How to Read Animal Studies: Limitations and Common Pitfalls

"Proven to build muscle" often turns out to mean "increased fibre size in rats". Animal research is essential to science, but it is also one of the most misused sources in supplement marketing. Our editors explain why animal studies are done, where their results break down when applied to people, and how to read them critically.

Why animal studies exist

Animal experiments allow scientists to study mechanisms, toxicity and early signs of effectiveness in ways that would be unethical or impossible in people. They can involve tissue sampling, genetic modification and controlled diets that no human volunteer would accept.

In the UK, such research is regulated under the Animals (Scientific Procedures) Act 1986. Researchers must justify their work and follow the principles of the 3Rs: Replacement, Reduction and Refinement. The NC3Rs, a national centre, promotes these standards.

Animal data are a key step in drug development. Regulators require toxicology studies before medicines are tested in humans, and mechanistic work in animals often generates the hypotheses later tested in clinical trials.

The problem is not the existence of animal studies but the leap from them to claims about humans. Many promising animal results never translate, and readers should treat them as a starting point rather than a conclusion.

Species differences that change the picture

Rodents differ from humans in metabolism, hormone regulation, lifespan and body composition. A mouse's metabolic rate per kilogram is many times higher than ours, which affects how quickly substances are broken down and eliminated.

Gene expression can differ too. A study by Seok and colleagues in 2013 found that genomic responses to inflammation in a common mouse model correlated poorly with those in humans. That kind of discrepancy can make a treatment look effective in mice but useless in people.

Laboratory animals are usually young, genetically identical and kept in controlled environments with standard diets. Humans vary in age, genetics, diet, sleep, stress and medication use, which can either amplify or erase an effect.

Muscle research is especially tricky. Rodent muscle responds differently to overload, and models such as synergist ablation or electrical stimulation do not replicate gym training. Results on fibre size may not mean anything for a person lifting weights three times a week.

The dose problem

Doses in animal studies are often far higher than what a person could realistically take. Converting doses is not as simple as scaling by body weight. Reagan-Shaw and colleagues described the body surface area approach, which typically gives a human equivalent dose much lower than a direct per-kilogram calculation would suggest.

Even with correct scaling, the route of administration matters. Many animal studies inject substances directly into the abdomen or bloodstream, bypassing digestion and first-pass liver metabolism. Oral supplements in humans may never reach comparable blood levels.

Duration is another trap. A few weeks in a mouse may represent a significant part of its life, while equivalent exposure in humans would take years. Long-term safety cannot be extrapolated from a short rodent study.

The table shows typical gaps between animal and human contexts.

Design flaws and publication bias

Animal studies vary in quality. Many lack randomisation, blinding of outcome assessors or sample size calculations. Without these, results are more likely to be exaggerated. The ARRIVE guidelines were developed to improve reporting, and an updated version was published in 2020.

Publication bias is a major concern. Positive results are more likely to be published than negative ones, so the literature may overstate how often a substance works. Perel and colleagues compared treatment effects in animal experiments with clinical trials and found frequent disagreement.

Hackam and Redelmeier looked at highly cited animal studies and found that only about a third were later replicated in human randomised trials. That figure is often quoted as a reminder of how uncertain translation can be.

Finally, conflicts of interest can shape how studies are designed and presented. Supplement companies sometimes fund animal research and then quote it on labels, while omitting the absence of human data.

A practical checklist for readers

When a news story or product page cites research, the first question is simple: was this done in people? If the answer is no, any claim of benefit should be treated as preliminary.

Next, look at the dose and route. If the study injected rats with an amount equivalent to dozens of capsules a day, the result tells you little about a normal supplement.

Useful questions to ask include:

1. Which species and model were used, and how close are they to the human situation?
2. Was the study randomised and blinded, and how many animals were involved?
3. What exactly was measured: a biomarker, tissue change or a meaningful outcome?
4. Has the result been replicated by independent groups?
5. Are there any human trials, and do they agree?
6. Who funded the research?

Animal research also matters for safety. If a substance shows toxicity in animals, that is a serious warning even before human data exist. Absence of harm in animals, however, does not guarantee safety in humans.

Conclusions of the editorial team

Animal studies are valuable for exploring mechanisms and identifying risks, but they cannot on their own justify claims about effects in people.

Species differences, unrealistic doses, unusual routes of administration and publication bias all widen the gap between laboratory results and real-world outcomes.

A sceptical but fair reader asks whether human data exist, whether results have been replicated and whether the dose makes sense. That habit filters out a large share of misleading marketing.

For more on this theme, see our articles on how to read a randomised controlled trial, on understanding meta-analyses, and on spotting misleading supplement claims.

References

1. Hackam DG, Redelmeier DA. Translation of research evidence from animals to humans. JAMA. 2006;296(14):1731–1732.
2. 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.
3. Reagan-Shaw S, Nihal M, Ahmad N. Dose translation from animal to human studies revisited. FASEB J. 2008;22(3):659–661.
4. 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.
5. Kilkenny C, Browne WJ, Cuthill IC, et al. Improving bioscience research reporting: the ARRIVE guidelines for reporting animal research. PLoS Biol. 2010;8(6):e1000412.
6. 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.
7. Ioannidis JPA. Why most published research findings are false. PLoS Med. 2005;2(8):e124.
8. https://steroidemeisterlegal.com/produkt-kategorie/orale-steroide/methandienone-tabletten/