Must Read Article on Bisphenol-A in the Public Library of Science

There’s a great piece of science journalism over on the Public Library of Science – Biology written by Liza Gross, highlighting Fred vom Saal with the University of Missouri and his work studying the developmental effects of endocrine disrupting chemicals, particularly the plastic monomer bisphenol-A (BPA).

Dr. vom Saal started his research with the risks of exogenous estrogens with diethylstilbestrol (DES). From 1938 to1971, U.S. physicians prescribed DES to pregnant women to prevent miscarriages and avoid other problems with pregnancy. At the time, it was thought that miscarriages and premature births occurred because some pregnant women did not produce enough estrogen naturally, and that the use of DES was safe. It is estimated that 5-10 million pregnant women and the children born of these pregnancies were exposed to DES.

However, what Dr. vom Saal is best known for is the controversy regarding the possible health risks from the endocrine effects associated with low-dose exposure to BPA. BPA is used as the monomer to manufacture polycarbonate plastic. Polycarbonate is used as the lining for most food and beverage cans, as dental sealants, and as an additive in other widely used consumer products. BPA is one of the highest-volume chemicals produced worldwide. Heating of containers to sterilize food, the presence of acidic or basic food or beverages, and repeated washing of polycarbonate containers have all been shown to result in an increase in the rate of leaching of BPA. According to the Centers for Disease Control (CDC), most everyone in the U.S. is exposed to trace levels of BPA, based on urinary monitoring data.

BPA is a textbook example of an environmental toxicology and health policy problem. It's a commercially important chemical substance. Nearly all of us are constantly exposed to low levels of it. The mechanisms for potentially adverse health effects are subtle and occur at low levels of exposure. The most vulnerable populations for these effects are young children, both during pre-natal development and in infancy. The kinds of biological effects observed in laboratory animals with low levels of exposure include stimulating the growth of prostate (in males) and mammary tissues (in females), potentially increasing susceptibility to carcinogenicity later in life; alterations in hormonol with effects such as early onset of sexual maturation; and neurobehavioral effects. Recently, one study provided indications that BPA might promote insulin resistance, a risk factor for diabetes. These effects can be observed at levels around 10-fold greater than levels of exposure in humans, which is not a comfortable margin of safety. As yet, adverse effects have not been observed in humans, though there hasn't been a concerted effort made to examine human populations.

Assessment of the human health risks from exposure to BPA gets really bogged down in arguments over the weight of evidence considered by different groups reviewing the evidence, disagreements about what really constitutes a significant adverse human health effect, the financial sponsorship of those groups, and arguments over differences in testing methods used by various investigators. Whether or not a low-dose estrogenic effect can be detected seems to depend on numerous factors including the strain of test animal used, purity of the chemicals used, timing of the dosing, composition of animal feeds and whether or not a positive control (a test group dosed with a known endocrine disruptor such as DES). The history of the scientific review of BPA further obscures the public debate. Ms. Gross chronicles the progression from government agency (the National Toxicology Program) concluding in 2001 there was credible evidence of low-dose estrogenic effects, to review by the Harvard Center for Risk Analysis report sponsored by the American Plastics Council which concluded that the evidence was weak that low-dose estrogenic effects were present in the animal studies, to the rebuttal co-authored by vom Saal in 2005 that called for a new risk assessment of BPA, to the 2006 follow-up by Gradient Corporation which amplified the conclusions from the HCRA report.

The article brings up the topic of the "expertise-for-hire" syndrome which produces scientific information that calls into question our understanding of potential occupational or environmental health risks, particularly when the studies in question indicate there may be adverse health effects. The outcome is the call to delay or defer risk-reduction activities until further studies can be conducted, because there isn’t sufficiently compelling evidence that a health risk exists and warrants action. David Michaels and Celeste Monforton, at George Washington University have written extensively on this topic (I blogged one of their articles a few years ago). More information on this subject of "manufactured uncertainty" can be found at the Project on Scientific Knowledge and Public Policy website. What the manufactured uncertainty tactic accomplishes is to provide a counter-point which responds nicely to the media's urgent need to create a simplistic, dramatic narrative of two opposing points of view; an approach that is used sell airtime, newspapers or magazines, but isn't terribly helpful for informing the public about an important environmental health policy issue.

The commentary co-authored in 2005 by vom Saal and endocrinologist Claude Hughes observed that most of the industry-funded studies concluded that adverse estrogenic effects were weak or non-existant, while the studies that were government-funded identified adverse effects. While there's no cry of foul play or anything like that, vom Saal suggested many investigators were unfamiliar with the methods used to study low-dose estrogenic effects:

“The moment we published something on bisphenol A, the chemical industry went out and hired a number of corporate laboratories to replicate our research. What was stunning about what they did,” vom Saal says with a mix of outrage and bemused disbelief, “was they hired people who had no idea how to do the work. Each of the members of these groups came to me and said, ‘We don't know how to do this, will you teach us?’”

There's always a risk in placing a lot of emphasis on findings from one lab that others can't replicate. However, a quick search of PubMed seems to show recent papers from multiple investigators observing low-dose estrogenic effects, so it seems that there is some reproducibility, one of the gold standards of scientific evidence. Another controversy with BPA research is that the source of funding seems to influence whether or not adverse effects are detected in an animal study. It's a complex matter. According to vom Saal and his co-author, Claude Hughes, the industry-funded studies lacked positive controls or used insensitive animal strains. Hughes does allow for the possibility of publication bias in the government-funded studies (i.e. do the studies with negative results get published?). To further complicate matters, there are various opinions regarding what gets considered as an adverse health effect in a laboratory animal, and how that is considered in a human health risk assessment. Coincidentally, NTP has recently published proceedings from a workshop on the relevance of rodent bioassays in assessing hormonally-induced reproductive tumors, which expresses a concern that currently available animal models aren't adequate for assessing these hormonally-induced effects.

Sorting through all of this must be tough for the non-specialist public and decision makers. The kinds of adverse effects being observed in laboratory animals sound like really bad things, but they seem to be pretty hard to detect and its difficult to say how they relate to human health. At the same time, how conclusive does the evidence from animal studies need to be before you start taking steps to reduce exposure? With DES, the evidence of adverse effects was first observed in humans, then confirmed with animal studies, which is not exactly the environmental health surveillance model you'd want to emulate. It comes down to a policy and philosophical judgment with the competing philosophies being "sound science", deferring action until we know or have fully assessed the health risks and the precautionary principle of taking protective action based on a more conservative interpretation of limited data. Such arguments can only be partially informed by the science.

I'm wondering if we could split the difference, while we're figuring out what the rodent data are telling us about BPA. There's probably enough evidence to argue that a sensible course would be to act now to restrict the uses with the highest potential for human exposure. These probably would be the food and beverage containers and dental appliances, where BPA is being leached and ingested. There are other uses that we probably don't have to worry about just this moment - how much BPA are you going to absorb through your fingertips from a polycarbonate compact disc? However, I'm not optimistic that a policy of moderation is going to have too many takers. As Jim Hightower said, "the only things in the middle of the road are yellow stripes and dead armadillos"; the plastics industry is going to stubbornly defend all uses; the endocrine disruption activists will stubbornly argue that we need to ban polycarbonates; both sides will have their experts, which prolong rather than resolve the scientific controversy; and the mainstream media gets to have conflict, narrative and drama, which might be great for attracting readers or viewers, but provide little for informing them.

Altogether this was a nice piece of science journalism, and as you can see, gave me a bit to think about. If you are new to the topic of BPA and confused by the sound bites you read in the mainstream media, this is a great place to start learning about this chemical and the associated environmental health issues.