Industry reacts to Consumer Reports' BPA report

Consumer Reports’ recent article on the presence of Bisphenol A in canned food drew widespread interest after it was published this past week. It also quickly drew critical comments from industry groups representing companies that manufacture or use BPA, a chemical whose safety is currently being reassessed by the U.S. Food and Drug Administration.

The groups took exception to some parts of the report that found nearly all of the 19 name-brand canned foods we tested contained this chemical, which is used in the linings of most food and beverage cans. They did not dispute the test findings of the BPA levels we measured in canned food. Rather, the discussion focused on our risk assessment of the effects of BPA, which was based on the scientific literature that has become available over the past 20 years.
 
Here’s a sampling of those reactions, along with a more detailed discussion of some of the research involved in the debate:
             
1. The American Chemistry Council issued a press release contending that our experts’ recommendations, which include calling for a ban on the use of BPA in all materials that come in contact with food, is “inconsistent with the conclusions of expert regulatory bodies worldwide, all of which have confirmed that BPA exposure levels are low and well within safety standards.”

That is exactly the issue. As our story makes clear, food safety experts at Consumers Union believe federal regulatory guidelines—which are the same as those set by the European Food Safety Authority—are outdated and fail to adequately protect consumers. The FDA’s own scientific advisory board also concluded that the agency’s assessment of BPA’s safety is inadequate. Hundreds of scientific studies have shown harm in animal studies from extremely low levels of BPA—levels that are ten to twenty thousand times lower than what the FDA considered as the basis of its safety assessment in 1988. And even some human studies show a link between elevated BPA levels and harmful effects such as diabetes and cardiovascular risk. Our test results show that consumers may be exposed to potentially harmful levels of BPA that could be reached through a few or multiple servings of the canned foods we tested.

2. The North American Metal Packaging Alliance released a similar public statement.  It cited a recently published study funded by the U.S. Environmental Protection Agency, which the can industry trade group says provides “strong new scientific evidence” that exposure to BPA at levels found in our test results is safe.

It is not surprising that the authors did not find effects from BPA because this study used a specific type of rat (Long-Evans) that has been previously shown to be insensitive or unresponsive to low-dose exposures to BPA and even typical birth-control dosages of synthetic estrogen, which was used as a control in the experiment. The insensitivity to both was confirmed again in this study. In other, more estrogenic-sensitive lab animals, BPA has been shown to cause adverse effects at BPA dose levels used in this study.
 
3. A blog posted by Trevor Butterworth, online editor of Statistical Assessment Service (STATS), questioned the scientific evidence used in our risk assessments. STATS says it is a non-profit, non-partisan  organization that acts as “a resource for journalists and policy makers on major scientific issues and controversies.” In his blog, Butterworth claimed that studies we cited as evidence of harm from BPA at low doses are irrelevant because they involved exposing lab animals to BPA via injection rather than orally.

In studies using adult lab animals, injecting BPA results in levels that are similar or slightly higher than those seen after the chemicals are administered orally, making those studies relevant. And a comprehensive study of the metabolism of BPA in newborn lab animals showed that there was no difference in the levels of free BPA based on the route of administration (oral versus injection). This suggests that for newborns, who are especially vulnerable to BPA’s health risks, the route of exposure matters even less than in adults.

It’s not the first time Butterworth has come to the defense of the BPA industry.  Earlier this year he also harshly criticized a prize-winning series of articles about BPA’s health risks by reporters at the Milwaukee Journal Sentinel.  The Milwaukee newspaper recently published a follow-up story describing a public relations blitz  by the BPA industry that “uses many of the same tactics—and people—the tobacco industry used in its decades-long fight against regulation.” 

The story includes a graph mapping a web of potential conflicts of interest in the battle over BPA and noted that STATS is affiliated with the Center for Media and Public Affairs, “a group which was paid by the tobacco industry to monitor news stories about the dangers of tobacco.” Tobacco lobbyists had a keen interest in the government's assessment of BPA because of concerns that a ban on the chemical would affect cigarette filters and plastic packaging, according to the newspaper account.

Consumer Reports examined STATS’  tax returns for  2005 through 2007, which confirm that relationship,  with the 2007 return stating that salary costs for STATS are shared with the Center for Media and Public Affairs. Other documents Consumer Reports has examined show STATS also has received funding from ExxonMobil, a major producer of benzene, one of the components used to manufacture BPA. ExxonMobil also makes a plastic food packaging film containing BPA.

The influence of industry over decisions about BPA’s safety and regulation is also detailed in a fascinating new report “The Politics of Plastics: The Making and Unmaking of Bisphenol A ‘Safety’”, published this week in the American Journal of Public Health.

Tracing the chemical’s history from its commercialization starting in the 1950s by producers such as General Electric, Shell Chemical, Dow Chemicals and Union Carbide through the present, the report notes that even though the government still adheres to a 20-year-old safety standard, some retailers and even BPA producers such as Sunoco now are responding to mounting concerns about the chemical’s safety. Six baby bottle manufacturers have announced that they are removing BPA from their products and Sunoco is asking its business customers to provide written confirmation that the BPA it sells them will not be used in food containers intended for children under the age of three. 
—Andrea Rock, Senior Editor
—Urvashi Rangan, PhD., Technical Policy Director, Consumers Union

Comments

[Raymond Foret]

All the legal smegal, one question, WHY USE MATERIAL THAT HAS ANY SUSPECT DETRIMENTAL QUALITIES. Money honey, nothing but MONEY. The same tactics used in the cigarette industry, delay deny, character assassinations,and when all else fails corporate associations. This is and has been a standard tactic used to deny any involvement in anything they could be held liable for. Bis phenol-a will not be removed from our food packaging without one hell of a fight. Why do they want to continue this practice, I have an idea, DO YOU.

[Amy]

Why all the hue and cry over this issue? Why should the anti-bpa camp be required to prove the chemical is harmful? Why can we just NOT use the stuff??

[Don]

I’m keeping an open mind whether the CR recommended exposure limit for BPA is valid and reasonable. But considering the criticism directed at the methods used by the FDA and the European Food Safety Authority, both of which have specifically reviewed the low dose animal studies on BPA as recently as 2008, I am surprised by the lack of transparency by Consumer Reports to reveal exactly how the CR lowest adverse animal dose was determined and who determined it.
The CR (undisclosed) food-safety scientists propose a maximum limit of 0.0024 microgram/Kg-bw/day based on certain adverse effects from “several animal studies” using an adverse dose of 2.4 microgram/Kg-bw/day. The FDA and EFSA have both re-examined BPA exposure and maintain that the reliable lowest adverse dose is 5 mg/Kg-bw/day. Certainly the FDA and EFSA have relied on many food-safety experts using the same data to arrive at their conclusions, so there seems to be a huge disparity of opinion as to which studies are relevant to human safety.
In determining that adverse dose, it is not clear which studies on which species/strains the CR scientists considered, nor what adverse effects (endpoints) were used and the relevance of these effects to human chemistry, nor the reproducibility of the studies, nor whether the lowest adverse dose is an average of several reports and endpoints (which ones?) or based on just the lowest adverse dose in a few studies. For all we know, the CR adverse animal dose could be based on one adverse effect seen at one dose in one animal in one study.
I realize that most readers are not interested in the details, but considering the huge disparity by a factor of 10,000 to 20,000 between the CR recommendation and the exposure limits currently allowed in the U.S., Canada, Europe and Japan, it would seem CR should be just as accountable and transparent as your readers would expect from those government agencies and should publicly reveal your methods and analyses of the data, perhaps as a separate document on your website.

[The editors]

In response to Don’s comment from 11/8 (see below):

To clarify, we are referring to the insensitivity of this model for clinically relevant doses of EE2 and therefore environmentally-related doses of BPA. While there was some response to higher levels of the control-estrogen used (17-alpha-ethinylestradiol or EE), there were no responses at lower levels that have shown harm in other animal models (e.g. 0.02-2ug/kg-bw/day). A clinically relevant dose of EE2 would be that found in a typical birth control pill (e.g. 20 ug – 50 ug). For a typical woman of 74 kg (CDC figure), that would mean a dose of 0.27ug/kg/day –0.68ug/kg/day. The lowest dose of EE2 that caused an adverse effect in the EPA-funded study was 5 ug/kg/day. In other words, the LE rat did not respond to clinically relevant doses of EE2. Another study investigating the effects of EE2 and BPA in LE rats found that the lowest dose of EE2 causing an effect in male LE rats was also 5 ug/kg/day.
It has been suggested that BPA is 10-1000 fold less potent than EE2 for several toxicological endpoints. Thus, the fact that the LE rats responded only at doses of EE2 at 5 ug/kg/day or higher, would mean that you would not expect to see effects of BPA (due to insensitivity of the model) until doses of 50ug/kg/day – 5,000ug/kg/day were reached or exceeded. So the fact that this LE model found negative results from the BPA doses tested (e.g 2, 20 and 200 ug/kg-bw/day), could simply mean that the LE model is insensitive to BPA at that level, not that BPA has no effect at that level. On the other hand, studies in mice investigating similar endpoints have demonstrated harmful effects at dose ranges that did not have an effect in this LE rat study.

[The editors]

Readers can find the entire BPA story from the December issue of Consumer Reports at this link:
http://www.consumerreports.org/cro/magazine-archive/december-2009/food/bpa/overview/bisphenol-a-ov.htm

[Don]

With regard to the discussion about whether taking money from a foundation (in this case the Scaife Foundation) constitutes an automatic bias or conflict of interest for a specific group (e.g. STATS) because the Foundation's income was founded on wealth from certain industries and/or from donations from certain companies or stock in certain companies, then one would have to apply the same logic to ALL such groups and grantees that are supported by ALL foundations.

So, for example, are all the research projects and universities supported by the Rockefeller Foundations over the years, for example, suspect because the foundation's wealth was based originally on Standard Oil money?
Or the Welch Foundation, founded based on fortunes made in oil and minerals, which supports chemistry research in several major universities.
Or the W.M. Keck Foundation established by the founder of Superior Oil Company, that supports science and medical research at many universities.
Or the Robert Wood Johnson Foundation, based on the fortunes from the founder of Johnson & Johnson, a major supporter of medical research.
Are all the projects and studies by all the professors and organizations that have been supported by these foundations automatically biased because their work was supported by these foundations which depended on wealth or income or donations from certain industries?
Of course, the answer is a resounding NO --- and my list of example foundations here certainly does not suggest in any way that they or their grantees research could be biased because of their past or current relationships. It is only to make a point.

The act of providing funding cannot be, itself, a valid or logical basis for assuming that such funding introduced any bias or conflict of interest unless one can prove a DIRECT LINK between the funding and the study/research results.

A link might be if the foundation asked the person/group to do the specific study in question. Another link might be if the foundation specified how the study was to be done. Another link might be if the foundation had any review or editing authority to change or restrict the publication of the work. And another link might be if the foundation restricted, reduced or stopped it's support of the person/group because it did not like the results of a particular study.

As far as I can tell, Consumer Reports has not provided any evidence of these or any other direct links in this case.

In the absence of clear evidence of these or similar direct links, then the accusation of bias or conflict of interest becomes simply a convenient way to deflect the discussion to avoid responding to criticism (whether valid or not) and to avoid actually defend one's own position or results.

[The editors]

STATS has posted a blog on its website regarding the information Consumer Reports provided about its funding sources. STATS stated:

"Apparently, because STATS received grants from the Sarah Scaife foundation (though not in recent years), and Scaife had stock in ExxonMobil, therefore STATS was being funded by ExxonMobil.
By the same logic, every employee for Consumer Reports should now disclose the nature of their stock investments and retirement accounts – just in case they turn out to be funded [sic] by the companies whose products they rate.”

It is inaccurate to say that STATS has not received grants from the Sarah Scaife Foundation in recent years. According to IRS Form 990 tax documents for the Sarah Scaife Foundation for the five most recent years available, the foundation provided $100,000 annually to STATS in 2008,  2007, 2006,  2005 and 2004. Those tax documents can be found here:

http://dynamodata.fdncenter.org/990s/990search/ffindershow.cgi?id=SCAI001

The attempt to question Consumer Reports’ funding is illogical. Consumer Reports does not rely on employees' individual stock investments and retirement accounts to fund the work it does, while the Scaife Foundation grants do indeed provide "general operating support" for STATS, according to the tax documents.

But, since they asked, Consumers Union, the nonprofit publisher of Consumer Reports, actually does require staff to annually fill out very rigorous conflict of interest disclosure forms precisely because we want to be sure staff members do not have a financial interest or stake in the products they rate. 

[KT]

Wait a minute, are you saying that because a foundation that contributed to STATS invests it funds in ExxonMobile, you are claiming that Exxon contributes directly to STATS. That is either very poor research or a flat out misrepresentation of the truth. Either way CU needs to retract that claim.
I, for one, am far more satisfied trusting the regulators on this then the many different interests that are misrepresenting the truth here, and now CU has to be put in that camp too.

[Jocobo]

The recent US EPA study fails to address the findings of earlier peer-reviewed studies have reported low-levels of BPA causing specific behavioral and reproductive effects in particular strains of rat. The recent US EPA study looked for the same effects, but used a different strain of rat under different experimental conditions. While their study demonstrated no observable effect for the strain of rat they used, they cannot conclude that the earlier results demonstrating adverse effects from BPA are invalid.

The objective of the studies was to determine whether BPA can cause an adverse effect in a test species at a relevant concentration. It is well known that different species, and different strains within a species, have different sensitivities to various toxicological end points. Imagine ten different studies on a chemical that each used a different strain of animal, looked at different end-points, or were conducted under different experimental conditions. In nine of the studies no toxic effect is reported, but in the tenth study specific toxic effects are observed. It would be nonsensical to conclude that the chemical is safe because nine of ten studies did not find any toxic effect. Science recognizes the importance of replicating results. If a peer-reviewed experiment shows a toxic effect, it is very useful to confirm or refute the results found in the original experiment. In order to refute the original experiment, an independent experiment should bed done which repeats the experiment under the same conditions, with perhaps a larger number of animals if needed to strengthen the statistics. Doing a different experiment on a different strain of rats or other species and getting different results does not refute the original experiment, it simply indicates what happens under those different conditions. Doing experiments on other strains or species, or under different experimental conditions still can bel very useful. Such experiments can demonstrate that the effects may be limited to some species or conditions. Conversely, they can demonstrate that the effects are generalizable to a broad range of conditions or species. This can be important for understanding the toxicological mechanisms and for guiding further study. Once the mechanisms of toxicity are understood, it may be possible to develop hypotheses regarding whether the effects are unlikely to occur in sensitive human populations. But, a single negative result, as reported in the US EPA study, does not do this.

The Stats.org web site (http://www.stats.org/stories/2009/top_epa_scientist_rejects_consumer_reports_bpa_claim_nov10_09.html ) includes a side bar box with quotes from an NTP report that discusses the relative sensitivity of different strains of rat. The NTP panel clearly recognize that different strains of rat have different responses to various estrogenic and antiestrogenic chemicals. At the end of the box is a quote stating that a there are no data to support a hypothesis (claimed by some researchers) that a particular strain of rat is insensitive to estrogens. This is interesting, but not germaine to the recent US EPA study. The US EPA study apparently chose the Long Evans rat strain because the sensitivity to ethinyl estradiol in this strain is similar to that in humans. Unfortunately, their study design which rejected other strains where effects were reported in favor of the LE strain as a human model, relies on an untested hypothesis that BPA and ethinyl estradiol interact with the same receptors in exactly the same way.

[Stephen]

Bravo Consumer Reports!
Thanks for bringing this situation to everyones attention.
Keep up the good work!

[The editors]

The funding sources for STATS are murky because the tax records don’t break out sources of donations to either STATS or its sister organization, The Center for Media and Public Affairs, both headed by STATS President Dr. Robert Lichter.
Media Matters.org lists sources of funding for STATS based on data other non-profit organizations provided in tax forms they are required to submit to the IRS. The largest donor to STATS that Media Matters’ data identified was the Sarah Scaife Foundation, based on records through 2007.
(http://mediamattersaction.org/transparency/organization/Statistical_Assessment_Service/funders)
We examined tax records to confirm the amounts donated to STATS by Pittsburgh-based Sarah Scaife Foundation and also to determine the source of money behind those donations. We found that from 2000 through 2008, the foundation contributed $700,000 to STATS. The Scaife Foundation’s largest single stock holding is ExxonMobil, totaling $17,961,750, according to its 2008 tax records, which can be found here:
http://dynamodata.fdncenter.org/990s/990search/ffindershow.cgi?id=SCAI001
STATS President Dr. Robert Lichter disputes our characterization of the financial link between STATS and ExxonMobil, as he makes clear in his comments below. We believe that non-profit organizations such as STATS that say they are independent and non-partisan should provide full disclosure of their funding sources so that it is easier for consumers to figure out who is financing the research and analysis such groups provide.

[Don]

I'm wondering if you actually read the literature you cite. In your response above about the type of L-E rat used in the recent EPA study, you state that the L-E is insensitive to both BPA and to synthetic estrogen and that "The insensitivity to both was confirmed again in this study." In fact the study did NOT show "insensitivity" for the ethinyl estradiol (the synthetic estrogen, EE2). This is very clear in the paper's abstract that EE2 had major effects on these rats: "EE2 (50 µg/kg/day) increased anogenital distance and reduced pup body weight at PND2, accelerated the age at vaginal opening, reduced F1 fertility and F2 litter sizes and induced malformations of the external genitalia (5 µg/kg). F1 females exposed to EE2 also displayed a reduced (male-like) saccharin preference (5 µg/kg) and absence of lordosis behavior (15 µg/kg), indications of defeminization of the CNS. BPA had no effect on any of the aforementioned measures." See: http://toxsci.oxfordjournals.org/cgi/content/abstract/kfp266v1
Perhaps you were confused by the previous controversy about the Sprague-Dawley rat used in some early studies, which apparently are less sensitive to estrogens.

[LEG]

You cite the following reference to support the use of sc studies in risk assessment "And a comprehensive study of the metabolism of BPA in newborn lab animals showed that there was no difference in the levels of free BPA based on the route of administration (oral versus injection). This suggests that for newborns, who are especially vulnerable to BPA’s health risks, the route of exposure matters even less than in adults."
However, this study lacked several critical measurements and has been rejected by regulatory agencies like the EFSA. Detailed descriptions of the limitations of this study are presented on the EFSA website and in the public comments on the NTP BPA draft monograph.
Ongoing studies by the NTP, NIEHS will resolve this issue finally, but the prelimary conclusions are reported to be divergent with those you cited.

[Robert Lichter]

In their response to Trevor Butterworth’s criticisms, Andrea Rock and
Urvashi Rangan devoted most of their efforts to trying to discredit STATS by linking it to industy. Luckily, this is a scientific debate, in which the best science will ultimately win, regardless of the rhetorical tactics adopted by some participants. Mr. Butterworth will respond further to the scientific issues. But it is necessary to reply immediately to the institutional allegations.

The Center for Media and Public Affairs (CMPA) was founded in 1984 to conduct social scientific studies of media coverage. The Statistical Assessment Service (STATS) was founded ten years later with a different mission – to improve the transmission of scientific and quantitative information to the public. They are independent non-profit organizations that share some administrative personnel and expenses; both are affiliates of George Mason University. The STATS websites describes them as “sister organizations.”

To avoid the appearance of any conflict of interest in its critiques, STATS does not accept industry funding. However, Rock and Rangen claim to have examined documents showing that STATS has received funding from ExxonMobil. Since no such funding has taken place to my knowledge, I call on the authors either to publicly release any such documents so that everyone can examine them or to retract this assertion.

The authors also cast aspersions on CMPA for having once done a study for a tobacco company. Specifically, in 1994 CMPA conducted a content analysis on media coverage of tobacco for Philip Morris. CMPA has done many such social scientific content analyses commissioned by groups in both the private and non-profit sectors. For example, the same year we studied media coverage of tobacco, we examined television's coverage of Hispanics for the National Council of La Raza.

The most salient question is whether, as the authors intimate, CMPA’s critical faculties were bought off by Big Tobacco money. As it happens, during the 1990’s we were actively involved in conducting and publicizing research on the risks of tobacco in a series of articles, newsletters, and public statements, including Congressional testimony.

The publications began appearing in 1993 (the year before Philip Morris commissioned a content analysis) and culminated in a 1999 Yale University Press book showing that cancer researchers not only “placed tobacco in a league of its own among cancer agents,” they believed the media understates the cancer risks associated with tobacco.

STATS also has an extensive history of criticizing the health risks of tobacco. For a lengthier explication of all this material, see http://stats.org/stories/2009/dorothy_meets_marlboro_aug27_09.html

Rhetorical devices aside, it is well to remember that taking industry money isn’t the same as doing industry’s bidding.

S. Robert Lichter
Professor of Communication
Director, CMPA and STATS
George Mason University

[nicholas J]

This is an excellent point by point rebuttal (and a key spot of investigative journalism) that takes apart weak and unconscionable lines of defense from financially interested parties that repeatedly demonstrate that they put maximization of profits before consumer health.