A Response to the ABPI's Letter to the Editor on the Use of Dogs in Predicting Drug Toxicity in Humans

Abstract

We are in total agreement with the ABPI that “informed debate amongst key stakeholders” is important. However, this should include animal protection groups, in addition to their suggested partnership between “the pharmaceutical industry, academia and the regulatory authorities.”

Unfortunately, as we have tried to explain, we believe that the ABPI authors have been too eager to play down the importance of our study. Far from being constructive, their critique is overly critical to the point of being unhelpful. Criticism must be fair and robust. When it has a weak scientific basis, it is unacceptable, especially when it is used as a basis to persist with the status quo, for which there is weak supportive evidence.

We acknowledge (as we did in our paper) that there are inherent and unavoidable caveats to our analysis. Notwithstanding this, our study is unprecedented in scale and type. Notably, it is much broader and more comprehensive than any of the studies cited by the ABPI, which involved a combination of: more limited (smaller and more restricted) sample sizes; exaggeration in terms of their supportive nature for dog studies; overlooking of evidence against dog studies; incomplete statistics, from which conclusions could not be drawn; incorrect statistical definitions and analyses; and questionable, or absent, human relevance. Conspicuously, none of these studies addressed any of the criticisms made by the ABPI of our study, on dosage and rare events, for example. In addition, the ABPI failed to provide constructive suggestions on how we could go about addressing these concerns. How does the ABPI suggest we address the problems of data limitations, dosage/exposure, and incidence of toxicity — considerations absent, or at least almost absent, in previous studies? Constructive criticism must include suggestions on how to go about resolving those issues. Arguably, given the restricted amount of data that are publicly available, at least some of those concerns cannot be addressed by third parties. The ABPI, and associated groups, could (and should) help by facilitating access to such information. Nevertheless, we were already in the process of analysing the data available to us in greater depth, the outcome of which will be published in the near future.

To reiterate: we did all we could do, with the data that were available to us, to produce a study of unprecedented scale and type. We are confident about, and as certain as we could be of, our analysis and our conclusions. They should not be overlooked because of their source, and because they are inconvenient to any stakeholders (including the Federation of Laboratory Animal Breeders Associations, which, not surprisingly, supports the ABPI's arguments). Should the ABPI be genuinely concerned that our analysis, using all the data and correct statistical methods at our disposal, is not sufficient, we refer it to our statement within our paper: “…if any pharmaceutical industry stakeholders have issues or concerns with our conclusions, we would encourage them to conduct further analyses by using their own proprietary data, and/or to facilitate such investigations by making available anonymised data, in accordance with the promotion of transparency encouraged by EU Directive 2010/63/EU, as well as to engage fully in constructive discussion and debate with us and our colleagues in animal protection organisations.” The only acceptable response by industry and regulators should be, at the very least, a desire for positive action to facilitate, or conduct, if possible, another study on a greater scale, involving the use of proprietary data. To ignore the most relevant findings, and/or focus on other inferred shortcomings of other elements of the study, is unacceptable, particularly when the upshot is to defend and perpetuate seriously flawed dog toxicology, with little or no substantiation.