From animals to humans – can we extrapolate data on harms?

July 13, 2017

Thirty-seven medicines have been withdrawn from the market because of harmful effects observed in animals. However, whether we should use evidence from animal research to make decisions on withdrawing human medicines from the market is not clear.

How do we know?

We conducted a systematic review to identify medicines that were withdrawn from the market because of harmful effects reported in animals. We then used a checklist adapted from the International Agency for Research on Cancer (IARC) criteria to judge the evidence used for making the withdrawals.

What did we find?

We found 37 medicines withdrawn from the market between 1963 and 2000. In almost two-thirds of cases (62%), possible risk of cancer was reported as the reason for withdrawal. In 80% of instances, the evidence used by drug regulators to make withdrawals was limited; and only one animal species was used as the evidence in 73%. On average, it took 2 years to withdraw the medicines from the market after the harms were reported in animals.

Only in five instances (14%) were further research conducted in humans to investigate whether there was an association between the withdrawn medicine and risk of the harm observed in the animals. Interestingly, in none of these cases was there a consistency when the research results between animals and humans were compared.

What are the implications of our findings?

Drug regulators are more likely to quickly withdraw medicines from the market when there are possible risks of cancers associated with their use.

We have not used evidence from animal research to make drug withdrawal decisions over the past 16 years or so. This implies that there have been improvements in how harms from medicines are examined in animals before such medicines are approved for use in humans. However, it may also indicate that drug manufacturers are not publishing data on harms in animal research when such results are unfavourable.

Regulators are still not that good at determining whether drug-adverse event associations observed in animals are likely to occur in humans. Our publication, therefore, made some recommendations to improve how to translate evidence of harms from animal research into humans:

  • Develop a registry of animal trials and ensure mandatory reporting to overcome negative publication bias;
  • Obtain data on harms from at least 2 animal species to improve prediction of risks to humans;
  • Conduct human studies when results of animal studies are inconclusive;
  • Undertake systematic reviews when suspected harms are reported in animals;
  • Adhere to animal toxicity guidelines in order to improve the quality of reporting;
  • Creation of an international body that reviews newer developments in animal testing

Dr Igho Onakpoya is a DPhil Student and Research Fellow in Evidence Based Practice and Pharmacovigilance.


Onakpoya IJHeneghan CJAronson JK. Postmarketing withdrawal of human medicinal products because of adverse reactions in animals: a systematic review and analysis. Pharmacoepidemiol Drug Saf. 2017 Jul 9. doi: 10.1002/pds.4256


Leave a Reply

Your email address will not be published. Required fields are marked *

* Checkbox GDPR is required


I agree