Cell, mouse, human, approved: how to tell how strong a peptide's evidence really is
Most compounds that look promising in a dish or a mouse never become medicines. A guide to the evidence ladder, the numbers behind it, and the gauge on every Synthos compound article.
- Cell
- Animal
- Human
- Approved
At a glance
- Evidence comes in rungs: cell studies, animal studies, human trials in three phases, regulatory approval, and monitoring after approval.
- In industry data from 2011 to 2020, a drug candidate entering its first human trial had a 7.9% chance of eventually being approved.
- A positive result on a lower rung is a reason to test further, not proof that something works in people.
- Warning signs: a literature built almost entirely in animals, or almost entirely by one research group.
- Every compound article in the Synthos Research Library shows an evidence gauge: CELL, ANIMAL, HUMAN or APPROVED.
"Studies show" can mean almost anything
In obese mice, a small fragment of human growth hormone called AOD-9604 increased fat burning and slowed weight gain [3]. It looked like a tidy story: the fat-burning part of a hormone, without the rest. Then it was tested properly in people. In a 24-week trial that randomised 502 participants, it failed to show a meaningful benefit on its main weight-loss endpoint [3].
Both results are real. They just sit on different rungs of the evidence ladder, and the higher rung wins. Knowing which rung a claim comes from is the single most useful skill for reading peptide research. Our peptides primer covers the chemistry; this guide covers the evidence.
The ladder, rung by rung
| Rung | What it is | What it can show | What it cannot show |
|---|---|---|---|
| Cell (in vitro) | Cells or tissues in a dish | A mechanism; whether a molecule can do something at all | What happens in a whole body, at realistic doses |
| Animal | Mice, rats, rabbits, sometimes larger animals | Effects in a living system; early safety signals | Whether the effect translates to humans |
| Phase 1 | Small human studies, often healthy volunteers | Safety, tolerability, how the body handles the drug | Whether it works |
| Phase 2 | Larger studies in patients | Early efficacy signals; dose finding | Firm proof of benefit |
| Phase 3 | Large, controlled, often multinational trials | Whether benefit outweighs harm in the target population | Rare or long-term harms |
| Approval | A regulator such as the FDA or EMA authorises use | That benefit outweighs risk for a specific use | That it works for anything else |
| Post-marketing | Surveillance and follow-up trials after approval | Rare side effects; confirmation of benefit | Nothing is ever final |
Each rung answers a narrower, harder question than the one below it. A peptide can look excellent in cells because cells cannot tell you about digestion, distribution, immune reactions or a placebo effect.
Why most mouse results never make it
The best numbers on attrition come from industry data. A 2021 analysis by the Biotechnology Innovation Organization, Informa Pharma Intelligence and QLS Advisors tracked 12,728 phase transitions across 9,704 development programmes from 2011 to 2020 [1]. Its findings:
- 52.0% of candidates passed phase 1.
- 28.9% passed phase 2, the biggest hurdle.
- 57.8% passed phase 3.
- 90.6% of applications filed with the FDA were approved.
Multiply those together and the overall likelihood of approval from phase 1 was 7.9%, after an average of 10.5 years [1].
Out of every 100 drug candidates that reach their first human trial, about 8 end up approved. The rest fail on the way, and they had already passed cell and animal testing to get there.
The view from the animal side is similar. When Toronto researchers followed 76 highly cited animal studies, only about a third (28) were later replicated in human randomised trials [2]. The usual explanations are that animals differ from people in metabolism and physiology, that animal disease models are simplified versions of human illness, and that positive animal studies are more likely to be published than negative ones.
Four compounds, four rungs
The Synthos Library includes examples from every part of the ladder.
AOD-9604: worked in mice, failed in people. Mouse studies showed increased fat oxidation and reduced weight gain, but the 24-week human trial in 502 people missed its primary endpoint [3]. A later review notes that this conclusion rests mainly on sponsor-reported results rather than a peer-reviewed efficacy paper, another reminder to ask where a claim was published [3].
SS-31 (elamipretide): failed, then approved for something narrower. In MMPOWER-3, a phase 3 trial in 218 adults with primary mitochondrial myopathy, it missed both primary endpoints; the difference in six-minute walk distance versus placebo was −3.2 metres (p = 0.69) [4]. Then, on 19 September 2025, the FDA granted it accelerated approval as Forzinity for Barth syndrome, a rare genetic disease, based on improved knee-extensor muscle strength, a measure the agency judged "reasonably likely to predict" patient benefit [5]. The company must run a randomised confirmatory trial [5]. Approval is always for a specific use.
Tesamorelin: the full ladder. Two phase 3, placebo-controlled trials enrolled 806 people with HIV and excess abdominal fat; after 26 weeks, visceral fat fell by 15.4% relative to placebo [6]. The FDA approved it on 10 November 2010 [7], for that indication only.
BPC-157: mostly animals. A 2026 systematic review found just three human studies, with 30 participants in total, all judged to be of poor methodological quality; the overall certainty of evidence was rated very low [8]. Underneath sits a large literature that is mostly animal work [8], which brings us to the next warning sign.
When most of the papers come from one lab
Independent replication is how science checks itself. A finding that only one group can produce may be real, but it has not yet survived that test.
For BPC-157, a 2025 analysis found that more than 80% of over 190 PubMed articles listed one of two researchers from a single research group as first or senior author, while independent labs had contributed only a handful of mostly short-term studies [9]. The authors argued that heavy reliance on self-replication increases the risk of confirmation bias [9]. That does not make the results wrong. It does mean the literature is less independent than its size suggests.
How to read a health headline
Exaggeration often starts before journalists get involved. In a study of 462 press releases from 20 leading UK universities, 36% overstated how animal findings applied to humans. When the press release exaggerated, 86% of the resulting news stories did too, compared with 10% when it did not [10].
A few questions cut through most headlines:
- What was studied? Cells, animals or people?
- How many? A 12-person pilot and a 1,000-person trial are different things.
- Was there a control group? Without one, improvement may be placebo, chance or natural recovery.
- What was measured? A blood marker or a scan is not the same as people feeling or living better.
- Who did it, and has anyone else? One lab, or several independent ones?
- Where was it published? A peer-reviewed paper, a press release or a social-media post?
The Synthos evidence gauge
Every compound article in the Synthos Research Library shows a four-step gauge: CELL · ANIMAL · HUMAN · APPROVED. It marks the highest rung reached for the use the article discusses.
- CELL: only in vitro studies.
- ANIMAL: studies in animals, with no published human trials.
- HUMAN: published human trials, from small pilots upwards.
- APPROVED: approved as a medicine by the FDA, the EMA or another major regulator, for some indication.
The gauge is a starting point, not a verdict. HUMAN can mean a single small pilot. APPROVED applies to a specific indication, so an approved peptide is not automatically supported for other uses. And the gauge says nothing about safety, which is why every article also has a section on what we don't know yet.
FAQ
What is the difference between phase 1, 2 and 3 trials?
Phase 1 tests safety in small groups, phase 2 looks for early signs of efficacy and the right dose, and phase 3 tests benefit and harm in large controlled trials. Phase 2 is where candidates most often fail [1].
If a peptide is approved for one condition, does it work for others?
Why do drugs that work in mice fail in humans?
Species differ in metabolism and physiology, animal models simplify human disease, and positive animal results are more likely to be published. Only about a third of highly cited animal studies were later confirmed in human randomised trials [2].
What does accelerated approval mean?
It is an FDA pathway for serious conditions that allows approval based on a measure judged reasonably likely to predict benefit, on condition that the company confirms the benefit in a further trial [5].
Research Library
Every Synthos lot is tested by an independent laboratory, and the certificate of analysis for that lot comes with your order. Ask your representative to see it before you pay anything.
Further reading elsewhere
Coverage and explainers from other publications, picked and checked by our team.
- Nature · Jun 2026Is the peptide craze backed by science? The promise behind the hype ↗A Nature news feature on why researchers think enthusiasm for unregulated peptides has run ahead of the human evidence.
- US FDA · Sep 2025FDA Grants Accelerated Approval to First Treatment for Barth Syndrome ↗The FDA's announcement on elamipretide (SS-31), approved on a measure judged reasonably likely to predict benefit, with a confirmatory trial required.
- Biotechnology Innovation Organization · Feb 2021Clinical Development Success Rates and Contributing Factors 2011-2020 ↗The industry report behind the 7.9% figure: how often drug candidates survive each phase of clinical development.
- Association of Health Care Journalists · Aug 2007Rats! Animal studies poor predictor of human effect ↗A short explainer for journalists on the JAMA analysis that found only about a third of highly cited animal studies held up in human trials.
References
- Thomas D, Chancellor D, Micklus A, et al. Clinical Development Success Rates and Contributing Factors 2011–2020. Biotechnology Innovation Organization, Informa Pharma Intelligence and QLS Advisors; February 2021. Cited for phase transition rates, the 7.9% likelihood of approval from phase 1 and the 10.5-year average timeline. BIO
- Hackam DG, Redelmeier DA. Translation of research evidence from animals to humans. JAMA. 2006;296(14):1731-1732. Cited for the share of highly cited animal studies later replicated in human randomised trials. PubMed
- Dominikowski A, Rękoś Z, Olejarz M, et al. The emerging landscape of performance-enhancing peptides modulating GH-IGF1 axis: bridging the gap between clinical evidence and patient self-administration. Front Endocrinol (Lausanne). 2026;17:1822475. Cited for AOD-9604's mouse results and its failed 24-week human trial (502 randomised). PubMed
- Karaa A, Bertini E, Carelli V, et al. Efficacy and safety of elamipretide in individuals with primary mitochondrial myopathy: the MMPOWER-3 randomized clinical trial. Neurology. 2023;101(3):e238-e252. Cited for the MMPOWER-3 results (n=218). PubMed
- US Food and Drug Administration. FDA grants accelerated approval to first treatment for Barth syndrome. Press release, 19 September 2025. Cited for the September 2025 accelerated approval of elamipretide for Barth syndrome and the confirmatory trial requirement. FDA
- Falutz J, Mamputu JC, Potvin D, et al. Effects of tesamorelin (TH9507), a growth hormone-releasing factor analog, in human immunodeficiency virus-infected patients with excess abdominal fat: a pooled analysis of two multicenter, double-blind placebo-controlled phase 3 trials with safety extension data. J Clin Endocrinol Metab. 2010;95(9):4291-4304. Cited for the pooled phase 3 results (n=806; visceral fat −15.4%). PubMed
- US Food and Drug Administration. Drugs@FDA: Egrifta (tesamorelin), application 022505, original approval 10 November 2010. Cited for the 10 November 2010 approval date. Drugs@FDA
- Poolton JT, Minett J, Kazmi A. Human evidence for BPC-157 in orthopaedic and regenerative medicine: a systematic review of efficacy and safety. Acta Orthop Belg. 2026;92(3). Cited for the three human studies (n=30) and very low certainty of evidence. PubMed
- Józwiak M, Bauer M, Kamysz W, Kleczkowska P. Reply to Sikiric et al. BPC 157 therapy: targeting angiogenesis and nitric oxide's cytotoxic and damaging actions. Pharmaceuticals (Basel). 2025;18(10):1451. Cited for the concentration of BPC-157 publications in one research group. PubMed
- Sumner P, Vivian-Griffiths S, Boivin J, et al. The association between exaggeration in health related science news and academic press releases: retrospective observational study. BMJ. 2014;349:g7015. Cited for exaggeration in press releases and news (462 press releases). PubMed
This article summarises published research for education and laboratory context. It is not medical advice. Synthos products are sold for laboratory research only and are not approved for human use.