Methods Suspect. Evidence of BPA Harm Swept Under the Rug?
In their August 2008 draft evaluation (PDF) on the safety of bisphenol A (BPA), the FDA used industry studies to reaffirm that an older, no observed adverse effect level (NOAEL) for humans. Their evaluation that BPA was safe flew in the face of hundreds of peer-reviewed studies that show low doses of BPA has potentially damaging affects on human development. In our last post we described how congress is closing in on the FDA, criticizing their methods for evaluating the safety of BPA and questioning FDA conflicts of interest.
Last week, 36 bisphenol A researchers also called into question the FDA's procedure for evaluating BPA science. In their published paper Myers et al analyzed an industry study authored by Tyl et al, (including authors from Bayer, Dow, ACC, TRI, and SABIC Innovative Plastics) paper 1, used by the FDA in their draft assessment, earned ten pages of criticism from Myers et al.
Neurobehavioral Affects Swept Under the Rug?
Two recent NIH reports, the 2008 NTP report, and Chapel Hill Consensus Statement indicate that the "greatest level of concern [for BPA] was directed towards possible neural and behavioral effects caused by BPA exposure in utero." Low dose bisphenol A is implicated in "changes in brain structure, brain chemistry and behavior represent the largest portion of the published low-dose BPA literature."
The NTP is supposed to advise the FDA on regulatory matters, but the FDA's draft report did reflect NTP conclusion that there was "some concern" about neurobehavioral effects. Instead, the FDA draft said that there was no evidence to support such a warning. Included in " documents on the FDA site is a research review of neurobehavioral studies contracted by The American Chemistry Council's (ACC) to Exponent 1, a consulting company in San Francisco, California. Exponent unsurprisingly found "no consistent adverse effects of perinatal exposures to low doses of BPA on neurobehavioral endpoints based on the 18 studies," a decision reflected in the FDA draft.
Good Laboratory Practice: "FDA's Misguided 'Gold Standard'"
The many, many low dose studies should convince anyone that BPA is not safe. Yet the chemistry industry keeps coming up with its own studies, one after another, which show the opposite results of non-industry scientists. The 2008 FDA draft gave the most weight to two industry studies that followed Good Laboratory Practice (GLP) standard.
Myers et al explain that the FDA designed the GLP standard to stop widespread private lab fraud in the 1970's, following a federal investigation of private lab practices. Scientists in the 1970's investigation needed to re-run 4000 tests by 235 companies and re-examine safety profiles of 15% of all U.S. pesticides on the market. Several men from one company were sent to prison for doctoring data. The Myers et al authors note:
"...fraudulent results were possible because contract lab studies used in the regulatory process are rarely subject to the checks and balances that peer-reviewed, replicated scientific findings undergoes."
The FDA's resultant industry GLP standards require extensive record-keeping to halt the type of fraud that Myers et al say is largely prevented in peer-reviewed research conducted under NIH grants. Another impediment of using GLP in academia is that GLP standards require large sample numbers of rats. 8000 rats were killed in a 2002 industry BPS study (a distressingly gross pile of dead rats). Such a large number of sacrificed animals would violate the animal care guidelines under the NIH grants.
GLP Trumps Good Methodology?
Among the supporting documents for the FDA draft here on the FDA site, you can find the ICF consulting product that Acronym Required mentioned in our last post, the neurobehavioural review contracted by ACC to the consulting company Exponent2, and various other reviews and communications.
There's obviously industry involvement in the FDA review. However, it is the government agency's job to consider the arguments of all constituents. I challenge anyone who doesn't know BPA research very well to make too much out of the content of these documents. However, given the FDA's conclusion you can's help infer that some of the industry sponsored research on the site influenced their decision. One you suspect that the FDA is influenced by industry, every document could look suspect.
For instance in one document ("Bisphenol A - Review of studies conducted by Vom Saal et al, Nagel et al, and MPI Research"), an FDA scientist, Dr. Sprando, compared low-dose studies on prostate development from Frederick Vom Saal's lab to an industry study where the scientists tried (and failed, with much public ado) to repeat Vom Saal's results. The MPI study used more animals in its experiments, which Sprando says makes the industry study "more powerful", an assertion that's not necessarily true.
More disconcerting, MPI's positive control failed. The FDA noted the lack of positive result, dismissed it, and concluded that all the results conflicted on neurobehavioral affects, therefore no decision was warranted.
Who To Trust?
On one hand you have several studies from Vom Saal's labs, as well as a later study from another academic lab showing negative affects of BPA on prostate development. On the other hand you have the MPI study sponsored by The Society of the Plastics Industry, Inc. which finds nothing. As with every other endpoint, the FDA compares NIH studies to an industry study, a "GLP" study, and favors or heavily weighs the industry result.
Myer's et al have concluded that the industry studies used by the FDA are "invalid":
"The fact that the U.S. regulatory community is willing to accept these industry-funded, antiquated and flawed studies as proof of the safety of BPA, while rejecting as invalid for regulatory purposes the findings from a very large number of academic and government investigators using 21st century scientific approaches, is of great concern."
"Industry research" shouldn't be code for "fraudulent", but it's difficult to read through these documents and not be suspicious. The GLP standard intended to lend credence to industry research which is not peer-reviewed, research that is subject to conflict of interest and historically littered with fraud. Now GLP is ironically being used for an important health decision on BPA to exclude over a hundred peer-reviewed studies.
Not only does this blemish industry research, it makes you wonder about academic research. Taxpayers invest in unbiased, peer-reviewed research on issues like whether BPA is safe for human consumption. The far wealthier chemical industry can fund a study showing some opposite result every time it sees something that might impede business. The FDA appears as if at times to be in cahoots with industry -- no? But what about the people? There's got to be a better way.
--------------------------------
1 Tyl et al: Two-Generation Reproductive Toxicity Study of Dietary Bisphenol A (BPA) in Mice. April, 2008: Toxicol Sci 104: 362 - 384.
2 Exponent is chemical consulting company located in San Francisco. Elizabeth Anderson, who is on the management team, was previously the president of Sciences International, the company fired for conflict of interest from the NIEHS bisphenol A contract, which we wrote about here and here, founded the journal Risk Analysis.
Leave a comment