Friday, April 28, 2006

The problem with animal models

Here is what I find one of the most powerful, effective arguments for replacing animal experiments with other methods of investigation.

It is an essay in logic of science and is written in a formal way, so for those who are not inclined to read it all I’ll try to summarize it in non-formal style here.

The authors of the essay, entitled Two Models of Models in Biomedical Research, Hugh LaFollette and Niall Shanks, make the crucial distinction between two uses of animals as models. These two models are normally confused in the general discourse, leading to the current difficulties in clarifying the question of usefulness of animal experiments (we are here letting aside the ethical question).

The two models are:

1) animals as models that are similar to the objects to be modelled, ie humans, functionally (HAMs). To understand what it means, think of the planetary model of the atom in physics. In the early stages of the atomic theory, when knowledge was limited, the solar system has indeed served as a useful tool, something known to help understand the unknown.
Or think of a spiral staircase as a model of DNA molecules.

In this way, though, the only use of a model is to inspire hypotheses, but, crucially, not to test them.

There are indeed demonstrable functional similarities between humans and our close biological relatives.

But there is a big difference between an animal model's being a good source of hypotheses and its being a good means to test hypotheses.

2) animals as models that are similar to humans causally (CAMs). Here is where the problem lies, because most biological phenomena do not follow a simple linear cause-effect pattern (deterministic), but are probabilistic: only in a certain percentage of cases the same effect will follow the same cause.

In fact, biomedical experiments on animals are doubly probabilistic: they involve not only the probabilistic causality within the (non-human) laboratory population, but also the probabilistic causality within the human population outside the laboratory.

In addition, there is an uncertainty about whether the results observed in the non-human animal population will be (statistically) relevant to the human biomedical phenomena of interest.
For this uncertainty to be small, there must be no causally relevant disanalogies between the test subjects and humans (the model and the thing modelled)..

The fact is: these important causal disanalogies exist.

Researchers who think non-human animals are good causal models (CAMs) of human biomedical phenomena believe human and non-human animal systems are causally similar because they are functionally similar.

But this is not so.

The same function in biological systems can be caused by entirely different mechanisms. Both birds’ and mammals’ lungs oxygenate blood (same function); but peribronchial lungs of birds, ventilated in a unidirectional fashion using a series of air sacs, and the alveolar lungs of mammals, ventilated in a tidal fashion using a diaphragm, differ considerably in structure and mechanism.

Functional similarity does not guarantee underlying causal similarity, nor does it make such similarity "probable".

This is predictable from Darwin’s evolution theory. Different organisms have evolved similar functions due to their phylogenetic proximity, but "descent with modification" means, in part, "modification of anatomical and physiological sub-systems, and the relations between them."

Resultant species differences are biologically significant. "The species is one of the basic foundations of almost all biological disciplines. Each species has different biological characteristics" (Mayr p. 331). Species differences, even when small, often result in radically divergent responses to qualitatively identical stimuli.

This is why, for example, even when species are phylogenetically close, as are the rat and the mouse, we cannot assume that the two species will react similarly to similar stimuli. Tests for chemically induced cancers in rats and mice yield the same results for only 70% of the substances tested. The figure drops to 51% for site-specific cancers.

Human mechanisms for metabolizing phenol are closer to the mechanisms in rats than to the mechanisms in pigs, despite the fact that humans are phylogenetically closer to pigs than to rats. And the carcinogenic effect of aflatoxin B is more similar in rats and monkeys than in rats and mice.

So, to reason that phylogenetic continuity implies underlying causal similarity is a fallacy.

As if all this were not enough, an additional complication is given by the fact that the various sub-systems of a biological organism interact with each other, thus multiplicating the number of variations of possible effects.

Biological objects are complex in the extreme: this is why the simple modelling method to test hypotheses that animal experimenters have imported from physics does not work in biomedical research.

Moreover, the differences between species will be greater and more difficult to compensate for exactly in those areas which interest animal researchers: this is the case, for example, of metabolic differences between species, which are centrally important in toxicological and teratological (effect on the fetus) investigations.

As one widely-used pharmacology text sums it up: "The lack of correlation between toxicity data in animals and adverse effects in humans is well known".

In short: different animal species are similar functionally but not causally, and that includes the human species. We tend to think they are similar because we only consider what is visible, ie the functions, not the underlying causal mechanisms.

Finally, the argument often used by animal experiments’ advocates, “It just works”, is subjected to such a potent critical analysis that it leaves it almost as naked as a tree in winter.

Partly, the authors here use an argument which is similar to the one explaining the post hoc, propter hoc fallacy used by Pietro Croce in his classical book Vivisection Or Science: A Choice To Make.

In a nutshell, simplified, it goes like this: how do we know that animal experimentation works?

Because, wait for this, after experimenting on animals, we then test the results of those animal experiments on humans!