Insight: Science for hire - Trial over plastic exposes disclosure deficit
NEW YORK (Reuters) - By 2012, Eastman Chemical seemed to be perfectly positioned when it came to producing plastic for drinking bottles. Concerns about a widely used chemical called bisphenol A (BPA) had become so great that Walmart stopped selling plastic baby bottles and children's sippy cups made with it and consumer groups were clamoring for regulators to ban it. Medical societies were warning that BPA's similarity to estrogens could disrupt the human hormone system and pose health risks, especially to fetuses and newborns.
Eastman, a specialty-chemicals company headquartered in Kingsport, Tennessee, had been selling Tritan, its trademarked hard, clear plastic, as an alternative to BPA for five years. It told prospective customers that Tritan was free of BPA and any other chemical that mimicked human hormones like estrogens. To support the claim, the company pointed to "independent third-party testing," whose results were reported in the journal Food and Chemical Toxicology, which is published by Elsevier.
As more manufacturers and retailers abandoned BPA, some wanted to make absolutely sure that Eastman's safety claims for Tritan had been reviewed independently. The Austin, Texas, office of upscale grocer Whole Foods, for instance, asked Eastman if it funded "any of these labs" that determined Tritan had no estrogenic properties, according to an email from Eastman chemist Emmett O'Brien and disclosed in a lawsuit.
"I mentioned that we did not and they were happy with the answer," O'Brien said in the email.
Eastman did not make O'Brien available for comment. Because an employee of Whole Foods "is under subpoena in this case," said spokeswoman Libba Letton, "we can't comment on it."
In fact, the four labs that tested Tritan for the peer-reviewed paper in the Elsevier journal received funding from Eastman that was not publicly disclosed. Also not reported by Eastman's marketing materials or the paper: The lead author of the study, who analyzed the data from the four labs, was paid by Eastman for that work.
The financial relationships Eastman had with four "independent" labs emerged from discovery in a lawsuit brought by the company against two chemical testing and consulting firms that challenged the safety of Tritan. In the case, which began this week in Austin, Eastman alleged that PlastiPure Inc and CertiChem Inc falsely portrayed Tritan as having hormone-disrupting properties similar to BPA in an effort to market their own services.
It isn't clear how or whether Eastman's role in sponsoring its pro-Tritan study will figure in the case. But the connection between Eastman and scientists who produced the favorable study raises questions among some scientists and businesses about the company's description of the research as independent.
Eastman spokeswoman Maranda Demuth said that although the company commissioned the labs to do the studies, it "had no role or input on the analysis of the tests or the results."
Elsevier confirmed to Reuters that prior to publication the paper's lead author and his co-authors - scientists from the four labs - disclosed to the editors that Eastman paid them to conduct the analysis and write it up. But the paper carried a note saying that the "authors declare no conflict of interest," and there was no "acknowledgements" section, where researchers typically indicate who funded their work.
As a result, readers had no way of knowing about the Eastman connection, information that scientists typically use to judge the merits of a study's methodology and conclusions.
The reason for the nondisclosure, said Tom Reller, vice resident of global corporate relations for Elsevier, was that "they (the authors) felt there was no conflict as Eastman had no part in the design, analysis or data interpretation." He said the journal's editors, who declined to comment to Reuters, accepted the authors' declaration, but he declined to explain whether it is authors or editors who get the final say in whether information about funding is disclosed via a "conflict" statement. It is not apparent how frequently conflicts of interest are reported in Elsevier journals.
The approach of letting authors decide if they have conflicts of interest that should be disclosed is embraced by some less prominent science and medical journals. The leading scientific journals, in contrast, tell authors that any financial relationship with an interested party in the research must be disclosed to readers. U.S. law requires that scientists disclose ties with pharmaceutical companies when they present research.
"It's not for authors to decide what a conflict is," said Dr. Jerome Kassirer of Tufts University, former editor of The New England Journal of Medicine and a longtime leader of journals' efforts to disclose such conflicts to their readers. (He is not involved in the Eastman lawsuit.)
"A conflict exists when there is a dual loyalty," said Kassirer, who believes both journal editors and authors are on the hook for being as transparent as possible regarding funding. Disclosing payment from an interested party and then telling the journal such compensation is not a conflict of interest "is a highly unusual way of doing things," he said.
CertiChem and PlastiPure, both founded by neuroscientist George Bittner of the University of Texas, published a paper in 2011 in the journal Environmental Health Perspectives (EHP) reporting that many non-BPA plastics such as Tritan have estrogenic activity, especially after being exposed to sunlight or washed with detergent. EHP is published by the National Institute of Environmental Health Sciences, part of the U.S. National Institutes of Health. Tritan is a polymer made of three molecular building blocks; none are BPA.
The authors disclosed that they are employed by, own stock in and have other financial ties to CertiChem and PlastiPure, as shown on the first page of the paper.
On the heels of this study, PlastiPure, a consulting firm that advises companies on ways to synthesize new or existing plastics to make them hormone-activity-free, told clients and asserted in brochures pitching its services that, based on the research reported in EHP, Tritan is not estrogen-activity-free.
Eastman's contrary claim, that Tritan is not estrogenic, was bolstered by the 2012 study in Food and Chemical Toxicology. Based on tests from two university and two commercial labs, the authors concluded that the three molecules from which Tritan is made "do not pose an ... estrogenic risk to humans."
Court documents show that Eastman paid lead author Thomas Osimitz, a toxicologist at consulting firm Science Strategies, $10,000 to assemble the data and write the paper. The remaining authors were scientists at the labs, and all received payment from Eastman, the company's lawyer Rick Harrison said in a pretrial hearing. The total value of the four lab contracts, which funded the scientists to conduct tests on Tritan, could not be determined.
"There are statements by the ... authors that they were funded by Eastman in participating and writing the paper," Harrison told Judge Sam Sparks. "Those have never been made public."
Gary Sayler, a co-author of the Food and Chemical Toxicology paper and director of the Center for Environmental Biotechnology at the University of Tennessee, one of the labs that provided data on Tritan, said Eastman had sponsored the testing that produced the data for the Osimitz study. Eastman acknowledges this.
"The work was done under contract to our research center," Sayler told Reuters. "Just normal salary and supplies. Nobody was supported in anything above their normal salary rate."
Researchers from the other three labs declined comment, citing the ongoing litigation, or did not respond.
Despite efforts by U.S. lawmakers and scientists to make research more transparent, how researchers, journals and companies interpret disclosure requirements varies widely, members of the research community say.
"Many journals don't even ask about conflict of interest," said Sayler, who is himself the editor of an environmental studies journal. "Normally if you just report the funding for research that's sufficient ... It's not a fiduciary responsibility on my part. We're not going to be enriched by this."
Documents in the Austin trial suggest that Eastman was involved in designing the Osimitz study even though it was conducted far from its headquarters and despite the authors' claims to the contrary. According to an expert report that he submitted to the Texas court, Eastman toxicologist James Deyo worked with one lab to "develop the proper dosage levels" of Tritan monomers to give to lab animals in the test, the goal of which was to determine whether Tritan mimics human hormones.
Asked about the discrepancy between Deyo's statement and Eastman's characterization of the research as "independent," spokeswoman Demuth said the company "is not providing additional comments beyond what is publicly available."
Deyo also said that he helped work out the "protocols" for the study - essentially, how it would be conducted. Both dosage levels and protocols can affect an experiment's outcome.
Eastman maintains that because the four labs are not part of the company per se, their work can accurately be described as independent and "third party."
Three days before the start of the trial, Osimitz told Reuters that the Elsevier journal's disclosure forms were "very confusing." He declined to comment further on the pre-publication vetting process, but said that he had just submitted to the journal a correction to the statement that "the authors declare that there are no conflicts of interest."
"Eastman paid for the work," he said. "There is no question about that."
(Reporting by Sharon Begley; editing by Prudence Crowther)