||vom Saal, F and C Hughes. 2005. An Extensive New Literature Concerning Low-Dose Effects of Bisphenol A Shows the Need for a New Risk Assessment. Environmental Health Pespectives, in press.
Bisphenol A in the news...
Excerpts from teleconference about bisphenol A
Audio clips (mp3) will open in new window.
FVS= Fred vom Saal, Ph.D. Professor, University of Missouri, Columbia and co-author with Claude Hughes of vom Saal and Hughes 2005.
GL = George Lucier, Ph.D. Former Associate Director, National Toxicology Program, National Institute of Environmental Health Sciences
FVS: How is BPA used?
FVS: Why are public health experts concerned?
FVS: Why are animal data relevant?
FVS: What are the effects?
FVS: How do the animal exposures compare to human exposures?
FVS: What is the chemical industry saying?
FVS: What about the Harvard Center for Risk Analysis?
GL: Why is a new risk assessment needed?
GL: How widespread is exposure?
GL: What about the review by the Harvard Center for Risk Analysis?
GL: What agencies are responsible for regulating BPA?
Listen to entire teleconference, including reporter Q&A
Vom Saal and Hughes provide an overview of rapidly accumulating evidence in the scientific literature that current health standards for bisphenol A (BPA), a molecule used widely in commerce to which virtually all Americans are exposed, should be strengthened dramatically to protect public health. They conclude that a flood of new scientific evidence on adverse effects of BPA requires a new, formal risk assessment by relevant US agencies; none has been conducted for over 15 years.
Indeed, the scientific results they summarize indicate that to keep exposure levels beneath those now known to cause adverse effects in laboratory studies, it may be necessary to restrict or even ban many common uses of bisphenol A.
Research over the past decade has established that BPA alters cellular function and disrupts developmental processes at exquisitely low levels, far beneath EPA's current 'reference dose' for the compound, and at levels to which many people are exposed daily, in the US and other countries. The list of human conditions that might be caused by BPA is long, and hence reducing exposures may provide dramatic health benefits. These could be rapid because BPA is not persistent.
Their analysis also reveals a striking pattern of bias in the reporting of research findings. In the last 7 years (through to December 2004), 115 relevant studies have been published. None of the 11 funded by industry reported adverse effects at low level, whereas 94 of 104 government-funded studies found effects. These many studies were conducted in academic laboratories in Japan, the US and Europe.
vom Saal and Hughes' conclusion about the need for a new risk assessment contradicts several recent reports from individuals or organizations associated with the chemical industry which concluded that existing evidence did not suggest BPA to be harmful.
What did they do? Vom Saal and Hughes summarize an extensive scientific literature which now documents many different effects of BPA.
The evidence comes from diverse university and government laboratories around the world, studying many different effects in a wide array of experiment types, including laboratory experiments with animals, experiments with cells in cell culture, and microarray analysis of changes in gene expression induced by exposure.
The scientific literature now provides overwhelming evidence that bisphenol A alters cellular signaling, fetal development and adult physiology and reproduction in animals at doses far beneath the current 'safe exposure' level established by the US, 50 µg/kg (50 ppb). This 'reference dose' is based on experiments conducted prior to 1988 indicating that BPA had no observable effect below 50 mg/kg (50 ppm).
Vom Saal's laboratory's first publication on low dose impacts of BPA reported effects at 2 ppb: male mice exposed in the womb grew up with enlarged prostates that were hypersensitized to further androgen stimulation.
What is bisphenol A? BPA is molecule invented in the 1930s as a synthetic estrogen. Polymer chemists subsequently learned that the individual 'monomers' of BPA could be combined in molecular chains--polymerized--to make polycarbonate plastic and other materials.
||BPA is now used in countless products, for example, to make polycarbonate water bottles like the wildly popular colored and tinted bottles sold under the trade name Nalgene.
Unfortunately, the ester bonds that link the BPA monomers together to make its polymerized form are not stable. They break up as the plastic ages. The degradation rate is highly sensitive to how the plastic is used. If heated or exposed to acid or basic solutions, BPA leaching rates increase. Multiple cycles in a dishwasher, for example, speed the ageing/leaching process.
BPA is also used to make a resin that lines food cans. While it is difficult to ascertain what percentage of food cans are lined with this resin, some estimates run as high as 80%.
These and many other uses of BPA guarantee that most people are exposed. Indeed, a recent CDC survey detected BPA in 95% of Americans sampled. Most troubling, the levels observed are within the range that animal and cell experiments indicate are sufficient to cause harm.
More background on BPA...
This result was highly criticized by industry scientists and lobbyists, who asserted that the results could not be replicated. As this review makes clear, not only has the specific result on prostate been replicated, but many effects at similar doses have been reported by many independent laboratories. A full list of the literature reviewed can be found on vom Saal's website. Some of the results have been described on the pages within this website. Below we highlight some examples.
Based on the current literature, plausible links have been suggested between bisphenol A and the human health conditions and disabilities listed below. The list (see below) is long.
What is industry saying?
Reporters writing about the vom Saal and Hughes paper interviewed industry representatives about their conclusions, especially the criticisms of the Harvard Center for Risk Analysis 2004 assessment of BPA and the biases associated with industry- funding of research.
The response to vom Saal and Hughes from industry representatives has been to assert that the quality of the studies showing no effects outweighs the quantity of the studies that show impacts.
for extreme industry comments
For example, interviewed by Marla Cone (Los Angeles Times), Steven G. Hentges, executive director of the polycarbonate business unit of the American Plastics Council, said "The sum of weak evidence does not make strong evidence." And interviewed by Elizabeth Weise (USA Today), Hentges said "You can have 1,000 studies, but if they're all weak, adding up weak evidence doesn't necessarily give you strong evidence of anything."
Unfortunately for Hentges, a close examination of the 11 studies that industry points to as proof of no harm reveals serious methodological errors and some actions that appear scientifically fraudalent. These points are fleshed out by vom Saal and Hughes. The most egregious violation of scientific procedure involved, in one of the papers industry cites most often, was the failure of the experiments to show an effect in their positive control group. While this failure was mentioned in the initial press releases announcing the result, it was omitted from the scientific publication that was published in a scientific journal. The failure of the positive control group means that the research was performed incompetently, not that the bisphenol A exposure didn't cause an effect.
And as vom Saal and Hughes point out, the Harvard Center for Risk Assessment, which has been notorious for its industry funding, did not review most of the relevant literature. They were contracted by the American Plastics Council to produce a very narrow review of the issue. In the process, they relied heavily upon some of the flawed industry studies. And most of the relevant literature has been published since HCRA closed its review process to new findings.
Claude Hughes, a coauthor along with vom Saal in this new EHP review, was also a participant in the HCRA review. It is telling that together with vom Saal he concludes that the HCRA effort is outdated.
An immediate question is how could one compound play a role in so many conditions. The traditional view in toxicology has been that a single contaminant might cause a few signature diseases, for example asbestos, which causes mesothelioma and lung cancer. Writing in the New York Times several years ago, Gina Kolata quoted Steven Safe, a Texas A&M toxicologist funded by industry for his prominent role downplaying possible risks of endocrine disruption in public debates in the late 1990's: "The list of illnesses that any chemical is known to cause is very short" (NYT, 11 April 2000).
New scientific findings on the ability of some compounds to alter gene expression at very low levels, however, are supplanting this outdated view. Some contaminants, including bisphenol A, can alter the expression of several hundred genes, with effects varying among specific tissues and also depending upon timing of exposure.
Bisphenol A, moreover, not only interacts with the traditional estrogen receptors within the cell nucleus, and thus alters the expression of many genes dependent upon estrogen signaling, it also stimulates calcium influx into the cell. Just like estradiol, BPA interacts with a cell membrane receptor, leading to the phosphorylation of a transcription factor, cyclic AMP response element binding protein (CREB). This signaling molecule alters the expression of genes involved in many different physiological processes, including brain growth, memory formation, the creation of fat cells, reproductive development, etc. BPA is just as powerful as the native hormone, estradiol, at provoking this response, and more powerful than the notorius drug diethylstilbestrol (DES). Hence it is anything but a 'weak estrogen.'
Given the ability of BPA to alter these very basic physiological functions, it is not surprising that it may be causing a wide array of adverse effects.
The following list provides examples of some of the health conditions and disabilities that animal and cell data currently suggest may be caused, at least in part, by bisphenol A at extremely low doses. (see vom Saal and Hughes for complete set of disease links and references):
Few epidemiological data are currently available to test for these effects in people. Lack of data should not be taken as evidence of safety, however. The relevant human research simply hasn't been done; indeed the analytical chemistry capable of measuring relevant levels in people has only recently been developed. As previously noted, a recent survey by the CDC has established that virtually all Americans are exposed to BPA within the range of exposures that the vom Saal and Hughes summary indicates is relevant. This further complicates epidemiological studies because, with everyone exposed, there can be no clean controls against which to compare cases.
As vom Saal comments in the international teleconference discussing this paper (see links to left), work on estrogen-like compounds conducted by many laboratories around the world, particularly by the US National Institute of Environmental Health Sciences, has shown time and again that animal studies of the effects of estrogenic substances are highly predictive of human impacts.
Impact of funding source
vom Saal and Hughes used library search engines to identify articles in the scientific literature that explored the impact of bisphenol A at levels at or beneath the current reference dose. Through December 2004, they found 115. Over half were published since 2002.
By examining acknowledgements and author affiliations in the articles, they determined which of those studies were funded by industry and which were funded by government. They then examined the impact of funding source on whether the study found an adverse effect or not. The results are displayed in the table below:
||While vom Saal and Hughes present no statistical analysis of this table, the association between source of funding and likelihood of finding no effect is highly significant (Chi-square = 54.5; p <<< 0.001).
Some of the weaknesses of the 11 industry-funded studies are readily apparent. As noted above ("What is industry saying"), one of the studies had its positive control fail, an indication that the whole experiment had failed, not that bisphenol A had not caused an effect. Another industry-funded study concluded BPA caused no effect, but an independent analysis of the experiment's data by scientists convened by the National Toxicology Program concluded that in fact there was an effect; in other words, the industry scientists had misreported their own results.
What does it mean? A wide range of scientific studies have reported adverse effects caused by bisphenol A at levels far beneath the level that regulatory agencies currently believe is safe, and at levels to which virtually all Americans are exposed. Many of these studies have been published within the last 2 years. They link bisphenol A to an extraordinary array of health impacts.
Unfortunately, relevant human studies have yet to be conducted. Laboratory and epidemiological studies of pharmaceutical agents working via the same mechanism of action, however, have consistently found that animal data are highly predictive of human impacts.
vom Saal and Hughes conclude that the current standards for BPA are outdated and that a new risk assessment should be conducted.
What might a new assessment conclude? The most sensitive endpoint yet discovered for BPA was reported by Wozniak et al., examining calcium influx and progesterone secretion stimulated by BPA interacting with a receptor on the surface of the cell membrane. Every BPA dose used in their experiments caused a measurable increase in calcium influx, including at a part-per-trillion level. For this endpoint, BPA is just as powerful at some doses as estradiol, and more powerful than the potent synthetic estrogen diethylstilbestrol, DES. Calcium influxes initiate a wide array of cellular signaling processes with many potential impacts on gene expression, development and a variety of physiological events. These results would indicate that BPA should not be used in any product or process that leads to any human exposure, in other words, BPA should be banned.
There would certainly appear to be enough information to justify steps by individuals to act on their own to reduce exposures. The widespread use of BPA, its unlabeled incorporation into many products, and its abundance in wastewater flowing into the nation's water systems, however, means that individual action will not be sufficient to avoid exposure.