How America became obsessed with BPA
“Their decision was absurd”
One of the other major discussions of the science around BPA in the paper also deserves close analysis, as it again shows how the focus of the paper’s investigation was virtually dominated by vom Saal and his collaborators’ perspective on BPA. This is an extract from a story published in the Journal Sentinel on Jan 23, 2008:
“Another study, funded by the National Institutes of Health, showed that mice exposed in utero had a higher propensity to have prostate lesions than animals that were not exposed. Again, this study was considered of only limited value because the animals were not exposed orally.
In response, Frederick vom Saal, a biologist at the University of Missouri-Columbia and a vocal critic of the panel, decided to test the panel's assumption.
‘Their decision was absurd,’ he said.
‘First of all, fetuses don't eat,’ he said. ‘Anything in maternal blood will freely cross the placenta. And unless the chemical is immediately cleared out of the mother's system, which it isn't, that blood will go immediately to the baby.’
In addition, fetuses and newborns lack, or express at low levels, the liver enzyme that deactivates the chemical.
‘This is not news,’ said vom Saal. ‘Pediatricians will tell you, babies are not little adults. They do not process chemicals the same way adults do.’
To demonstrate this, vom Saal and fellow researchers Wade Welshons and Juliet Taylor exposed 3-day-old female mice to bisphenol A. They separated the mice into four groups.
Two groups were exposed to the chemical through the mouth - one group received a high dose, the other a low dose. Two other groups received injections - again, one high and one low.
Animals were killed at intervals over the next 24 hours, and concentrations of bisphenol A in the blood were measured.
The team found no difference between animals that had received the chemical orally or via injection.
‘It wasn't just that there was no difference,’ said vom Saal. ‘It was exactly the same.’
Vom Saal said that both people and rodents have this particular enzyme, and in both cases, fetuses and newborns do not express it at the same level as adults.
This new research has the potential to upset the panel's findings, said Gail Prins, a researcher at the University of Illinois at Chicago who has been critical of the panel's report.
The bottom line
She said that what really matters in these studies is the concentration of biologically active bisphenol A in the blood, irrespective of how it got there. Animals should have concentrations that are similar to what isfound in people, because that is what is relevant in these studies.
However, Robert Chapin, the chairman of the panel, and an executive at Pfizer, said the new research ‘stands in contrast to a number of other studies that show the opposite.’ He said it was those other studies that ‘led us to the logical conclusion we reached.’
When asked to supply the citations for those studies, he said he could not remember them offhand. He also said that if other scientists could replicate vom Saal's work ‘and provide a rational explanation for the sudden shift,’ the panel would reconvene and reconsider its position.
Prins said her lab will take up that challenge immediately.
The Journal Sentinel reviewed the panel's report and found several studies that showed differences betweenoral and non-oral exposures in adult animals, but none that looked at newborns.
When asked to respond to the Missouri study, L. Earl Gray Jr., an Environmental Protection Agency toxicologist and a member of the panel, forwarded a study funded by the American Plastics Society. He said the study, which was reviewed by the panel, suggested that newborn mice have enough of the liver enzyme to deactivate bisphenol A at low doses.
However, the authors of the study, who were from Dow Chemical, reported that 4-day-old mice had a 10- to 18-fold higher concentration of biologically active bisphenol A in their blood than adults - a finding that vom Saal and Prins say supports their contention.
‘They had this information right there,’ said Prins. ‘Yet, they ignored it.’”
First, Chapin is not an executive at Pfizer; he is a researcher – the head of the Screening and Investigative Lab in the Developmental and Reproductive Group, where he leads a group of investigative deve-lopmental toxicologists and works on mechanisms of male and female reproductive toxicity and fetal developmental toxicology. Before that he was at the National Institute of Environmental Health Sciences, working on mechanisms of male reproductive toxicity. By labeling him an executive it makes it seem as if he is on the business side of Pfizer and not someone with real scientific expertise.
Second, in asking Gail Prins to comment on the validity of vom Saal’s critique, the Journal Sentinel neglected to mention that vom Saal was co-investigator with Prins on a five year study funded by the National Institute of Health’s NIEHS on “Developmental Toxicity of Environmental Chemicals”, R01 ES11283. Prins was also a signatory on the Chapel Hill Consensus statement on BPA.
This again raises the issue of cherry-picking seemingly independent experts to comment on the validity of the science or arbitrate a point of dispute in the narrative, when those experts have specific links and perspectives to the science and scientist in question. In effect, Prins is validating her own position in being asked to validate vom Saal. That’s fine, but the Journal Sentinel should have disclosed that they are co-researchers.
The first substantive issue with the passage is that studies showing the effects of BPA in mice need to be treated with caution. As the European Food Safety Authority cautioned, caution that extrapolating results from toxicity studies on mice was problematic because
“The much higher levels of oestradiol, oestrone and oestriol sustained by the human fetus as compared to the mouse fetus suggest a much lower sensitivity of humans as compared to rats or mice to possible additive effects of weak oestrogens such as BPA (Witorsch, 2002b).”
With respect to the idea put by vom Saal that "Pediatricians will tell you, babies are not little adults. They do not process chemicals the same way adults do." Willhite noted the following via email:
"Adult rats and mice are not just little people - just like people are not just big rats or mice. While it is true that infants and children are not just 'little adults', it is also true that baby rats and baby mice are not baby people.
When we compare the BPA rodent data to the human, we must take into account not only the rates of conjugation but also the ability to eliminate the BPA-glucuronide product of metabolism in the urine; that is why we have integrated methods called "physiologically-based pharma-cokinetic models" (like that published for BPA during 2007 in Toxicology and Applied Pharmacology Volume 224, pages 182-191). One cannot just "cherry-pick" the data and make interspecies (or even intraspecies) comparisons based on a single factor like a difference in glucuronidation efficiency.”
Willhite pointed out that there are, therefore, two questions to consider. The difference between rodents and humans in how BPA is metabolized and the difference between people of different ages, genders and races. With the former there are always going to be conditions particular to each chemical, but, writes Willhite,
“[W]e do know that the activity of UDP-glucuronsyltransferase in humans is twice that of the rat. This contributes to the often more rapid (not slower) rate of detoxification of drugs and other chemicals that are processed by phase II conjugation reactions in people than in rats (see Drug Metabolism and Disposition Volume 33, pages 1513-1520, 2005)."
As to the second question he continues:
“Assessing the differences in conjugation capacity between infants, young children and adults must rely on studies with various therapeutic drugs that are metabolized by what are called phase II reactions (as described above and on pages 79-81 of the 2008 review by NSF International published in the Journal of Toxicology and Environmental Health B Volume 11). Since the BPA-glucuronide conjugate is eliminated in urine, one must also remember than the infant rate of glomerular filtration (a measure of kidney function) is less than that in the adult so this can prolong the rate of elimination of some drugs and other chemicals in babies compared to adults. This does not mean that a chemical is "more" or "less" toxic in infants than adults, only that the rate of elimination is different.
In humans and rodents, this conjugation (detoxification) is carried out by a set of enzymes called UDP-glucuronosyltransferase. In humans, there is a progressive increase in the activity of this enzyme system with increasing age. Among children less than 11 years of age, the activity of this enzyme system is less than that of adults, but when the rate of its activity is compared on a body surface area basis (a correlation called allometric scaling that can be used to compare differences in metabolic activity between species - say between an elephant and a hummingbird - or between members of the same species) there is actually no difference in enzyme conjugation activity between any of the pediatric age categories and that in the adult (see Drug Metabolism and Disposition Volume 34, pages 2097-2102, 2004).
In very young children, these metabolic pathways are not fully mature until ~3 months of age (see Food Additives Contamination Volume 15, pages 45-51, 1998). As the infant ages from premature to full term, studies with various drugs show the premature infant is about 4 times less efficient than the adult and the newborn full term infant is about 3 times less efficient than the adult; but for toddlers at 1 year the rates are equivalent to those in the adult (Developmental Pharmacology and Therapeutics Volume 16, pages 71-77, 1991). But this is only part of the difference.
If one then examines a broad range of chemicals [as was done by the Connecticut Department of Health and the U.S. EPA], it can be seen that the overall differences between adults and infants in their rates of conjugation and urinary elimination disappear by 2 to 6 months of age. Even more to the point for children ages greater than 6 months, the rates of elimination of some chemicals can actually be faster than in adults (see G. Ginsberg, D. Hattis, B. Sonawane, A. Russ, P. Banati, M. Kozlak, S. Smolenski and R. Goble. 2002. Evaluation of child/adult pharmacokinetic differences from a database derived from the therapeutic drug literature. Toxicological Sciences Volume 66, pages 185-200). These factors are specific to the particular chemical under consideration - so one cannot make helter-skelter generalizations about the differences in conjugation and elimination rates between adults, infants and children. It depends on the particular substance - be it a therapeutic drug or an environmental chemical.
Often more important than whether you consider the age of a person, it is their individual genotype that dictates the efficiency of glucuronidation – but this again depends on the specific chemical. One of the best examples happens to be that for estradiol, the most potent physiologic form of the estrogens. Human interindividual differences in conjugation of estradiol vary up to 16-fold (see British Journal of Clinical Pharmacology Volume 64, pages 458-468, 2007) – a value greater than the difference between that of an infant and that of an adult. In the case of BPA, however, we know from studies in Korean adults (see Environmental and Molecular Mutagenesis Volume 47, pages 571-587, 2006) that there are no metabolic differences between males and females or that depend on lifestyle factors (e.g., tobacco use, ethanol consumption) or differences in expression of enzyme polymorphisms (differences in UDP-glucuronosyltransferase and sulfotransferase) that influence BPA elimination. Therefore, when deriving a reference dose for BPA the standard application of a 10-fold interspecies uncertainty factor encompasses nearly all people.
So at this point with BPA, we don't need to kill more mice. We need the urinary BPA data for infants, then we can compare the daily exposures for infants and other age groups to the oral reference dose calculated from the available and very robust database that already exists in animals to the human level of daily BPA exposure, and then determine the quantitative margin of exposure - as has been done by Dekant and Volkel (Dekant, W. and W. Volkel 2008. Human exposure to bisphenol A by biomonitoring: Methods, results and assessment of environmental exposures, Toxicology and Applied Pharmacology, Volume 228, pages 114-134.) for all groups - except for of course the missing infant exposure assessment.
That data has begun to materialize, and it appears to give even further credence to the arguments that the risk of BPA is best assessed from oral exposure rather than injection.