears relatively safe over about a year in healthy older adults, with adverse events similar to placebo,12 but it has not improved key targets in some trials14 and has no human mortality data. A 2025 review concluded the evidence is insufficient to call it a proven anti-aging therapy.15 Any use would be off-label and a clinician’s call.

What’s the difference between healthspan and lifespan?

Lifespan is how long you live; healthspan is how long you stay healthy and functional. Many interventions plausibly affect one without proof for the other, and much marketing blurs the two. Improving a biomarker relates to both only indirectly.3032

Are NAD+ boosters worth taking?

They dependably raise NAD+-related markers and are generally well tolerated,1617 but a 2023 meta-analysis and a 2025 review found their effects on functional, metabolic, and aging outcomes inconclusive.1619 You’d be paying to move a number without demonstrated real-world benefit.

What actually reduces mortality?

The most consistent human evidence favors not smoking, regular physical activity, and a Mediterranean-style diet, each independently linked to lower all-cause mortality.234 Managing standard cardiometabolic markers — glucose, lipids, inflammation — also tracks with better healthspan and lifespan.33

Where this leaves us

Longevity science is real, active, and mostly earlier-stage than its marketing implies. The interventions with the best human evidence are the ones no one can patent: not smoking, moving regularly, eating well, and keeping ordinary cardiometabolic markers in a healthy range. The compounds that dominate the conversation — NAD+ boosters, resveratrol, and to a lesser extent metformin and rapamycin — range from unproven to overhyped, even when the underlying biology is genuinely interesting.

A useful rule for the next claim you meet: ask what outcome was measured, in what species, and in what kind of study. If the answer is a biomarker in a mouse, treat it as a lead rather than a conclusion. The most evidence-based posture toward longevity today is patience — do the proven things now, and let the promising things earn their proof.

Sources

  1. Frontiers in Aging, 2024: Climbing the longevity pyramid: overview of evidence-driven healthcare prevention strategies for human longevity
  2. JAMA/Archives of Internal Medicine, 2004: Mediterranean diet, lifestyle factors, and 10-year mortality in aging women
  3. PLOS ONE, 2014: Effect of Sustaining Lifestyle Modifications on Recurrent Coronary Artery Disease and Mortality
  4. PMC, 2025: Medical Nutrition Therapy (MNT) Evidence Update
  5. Ageing Research Reviews, 2017: Metformin reduces all-cause mortality and diseases of ageing independent of its effect on diabetes control
  6. Cells, 2019: Metformin as Anti-Aging Therapy: Is It for Everyone?
  7. Advancements in longevity pharmacology research, 2026: Review of longevity pharmacology (mechanistic progress outpacing clinical translation)
  8. PMC, 2025: Reporting quality, effect sizes, and biases for aging interventions: a methodological appraisal of the DrugAge database
  9. Aging Cell, 2023: Comparison of long-term effects of metformin on longevity between people with type 2 diabetes and matched non-diabetic controls
  10. Aging Cell, 2022: Metformin has heterogeneous effects on model organism lifespans and no significant overall effect in mice
  11. Current Diabetes Reports, 2016: Metformin in the aging population: safety considerations and adverse effects
  12. Aging / Geroscience (PEARL trial), 2025: Influence of rapamycin on safety and healthspan metrics after one year
  13. The Lancet Healthy Longevity, 2024: Targeting ageing with rapamycin and its derivatives in humans: a systematic review
  14. Journal of Cachexia, Sarcopenia and Muscle, 2026: Exercise and Weekly Sirolimus (Rapamycin) in Older Adults: RAPA
  15. Clinical Interventions in Aging, 2025: What is the clinical evidence to support off-label rapamycin use for healthy aging?
  16. Advances in Nutrition, 2023: Dietary Supplementation With NAD+-Boosting Compounds in Humans: A Systematic Review and Meta-analysis
  17. Nature Communications, 2018: Chronic nicotinamide riboside supplementation is well-tolerated and elevates NAD+ in healthy middle-aged and older adults
  18. npj Aging, 2022: Chronic nicotinamide mononucleotide supplementation elevates blood NAD+ and alters muscle function in healthy older men
  19. Trends in Endocrinology & Metabolism, 2025/2026: NAD+ supplementation for anti-aging and wellness
  20. PubMed, 2025: Impact of Resveratrol Supplementation on Human Sirtuin 1
  21. Aging Cell, 2012: The effect of resveratrol on longevity across species
  22. PubMed, 2009: Pharmacokinetic and safety profile of trans-resveratrol
  23. Nutrients, 2021: A Randomized, Double-Blind, Dose-Ranging Pilot Trial of resveratrol with piperine
  24. Nature Medicine, 2023: Senolytic therapy in mild Alzheimer’s disease: a phase 1 feasibility trial
  25. EBioMedicine, 2019: Senolytics decrease senescent cells in humans: Dasatinib plus Quercetin in diabetic kidney disease
  26. Aging (Albany NY), 2019: The epigenetic clock as a predictor of disease and mortality risk: a systematic review and meta-analysis
  27. Ageing Research Reviews, 2021: A systematic review of factors associated with epigenetic age acceleration
  28. Aging, 2018: An epigenetic biomarker of aging for lifespan and healthspan (DNAm PhenoAge)
  29. PubMed, 2025: Associations between five indicators of epigenetic age acceleration and mortality in US adults
  30. PubMed, 2025: Biological versus Technical Reliability of Epigenetic Clocks
  31. PubMed, 2026: Longitudinal changes in epigenetic clocks predict survival
  32. EBioMedicine, 2021: Clinical biomarkers and associations with healthspan and lifespan
  33. EBioMedicine, 2021: Clinical biomarkers and associations with healthspan and lifespan (full analysis)
  34. PMC, 2025: Tretinoin for Photodamaged Facial Skin: Systematic Review and Meta-analysis
  35. PubMed, 2025: Comparative efficacy of topical interventions for facial photoaging
  36. PubMed, 2024: Cosmeceuticals for antiaging: a systematic review of safety and efficacy
  37. PubMed, 2022: Evidence for the Efficacy of Over-the-counter Vitamin A Cosmetic Products in Skin Aging
  38. PubMed, 2010: Improvement of photoaged facial skin in middle-aged adults with topical retinol
  39. PubMed, 2018: Prospective, randomized, double-blind assessment of bakuchiol and retinol for facial photoageing
  40. PubMed, 1998: A double-blind randomized clinical trial of a 5% glycolic acid cream in photoaging
  41. PubMed, 2012: Over-the-counter topical skincare products: a review of the literature

Leave a Comment

Your email address will not be published. Required fields are marked *

Scroll to Top