The Science of Longevity: A Comprehensive Overview
What the field actually knows, what it is testing, and what it has not shown. An orientation to longevity science without the enthusiasm or the cynicism.
The Short Answer
Longevity science studies why organisms age and what modifies the rate at which they do. Its most robust findings are that aging is malleable in every model organism tested, that a small number of pathways governing nutrient sensing and cellular maintenance are involved across species, and that in humans the interventions with the strongest evidence remain behavioural rather than pharmacological. The field's central practical distinction is between lifespan, meaning years lived, and healthspan, meaning years lived in good function, and the gap between them in high-income countries is roughly a decade, which is arguably the more urgent target.
Lifespan, Healthspan and the Shape of the Problem
Life expectancy at birth in high-income countries has roughly doubled since the mid-nineteenth century, driven overwhelmingly by reductions in infant and infectious mortality rather than by any change in the rate of aging. Maximum human lifespan has moved very little, and the survival curve has become more rectangular: more people reach old age, and then die within a compressed window.
Healthspan has not kept pace. Estimates of years lived in poor health late in life cluster around a decade in most high-income countries, and in several the gap has widened as life expectancy rose. The practical implication is that adding years is a less valuable objective than compressing the period of dysfunction at the end, which is the concept of compression of morbidity.
This reframing matters because it changes what counts as success. An intervention that extends life without extending function is not obviously desirable. An intervention that maintains function without extending life is valuable to almost everyone.
Why Organisms Age at All
Aging is not a designed process and it is not universal in the same form. Understanding the evolutionary logic explains why intervention is possible at all.
Mutation accumulation. Natural selection acts weakly on traits expressed after reproduction, so harmful late-acting variants persist because selection barely sees them.
Antagonistic pleiotropy. Some variants that benefit early life impose costs later. Selection favours the early benefit, and the late cost is carried.
Disposable soma. Energy allocated to reproduction cannot be allocated to somatic maintenance. Under resource constraint, organisms that reproduce sooner outcompete those that maintain their bodies indefinitely.
The consequence is optimistic rather than fatalistic. Aging is not a programme running to completion; it is what happens when maintenance is under-resourced relative to damage. Where maintenance is upregulated experimentally, aging slows, which is exactly what caloric restriction and mTOR inhibition demonstrate across species.
Comparative biology reinforces this. Naked mole rats live decades with negligible increases in mortality rate. Some bivalves live centuries. Bowhead whales live over two hundred years with far more cells than humans and no corresponding cancer burden. Aging rates vary enormously across species, which means they are set by biology that can in principle be understood.
What Works in Animals
The animal literature is where causal claims can be made, and it is more disciplined than public discussion suggests. A programme in the United States tests candidate compounds for lifespan extension in genetically heterogeneous mice across multiple independent sites, specifically to weed out single-laboratory artefacts. Its results are instructive.
Compounds that extended lifespan in that programme include rapamycin, robustly and at multiple doses, acarbose, particularly in males, 17-alpha-estradiol in males, and canagliflozin in males. Compounds that did not extend lifespan there, despite considerable public enthusiasm, include resveratrol, nicotinamide riboside, and several other popular supplements.
Caloric restriction extends lifespan across yeast, worms, flies and rodents, though the magnitude varies by strain and diet composition, and in non-human primates the results depend heavily on the control diet. Genetic interventions reducing growth hormone and IGF-1 signalling produce the longest-lived mice known.
The lesson is not that supplements are useless. It is that public enthusiasm and rigorous replication have diverged sharply in this field, and that a compound with a good mechanistic story and no lifespan effect in a well-controlled programme deserves scepticism.
What Works in Humans
No intervention has been demonstrated to extend human lifespan in a randomised trial, because such a trial would take decades and has never been run. What exists is strong evidence for interventions that reduce mortality and preserve function.
- Not smoking. The largest single modifiable factor, associated with roughly a decade of life expectancy difference.
- Cardiorespiratory fitness. The association between fitness and all-cause mortality is among the strongest in epidemiology, larger than for most single clinical measures, and it is dose-responsive across the range.
- Muscle mass and strength. Grip strength and muscle mass independently predict mortality and functional independence.
- Metabolic health. Glycaemic control, visceral adiposity and lipid handling are individually and collectively associated with outcomes, and all are modifiable.
- Sleep. Duration and regularity both associate with mortality, with a U-shaped curve and regularity emerging as an independent factor.
- Dietary pattern. Patterns emphasising plants, legumes, fish, whole grains and unsaturated fats associate consistently with lower mortality. Specific nutrient claims fare much worse than pattern-level evidence.
- Social connection. Social isolation associates with mortality at magnitudes comparable to major physical factors, which is consistently underweighted in optimisation-focused discussion.
- Blood pressure control. One of the few areas where randomised intervention data on hard outcomes is abundant.
None of these is novel. All of them outperform anything currently available in a capsule, and any honest account of longevity science has to lead with that rather than bury it.
What Is Being Tested
The pharmacological pipeline is real and worth following without assuming it is available.
Rapamycin and rapalogs. The most convincing animal evidence of any compound. Human trials in older adults have examined immune function and other endpoints, with intermittent low-dose regimens used to limit immunosuppressive and metabolic effects. Use outside trials happens and is not evidence-based.
Metformin. Observational data suggested favourable outcomes in people taking it for glycaemic control, and a dedicated trial designed to test whether it delays multiple age-related conditions has been proposed and repeatedly delayed by funding. Notably, metformin may blunt some exercise adaptations, which complicates its use in fit populations.
Senolytics. Early-phase human trials in progress. Mechanistically compelling, human-unproven.
GLP-1 receptor agonists. Substantial effects on weight and cardiometabolic outcomes, with active investigation of whether the benefits extend beyond those mechanisms. Muscle mass preservation during weight loss is the open question most relevant to longevity.
Partial reprogramming. The most scientifically radical direction, currently animal-only, with tumour risk as the central obstacle.
Plasma-based and circulating factor approaches. Early and unresolved.
How to Read This Field Without Being Misled
Five habits separate useful reading from noise.
Check the organism. A striking result in worms, flies or mice is a hypothesis about humans, not a finding in humans. This single check filters most overclaiming.
Check the endpoint. A change in a biomarker is not a change in an outcome. Many interventions move markers and fail on outcomes.
Check the comparator. Caloric restriction results in primates diverged largely because the control groups differed. What an intervention is compared against determines what the result means.
Check who is selling. A large share of the most quoted supplement research is funded by parties with a commercial interest. It is not therefore wrong, and it is not therefore reliable.
Check the effect size against the boring baseline. An intervention offering a small improvement in a biomarker is not comparable with the mortality association of cardiorespiratory fitness. Order of magnitude matters more than novelty.
The AEONNN Perspective
AEONNN's position in this field is deliberately narrow: it organises what is known into a coherent personal picture, and it is explicit about confidence. That means the Pillar Matrix leads with the systems where evidence and modifiability are highest, and supplement intelligence sits within that structure rather than in front of it. A platform that presented capsules as the centre of longevity practice would be misrepresenting its own evidence base.
The Evidence Levels exist for the same reason. Level A means clinically substantiated, Level B emerging, Level C experimental and educational. Most of the compounds that dominate longevity conversation sit at B or C, and a member deserves to see that rather than to infer it.
The lifespan and healthspan distinction is also why continuity is the product rather than a single assessment. Compressing the period of dysfunction at the end of life is the result of decades of maintained behaviour across ten systems, not of an intervention chosen once. A system that persists, adapts and keeps a member aligned over years is addressing the actual mechanism of the outcome people want.
Pillar Matrix mapping
Database Matrix layers
- Evidence Layer (PubMed, Cochrane, ClinicalTrials.gov)
- Meta / Consensus Layer (JAMA, BMJ, specialty society positions)
- Population Layer (UK Biobank, NHANES)
- Mechanistic Layer (KEGG, Reactome, UniProt)
- Innovation Layer (bioRxiv preprints, patent filings)
Frequently Asked
What is the difference between lifespan and healthspan?
Lifespan is years lived. Healthspan is years lived in good function. In high-income countries the gap is roughly a decade, and compressing that period of dysfunction is arguably a more valuable objective than adding years.
Has anything been proven to extend human lifespan?
No intervention has been demonstrated to extend human lifespan in a randomised trial, because such a trial would take decades. Strong evidence exists for interventions that reduce mortality and preserve function, led by not smoking, cardiorespiratory fitness, metabolic health and blood pressure control.
Which compounds extended lifespan in rigorous animal testing?
In a multi-site programme testing compounds in genetically heterogeneous mice, rapamycin extended lifespan robustly, along with acarbose, 17-alpha-estradiol and canagliflozin, with sex differences. Resveratrol and nicotinamide riboside did not extend lifespan in that programme.
Why do some animals live so much longer than others?
Aging rates are set by biology that varies enormously across species. Naked mole rats show negligible increases in mortality rate with age, some bivalves live centuries, and bowhead whales exceed two hundred years. That variation is why aging is considered malleable in principle.
Is caloric restriction proven in humans?
A two-year randomised trial in healthy non-obese adults found favourable changes in several markers, and a secondary analysis found a slower pace of aging on a third-generation epigenetic measure. It has not been tested against human lifespan.
What is the single most valuable thing for longevity?
Not smoking, followed by cardiorespiratory fitness. The association between fitness and all-cause mortality is among the strongest in epidemiology and is dose-responsive across the whole range.
Why is social connection included in longevity science?
Because social isolation associates with mortality at magnitudes comparable to major physical factors in large meta-analyses. It is consistently underweighted in optimisation-focused discussion.
Evidence and review
Any dosage ranges cited here reflect the ranges used in published human trials, not personal recommendations. Evidence in this field moves, so this article is reviewed quarterly and carries its last-updated date above. Nothing here is intended as medical advice, and supplementation should be discussed with a qualified clinician, particularly alongside prescribed medication or an existing condition.