Context: First synthesized in 1891, bisphenol A is now used in a wide array of consumer and industrial products, including polycarbonate plastic and expoxy resins used to line food cans and coat children's teeth. Its ubiquitous use has lead to widespread exposure.
In the 1930s, medical experts discovered that BPA was estrogenic, but the assays of the day indicated it was quite weak compared to human estrogen and to the newly synthesized drug, diethylstilbestrol.
Studies of BPA toxicity in the 1980s showed that at relatively high levels (50ppm) it caused weight loss in rodents. Lower doses were not tested.
Beginning in the late 1990s, researchers began to look at BPA effects at much lower levels of exposure. These studies were launched when scientists from the University of Missouri discovered that BPA behaved differently in serum than the human estrogen, estradiol. Most circulating estradiol is bound by serum-binding proteins and hence biologically unavailable. Very little BPA, however, is bound. This led the Missouri scientists to predict that BPA would have effects at levels below those considered possible based simply on its weak estrogenicity. Their experiments confirmed that prediction, and since then many studies have been published showing BPA effects far beneath the level used to establish the current EPA safety threshold.
In 2005 scientists from the University of Texas discovered even more potent effects of BPA, in effects that work through receptors on the surface of the cell membrane. These impacts can be observed in solutions as low as 0.23 parts per trillion. Through this mechanism, BPA is just as powerful as estradiol.
In 2007, a group of the world's leading experts on BPA research published a comprehensive set of reviews of the BPA literature. Their reviews included new analyses which led them to conclude: (1) that median levels of BPA in human serum in countries like the US, Germany and Japan are significantly above those sufficient to cause adverse effects in animals; and (2) that the only way that human serum levels could be that high is if current human exposure is above the safety standard that has been adopted by the US and Europe.
What did they do? Calafat et al. analyzed urine gathered by the National Health and Nutrition Survey (NHANES) in 2003-2004. This survey is carefully designed to obtain a representative sample of Americans. They measured bisphenol A (BPA) and 4-tertiary-octylphenol (tOP) levels in 2517 samples. Their sensitive methods allowed them to detect BPA levels as low as 0.4 μg/L and tOP levels as low as 0.2 μg/L. In their statistical analysis of the data, they compared levels in different ethnic groups, ages and among different household income groups.
What did they find? They detected BPA in 92.6% of the people sampled (none younger than 6 yrs old) and tOP in 57.4%. The geometric mean concentration for BPA was 2.6 μg/L while the 95th percentile concentration was 15.9 μg/L. They did not calculate a geometric mean for tOP because fewer than 60% of samples had detectable levels. The 95th percentile for tOP was 2.2 μg/L.
Levels of BPA were significantly lower in Mexican Americans than in non-Hispanic blacks (p = 0.006) and non-Hispanic whites (p= 0.007). They found no difference between non-Hispanic blacks and non-Hispanic whites (p = 0.21).
Females had higher levels than males (p = 0.043). People in the lowest category of household income (<$20,000 per year) had higher levels than those with higher household income (>$45,000 per year).
Children had higher levels than adolescents (p < 0.001) who had higher levels than adults (p < 0.003).
What does it mean? Because of the relatively short half-life of BPA in people following ingestion, these data indicate that Americans are exposed continuously to BPA.
The levels reported by Calafat et al. indicate the exposure for many people is above the EPA's reference dose of 50 μg/kg/day, according to calculations reported in a comprehensive examination of human exposures published in August 2007. The reference dose is an exposure level calculated by the EPA that is, in theory, low enough to be safe. Many studies have been published in the last 10 years indicating that the current reference dose is dramatically out of date and far too high. Indeed, a careful analysis of the data available on human and animal metabolism of BPA indicates that levels in people today are higher than those sufficient to cause a wide range of adverse effects in animals.
Calafat et al.'s discovery that children have the highest levels of exposure is troubling, for two reasons. First, most of the experimental work with animals shows that exposure during development is the most problematic. Second, none of the children sampled were younger than 6 because of the way the NHANES is designed. Exposure to infants and toddlers may be much higher: Many are fed milk in polycarbonate bottles which have been warmed in water or microwaved. Heat increases the rate of BPA leaching. Baby formula can be another source of BPA because of leaching from the cans in which it is sold.
Relatively greater exposure in low income households reveals an unexpected health risk that could be confounding epidemiological studies of disease endpoints associated with BPA on the basis of animal studies, for example, obesity.