PepGenexShop our research products

Animal-to-human evidence

Animal-to-human evidence is the question of whether, and how far, a finding observed in a living non-human animal holds in humans. Van der Worp and colleagues write that animal experiments have contributed much to the understanding of disease mechanisms, but that animal studies do not predict with sufficient certainty what will happen in humans, which is why clinical trials remain essential.[1]

How often have animal findings matched human findings?

Olson and colleagues reported in 2000 on a survey of 12 pharmaceutical companies covering 150 compounds with 221 human toxicity events identified during clinical development. The true positive concordance rate with animal toxicity studies was 71% for rodent and non-rodent species together, 63% for non-rodents alone and 43% for rodents alone.[2]

In the same survey, concordance was highest for hematological, gastrointestinal and cardiovascular toxicities and lowest for cutaneous toxicity. Where animal studies identified a concordant toxicity, 94% were first observed in studies of 1 month or less in duration.[2]

Van der Worp and colleagues cite a review of animal studies published in seven leading scientific journals of high impact, in which about one-third of the studies translated at the level of human randomised trials and about one-tenth of the interventions were subsequently approved.[1]

What did a systematic comparison of animal and clinical studies find?

Perel and colleagues reported in 2007 a systematic review of animal studies of six interventions for which clinical trials had given unambiguous evidence of benefit or harm, extracting data on study design, allocation concealment, number of randomised animals, type of model, intervention and outcome.[3]

Agreement was mixed. Tirilazad was associated with a worse outcome in clinical trials in acute ischaemic stroke, while in animal models it reduced infarct volume and improved neurobehavioural scores. Antifibrinolytics reduced bleeding in clinical trials, while the animal data were inconclusive.[3]

The authors concluded that discordance between animal and human studies may be due to bias or to the failure of animal models to mimic clinical disease adequately.[3]

What design problems reduce the reliability of animal studies?

Van der Worp and colleagues describe internal validity as the extent to which differences observed between groups of animals can, apart from random error, be attributed to the intervention under investigation. They list randomisation, allocation concealment and blinding as the means of reducing the bias that threatens it.[1]

In systematic reviews of animal studies across several disease models, they report that generally about a third or less of the studies reported random allocation, fewer reported concealment of allocation or blinded measurement of outcome, and a priori sample size calculations were reported in 0% to 3% of studies.[1]

Kilkenny and colleagues describe a survey of 271 randomly chosen articles in which only 59% stated the hypothesis or objective of the study together with the number and characteristics of the animals used, and most did not report using randomisation (87%) or blinding (86%).[4]

Why can a well-run animal study still fail to translate?

Van der Worp and colleagues write that even when the design and conduct of an animal study are sound, translation may fail because of disparities between the model and the clinical situation, which they call reduced external validity.[1]

The common causes they list include inducing a disease in animals that are young and otherwise healthy when in people it mainly occurs in older individuals with other conditions, studying a homogeneous group of animals rather than a heterogeneous human population, using animals of one sex only, using models with insufficient similarity to the human condition, and differences in outcome measures and in the timing of outcome measurement.[1]

In their stroke example, over 95% of the studies were performed in rats and mice, the animals were almost invariably young, and outcome was usually measured at 1 to 3 days in animal models compared with 3 months in people.[1]

How does publication bias affect the animal literature?

Van der Worp and colleagues describe a meta-analysis of 525 publications from systematic reviews of 16 interventions tested in animal studies of acute ischaemic stroke, which suggested that publication bias might account for around one-third of the effect reported. Only ten of those publications (2%) did not report at least one significant effect.[1]

The ARRIVE guidelines were developed using the CONSORT statement for clinical trials as their foundation, as a checklist of the minimum information that publications reporting animal research should include.[4]

Van der Worp and colleagues state that evidence from a single laboratory, or obtained in a single model or species, is probably not sufficient, and propose systematic review and meta-analysis of all available preclinical evidence before clinical trials are started.[1]

How does PepGenex Science present animal evidence?

PepGenex Science uses the evidence type Animal for studies in living non-human organisms. A result in an animal model is described as an animal model result, attributed to the study that reported it, and grouped separately from in vitro and human findings.

Where the evidence for a compound is only in animals, the compound profile says so. An animal finding is not presented as a human finding.

Limitations

Concordance figures depend on how they were gathered. The database of Olson and colleagues covered only compounds for which human toxicities had been identified during clinical development, and the review by Perel and colleagues covered six interventions chosen because the clinical-trial evidence for them was unambiguous.[2],[3]

Van der Worp and colleagues note that bias in animal studies has been tested most extensively in acute ischaemic stroke, and that the relative importance of the various possible sources of bias is not yet known.[1]

This page describes what the cited documents say. It is not a statement that any PepGenex material has been manufactured, tested or released under any of them.

Compound profiles

Compound profiles whose recorded evidence includes both animal and human studies.

References

  1. van der Worp HB, Howells DW, Sena ES, et al. Can animal models of disease reliably inform human studies? PLoS Med. 2010;7(3):e1000245. PMID 20361020 · DOI 10.1371/journal.pmed.1000245 · PMC2846855
  2. Olson H, Betton G, Robinson D, et al. Concordance of the toxicity of pharmaceuticals in humans and in animals. Regul Toxicol Pharmacol. 2000;32(1):56-67. PMID 11029269 · DOI 10.1006/rtph.2000.1399
  3. Perel P, Roberts I, Sena E, et al. BMJ. 2007;334(7586):197 (title withheld on this site; see the publication record). PMID 17175568 · DOI 10.1136/bmj.39048.407928.BE · PMC1781970
  4. Kilkenny C, Browne WJ, Cuthill IC, Emerson M, Altman DG. Improving bioscience research reporting: the ARRIVE guidelines for reporting animal research. PLoS Biol. 2010;8(6):e1000412. PMID 20613859 · DOI 10.1371/journal.pbio.1000412 · PMC2893951