Bisphenol A

I didn't bother to freak-out when the panic hit over the latest poison-du-jour, 2,2-BIS(4-HYDROXYPHENYL)PROPANE (bisphenol A). Statistical experience, plus the smattering of toxicology I have to know to do my job, has taught me to be cautious regarding chemical scares. This definitely turns out to be the case here.

Courtesy of George Mason University, STATS provides a helpful (if lengthy) primer on the current research on BPA. For those who don't want to delve into 27,000 words, I will summarize:

BPA is a chemical that deserves some level of attention because it can mimic the effects of estrogen in the human body. A number of studies in mice showed effects upon their reproductive systems. So, larger studies were commissioned by agencies such as the United States Food and Drug Administration and the European Food Safety Authority. These agencies have repeatedly discovered no harmful effects at many thousands of times the doses that anyone has ever been exposed to by plastic food containers.

Furor erupted. One scientist, Frederick vom Saal, claims that the government-mandated studies must be biased, because they were funded by companies that produce products containing BPA. The agencies countered by claiming that vom Saal's work did not meet their standards, and lacked statistical power and rigor. vom Saal uses the press to spread his story, while the agencies rely on their bureaucratic inertia to quell opposition.

However, this is not something that really requires a great deal of scientific or statistical expertise to understand. The basic differences between the studies of vom Saal and the corporate studies were sample size and route of exposure.

Sample size is easy, bigger samples are almost always better, and even better yet is when you introduce extra variability (such as making sure each animal comes from a different litter, rather than using several from each litter). This all costs a great deal of money, so the corporate researchers have the advantage here. The FDA made them do it, but nevertheless.

Route of exposure is really the big hitter. vom Saal and most of the other studies to show an effect used subcutaneous injection, whereas the corporate studies featured oral exposure. The FDA and the EFSA prefer this method because it mimics the route of exposure in the general population. There is a subtle point here, because BPA is broken down by the digestive system. This is something you wouldn't know unless you are an expert (as I am not), but in general most poisons are not effectively absorbed by eating them in comparison to direct injection. The oral studies in mice showed no effect at the doses studied, while injecting the same amount in a mouse does show an effect.

This is not particularly surprising, for the reason noted above. But it does mean that the oral studies are more appropriate for determining whether there is a public health issue, which is precisely the position of the FDA and EFSA. Furthermore, the doses in question are astoundingly small. The baseline, No Observed Adverse Effects Limit (NOAEL) was 5 mg/kg of bodyweight. The actual measured exposure is only .0001 mg/kg of bodyweight. Let us not forget that the NOAEL is intended to be ingested every single day of an entire human lifetime.

5 mg/kg of body weight causes no effect orally, but effects subcutaneously, however neither is even close to the actual amount you get out of a water bottle. Therefore, I keep drinking out of my Nalgene bottle.

My prediction is that in a couple of years someone will measure infinitesimal amounts of chromium in all the stainless steel water bottles you see these days, and a new panic will be born.