Scientists were also uncovering links between endocrine-disrupting chemicals known as phthalates and health problems, including genital abnormalities and infertility in humans. These chemical additives were commonly found in soft, pliable plastics, such as those used in pacifiers and baby bottle nipples. In 2008, Congress passed a law banning six types of phthalates in children's products. As concerns about BPA hit the mainstream, Congress also launched an investigation into the industry's efforts to manipulate science and regulation, and a number of states proposed BPA bans.
In 2009, the BPA Joint Trade Association—which included the American Chemistry Council, Coca-Cola, and Del Monte, among others—gathered at the Cosmos Club, a members-only retreat in Washington, DC's Dupont Circle. According to meeting minutes leaked to the Milwaukee Journal Sentinel, the group explored messaging strategies, "including using fear tactics (e.g., 'Do you want to have access to baby food anymore?')." The "'holy grail' spokesperson," attendees agreed, was a "pregnant young mother who would be willing to speak around the country about the benefits of BPA."
Even as the industry crafted defensive talking points, some companies began offering BPA-free alternatives. But they often didn't bother testing them for other potentially toxic compounds or synthetic hormones. Nor did they have to: Under US law, chemicals are presumed safe until proven otherwise, and companies are rarely required to collect or disclose chemical-safety data. Michael Green, the Center for Environmental Health director who worried about his daughter's sippy cup, says this results in a "toxic shell game": Corporations that come under pressure to root out toxins often replace them with untested chemicals, which sometimes turn out to be just as hazardous. "It's an unplanned science experiment we're doing on our families," Green told me when I visited him at his Bay Area home, where Juliette, now 5, was padding around in a pink princess costume.
One of the most popular BPA-free options, especially among companies catering to families and health-conscious consumers, was Tritan, a clear, sturdy, heat-resistant plastic that Eastman rolled out in 2007. (Eastman also produces the chemical that sullied the drinking water of 300,000 West Virginians in January.) A company founded by alternative medicine guru Dr. Andrew Weil launched a line of Weil Baby bottles made from Tritan, which it touted as "revolutionary" and "ultra-safe" material. Thermos began churning out Tritan sippy cups, decorated with Barbie and Batman. With more and more consumers demanding BPA-free products, Nalgene, CamelBack, Evenflo, Cuisinart, Tupperware, Rubbermaid, and many other companies also worked Tritan into their production lines.
Eastman, a $7 billion company that was spun off from Eastman Kodak in the 1990s, assured its corporate customers that it had done extensive safety testing on Tritan. But its methods were questionable. According to internal Eastman documents, in 2008 Eastman signed a two-year contract with Sciences International, another product defense firm that had played a key role in the tobacco industry's scientific misinformation campaign. On Sciences' advice, Eastman then commissioned a study that used computer modeling to predict whether a substance contains synthetic estrogens, based on its chemical structure. The model suggested that one of Tritan's ingredients—triphenyl phosphate, or TPP—was more estrogenic than BPA.
Eastman, which never disclosed these findings to its customers, later commissioned another study, this one involving breast cancer cells. Again, the initial results appeared positive for estrogenic activity. In an email to colleagues, Eastman's senior toxicologist, James Deyo, called this an "oh shit moment."
Cell culture tests for estrogenic effects generally involve soaking plastic in alcohol or salt water, then exposing cells to various concentrations of the chemicals that seep out. After Deyo informed the lab that its findings must "be worded very well relative to the lack of" estrogenic activity, it issued a report that only counted data from the lowest concentrations—even though this violated the lab's testing guidelines, and made the results appear negative when they weren't. "The lab ignored its own criteria and misrepresented its findings," says Michael Denison, a professor of toxicology at the University of California-Davis who evaluated the document.
Eastman wasn't the only company testing Tritan. In 2009, Bittner's PlastiPure, which was searching for nonestrogenic alternatives to recommend to clients, began vetting products made with it and found that some had even more estrogenic activity than their BPA-laden counterparts. PlastiPure's CEO, Mike Usey, says CertiChem disclosed this to clients, but many chose Tritan anyway.
"It's an unplanned science experiment we're doing on our families," Michael Green says.
This was part of a broader pattern of indifference. According to Usey, hundreds of manufacturers—including most of the big baby bottle makers—contacted CertiChem to inquire about testing their BPA-free products for estrogenic chemicals, but few actually followed through. "Their position was: Until consumers are demanding nonestrogenic products, there's no reason to be an early adopter," Usey explains. "They want to delay as long as they can, because they know any transition will cost them." In some cases, manufacturers paid for testing, then never collected the findings. "They didn't want to know the results because there's liability in knowing," Usey says. "They're right in the sense that you don't want to know if you're not going to fix the problem."
Despite its "oh shit" findings, by 2010 Eastman began to produce marketing materials claiming that Tritan was free of all synthetic estrogens. One section of its website featured the tagline "Safety is our key ingredient" along with photos of smiling children eating and drinking out of plastic containers. The site claimed "third-party research" had shown Tritan to be free of estrogenic activity, but when corporate customers tried to verify this information, Eastman grew cagey. In early 2010, Philips Avent, a top producer of baby bottles and sippy cups, inquired about having an outside lab run testing on Tritan. Eastman's senior chemist Emmett O'Brien fired off an email to colleagues, saying, "We need to [do] everything possible to convince the customer NOT to do EA [estrogenic activity] testing." Philips was persuaded. But, according to testimony from Eastman executives, that same year Nestlé vetted Tritan, and found it leached synthetic estrogen. (Frédérique Henry, a spokeswoman for Nestlé, acknowledges the company tested Tritan but denies the results were positive.) Nestlé has nevertheless continued using Tritan in some of its water bottles.
Bittner and Usey, meanwhile, decided to go public. "As long as the consumer demand wasn't there, product manufacturers felt we were selling them a problem rather than a solution," Usey explains. "We saw this as the only way forward." Bittner's companies, which have received more than $8 million in NIH funding, began working with Jordan, the Georgetown professor, on a paper for publication. In the fall of 2010, Usey attended the ABC Kids Expo, a children's product extravaganza in Las Vegas, and handed out flyers with a graph showing how various products that were marketed as nonestrogenic stacked up in CertiChem's tests. The most estrogenic among them, Weil Baby bottles, were made from Tritan. (The company referred Mother Jones to a press release on its website stating that it "remains confident that Tritan is safe.")
Soon Eastman's customers began inquiring about CertiChem's findings. For the most part, Eastman convinced them to disregard Bittner's claims. At one point, O'Brien met with Whole Foods executives. They were considering replacing their polycarbonate bulk food bins with ones made from Tritan, even though Bittner had previously informed them that the product was estrogenic. According to a memo O'Brien later wrote, when the subject came up, he responded by attacking Bittner, whom he called "shady," and his test results, which he alleged were "very questionable." The Whole Foods executives later pressed O'Brien about the other tests carried out on Tritan.
The chemist claimed, falsely, that they were performed by independent scientists with no funding from Eastman and hadn't turned up any evidence that Tritan leached synthetic estrogens. Whole Foods—which declined to comment for this story—plowed ahead and installed Tritan bins in many of its 270 US stores.
Eastman refused to answer questions for this story, but it released a written statement saying that it had "paid the labs for their time and expertise and not for a particular conclusion," and remained "confident in the testing and safety of Tritan."
In March 2011, the Environmental Health Perspectives paper by Jordan and researchers from CertiChem and PlastiPure appeared online. They'd tested 455 store-bought food containers and storage products, including several made from Tritan. The results? Seventy-two percent leached synthetic estrogens. And every type of plastic commonly used in food packaging (polypropylene and polystyrene, for example) tested positive in some cases, which suggested there was no surefire way to avoid exposure.
Other scientists have also found evidence of estrogen-mimicking chemicals in BPA-free plastics. In 2009, two German environmental toxicologists tested PET, a plastic commonly used in water bottles, on a strain of mud snails that produce more embryos when exposed to synthetic estrogen. Snails reared in PET bottles produced twice as many as those reared in a glass culture dish.
These studies don't identify which estrogenic chemicals are leaching from BPA-free plastics, but many of these products are known to contain phthalates or bisphenol S (BPS), a chemical cousin of BPA that plastic makers frequently use in its place. Cell-culture tests suggest that BPA and BPS have similar effects.
In other cases, little may be known about the specific health effects of the chemicals involved, but a 2012 literature review by 12 prominent scientists found there is "substantial evidence" that endocrine-disrupting chemicals generally harm human health. "We know that there's a cost when we mess with the levels of these hormones in our bodies, regardless of how we do it," says the study's lead author, Laura Vandenberg, a professor of environmental health sciences at the University of Massachusetts-Amherst. "Even small changes early in life can alter brain and organ development and set us up for disease later on."
Every type of plastic commonly used in food packaging tested positive in some cases, which suggested there was no surefire way to avoid exposure.
The month after Bittner's study appeared, the American Chemistry Council contacted Chris Borgert, the former tobacco industry scientist who stymied the EPA's Endocrine Disruptor Screening Program. According to internal emails, the council and the Society of the Plastics Industry offered to pay him $15,000 to write a brief letter to the journal's editor refuting CertiChem's study, and to enlist another scientist to sign on. Their letter argued that CertiChem's findings were "unconvincing"; just because a substance behaved like estrogen in a culture dish didn't mean it would do so in animals or humans.
At the same time, Eastman laid plans to sue CertiChem and PlastiPure for false advertising. Expecting that Bittner would lash out after being served papers, the company launched a preemptive PR blitz. "By proactively promoting Tritan safety," an internal memo noted, "it will put PlastiPure in a position to have to prove Eastman wrong." The company also paid a scientist named Thomas Osimitz $10,000 to author a research paper on Tritan. While Osimitz was ostensibly working independently, Deyo, the Eastman toxicologist, micromanaged the process, from designing the study to writing the introduction. Deyo's study design virtually guaranteed estrogenic activity wouldn't be found. For example, he opted to use the hormone-insensitive Charles River Sprague Dawley lab rat. Rather than testing Tritan itself, he instructed Osimitz to test only some Tritan ingredients—TPP, the one that had raised red flags in the computer-modeling study, was not included. (The European Union has since classified the compound as a suspected endocrine disruptor.)
In June 2012, Osimitz's paper—finding that Tritan was not estrogenic—appeared in Food and Chemical Toxicology, an industry-friendly journal. Its editor, A. Wallace Hayes, was previously vice president of biochemical and biobehavioral research at R.J. Reynolds, which led the attack against science linking secondhand smoke to human health problems.
Scientific journals generally require authors to disclose any conflicts of interest. But the Food and Chemical Toxicology article made no mention of Eastman's role in the study. According to internal Eastman emails, the company was also aiming to hire Osimitz to author a second paper, again with "no…mention of Eastman." As Deyo noted, "credibility is somewhat enhanced if it is not 'Eastman' authors."
Once its own data had been published, Eastman set out to bury Bittner's findings. In August 2012, the company sued CertiChem and PlastiPure, which it claimed were spreading false information about Tritan to generate demand for their own services. Eastman's lawyers asked the judge to bar both firms from ever claiming Tritan was estrogenic—or saying that cell-based tests could detect estrogenic activity, even though scientists routinely use them for this purpose. For decades, scientists have relied on the same breast cancer cell line Bittner's lab uses, MCF-7, to screen for estrogenic activity. According to UMass' Vandenberg, these cells have proven "remarkably good at telling us if compounds found in plastics and personal care products mimic estrogen" and their "failure rates are minuscule."
On July 15, 2013, Bittner squared off against Eastman at a federal courthouse in Austin. The company's attorneys went in hard. Specifically, they claimed running a company that tested products for estrogenic activity, as well as one that helped companies find nonestrogenic alternatives, created a conflict of interest. (Bittner counters that he's no more conflicted than a doctor who both diagnoses and treats patients.) But they didn't directly challenge the validity of Bittner's findings. Instead, they leaned on the questionable industry claim that tests based on human cells aren't sufficient to establish estrogenic activity.
Eastman's star witness, Chris Borgert, made the case that animal studies—which the industry had also fought to undermine—were a more telling indicator. But even they were not "in and of themselves" definitive. For the result to be relevant, the effects had to be demonstrated "in an animal, at least, and then on to humans." There was no mention of the ethical and legal barriers to testing on humans. And the judge barred Bittner's lawyers from mentioning Borgert's tobacco industry ties, which Eastman argued were "prejudicial." This left the jury ill-equipped to gauge his credibility.
Borgert's testimony may have done less damage than other factors. Bittner's lawyers struggled to explain the science to jurors, and Bittner grew testy on the stand. Welshons, who'd designed CertiChem's tests, testified in a deposition—just as he'd told the NIH—that Bittner had misrepresented some data in a brochure. Bittner's attorneys managed to block his testimony from being introduced. But, Bittner says, his attorneys balked at presenting key evidence, such as figures on CertiChem's NIH funding, because it might have made Welshons' testimony admissible. Bittner also maintains that his rift with vom Saal and Welshons made it difficult to recruit witnesses.
Still, several prominent scientists testified for CertiChem, including UC-Davis' Michael Denison, who coinvented a widely used test for estrogenic activity using human ovarian cells. Denison testified that he'd tested 27 samples of Tritan for estrogenic activity using this method and registered positives across the board.
But the most remarkable data might have come from none other than Wade Welshons. In the run-up to the trial, the University of Missouri scientist, who expected to prove Bittner wrong, began testing Tritan products in his lab. To his surprise, he wound up confirming CertiChem's findings. "It doesn't matter what I think of them personally," Welshons told me. "If they're right, they're right, and many of my objections no longer matter."
Welshons' findings never made it into court, however, and when the jurors returned their verdict in late July, they found against Bittner's companies on counts of false advertising and unfair competition. They also concluded Tritan was not estrogenic. Their rationale, according to postverdict interviews, echoed Eastman's claims that estrogenic activity could not be established solely through cell-based tests. In his final ruling, the judge also noted that the "jury was likely unimpressed with Dr. Bittner's combative demeanor." And he upbraided both sides for failing to explain the science in terms jurors could understand. In the end, he barred Bittner's companies from ever talking about their Tritan findings, at least in a commercial setting. But he refused to stop the companies from asserting that their tests could detect synthetic estrogens.
The long legal battle has depleted CertiChem and PlastiPure's coffers—"We've laid off half of our staff," Usey told me. "It has pretty much crushed us"—and emboldened Eastman. After I began raising questions about Tritan, Rick W. Harrison, an attorney for the chemical giant, inadvertently copied me on an email about Eastman's damage control strategy. "If this somehow gets picked up by mainstream media—Oprah or NY media—Eastman sends Lucian [Boldea, the vice president of Eastman's specialty plastics division] or whoever on the show prepped with the verdict, order and judgment and express surprise and indignation that these issues are still being raised after three years of litigation," he wrote. "The court/jury has spoken and spoken loudly."
The industry, meanwhile, has revived its campaign to downplay the dangers of BPA. A month after the Eastman case concluded, the American Chemistry Council relaunched its pro-BPA website, FactsAboutBPA.org. The section on infant health suggests that BPA isn't harmful, even to premature babies. "They're reverting back to exactly the arguments they were making in 1998," says vom Saal. "It's as if the last 15 years didn't happen."
US regulators also have continued to ignore the mounting evidence linking BPA and similar chemicals to human disease, even as bans have cropped up around the world. Although more than 90 studies examining people with various levels of exposure suggest BPA affects humans much as it does animals, the FDA recently announced that its research "supports the safety of BPA" in food containers and packaging. And the EPA program that was supposed to screen some 80,000 chemicals for endocrine disruption hasn't fully vetted a single substance. In 2010, the agency sought White House approval to add some endocrine-disrupting chemicals that are commonly found in plastic—among them BPA, phthalates, and a class of compounds known as PBDEs—to its "chemicals of concern" list because it found they "may present an unreasonable risk to human health." This would have required chemical makers to share safety-testing data with federal regulators. The proposal languished until last September, when the EPA quietly withdrew it, along with a proposed rule requiring manufacturers to disclose safety data on chemicals in their products.
Still, Bittner isn't giving up the fight. When I visited CertiChem's office in Austin recently, he was sitting barefoot at a conference table surrounded by sippy cups and heaps of lab notebooks. CertiChem and PlastiPure were planning to appeal the Eastman ruling (they've since done so) and were working with Denison on data for new papers, one on estrogenic activity in plastic resins, which are used to make plastic products and contain fewer additives that can skew results. Bittner called up a series of graphs on the overhead projector, showing the results for several new BPA-free plastics that he had tested for estrogenic activity. He raked his laser pointer over a graph displaying the results for Tritan. The line curved up steeply. "Eastman won the battle," he said. "But that doesn't mean it will win the war."
Update (3/3/14): After this story went to press, the US Food and Drug Administration published a paper finding that BPA was safe in low doses. However, due to laboratory contamination, all of the animals—including the control group—were exposed to this chemical. Academic scientists say this raises serious questions about the study's credibility. Stay tuned for more in-depth reporting on the FDA's most recent study.