Skip to content
Aminoscope
← Research
Longevity

Cardarine (GW-501516): the evidence, and why the drug was abandoned

It is not a SARM — it is a PPARδ agonist. The rodent endurance data is real but narrower than quoted, the human data is entirely about lipids, and GlaxoSmithKline stopped the programme after two-year rodent studies produced tumours.

Priya Anand11 min read
Cardarine is a PPARδ agonist, not a SARM — and its development ended in discontinuation, not approvalTWO DIFFERENT RECEPTORScardarine (GW-501516)PPARδ agonistPPARδnuclear lipid-sensing receptorfatty-acid oxidation genesSARMsostarine, RAD-140, LGD-4033androgen receptortestosterone’s receptoranabolic muscle genesno androgenactivityA PROGRAMME THAT ENDED, NOT A DRUG THAT LAUNCHED2001designed at GSK2006–12human lipid trials2009104-week rodenttumour findingsdevelopment stops2013WADA health alerttodayapproved nowhereTHE MECHANISM IS REAL. THE REASON IT STOPPED IS ALSO REAL.

Cardarine — GW-501516, sometimes sold as endurobol — is the compound behind the “exercise in a pill” headline. It is also one of the most consistently miscategorised substances in the performance market: it sits on the same vendor pages as ostarine and RAD-140, it is described in the same breath as them, and it is not a SARM. It never was. The other thing the vendor pages leave out is why you can buy it from a research-chemical site but not from a pharmacy: GlaxoSmithKline stopped developing it after long-term rodent studies produced tumours, and no company has taken it forward since. Both facts are load-bearing, and both are documented. Here is the whole record.

That meter is specifically about endurance, because that is what cardarine is bought for. The lipid programme is a separate story and it did reach Phase 2 in humans before it was abandoned — we cover exactly what those trials found below.

The correction that matters most: cardarine is not a SARM

SARMs — selective androgen receptor modulators — do what the name says: they bind the androgen receptor, the same receptor testosterone uses, with the goal of driving muscle growth in some tissues while sparing others. Ostarine, RAD-140 and LGD-4033 all work this way. We break that class down separately in peptides vs SARMs.

Cardarine does none of that. It was developed at GlaxoSmithKline using combinatorial chemistry and structure-based drug design as a potent, subtype-selective agonist of PPARδ — peroxisome proliferator-activated receptor delta, a nuclear receptor that acts as a dietary lipid sensor and switches on genes for fatty-acid handling.[1] In the founding 2001 paper it raised expression of the reverse cholesterol transporter ABCA1 and, in insulin-resistant obese rhesus monkeys, produced a dose-dependent rise in HDL cholesterol with falls in small dense LDL, triglycerides and fasting insulin.[1] Androgen receptor: not involved, not targeted, not implicated.

This is not a pedantic distinction. It changes what you should expect and what you should worry about. Cardarine will not suppress your own testosterone the way a SARM can, because it never touches that axis. It also will not build muscle through that axis. And its actual risk profile — which is the reason it exists only as a research chemical — comes from a completely different place than a SARM’s.

Cardarine and SARMs are sold together and share nothing mechanistically: different receptor, different drug class, different risk story.
Cardarine (GW-501516)SARMs (ostarine, RAD-140, LGD-4033)
ReceptorPPARδ — a nuclear lipid-sensing receptorAndrogen receptor — testosterone's receptor
Drug classMetabolic modulator / PPARδ agonistNon-steroidal androgen receptor modulator
Primary intended effectShift fuel use toward fatty-acid oxidation; raise HDL, lower triglyceridesIncrease lean mass and strength via androgenic signalling
Effect on natural testosteroneNo androgen-receptor activity, so no direct suppression of the HPG axisSuppression of endogenous testosterone is a recognised class effect
What ended / limits developmentTwo-year rodent carcinogenicity findings; programme discontinuedLiver injury and cardiovascular signals; no SARM is FDA-approved either
WADA statusProhibited at all times under S4, hormone and metabolic modulatorsProhibited at all times under S1, anabolic agents
Cardarine and SARMs are sold together and share nothing mechanistically: different receptor, different drug class, different risk story. Oliver 2001, PNAS — PMID 11309497; Van Wagoner 2017, JAMA — PMID 29183075; WADA Prohibited List

“Exercise in a pill” — what the rodent work actually showed

The claim traces to one famous paper: Narkar and colleagues, in Cell in 2008, testing pathway-specific drugs on treadmill endurance in mice.[2] It is worth reading the finding precisely, because the popular version has drifted a long way from it.

What the paper reported is that the PPARβ/δ agonist and exercise training synergistically increased oxidative myofibres and running endurance in adult mice. The drug amplified training. It did not substitute for it. The result everyone quotes — sedentary animals running 44% further with no training at all — came from the other arm of the same study, four weeks of AICAR, an orally active AMPK agonist. AICAR is a different molecule with a different mechanism. Attributing its sedentary-mouse result to cardarine is the single most common error in writing about this compound.[2]

The underlying biology is nonetheless coherent: PPARδ activation upregulates the machinery of fatty-acid oxidation in skeletal muscle, remodels fibre type toward oxidative slow-twitch, and overlaps substantially with the transcriptional response to endurance training — which is exactly why the “exercise pill” literature keeps returning to this receptor.[3] It is a real target. The gap is between a target and a drug. For a newer compound making a very similar promise from a different receptor, and running into a very similar evidence gap, see SLU-PP-332.

What was actually tested in humans

Cardarine did get into people — hundreds of them — and the trials were competently done and published in real journals. None of them was about endurance.

  • First-in-human, 2007. Twenty-four healthy volunteers, hospitalised and deliberately sedentary, given placebo (n=6), 2.5 mg (n=9) or 10 mg (n=9) daily for two weeks. HDL cholesterol rose in both drug groups while falling 11.5% in placebo; triglycerides trended down at 10 mg and post-fat-meal triglyceride clearance improved.[4] Note the design: the very first human study removed exercise on purpose.
  • Metabolic syndrome, 2008. Six moderately overweight men per arm, 10 mg daily for two weeks against a PPARα comparator and placebo. GW-501516 cut fasting triglycerides 30%, apolipoprotein B 26%, LDL cholesterol 23% and insulin 11%, reduced liver fat 20% and urinary isoprostanes (a global oxidative-stress marker) 30%, and increased skeletal-muscle CPT1b expression — direct evidence the fat-oxidation mechanism operates in humans.[5]
  • The largest trial, 12 weeks. 268 patients with low HDL cholesterol randomised to 2.5, 5.0 or 10.0 mg or placebo. HDL cholesterol rose up to 16.9% and apoA-I 6.6%; LDL cholesterol fell 7.3%, triglycerides 16.9%, apoB 14.9% and free fatty acids 19.4%. A smaller companion study (n=37) showed fewer VLDL and IDL particles and more medium and large HDL particles.[6] This is the high-water mark of the human evidence, and it is a lipid result.
  • Mechanism in dyslipidaemic men, 2011. Thirteen men with central obesity, randomised double-blind crossover, 2.5 mg/day for six weeks. Stable-isotope kinetics showed the triglyceride effect came from faster VLDL clearance and lower apoC-III production, and the HDL effect from increased apoA-II production.[7]

Add it up: the longest human exposure on record is twelve weeks, the population was dyslipidaemic or metabolically unhealthy adults, and every primary endpoint was a lipid or a lipoprotein. That is a legitimate Phase 2 signal for a cardiometabolic drug. It is not evidence about athletic performance, and it is not long-term safety data.

The rodent column carries both the endurance promise and the reason development stopped. The human column contains a real lipid result and nothing else.
EndpointRodent / preclinicalHuman
Running enduranceYes — increased, but in combination with exercise trainingNever measured, at any dose or duration
VO₂ max, time trial, work capacityNot the endpoint usedNever measured
Fatty-acid oxidation / fibre typeYes — oxidative myofibre shiftIndirect: increased muscle CPT1b, more meal fat oxidised (n≈6 per arm, 2 weeks)
HDL cholesterolYes — dose-dependent rise in obese rhesus monkeysYes — up to +16.9% over 12 weeks (n=268)
TriglyceridesYesYes — −16.9% (12 weeks); −30% at 10 mg (2 weeks)
Liver fatYes−20% in a 2-week study of 6 men
Cardiovascular eventsNot assessedNever assessed — no outcome trial exists
Tumour formationYes — tumours after 104 weeks of dosing in rats and miceNo data whatsoever; longest human exposure is 12 weeks
Long-term safetyThis is precisely what ended the programmeNone
The rodent column carries both the endurance promise and the reason development stopped. The human column contains a real lipid result and nothing else. Narkar 2008, Cell — PMID 18674809; Risérus 2008, Diabetes — PMID 18024853; Olson 2012, ATVB — PMID 22814748; Mitchell 2019, Pulm Circ — PMID 30351241

The endurance claim, checked from the other side

“No human endurance trial exists” is a negative claim, so we went looking for one rather than assuming. Searching the published literature for GW-501516 against exercise capacity, VO₂ max and athletic performance returns nothing in humans. A search of ClinicalTrials.gov returns exactly three registered GW-501516 studies, all sponsored in the GSK era: two lipid trials and one study of body fat and inflammation. None has a performance endpoint. There is no registered, completed or terminated human trial of cardarine and endurance.

Two near-misses deserve to be named, because they are the closest anything has come and because neither counts:

  • Mavodelpar (REN001) — a different PPARδ agonist — was tested in a randomised, double-blind, placebo-controlled Phase 2 trial of 213 patients with primary mitochondrial myopathy, with change in distance walked in a 12-minute walk test as the primary endpoint over 24 weeks. That is a genuine human exercise-capacity endpoint for the receptor class. It does not transfer to cardarine: different molecule, a rare disease population with impaired mitochondrial function rather than healthy athletes — and it did not work. Posted results show a mean improvement of 26.75 m on drug versus 30.89 m on placebo.[20] If anything, it is evidence against the extrapolation people make.
  • Bocidelpar (ASP0367), another selective PPARδ modulator, reached humans in a double-blind Phase 1 study of single and multiple ascending doses in healthy adults, lasting 1 and 14 days respectively.[19] Phase 1 measures blood levels and tolerability, not performance, and again it is not cardarine.

So the honest statement is narrow and firm: every endurance claim made for cardarine rests on rodent treadmill data, and on rodent treadmill data in which the drug worked alongside training rather than instead of it.

The cancer question, stated precisely

Two independent lines of published, peer-reviewed work show PPARδ activation by GW-501516 promoting tumour growth in animals, and they are worth citing specifically rather than gesturing at:

  • In Apcmin mice, which are genetically predisposed to intestinal polyposis, GW-501516 significantly increased the number and size of intestinal polyps — most strikingly a fivefold increase in polyps larger than 2 mm.[10]
  • In carcinogen-initiated mice, GW-501516 drove highly metastatic squamous cell carcinomas of the forestomach within two months. The authors describe the model as one in which tumorigenesis is dependent on the PPARδ agonist.[11]

Now the counterweight, because being honest cuts both ways. The role of PPARβ/δ in carcinogenesis is genuinely contested in the field: a 2015 review in Trends in Endocrinology & Metabolismopens by noting that conflicting studies indicate PPARβ/δ both inhibits and promotes tumorigenesis, and that context-specific mechanisms may explain the contradiction.[12] Some laboratories report PPARβ/δ activation suppressing tumour growth. This is not a settled mechanism.

Three things follow, and they should be held together. First, a 104-week rodent bioassay at pharmacology-driven doses is not the same as a human dose over a human timeframe, and the standard cautions about species differences in rodent carcinogenicity apply. Second, we have no human cancer data on cardarine at all, and we never will unless someone runs a study nobody is going to run — twelve weeks is the longest anyone has ever taken it under observation. Third — and this is the part the debate does not rescue — the finding was severe enough that a sponsor walked away from a late-stage cardiometabolic asset with a genuine efficacy signal, and no company has picked it up in the two decades since. That is an unusually loud market signal about a molecule that had already cleared Phase 2 on its primary endpoint.

Regulatory status, sport, and what is actually in the bottle

Cardarine is not approved as a medicine anywhere in the world, and it never reached market. It is not a lawful dietary-supplement ingredient either. In a warning letter issued in December 2025, the FDA told a vendor that products on its site — naming “Cardarine GW501516” alongside ostarine, ibutamoren and ligandrol — are unapproved new drugs under section 505(a) of the Federal Food, Drug, and Cosmetic Act, and noted that the firm marketed the whole group under the SARM label.[13] That is the regulator documenting the same category error this page opened with.

In sport it is prohibited at all times — in and out of competition — under section S4 of the WADA Prohibited List, hormone and metabolic modulators, where PPARδ agonists are named explicitly with GW1516/GW501516 given by its full chemical name.[14] In March 2013 WADA took the unusual step of issuing a public health alert about it, warning athletes that the compound had been withdrawn from development because of serious toxicities found in preclinical studies and stating that clinical approval had not and would not be given.[15] Anti-doping bodies almost never editorialise about the health of people trying to cheat; that they did here is itself information. Detection is not an obstacle either — the sulfoxide and sulfone metabolites were characterised and built into routine doping controls back in 2010.[16] We cover the wider prohibited-substance landscape in banned peptides and WADA.

Then there is the question of what you would actually be taking. In a JAMA analysis of 44 products bought online and sold as SARMs, only 52% contained a SARM at all; another 39% contained a different unapproved drug, and GW-501516 was one of the compounds found in that undeclared group. Across the sample, no active compound at all was detected in 9% of products, substances not listed on the label were present in 25%, and the amount found matched the label in only 41%.[17] A five-year European market-surveillance study of 324 seized samples across 13 countries found cardarine among the top five molecules detected, mostly from illegal distribution, with the majority containing active — and sometimes over-stated — doses.[18] The practical translation: the label is close to meaningless, and the two failure modes run in opposite directions.

The honest bottom line

Cardarine is a well-designed molecule with a mechanism that works. PPARδ activation really does shift fuel use toward fatty-acid oxidation, and that really does show up in humans as measurable changes in muscle gene expression, meal fat handling and blood lipids.[5][6] The lipid result — HDL up, triglycerides down, apoB down, over twelve weeks in 268 patients — is the kind of Phase 2 signal that normally buys a compound a Phase 3.[6]

It did not get one. The programme ended because two-year rodent studies produced tumours, and published animal work independently shows the same agonist accelerating tumour growth in two different models.[8][10][11] Set against that: no human trial has ever tested the thing cardarine is actually bought for, the strongest rodent endurance result required training alongside the drug rather than replacing it, and the closest human exercise-capacity trial in the receptor class — a different, better-characterised PPARδ agonist — came out no better than placebo.[2][20]

So the ledger reads: an unquantified but non-trivial cancer signal, on the debit side, against a benefit that has never been demonstrated in a human being on the credit side. There is no dose that makes that trade work, which is why this page does not contain one. If the underlying goal is endurance, the interventions with actual human outcome data are unglamorous and available — structured training, and the nutritional ergogenics that have been through real trials.

This article is research information, not medical advice, and it is not a protocol. Cardarine (GW-501516) is an unapproved investigational drug whose development was discontinued after rodent carcinogenicity findings; it is not approved for human use in any country, it is not a lawful dietary-supplement ingredient, and the FDA has issued warning letters to companies selling it. It is prohibited at all times in sport, and a positive test carries the sanctions attached to a non-specified substance. Products sold online are frequently mislabelled in both directions. Nothing here should be read as a recommendation to obtain or use it; if you have taken it, tell a clinician rather than a forum.

Reviewed against primary sources by the Aminoscope desk

Frequently asked

Is cardarine a SARM?
No. Cardarine (GW-501516) is a PPARδ agonist — it activates peroxisome proliferator-activated receptor delta, a nuclear receptor that senses lipids and switches on genes for fatty-acid handling. SARMs such as ostarine, RAD-140 and LGD-4033 act on the androgen receptor, the same receptor testosterone uses. The two have no mechanistic overlap: cardarine has no androgen-receptor activity at all. It is grouped with SARMs purely because it is sold on the same websites, and the confusion is so entrenched that the FDA's own warning letters describe vendors marketing 'Cardarine GW501516' under the SARM heading. The distinction matters practically: cardarine will not suppress your natural testosterone the way a SARM can, and it also will not build muscle through that pathway.
Does cardarine cause cancer?
In rodents, prolonged dosing produced tumours, and that finding ended the drug's development. GlaxoSmithKline discontinued research on GW-501516 after preclinical rodent studies linked it to widespread tumour development, and the peer-reviewed account states that the compound causes cancer in rats and mice after 104 weeks of dosing. Independent published work points the same way: GW-501516 produced a fivefold increase in intestinal polyps larger than 2 mm in Apc-min mice, and drove metastatic forestomach carcinomas within two months in carcinogen-initiated mice. Two honest caveats: the two-year carcinogenicity studies themselves were presented as sponsor conference abstracts in 2009 and were never published in full, so no public paper lists every affected organ; and the broader role of PPARβ/δ in cancer is genuinely contested, with some laboratories reporting the opposite effect. There is no human cancer data on cardarine — the longest anyone has taken it under observation is twelve weeks. What can be said plainly is that a sponsor abandoned a late-stage asset with a real efficacy signal because of this finding, and nobody has picked it up since.
Does cardarine actually work for endurance?
No human being has ever been tested. There is no published human trial of cardarine measuring endurance, VO₂ max, time-trial performance or work capacity, at any dose, for any duration, and ClinicalTrials.gov lists only three registered GW-501516 studies — all GSK-era lipid or body-fat work. The entire endurance case rests on the 2008 Cell paper that made the compound famous, and that paper is narrower than it is usually quoted: the PPARδ agonist increased running endurance in mice when combined with exercise training. The frequently cited '44% further without training' result belongs to AICAR, a different drug tested in the same study. The nearest human evidence in the receptor class is a Phase 2 trial of a different PPARδ agonist, mavodelpar, in 213 patients with mitochondrial myopathy, which used a 12-minute walk test as its primary endpoint over 24 weeks — and it came out no better than placebo.
Is cardarine legal?
It is not approved as a medicine in any country and never reached market, and it is not a lawful dietary-supplement ingredient. The FDA has issued warning letters to vendors stating that products marketed as SARMs including 'Cardarine GW501516' are unapproved new drugs under section 505(a) of the Federal Food, Drug, and Cosmetic Act. It is typically sold under 'research chemical, not for human consumption' labelling, which is a legal posture rather than a safety one. In sport it is prohibited at all times, in and out of competition, listed by name under section S4 of the WADA Prohibited List as a PPARδ agonist, and its urinary sulfoxide and sulfone metabolites have been part of routine doping controls since 2010.
What are you actually buying when you buy cardarine online?
Frequently not what the label says. In a JAMA analysis of 44 products purchased online and sold as SARMs, only 52% contained a SARM, 39% contained a different unapproved drug — GW-501516 among them, in products that did not declare it — no active compound at all was found in 9%, undeclared substances were present in 25%, and the amount matched the label in only 41%. A five-year European market-surveillance study of 324 seized samples across 13 countries found cardarine among the five most commonly detected molecules, mostly from illegal distribution, with most samples containing active and sometimes over-stated doses. So the two failure modes run in opposite directions: you may be taking nothing, something else entirely, or considerably more than you intended.

Sources

  1. [1] Oliver WR Jr, Shenk JL, Snaith MR, et al. (2001). A selective peroxisome proliferator-activated receptor delta agonist promotes reverse cholesterol transport. Proc Natl Acad Sci U S A. PMID 11309497
  2. [2] Narkar VA, Downes M, Yu RT, et al. (2008). AMPK and PPARdelta agonists are exercise mimetics. Cell. PMID 18674809
  3. [3] Li S, Laher I. (2015). Exercise Pills: At the Starting Line. Trends Pharmacol Sci. PMID 26439443
  4. [4] Sprecher DL, Massien C, Pearce G, et al. (2007). Triglyceride:high-density lipoprotein cholesterol effects in healthy subjects administered a peroxisome proliferator activated receptor delta agonist. Arterioscler Thromb Vasc Biol. PMID 17110604
  5. [5] Risérus U, Sprecher D, Johnson T, et al. (2008). Activation of peroxisome proliferator-activated receptor (PPAR)delta promotes reversal of multiple metabolic abnormalities, reduces oxidative stress, and increases fatty acid oxidation in moderately obese men. Diabetes. PMID 18024853
  6. [6] Olson EJ, Pearce GL, Jones NP, Sprecher DL. (2012). Lipid effects of peroxisome proliferator-activated receptor-δ agonist GW501516 in subjects with low high-density lipoprotein cholesterol: characteristics of metabolic syndrome. Arterioscler Thromb Vasc Biol. PMID 22814748
  7. [7] Ooi EM, Watts GF, Sprecher DL, Chan DC, Barrett PH. (2011). Mechanism of action of a peroxisome proliferator-activated receptor (PPAR)-delta agonist on lipoprotein metabolism in dyslipidemic subjects with central obesity. J Clin Endocrinol Metab. PMID 21816786
  8. [8] Mackenzie LS, Lione L. (2013). Harnessing the benefits of PPARβ/δ agonists. Life Sci. PMID 24184294
  9. [9] Mitchell JA, Bishop-Bailey D. (2019). PPARβ/δ a potential target in pulmonary hypertension blighted by cancer risk. Pulm Circ. PMID 30351241
  10. [10] Gupta RA, Wang D, Katkuri S, Wang H, Dey SK, DuBois RN. (2004). Activation of nuclear hormone receptor peroxisome proliferator-activated receptor-delta accelerates intestinal adenoma growth. Nat Med. PMID 14758356
  11. [11] Pollock CB, Rodriguez O, Martin PL, et al. (2010). Induction of metastatic gastric cancer by peroxisome proliferator-activated receptor δ activation. PPAR Res. PMID 21318167
  12. [12] Peters JM, Gonzalez FJ, Müller R. (2015). Establishing the Role of PPARβ/δ in Carcinogenesis. Trends Endocrinol Metab. PMID 26490384
  13. [13] U.S. Food and Drug Administration (2025). Warning Letter: Dynamic Health Group dba SARMS AMERICA (MARCS-CMS 719257) — products marketed as SARMs, including "Cardarine GW501516", are unapproved new drugs. FDA Office of Regulatory Affairs. Source
  14. [14] World Anti-Doping Agency (2026). World Anti-Doping Code International Standard: Prohibited List — S4.4 Metabolic Modulators, peroxisome proliferator-activated receptor delta (PPARδ) agonists (GW1516, GW501516). WADA. Source
  15. [15] World Anti-Doping Agency (2013). WADA issues alert on GW501516. WADA news release, 21 March 2013. Source
  16. [16] Thevis M, Möller I, Thomas A, et al. (2010). Characterization of two major urinary metabolites of the PPARdelta-agonist GW1516 and implementation of the drug in routine doping controls. Anal Bioanal Chem. PMID 19946680
  17. [17] Van Wagoner RM, Eichner A, Bhasin S, Deuster PA, Eichner D. (2017). Chemical Composition and Labeling of Substances Marketed as Selective Androgen Receptor Modulators and Sold via the Internet. JAMA. PMID 29183075
  18. [18] Barrios MM, Deconinck E, Vanhee C, et al. (2025). SARMs, Metabolic Modulators and Growth Hormone Secretagogues in Suspected Illegal Medicines, Bought as Sport Performance Enhancers: A Retro- and Prospective Study Within the GEON. Drug Test Anal. PMID 40551438
  19. [19] Ito M, Tauscher-Wisniewski S, Smulders RA, et al. (2022). Single- and multiple-dose safety, tolerability, pharmacokinetic, and pharmacodynamic profiles of ASP0367, or bocidelpar sulfate, a novel modulator of peroxisome proliferator-activated receptor delta in healthy adults: Results from a phase 1 study. Muscle Nerve. PMID 34642949
  20. [20] Reneo Pharma Ltd (STRIDE) (2023). A Double-blind, Placebo-controlled Study to Evaluate the Efficacy and Safety of 24 Weeks Treatment With REN001 (mavodelpar) in Patients With Primary Mitochondrial Myopathy — NCT04535609, posted results: 12-minute walk test change 26.75 m (drug) vs 30.89 m (placebo). ClinicalTrials.gov. Source

Compare it

More in Longevity