Investigator brochures (IBs), or the documents in which drug makers summarise animal data on the efficacy of new drugs, often have large gaps, reports a new study. This means that institutional review boards (IRBs) and drug regulators who rely on these documents to green light human trials are doing so based on inadequate information, putting human participants at risk. The study, ‘Preclinical efficacy studies in investigator brochures: Do they enable risk–benefit assessment?’, was published this month in PLOS Biology.
Animal studies should be of high quality to enable IRBs to make good decisions. This means that the sample sizes for such studies should ideally be calculated in advance, and the studies should be randomised and blinded to remove bias. To understand how well IBs perform on these counts, a team of bioethicists, led by Daniel Strech of Germany’s Hannover Medical School, dug into 109 IBs submitted to three IRBs at German medical institutions. These IBs included data from 708 animal trials. Strech and his collaborators found that of the 708 trials, only 5% included information on sample size, randomisation and blinding, and 89% of the animal studies were never published. Finally, 82% reported positive outcomes.
These findings are troubling because they indicate that IRBs and regulators are basing critical decisions on shoddy data. While the lack of sample size information in IBs, for example, doesn’t mean the drug makers didn’t calculate sample sizes, IRBs and regulators have no way of knowing the truth. Worse, since most of the animal data is unpublished, the IRBs cannot tell if data had been peer reviewed or not. The authors also argue that the large percentage of positive findings could mean that drug makers are cherry-picking data.
The poor quality of IBs means that humans in Phase 1 and 2 trials are being exposed to unnecessary risks. Such risks can be minimised in several ways, the researchers say. For one, drug makers should only report those animal trials that have been registered in advance, so that negative data are not concealed. Also, information on other bias-reducing tactics like blinding should be included in IBs in a tabular form, so that IRB members, who are often laypeople, can grasp it easily.
Unless the quality of reporting in IBs improves, using them to give the go-ahead to human trials is unethical, the study authors say.
COMMents
SHARE