Anti-aging treatment in China

Six unrelated things are sold under one phrase, and they do not share an evidence base. Chinese regulation cuts the shelf in half in a way almost nothing written about it explains: cell therapy is licensed against a disease and never against age, and everything else on the shelf is untouched by that rule. Here is which half is which, and what each is worth.

The short answer

“Anti-aging treatment” is not a category of medicine. It is a shelf, and at least six unrelated products sit on it: lifestyle and physiology, diagnostics, aesthetic medicine, hormones, molecules and supplements, and cell therapy. They have nothing in common except the promise on the label, and they differ enormously in how much human evidence stands behind them — from “overwhelming” at one end to “nine patients in a pilot study” at the other.

In mainland China the shelf gets split down the middle by regulation, and the split is unusually clean. Cell and gene therapy is licensed against a named disease and never against age, under the frame that State Council Order No. 818 put in force on 1 May 2026. Everything else — the screening battery, the non-cell infusion and apheresis protocols, the heat and oxygen work, the sleep and food and movement side — is ordinary licensed hospital practice and carries no such restriction. So the honest map of anti-aging treatment in China has two halves, and knowing which half you are standing in tells you more than any brochure will.

What follows is that map: what each class of intervention actually is, what it has and has not shown in humans, what a biological-age score is worth, and what can lawfully be obtained here.

Six things wearing one name

Most disappointment in this market is manufactured in the gap between column one and column three. A clinic sells you class six with the evidence base of class one quoted underneath it, and the two are never connected on the page.

Why the phrase does not survive contact with Chinese law

This is the part that surprises people who arrive expecting the opposite. China is frequently described abroad as somewhere you can buy what your own regulator will not let you have. For cell therapy, since 2026, that is simply not the case, and the rule is easy to state: a licensed program is licensed against a diagnosed condition. Age is not one. A provider anywhere on the mainland offering a stem-cell infusion to a healthy person as an anti-aging purchase is not operating under the pilot zone’s licence, whatever the brochure says, and the National Health Commission has specifically warned against “anti-ageing injection” marketing. The full mechanism — the 2015 research-only rules, what Order 818 changed, and how to tell a lawful program from a clinic advertisement — is set out in the legality guide rather than repeated here.

The constructive half matters just as much, and it is routinely lost in the reporting. Nothing in that rule touches the non-cell work. Therapeutic plasma exchange, mitochondrial and NAD+ protocols, whole-body hyperthermia, hyperbaric oxygen, functional infusion: these are ordinary procedures delivered by licensed hospitals and they need no disease indication at all. They also, and this cuts the other way, carry none of the scaffolding that Order 818 imposes on a cell program — no filed indication, no published candidacy criteria, no release-tested product. A protocol that needs no licence is not thereby a protocol that has been shown to work. It simply means the regulator was never asked the question.

What the evidence actually supports

The best-evidenced interventions are the unglamorous ones

Nothing sold under the anti-aging label has evidence approaching that for cardiorespiratory fitness, muscle mass, sleep, diet and not smoking. That is not a platitude used to fill a page; it is the reason a serious program measures those things first and builds everything else on top of them. The nearest thing to a controlled human trial of an aging intervention remains CALERIE, the two-year randomised study of sustained caloric restriction in healthy non-obese adults (Kraus and colleagues, The Lancet Diabetes & Endocrinology, 2019), which improved cardiometabolic risk factors; a later analysis of the same cohort (Waziry and colleagues, Nature Aging, 2023) reported a roughly two to three per cent slowing of one pace-of-aging measure. That is a real result, and it is also a surrogate endpoint from a single trial — worth knowing precisely because it is the high-water mark of the field, not a footnote to something stronger.

The molecules

Metformin. An old, cheap, well-characterised diabetes drug with epidemiological signals that have driven a decade of interest. The trial designed to test the question directly — TAME, Targeting Aging with Metformin, led by Nir Barzilai — has been designed and discussed for years and has not reported. A pilot in older adults (MILES) found transcriptomic changes; a separate randomised study (Konopka and colleagues, 2019) found metformin blunted the improvement in insulin sensitivity and mitochondrial respiration that exercise training would otherwise have produced. It is not approved as an anti-aging drug anywhere, and the interaction with the best-evidenced intervention in the field is a reason for caution rather than a detail.

Rapamycin and analogues. The strongest pre-clinical case in the field: rapamycin extended lifespan in genetically heterogeneous mice in the National Institute on Aging’s Interventions Testing Program, even when started late in life (Harrison and colleagues, Nature, 2009). The human record is thinner and about immune function, not lifespan — a randomised trial of the analogue everolimus improved influenza-vaccine response in adults over sixty-five (Mannick and colleagues, Science Translational Medicine, 2014) — and subsequent off-label low-dose studies in healthy adults have been small and short. No human study has a lifespan endpoint, and none will for a long time.

Senolytics. The idea — clear senescent cells and the tissue behaves younger — is elegant and well supported in animals. The human data are two open-label pilots, of fourteen patients with idiopathic pulmonary fibrosis and nine with diabetic kidney disease, both published in EBioMedicine in 2019 by Justice and Hickson and colleagues respectively. Those are feasibility studies. Treating them as evidence of an anti-aging effect in a healthy person is a category error that the supplement market makes daily.

NAD+ precursors. Nicotinamide riboside and NMN reliably raise circulating NAD+ in humans; the question is whether anything follows from that, and the functional trial record is where it gets thin. Because this is the single most-marketed molecule class in longevity medicine, it has a guide of its own, including what intravenous NAD+ does and does not add over the oral precursors, and why “mitochondrial therapy” names two completely different things.

Hormones

Growth hormone is the cautionary tale of this field. The study that started the industry — Rudman and colleagues in the New England Journal of Medicine in 1990 — enrolled twelve men, ran six months, and reported changes in lean mass and fat mass. It was never a longevity result. A systematic review of growth hormone in the healthy elderly (Liu and colleagues, Annals of Internal Medicine, 2007) concluded that the small body-composition changes came with meaningfully increased rates of soft-tissue oedema, joint pain, carpal tunnel syndrome and impaired fasting glucose, and did not support its use. Thirty-five years on, the marketing still cites the 1990 paper.

Menopausal hormone therapy is a different matter and deserves not to be lumped in. It is an effective, legitimate treatment for menopausal symptoms whose risk profile depends heavily on age at initiation and time since menopause — the reading that emerged after the Women’s Health Initiative results in 2002 were widely over-generalised. That makes it a treatment for a condition, prescribed and monitored by a physician who knows your history. It does not make it an anti-aging drug, and nobody should be sold it as one.

Blood, heat and oxygen

Three procedures have become fixtures of the longevity-clinic menu, and each has a real literature attached to a claim narrower than the one being sold. Therapeutic plasma exchange has strong evidence where a circulating target is named, and a single small unreplicated human longevity study. Hyperthermia has randomised evidence as an oncology adjunct and a hormesis argument borrowed from sauna cohort data. Hyperbaric oxygen rests, for this purpose, largely on one 2020 Israeli study of telomere length and senescent-cell fraction. Each is covered properly in its own guide, including the risks, which are not zero for any of the three.

Biological age, and what a clock is worth

Almost every anti-aging program now hands over a number: you are forty-eight, your biological age is thirty-nine. It is the most persuasive object in the industry and the one that most needs explaining.

These scores come from epigenetic clocks — statistical models trained on DNA-methylation patterns, from Horvath’s 2013 multi-tissue predictor through the later PhenoAge and GrimAge mortality-trained clocks to pace-of-aging measures such as DunedinPACE. They are genuine science and they do predict mortality and disease risk at the level of populations. Three things follow that the number on the page does not tell you.

They are surrogate endpoints. No randomised trial has shown that moving a clock with an intervention changes how long an individual lives, or even how long they stay well. A clock is a description of a state, validated as a correlate; treating it as a target assumes the causal step that has not been demonstrated.

They are noisier than they look. Test–retest reliability has been a known weakness of the first-generation clocks — run the same sample twice and the answer moves — which is why principal-component versions were developed specifically to address it (Higgins-Chen and colleagues, Nature Aging, 2022). A three-year improvement may be a three-year improvement, or it may be the assay.

They compress away the information you came for. A single composite score is built by discarding most of what the underlying panel knows. That is a defensible trade if you want one figure for a marketing page; it is a bad trade if you want to act on anything. It is also why the four-day assessment hands over the marker-by-marker figures and the imaging reads rather than a headline age, and says so plainly.

What is actually available here

Which brings the question back to what a guest on this island can obtain, and the answer follows the split set out above rather than cutting across it.

The four-day retreat, from $2,700, is a screening battery run inside a licensed hospital environment, read by a specialist panel, turned into a written plan, and followed for twelve months. That is the diagnostics-led and lifestyle end of the shelf, delivered properly, and it is the part of this that is defensible without qualification.

The non-cell longevity protocols — plasma exchange, the mitochondrial and NAD+ course, whole-body hyperthermia, hyperbaric oxygen, functional infusion — are available as session-based hospital procedures with no disease indication required, on the understanding set out earlier: no licence requirement also means no filed criteria, and the evidence for each is what its own guide says it is, no more.

Licensed cell therapy runs on the other half of the split entirely. It proceeds where the assessment finds a condition that has a filed indication, or where a guest arrives holding a diagnosis or prescription from their own physician. It is not available as an anti-aging purchase, it is never sold from a menu, and what a hospital quote is built from — and why there is no published price list for it — is its own subject.

The gap between that and an Alpine longevity week, which is published at roughly $39,000 to $115,000, is the subject of a separate country-by-country comparison, and the honest version of it is not a claim that one is better.

Six questions for any anti-aging clinic

These work in any country and they are deliberately awkward. A good provider answers all six in a sentence each.

1. Which of the six classes is this, actually? If the answer moves between classes as you press on it — a supplement defended with exercise data, a cell infusion defended with a diagnostics argument — that movement is the finding.
2. What is the regulatory route you are operating on, for this specific intervention? Ordinary practice, a clinical trial, or a special pathway. “It is legal here” is not an answer to that question.
3. What is the endpoint, and was it measured in humans? Ask for the trial, the number of participants and the duration. A mouse lifespan result and a fourteen-person pilot are both legitimate science and neither is a reason to buy.
4. Who reads my results, and what are their qualifications? The difference between a panel of specialists and an algorithm with a physician signature is not visible on the report.
5. What happens if something is found? The single most important question, and the one that separates a hospital-attached program from a resort. Referral out, on a foreign trip, is worse than it sounds.
6. What does the price include, and what is the next invoice? Ask what is excluded, in writing, before anything is scheduled.

A longer version of the same logic, built for comparing providers across countries rather than assessing one, is the eight-question list in the comparison guide.

What this page is not claiming

No program anywhere — here, in Switzerland, in Japan, in the United States — has demonstrated that it extends life. There is no randomised controlled trial of a longevity program with a lifespan or all-cause-mortality endpoint in any country, and the measures the industry reports in place of one are surrogates. What a serious program can defensibly offer is earlier detection, a correct baseline, access to licensed treatment if something is found, and a plan specific enough that you will actually follow it. Everything beyond that line is ahead of the evidence, at any latitude and at any price.

Frequently asked

What is the best anti-aging treatment?

The one with by far the strongest human evidence is not sold in a clinic: cardiorespiratory fitness, muscle mass, sleep, diet and not smoking. Every credible programme builds on those and measures them first. Beyond that, the honest ranking is by evidence class rather than by price. Diagnostics and lifestyle medicine are well founded. Aesthetic medicine works at what it claims, which is appearance. Hormone therapy is a treatment for a diagnosed condition. The molecule class — NAD+ precursors, metformin, rapamycin, senolytics — is mechanistically interesting and supported in humans only by small, short studies on surrogate endpoints. Cell therapy is real medicine inside a filed disease indication and is not an anti-aging purchase. Any clinic that will not tell you which class it is selling you is answering a different question.

Can you get anti-aging stem cell therapy in China?

No. Cell therapy in mainland China is licensed against a named disease indication and never against age, under the frame State Council Order No. 818 put in force on 1 May 2026, and the National Health Commission has specifically warned against “anti-ageing injection” marketing. Licensed cell therapy proceeds where a diagnosis exists — found by the assessment, or brought with you as a prescription or diagnosis from your own physician. A provider on the mainland offering a stem-cell infusion to a healthy person as a rejuvenation purchase is not operating under the pilot zone's licence, whatever the brochure says. The non-cell longevity protocols are a separate matter entirely and carry no such restriction.

Is anti-aging treatment in China legal?

It depends entirely on which class of treatment is meant, which is why the question is usually answered badly. Diagnostics, lifestyle medicine, aesthetic medicine, hormone therapy prescribed for a diagnosed condition, and the non-cell longevity protocols such as plasma exchange, hyperthermia, hyperbaric oxygen and NAD+ infusion are ordinary licensed hospital practice and are lawful. Cell and gene therapy is lawful only inside licensed institutions against filed disease indications; sold as a general anti-aging service to a healthy person it is not. The dividing line is the disease indication, not the substance, and the legality guide on this site sets out the mechanism in full.

How much does anti-aging treatment in China cost?

The four-day retreat described on this site is offered from about $2,700 all-in, covering the screening battery, the specialist panel that reads it, the villa, meals, transfers and twelve months of follow-up. Published Alpine longevity programmes sit at roughly $39,000 to $115,000 for a week, which is the comparison most people are working from. Licensed treatment programmes are a different kind of purchase and are never sold from a menu: each is priced by the hospital that would deliver it and confirmed in writing before anything is committed to, because what a course costs depends on what is being treated and how. Be sceptical of any online price table for this category, including comparisons that flatter us — the packages are not comparable line by line.

Does anti-aging treatment actually work?

That depends on what “work” is being claimed. No programme of any kind, in any country, has demonstrated that it extends life: there is no randomised controlled trial of a longevity programme with a lifespan or all-cause-mortality endpoint anywhere. What is demonstrated is narrower and still valuable — that fitness, strength, sleep and diet change disease risk substantially; that thorough screening finds treatable things earlier; that specific treatments work for specific diagnosed conditions. The surrogate measures the industry reports instead, including epigenetic age clocks, describe a state and have not been validated as targets. Earlier detection, a correct baseline and a plan you will follow is a defensible claim. Reversing aging is not.

What is a biological age test, and is it accurate?

It is a statistical model, usually trained on DNA-methylation patterns, that returns a single number from a blood sample — the family running from Horvath's 2013 multi-tissue clock through the mortality-trained PhenoAge and GrimAge clocks to pace-of-aging measures such as DunedinPACE. At population level they genuinely predict mortality and disease risk. At the level of one person and one result, three caveats matter: they are surrogate endpoints, with no trial showing that moving the number changes an individual's outcome; test–retest reliability has been a known weakness of the first-generation clocks, which is why principal-component versions were built to address it; and a composite score discards most of what the underlying panel knows. A marker-by-marker report is less persuasive and more useful.

What is the difference between anti-aging treatment and a longevity program?

Mostly vocabulary, and the vocabulary is commercial rather than clinical. “Anti-aging” is the older retail term and leans toward a single intervention sold as a product — an injection, a molecule, a facial protocol. “Longevity programme” is the newer term and usually implies measurement first: a panel, imaging, a plan, follow-up. Neither word is licensed or policed, so neither tells you anything reliable on its own. The two questions that do are what is being measured and who reads it, and what the provider is legally permitted to deliver if something turns up.

Or download the Retreat Dossier (PDF) → — itinerary, inclusions, programs, practicalities.