The field of toxicity testing is an urgent humanitarian problem,
for it regularly involves a finite and large incidence of distress
which is often considerable and sometimes acute.
In the latter part of Chapter 5 of The Principles of Humane Experimental Technique,1 on Replacement, Russell and Burch described the question of progress in the field of toxicity testing as a particularly serious problem, since “this is one use [of animals] which is an urgent humanitarian problem, both numerically and in terms of severity, for it regularly involves a finite and large incidence of distress which is often considerable and sometimes acute.”
They recognised the widely-held view that toxicity testing must be conducted with mammals, on the basis of agreement with the high-fidelity argument, because of “the importance rightly attached to the need for safeguarding human patients against the toxic side-effects of drugs”. However, they were also careful to point out that high-fidelity arguments cut both ways. Other mammals may be superficially similar to humans, but “the mechanisms of excretion and detoxification in non-human mammalian species frequently differ from our own”. This objection, they said, “is met in practice by erring on the side of caution, and by using more than one mammalian species; it cannot be fully met, for there may always be metabolic peculiarities specific to man”. Put more forcefully, it can be argued that toxicity testing in laboratory animals is an obvious example of the risks inherent in the high-fidelity fallacy, especially when the insistence on the use of a rodent species and a non-rodent species is based on the misperception that using two inadequate models is better than using only one.
In a particularly important paragraph, they refer to dosage: A very large number of substances are toxic at a high enough dosage, so “the important concept is the therapeutic index of a drug — that is, the ratio between its toxic dose and its therapeutically effective dosage. If this ratio is great, the [use of the] drug or preparation is prudent, since it allows for the wide variations between human individual patients in sensitivity to the toxic effects”. Toxicity testing would be a less serious humanitarian problem, if more care were to be taken to administer only rational doses of chemical and products in general, based on reasonable estimations of likely exposure.
Russell and Burch made a distinction between bioassay and toxicity testing: bioassay being “the detection and quantitative estimation of a known activity or principle in a relatively impure preparation, usually, though not invariably, assessed by comparison with a preparation of standard purity and potency”; while toxicity testing “may mean the assessment of various unknown or unpredictable special activities with the general property in common of toxicity in higher animals”.
They made a further useful distinction between two types of toxicity testing. The first of these, somewhat akin to bioassay, is routine testing for specific toxic effects which are well-known. They saw this as an activity where in vitro methods could be particularly useful.
The second type of testing particularly applies to new chemicals or drugs, where “the kind of toxic effect which may arise is virtually unknown”. This is the focus of insistence that high-fidelity mammalian models must be used, to reveal the unexpected, whereas non-animal in silico and in vitro, i.e. alternative, methods are designed to answer specific questions.
Their comment, that “in theory, we should be able to classify all the ways in which toxic and lethal symptoms are produced”, must be considered embarrassing in 2014 — 55 years after The Principles was published — since, despite all the efforts of toxicologists and other scientists, the suffering of huge numbers of laboratory animals, considerable damage to many patients and to the environment, and all the associated costs, we are very far from achieving that goal. We lack sufficient understanding of the main mechanisms involved in inducing toxic effects, and effective ways of predicting the likelihood of their occurrence in particular circumstances. It is true that there are hopeful signs that human-relevant and human-focused, non-animal testing strategies will eventually become available for the efficacy and toxicity testing of new medicines, but the continuation of stubborn and powerful defence of the need to rely on rodents, dogs and non-human primates represents a ball and chain around the ankle of progress. Chapter 5 of The Principles should be required reading for all concerned.
1 Russell, W.M.S. & Burch, R.L. (1959). The Principles of Humane Experimental Technique, xiv + 238pp. London, UK: Methuen.