Longevity Medicine: What's Proven vs Hyped

The honest starting point

"Longevity medicine" promises to slow or reverse aging. The uncomfortable truth is that almost none of the interventions sold under that banner have been shown to extend human lifespan, and most have not even been shown to improve human healthspan. The biology is often plausible. The human proof is usually missing.

This page grades the toolkit the way we wish more clinics would: by asking a single question for each intervention. Has it changed a hard human outcome in a randomized controlled trial (RCT), or is the case built on animal models, mechanisms, and surrogate biomarkers? The honest answer reorders the field dramatically — and it tends to put cheap, well-validated basics like cardiorespiratory fitness (see VO2 max and longevity) above most of the expensive supplements and clinic protocols.

The evidence hierarchy we use

Not all evidence is equal. From strongest to weakest:

1. 1.Hard human outcomes in RCTs — randomized trials measuring events people care about (death, heart attack, stroke).
2. 2.Surrogate human outcomes in RCTs — randomized trials measuring markers believed to predict those events (blood pressure, insulin sensitivity, aging biomarkers).
3. 3.Observational human data — associations in populations, confounded and hypothesis-generating only.
4. 4.Animal and mechanistic data — lifespan extension in mice, or a plausible pathway in a dish.

The longevity industry routinely markets level-4 evidence as if it were level-1. Our grades push back.

What actually has human hard-outcome RCTs

GLP-1 receptor agonists (semaglutide). This is the one place the evidence is genuinely strong. In the SELECT trial — roughly 17,600 overweight or obese adults without diabetes — full-dose semaglutide cut major adverse cardiovascular events by about 20% over several years¹. That is a real, randomized, hard-outcome benefit, the kind of result the rest of the longevity field can only aspire to. But read the fine print: SELECT proves cardiovascular risk reduction, not lifespan extension or "anti-aging." Reduced heart attacks and strokes in a high-risk population is meaningful and may well translate into longer life, but the trial was not designed to measure aging, and inferring "slowed aging" from a cardiovascular endpoint is exactly the kind of leap this site exists to flag. It is the strongest card in the deck, and it is still not proof of slowed aging. We unpack the full trial set — SELECT, FLOW, STEP-HFpEF, and STEP 9 — and where the "longevity" leap creeps in, in GLP-1s for healthspan & longevity: the evidence.

A grade-at-a-glance

If we collapse the toolkit into one ranking by strength of human evidence, the order is uncomfortable for the industry: GLP-1 (hard-outcome RCT) sits alone at the top; caloric restriction follows on surrogate RCTs; NAD+ precursors trail with reliable biomarker engagement but modest, mixed outcomes; rapamycin, metformin, and senolytics cluster near the bottom on animal or mechanistic data with no completed human longevity RCT; and growth hormone sits below zero — actively cautioned against. Nothing on this list has been shown to extend human lifespan. Keep that hierarchy in mind every time a clinic presents its menu as uniformly "evidence-based."

What has decent human surrogate data

Caloric restriction. The CALERIE Phase 2 trial randomized non-obese adults to roughly 12% sustained calorie restriction for two years. It was feasible and improved cardiometabolic risk markers — predictors of healthspan, not healthspan itself². Mechanistic follow-up found CR altered thymic and adipose immunometabolism in plausibly beneficial ways³. This is the best human RCT data in the longevity field, but it sits at level 2: surrogate endpoints, not demonstrated longer life.

NAD+ precursors (NMN, NR). These reliably do one thing: raise blood NAD+. A placebo-controlled crossover showed oral nicotinamide riboside roughly doubles NAD+ in older adults and is well tolerated — but the signal on blood pressure and arterial stiffness was only suggestive, not significant⁴. A separate RCT found NMN improved muscle insulin sensitivity in prediabetic women⁵, and a small trial reported better aerobic capacity in amateur runners⁶. Against those positives, a physiologic study confirmed NAD+ rose but found no improvement in muscle function, aerobic capacity, or metabolism⁷, and a meta-analysis found no consistent, clinically meaningful effect on metabolic-syndrome parameters⁸. Honest grade: target engagement is real, outcomes are modest and mixed. We dig deeper in do NAD+ and peptides actually extend lifespan?, and walk through the trial-by-trial gap between hype and data in NAD+ for longevity: what the trials actually show.

What is animal or mechanistic-only

Rapamycin / mTOR inhibition. Rapamycin is the most reproducible pharmacologic lifespan extender in animals — but no rapalog has been shown to extend human lifespan or healthspan, and the human evidence base is early⁹. The PEARL trial, a decentralized one-year RCT of low-dose rapamycin in healthy adults, found acceptable safety and a few modest subgroup signals but missed its primary endpoint (visceral fat) and showed no demonstrated effect on aging biomarkers¹⁰ — see what the PEARL trial actually showed. Strong animal data, absent human-longevity proof — we separate the hype from the evidence in rapamycin for longevity: hype vs evidence.

Metformin. The case rests on observational data — one large study found metformin-treated diabetics had survival comparable to matched non-diabetic controls¹¹ — plus mechanistic mapping onto the hallmarks of aging¹². The proposed TAME trial was designed precisely because that proof does not yet exist¹³. We unpack the metformin case in full — including the exercise-adaptation caveat — in metformin for longevity: the TAME trial evidence, and cover both drugs together in rapamycin & metformin for longevity: the evidence.

Senolytics (dasatinib + quercetin, fisetin). Tiny early pilots show dasatinib plus quercetin can reduce senescent-cell burden in humans¹⁴ and improved some physical-function measures in a 14-person idiopathic pulmonary fibrosis study¹⁵. Reviews stress that human data remain limited to small early-phase pilots with unproven healthspan efficacy¹⁶. Proof of mechanism, not proof of benefit. We grade the prescription flagship in full — and why it's a chemo drug plus a flavonoid, not a supplement stack — in dasatinib + quercetin: how far along is the flagship senolytic?. The over-the-counter senolytic most people actually buy is the flavonoid fisetin — strong in mice, near-zero in human longevity data; we grade it in fisetin as a senolytic: what the evidence shows.

What is cautioned against

Growth hormone. GH is marketed as rejuvenation; the evidence runs the other way. A landmark systematic review found that in healthy older adults, GH produced small body-composition changes but no proven functional benefit and significantly more adverse events — edema, joint pain, gynecomastia, glucose intolerance¹⁷. Mechanistically, across species, lower GH/IGF-1 signaling is associated with longer life, not shorter¹⁸. GH is not a longevity therapy; it is a risk. The same caution extends to the GH-axis "peptides" sold alongside it — secretagogues like MK-677, GHRH analogs like tesamorelin — which we grade category by category in peptides for longevity: what's real and what's marketing.

The weakest tier: DTC and IV "longevity clinics"

Direct-to-consumer and concierge clinics selling IV "longevity" drips — high-dose vitamins, glutathione, NAD+ — sit on the thinnest evidence of all. A critical review concluded the evidence comes mostly from disease-specific or aesthetic contexts, that placebo-controlled trials are scarce and conflicting, and that validated aging biomarkers are rarely even measured¹⁹. We cover this in are longevity clinics worth it?, and we compare it against the other provider models — concierge clinics, DTC lab memberships, and single-product Rx — in longevity clinics vs lab memberships vs Rx telehealth.

How to judge a longevity provider

The field still largely lacks completed human healthspan or lifespan RCTs, and the methodology to run them is only now maturing²⁰. Use that reality as your filter:

• •Does the provider distinguish proven from plausible? A good clinic will tell you semaglutide has hard-outcome data and rapamycin does not. A clinic that pitches every item on its menu as equally validated has already failed the test.
• •Do they cite human RCTs or mouse studies? Animal lifespan data is interesting, not actionable. The most reproducible mouse result in the field — rapamycin — still has no human longevity RCT behind it.
• •Do they measure validated outcomes, or sell unmeasured "optimization"? Vague "longevity panels" without validated aging biomarkers are a red flag — and even the headline "biological age" number from an epigenetic-clock test is noisy and unproven as a personal target. If a clinic can't tell you what endpoint improved, it can't tell you the treatment worked.
• •Are they honest about risk? Anyone pitching growth hormone for anti-aging is ignoring the evidence — and the underlying biology, which points the opposite way.
• •Does the pricing match the evidence? Premium prices for IV drips and infusions with no human outcome data should give you pause, not a sense of exclusivity.

For our independently graded shortlist of providers that pass these tests, see our longevity clinic rankings. And if you're weighing whether to hire a clinician at all, see what is a longevity doctor (and do you need one)?.

Bottom line

The longevity toolkit is mechanistically rich and clinically humble. One intervention (GLP-1) has robust human hard-outcome data — for cardiovascular risk, not aging. Caloric restriction has the best human RCT data, but only on surrogates. Rapamycin, metformin, and senolytics have compelling animal or mechanistic support and essentially no completed human longevity RCTs. Growth hormone is cautioned against. Grade accordingly, and be skeptical of anyone who doesn't.