Science for Sale?

The tendency of industry-sponsored studies to find new drugs much safer and more effective than independent studies do has been documented time and again (for instance, here), so it should come as no surprise when industry-funded studies of toxic chemicals give them a cleaner bill of health than independent studies do. In the case of bisphenol A, the molecule from which the durable, transparent plastic polycarbonate is made, the divergent conclusions are especially stark, and an analysis in the April 16 issue of Chemical and Engineering News probes why.

Bisphenol A has been the target of consumer advocates' ire because it is what's called a synthetic estrogen. That means its physiological effects mimic those of this female hormone. Since bisphenol A has been found in products from baby bottles to food and beverage containers, that's potentially worrisome: nearly everyone in the U.S. and other industrialized countries has been exposed to it. The U.S. Centers for Disease Control and Prevention finds it in 95 percent of urine samples.

Industry-paid studies conclude we shouldn't worry. Of the 13 studies of bisphenol A underwritten by chemical companies, none reported any adverse effects on the lab animals or tissue samples exposed to it, C&EN finds. But of the studies funded by government, 153 found such effects (including obesity, cancer and insulin resistance).

How can the results diverge so much? You might think that studying whether a chemical is harmful entails feeding it to lab animals and seeing what happens. But there are all sorts of ways to stack the deck. One is to pick a lab animal that is insensitive to estrogen. Two large studies of bisphenol A that were funded by industry used a strain of rat called Sprague-Dawley, which is well-known to be insensitive even to powerful estrogens. No surprise, then, that the rats showed no ill effects of exposure to bisphenol A, a weaker estrogen.

One can hope that people, too, are insensitive to weak estrogens, but there is no empirical evidence of that. In fact, several of the effects of bisphenol A that have been documented in rodents include those that have recently been rising in people. The incidence of prostate cancer increased 85 percent from 1975 to 2002, for instance, and childhood obesity (some cases of which have been traced back to insulin resistance) has more than quadrupled since the 1960s. Bisphenol A causes prostate cancer in adult rats exposed to it as pups, as well as insulin resistance and obesity in rats whose pregnant mothers were exposed to it.