Sington's Aspirin
a BBC Science Producer reveals his scientific ignorance

David Sington is, or was, one of the BBC Horizon production team who answered complaints against the bias of Nuclear Nightmares. This highly biased 2006 documentary pissed all over the idea that radioactivity from Chernobyl had caused any observable health effects. One of the grounds of complaint was that Horizon's journalists had a poor grasp of the science, and Mr. Sington proved the point by trying illustrate the implausibility of supra-linear dose responses. To do this he invoked a hopelessly garbled analogy with aspirin:

It is hard to see why taking one aspirin a day for 100 days might be more dangerous than taking all 100 in one day, yet that is what supra-linearity requires — that receiving a tiny dose of radiation every day is actually more dangerous than getting the total dose all at once.

He has revealed his ignorance with two errors. One is that dose protraction (a little every day) has, in principle, nothing to do with it. 1 The other is that he hasn't understood the relevance of heterogeneity - uneven distribution of energy within tissues. 2

His analogy would work if all radiation exposures could validly be averaged. Aspirin, being soluble, can be averaged. Some sources of radiation can be averaged as well, but many can not; hot particles are only one of many examples (see these caveats from ICRP about other kinds of exposure. Incidentally, they are predominantly anthropogenic.). The analogy requires us to imagine a form of aspirin that stays as a discrete tablet after you swallow it. It must also be able to migrate, still intact, to a susceptible target tissue where it delivers all its dose locally. Only under such circumstances is Sington's idea worth a light.

If, under these circumstances, the hypothetical aspirin had enough fire-power to damage a target equivalent to 1% of the body, its effects on that tissue would be equivalent to Sington's 100 aspirins. If a whole body dose of 100 aspirins caused one fatal cancer then the one localised aspirin would cause one fatal cancer; that is, it would have the same effect dose for dose. This is a more relevant analogy and, though it is still simplistic, you can see that a pharmaceutical industry which, thinking there was only one form of aspirin, dispensed both sorts indiscriminately wouldn't be able to understand how one aspirin could cause the cancer — the dose is 100 times too small, it must be something else. This is the same problem as the radiation protection community faces. And if a campaign group started banging a drum about supralinear dose responses to aspirin, the pharmaceutical industry and the Health Protection Agency, and some doctors, and a lot of ignorant smartarse journalists (and possibly Ben Goldacre) would think the campaigners were nuts, spreading aspirophobia, depriving the sick of a useful drug, and doing reputational damage to a lot of kindly, white-haired pharmacists.

In fact the apparent supralinearity is an artefact of a false conceptual model. If radiation risk diagrams used "ionisation density" as dose units we wouldn't see the dose response curve taking unexpected supralinear, biphasic and polymodal swerves, all of which are observed under various conditions. For many health endpoints (particularly those that depend on double strand breaks) dose response might well follow a straight line.

Richard Bramhall
Low Level Radiation Campaign

---

Footnotes

1 Split dose experiments show dose protraction does play a role in inducing mutation, probably because the first dose initiates cell repair which subsequent doses interfere with. This was proposed by Busby and is now the basis of some cancer treatments, but it is not at all the same thing as claiming that dose protraction is necessary to demonstrating the existence of supra-linear effects. To do that one only needs epidemiological or experimental findings which show a greater proportional (i.e. dose for dose) effect at lower doses. There are many such results.

2 For the sake of keeping this simple, I am leaving aside considerations of epigenetic effects (genomic instability and the bystander phenomenon) which make nonsense of linear dose response assumptions because they mean that the target for induced effects is necessarily far greater than the target for radiation tracks.