Skip to Main Content

Government & Policy

April 16, 2007
Volume 85, Number 16
p. 38


Bisphenol A On Trial

Only an unbiased panel with appropriate expertise can resolve apparently conflicting results of health studies

Bette Hileman

In industrialized countries, nearly everyone is exposed to bisphenol A (BPA), a weak synthetic estrogen. Most of the 2 billion to 3 billion lb produced in the U.S. each year is used to make polycarbonate food containers, baby bottles, refillable water containers, compact discs, and resins that line metal food and beverage cans. The Centers for Disease Control & Prevention has found measurable levels—0.4 to 8.0 parts per billion—of BPA in 95% of U.S. urine samples.


Although BPA is only one of thousands of synthetic chemicals the U.S. population encounters, it may be the most important for the developing fetus and perhaps young children. Every time children consume food or drink from a can, they are exposed to a small amount of BPA. Every time babies drink from clear polycarbonate plastic bottles, they ingest the chemical.

Because of this widespread exposure, researchers have conducted many studies of the health effects of low doses of BPA. Among government-funded experiments on lab animals and tissues, 153 found adverse effects and 14 did not. Contrarily, all 13 studies of BPA funded by chemical corporations reported no harm. The studies indicating harm report a variety of deleterious effects in rodent offspring exposed in the womb: abnormal weight gain, insulin resistance, prostate cancer, and excessive mammary gland development.

What can explain the vast discrepancies in the findings of government-funded and industry experiments? In a commentary in Environmental Health Perspectives, University of Missouri biologist Frederick S. vom Saal points to several reasons low-dose studies failed to find adverse effects. One is the strain of rodent used. For example, two large industry-funded studies used the Sprague-Dawley rat, a strain known to be insensitive even to strong, well-characterized estrogens, such as diethylstilbestrol. So it is not surprising that these rats show no response to the weaker estrogen BPA. Many other rodent strains are far more sensitive, vom Saal wrote.

Another reason is that different batches of the feed used in several industry studies had highly variable estrogenic activities. Phytoestrogens, such as genistein in soy, as well as other estrogenic components, can be present in variable amounts in different batches, vom Saal reports. In some studies, estrogenic substances in the feed may have masked the effects of BPA, he says.

In light of the potential health effects of BPA exposure and the inconsistent study outcomes, it is especially important that an unbiased panel with no conflicts of interest and with a detailed knowledge of the field evaluate the literature on BPA, consider the weight of evidence in regard to adverse effects, and choose valid studies to include in its report.

Bias is the reason for the recent outcry from environmental groups and lawmakers when they became aware that outside contractor Sciences International (SI) was deeply involved in the National Toxicology Program's assessment of BPA (C&EN, March 12, page 13). SI helped select the scientific panel, reviewed the literature, and wrote the original draft report—essentially a literature review—on BPA for NTP's Center for the Evaluation of Risks to Human Reproduction. SI has worked for two BPA manufacturers, Dow Chemical and BASF. NTP is housed at the National Institute of Environmental Health Sciences (NIEHS).

It would seem that a scientific literature review would be an objective, cut-and-dried exercise where conflicts of interest could have little influence. But reality is almost the opposite. For example, researchers in the field know that the biologically active fraction of total circulating BPA is the part that is not bound to plasma proteins, because protein-bound BPA cannot diffuse through cell membranes. So it is the levels of unbound BPA that must be compared when evaluating study results. But some sections of the original draft report on BPA compared results of studies that were measuring total BPA to others that were measuring only the unbound fraction. Another source of bias is omission of critical information from a review.

Because BPA causes adverse effects in rodents that are almost identical to some of the health problems that have recently increased in human populations, many researchers believe BPA may be partially responsible. The incidence of human prostate cancer rose 85% from 1975 to 2002, and insulin resistance, which leads to type 2 diabetes, became a much more prevalent health problem. The incidence of childhood obesity has more than quadrupled over the past 40 years.

When newborn rats are exposed to low doses of BPA, they develop early-stage prostate cancer as adults. Low doses of BPA also cause insulin resistance in lab animals. Several of the most alarming studies released recently link obesity with BPA. After vom Saal and other researchers exposed pregnant rodents to low levels of BPA, their pups gained weight rapidly and stayed overweight for the rest of their lives, while the control animals grew normally.

A similar phenomenon has been observed recently in humans, says Jerry Heindel of the NIEHS. Some full-term infants gain weight abnormally fast and stay obese throughout childhood. He believes that in utero exposure to BPA, or to a mixture of environmental chemicals, may be playing some role in childhood obesity. High-calorie diets and lack of exercise are certainly important but may not be the only factors leading to excessive weight gain in children.

It seems urgent to investigate whether exposure to certain chemicals in the womb is one possible cause of prostate cancer, insulin resistance, and childhood obesity.

Views expressed on this page are those of the author and not necessarily those of ACS.

Chemical & Engineering News
ISSN 0009-2347
Copyright © 2011 American Chemical Society