The Lancet has published an open letter to British Prime Minister David Cameron from 21 people. The letter is essentially about the use of animal models to predict adverse drug reactions (ADRs) in humans. I have written volumes on this so will not go into detail here except to say that animals cannot predict adverse drug reactions or other drug properties such as efficacy or ADME in humans. There is both theory and empirical evidence supporting this. For the most part, I have no problem with the way the self-described clinicians and scientists present the problem with animal models and ADRs.
I do however have a major problem with the following from their letter:
We call on the UK Government to initiate a comparison of a set of human-biology-based tests with those currently used, as proposed in the Safety of Medicines Bill 2010—11, to see which are more effective for predicting the safety of medicines for patients.
This request reveals an utter ignorance of current science (or perhaps ulterior motives not immediately recognizable from the letter as not all the signatories are listed).
1. Drug reactions take place at a level of organization higher than that which allows trans-species extrapolation. Even a superficial understanding of complex systems will allow one to appreciate that a different species will not be able to predict perturbations to the system of another species on that level of organization.
2. Once again we are back to that pearl of wisdom by Theodosius Dobzhansky: "Nothing in biology makes sense except in the light of evolution." Even a moderate understanding of evolutionary biology forces one to acknowledge that animals are expected to react differently because of different evolutionary histories.
3. There currently exists enough empirical evidence to slam the door on the animal models and prediction question. I have covered some of this evidence in the books and blogs. There are volumes more. More such studies are not needed. Moreover, this evidence can now be placed into the context of theory provided by complexity and evolution.
4. Tests for human response to drugs and disease either satisfy the criteria for scientific predictability or they do not. If animal tests have a positive predictive value (PPV) in the 0.3 neighborhood and in vitro tests have a PPV in 0.5 neighborhood, neither qualifies as predictive and neither offers industry much help in deciding what chemicals to develop. Neither test informs or predicts. One does not need direct comparisons to decide which test or set of tests is better. Each test or set of tests can be judged independently. In this context, none of these tests per se qualify as predictive and small differences in PPV and NPV between animal tests and in silico and or in vitro tests are irrelevant.
5. Finally, there is not one test or one series of tests for ADRs or efficacy or any other drug property. There are many tests and many properties that are tested for. Toxicity alone probably has around 15 requirements; for example, neurotoxicity, renal toxicity, developmental toxicity, skin and so forth. The animal tests for these are not effective but the testing process per se does not work either. Industry needs better tools! They acknowledge this and to their credit are working on developing better tools. I agree with the authors that animal models cannot predict human response to ADRs but the situation is not going to be corrected by demanding irrelevant direct test-to-test evaluations. This helps neither humans nor animals and reveals, among other things, a lack of understanding of the drug development process.
The proposal in the letter exemplifies what I have stated in this blog many times and that has offended many in the animal protection movement: loving animals does not make you a scientist or scientifically competent. In order to competently address the use of animals in research and science today, an understanding of science in general, the philosophy of science, complex systems, evo devo, evolutionary biology, and a thorough knowledge the current empirical evidence, in addition to other facts is required. Those that cavalierly comment and publish on issues beyond their understanding and competence cause real harm.
I call upon The Lancet and the signatories to make available all the names of those that signed this letter. (The “WebExtra Content” on The Lancet’s site does not currently contain these names nor does the Safer Medicines Campaign website.)