Why most peptide research is preclinical, and what that means for interpreting it

In short: Preclinical research is laboratory work done before a compound is tested in people: cell studies (in vitro) and animal studies (in vivo). Most research peptides have never been through a full clinical trial programme, so their literature stops at this stage. Preclinical results often fail to carry over to people; they describe a model, not a human outcome.

Most published work on the peptides sold for laboratory research comes from cell cultures and animals. This article explains why, and how to read that kind of evidence. It builds on our guide to how to read a peptide study, which sets out the evidence labels we use for cell, animal and human research.

What does preclinical research mean?

Preclinical research is the laboratory and animal work that comes before any testing in people.

The US Food and Drug Administration (FDA) lists two types, in vitro and in vivo. In vitro studies are done in cells, tissues or molecules outside a living organism; in vivo studies are done in living animals. Their purpose, the FDA says, is to find out whether a compound could cause serious harm before anyone is exposed to it. The FDA notes that these studies are usually not very large, and that researchers review the findings afterwards to decide whether testing in people should go ahead 1.

“Clinical research”, by contrast, means studies done in people. It usually runs from small phase 1 studies to large phase 3 studies 2.

Illustration: a cell culture plate under a microscope, the setting for much in vitro peptide research.

Why do most research peptides have only preclinical evidence?

Human trials are usually run by a company developing a medicine, and most research peptides either never had such a sponsor or had a programme that stopped early.

Peptides as a class do progress through trials to approval. More than 80 peptide drugs have reached the market, according to a 2021 perspective by Muttenthaler et al. 3. What separates those compounds from most research peptides is the path through sponsored trials.

That path is long and narrow. In a 2003 analysis, Contopoulos-Ioannidis et al. followed 101 basic science papers from 1979 to 1983 that claimed a promising new therapy or preventive. By 2002, 27 had led to at least one published randomised trial and 5 to a licensed clinical use. Industry involvement was the strongest predictor of a finding reaching human testing. Findings not tested in people within 10 to 12 years were unlikely to be tested later 4.

Even compounds that enter human trials usually stop. The FDA estimates that about 70% of drugs move on from phase 1, about 33% from phase 2 and 25 to 30% from phase 3 2. Wong et al. analysed trial records for more than 21,000 compounds from 2000 to 2015 and estimated that 13.8% of development programmes entering phase 1 eventually reached approval 5.

Research peptides show the same pattern:

• BPC-157. A 2025 systematic review by Vasireddi et al. included 36 studies, of which 35 were preclinical and one was a clinical study. The authors found no clinical safety data 6. Our BPC-157 research review covers that literature.
• CJC-1295 and ipamorelin. ClinicalTrials.gov lists no trial of either compound beyond phase 2 7. Our review of CJC-1295 and ipamorelin sets out the registered trials, including a completed trial with no published results.

How often do preclinical findings hold up in people?

Often they do not: studies that compare animal and human results find frequent disagreement.

In a 2007 systematic review, Perel et al. looked at six treatments whose effects in clinical trials were clear, and compared them with the animal studies of the same treatments. The animal and human results agreed for some and disagreed for others. In two cases, the animal studies reported favourable effects that clinical trials did not reproduce, and one of those treatments was associated with worse outcomes in patients. The authors suggested that the gap may come from bias in the animal studies or from animal models that do not mimic the human disease 8.

Species differences are part of the problem. In a 2013 study, Seok et al. compared gene activity in people with acute inflammatory conditions of different causes with gene activity in the matching mouse models. The human conditions produced highly similar responses. The mouse responses matched the human ones poorly and were close to random for many genes 9.

Why can preclinical results mislead?

Three features of the preclinical literature can make results look stronger or more general than they are: weak study design, selective publication and reliance on a few laboratories.

Weak design. Randomisation (allocating animals to groups by chance) and blinding (hiding group allocation from the people assessing outcomes) reduce bias. In a 2015 study of 2,671 animal efficacy papers, Macleod et al. found that 24.8% reported randomisation, 29.5% reported blinded outcome assessment and 0.7% reported a sample size calculation. In a random sample of 146 in vivo papers from PubMed, none reported a sample size calculation 10.

Selective publication. Publication bias means that studies with positive results are more likely to be published. Sena et al. examined 525 publications of animal stroke studies in 2010 and found that only 1.2% reported no significant finding at all. Their analysis suggested that publication bias accounted for about a third of the reported efficacy, and that about 14% of experiments went unreported 11.

Few laboratories. When most papers on a compound come from one research group, the findings may not have been independently replicated. We note this in each compound review where it applies.

How should you read a preclinical peptide study?

Read it as a description of what happened in one model under set conditions, and check how it was designed and reported.

A short checklist:

• Model. Is it a cell line, isolated tissue or a living animal? Which species, strain and sex, and how was the disease or injury produced?
• Comparison. Was there a control group, and what did it receive?
• Bias controls. Does the paper report randomisation, blinded assessment and a sample size calculation?
• Reporting standard. The ARRIVE 2.0 guidelines list the “ARRIVE Essential 10”, the minimum items a report of animal research should contain 12. Missing items make a study harder to judge.
• Replication. Have other laboratories reported the same finding in other models?
• Human data. Is there a registered trial, and were its results published? Registries such as ClinicalTrials.gov and the ANZCTR show what was planned 7, 13.

Illustration: many sealed vials narrowing to a few, suggesting how few preclinical candidates reach human testing.

The wording of a finding matters too. “In a rat tendon model, the treated group showed X compared with controls” is a preclinical result. It is not evidence of the same result in people. Our guide to how to read a peptide study explains how cell, animal and human evidence are ranked, and why cell and animal work sits outside the NHMRC evidence levels, which grade human study designs.

Frequently asked questions

What is the difference between in vitro and in vivo research?

In vitro research is done outside a living organism, for example in cultured cells, tissue samples or purified proteins. In vivo research is done in a living animal. The FDA describes both as the two types of preclinical research carried out before a compound is tested in people (FDA). Ex vivo work, on tissue removed from an animal, sits between the two.

Do animal study results predict results in people?

Not reliably. A 2007 systematic review by Perel et al. compared animal studies with clinical trials for six treatments and found agreement for some and disagreement for others, including one treatment that did well in animals but was associated with worse outcomes in patients (Perel et al., 2007). Animal results show what happened in a model, and need human studies to confirm them.

What are the ARRIVE guidelines?

ARRIVE (Animal Research: Reporting of In Vivo Experiments) is a checklist of what a paper on animal research should report, first published in 2010. The 2020 update, ARRIVE 2.0, splits the items into the “ARRIVE Essential 10”, the minimum set, and a “Recommended Set” that adds context (Percie du Sert et al., 2020). Many journals and funders endorse it.

What proportion of drug candidates reach approval?

A minority. Wong et al. analysed clinical trial records from 2000 to 2015 and estimated that 13.8% of drug development programmes entering phase 1 eventually reached approval (Wong et al., 2019). The FDA estimates that about 70% of drugs move on from phase 1 and about 33% from phase 2 (FDA).

How can I check whether a peptide has been tested in human trials?

Search the trial registries. ClinicalTrials.gov holds most international trials, and the Australian New Zealand Clinical Trials Registry covers trials run in Australia and New Zealand. Search by compound name and code names, then check each record’s phase, status and whether results were posted or published. A registered trial is not the same as a published result.

References

1. US Food and Drug Administration. The drug development process. Step 2: preclinical research. https://www.fda.gov/patients/drug-development-process/step-2-preclinical-research. Accessed 28 September 2026. [regulator information page, US]
2. US Food and Drug Administration. The drug development process. Step 3: clinical research. https://www.fda.gov/patients/drug-development-process/step-3-clinical-research. Accessed 28 September 2026. [regulator information page, US]
3. Muttenthaler M, King GF, Adams DJ, et al. Trends in peptide drug discovery. Nat Rev Drug Discov. 2021;20(4):309-325. doi:10.1038/s41573-020-00135-8. PMID: 33536635. [perspective review]
4. Contopoulos-Ioannidis DG, Ntzani E, Ioannidis JP. Translation of highly promising basic science research into clinical applications. Am J Med. 2003;114(6):477-484. doi:10.1016/s0002-9343(03)00013-5. PMID: 12731504. [cohort study of published research]
5. Wong CH, Siah KW, Lo AW. Estimation of clinical trial success rates and related parameters. Biostatistics. 2019;20(2):273-286. doi:10.1093/biostatistics/kxx069. PMID: 29394327. PMCID: PMC6409418. [analysis of clinical trial databases]
6. Vasireddi N, Hahamyan H, Salata MJ, et al. Emerging Use of BPC-157 in Orthopaedic Sports Medicine: A Systematic Review. HSS J. 2025;21(4):485-495. doi:10.1177/15563316251355551. PMID: 40756949. PMCID: PMC12313605. [systematic review]
7. ClinicalTrials.gov. Searches of the intervention field for “CJC-1295” and “ipamorelin”. https://clinicaltrials.gov/. Accessed 28 September 2026. [trial registry]
8. Perel P, Roberts I, Sena E, et al. Comparison of treatment effects between animal experiments and clinical trials: systematic review. BMJ. 2007;334(7586):197. doi:10.1136/bmj.39048.407928.be. PMID: 17175568. PMCID: PMC1781970. [systematic review]
9. Seok J, Warren HS, Cuenca AG, et al. Genomic responses in mouse models poorly mimic human inflammatory diseases. Proc Natl Acad Sci U S A. 2013;110(9):3507-3512. doi:10.1073/pnas.1222878110. PMID: 23401516. PMCID: PMC3587220. [comparative genomic study (human and mouse)]
10. Macleod MR, Lawson McLean A, Kyriakopoulou A, et al. Risk of Bias in Reports of In Vivo Research: A Focus for Improvement. PLoS Biol. 2015;13(10):e1002273. doi:10.1371/journal.pbio.1002273. PMID: 26460723. PMCID: PMC4603955. [meta-research study]
11. Sena ES, van der Worp HB, Bath PM, et al. Publication bias in reports of animal stroke studies leads to major overstatement of efficacy. PLoS Biol. 2010;8(3):e1000344. doi:10.1371/journal.pbio.1000344. PMID: 20361022. PMCID: PMC2846857. [meta-research study]
12. Percie du Sert N, Hurst V, Ahluwalia A, et al. The ARRIVE guidelines 2.0: Updated guidelines for reporting animal research. PLoS Biol. 2020;18(7):e3000410. doi:10.1371/journal.pbio.3000410. PMID: 32663219. PMCID: PMC7360023. [reporting guideline]
13. Australian New Zealand Clinical Trials Registry (ANZCTR). https://www.anzctr.org.au/. Accessed 28 September 2026. [trial registry]