The concept of "extrapolation " from animals to humans, and the philosophy of Karl Popper

Michael FW Festing
MRC Toxicology Unit, Hodgkin Building, University of Leicester, PO Box 138, Lancaster Road, Leicester LE1 9HN, UK.

Those involved in research on animals need a clear understanding of the underlying philosophy if they are to justify such use. It seems to be accepted wisdom that one step in animal research involves "extrapolation' to humans. Unfortunately, nobody has yet explained exactly how this is done, though there seems to be a feeling that if only the animals were more like humans, such extrapolation would be easier. In fact, animals are different from humans in many ways. How can such extrapolation ever be logically justified? The answer is that it can not. But that presents no problems because we do not do it. The whole concept of extrapolation is based on the outdated philosophy of inductive reasoning whereby it was thought that we develop scientific laws as a result of observing repeated events. The alternative, deductive philosophy of Sir Karl Popper is now widely accepted by working scientists, and provides a more secure foundation for the rational use of animals in research. According to Popper, we gain new knowledge by formulating hypotheses, based on current knowledge, which we then test. These hypotheses can sometimes be shown to be false but they can never be shown to be true. Failure to falsify a hypothesis leads to its provisional acceptance. The best hypotheses are those which are most easily be falsified. According to Popper the method of science involves "bold conjectures and severe and ingenious attempts to refute them". Applying this philosophy to the toxicity testing of a new compound it is clear the toxicologist does in fact formulate and test a series of hypotheses. One common hypothesis is that the compound is not genotoxic. If it is, we know that it may be a human carcinogen. This can be tested using a range of in-vitro techniques such as the Ames test of mutagenicity to Salmonella. A positive result in one or more of these tests would lead to a rejection of the hypothesis. Negative results lead to a provisional acceptance, possibly with the formulation of a new hypothesis that it is not a mammalian carcinogen. This could be tested using rats. No extrapolation is involved. The experiment is interpreted before it is carried out, based on current understanding of mammalian biology. If it is non-carcinogenic in rats, then it is provisionally accepted as being non-carcinogenic in mammals, although whether testing only. in rats is sufficiently rigorous must also be considered. In no case is there any step which resembles extrapolation except where there is true mathematical extrapolation of a dose-response curve to some no effect level for the test species. We do not extrapolate from animals to humans, we have never done so, and it is time that we abandon the whole concept which can only do us a disservice. What we actually do is to formulate and test hypotheses, and interpret the results of our animal experiments.
Back to the Scand-LAS 1998 Symposium programme.
This page is provided by the server of the Laboratory Animal Unit, at the Norwegian College of Veterinary Medicine, Oslo.