- One of the most disturbing trends in science reporting is what the @NYTimes calls “single-study syndrome.”
- Single-study syndrome was in sharp display after the publication of a study on BPA by biologist Michael Baker.
- The bizarre journalistic reaction to Michael Baker’s BPA study underscores a distressing turn in science reporting.
- According to critics, chemicals like BPA can be blamed for everything from cancer to obesity to diabetes.
One of the most disturbing trends in science reporting is what The New York Times’ Andrew Revkin calls “single-study syndrome”— the increasing tendency of reporters and non-governmental organizations (NGOs) to trumpet research that supports a pre-determined perspective, no matter how tenuous—or dubious—a study might be.
That tendency was in sharp display recently after the publication of a study on Bisphenol A by University of California structural biologist Michael Baker in the online journal PLOS ONE. The chemical—a ubiquitous additive found in common plastic products and used as an ingredient in epoxy liners in cans that prevent spoilage and botulism—has been in the crosshairs of a small but determined group of university researchers, activist NGOs and journalists.
Critics of BPA believe it causes a range of disorders, particularly in pregnant women, developing fetuses and infants. They call it an “endocrine disruptor,” a term that gradually crept into mainstream use to describe synthetic chemicals that “mediate” hormonal activity, as do many natural substances, such as soy. BPA, they claim, is one chemical in a rogue’s gallery that includes the common weed killer atrazine, plasticizer phthalates and chemicals used to produce resins and detergents.
As often happens when a new study on BPA or “endocrine disruption” is published, a chorus of what seems like coordinated reactions materializes on the web. Within hours of the release of the Baker study, cyberspace exploded. The University of California-San Diego Health System’s news release, breathlessly titled, “BPA’s Real Threat May Be After It Has Metabolized,” was picked up verbatim by thousands of websites. It sparked uncritical stories by Chemical & Engineering News, Yahoo!, numerous other reputable news outlets and hundreds of speculative pieces by chemical-fearing NGOs with headlines ranging from “BPA is Bad to the Bone, Now We Know Why,” to “BPA Can be passed from moms to kids at birth” and “New studies add fuel to concerns over BPA."
The news reports shared a common theme—asserting or at least suggesting that the Baker study was a smoking gun and perhaps the final scientific word that activists have been awaiting to confirm their long-held belief that BPA is harmful and should be banned or sharply restricted. The irony is that the Baker study, when analyzed, does not support that view. Rather, it provides additional confirmation of the unlikelihood that BPA or many other so-called “endocrine disrupting” chemicals pose serious health threats.
The bizarre journalistic reaction to the Baker study underscores a distressing turn in “science reporting”—a disconnection between what agenda-seeking NGOs and their partners in the popular media report versus what the empirical evidence suggests. BPA antagonists range from chemophobic “natural” activist groups (Healthy Children Healthy World, Joseph Mercola) and advocacy media organizations (Mother Jones, Grist) to popular environmental organizations (Environmental Working Group, Natural Resources Defense Council). They share a worldview with some high-profile journalists—most notoriously the columnist Nicholas Kristof of The New York Times. Kristof has made “endocrine disrupting” chemicals one of his signature issues—much to the embarrassment of top-flight science journalists. This loose but very real coalition also includes reporters at many otherwise reputable news organizations who are not necessarily locked into a view about BPA or chemicals in general but, based on their reporting, appear to have only a superficial understanding of scientific risk.
According to these critics, chemicals like BPA can be blamed for everything from cancer to obesity, retarded neurological development, diabetes, cardiovascular disease, infertility and other disorders related to sexual development. These are catastrophic allegations, which if supported by the evidence warrant immediate and decisive restrictions. But in the case of BPA, the very opposite is happening. Since 2007, there have been more than a dozen comprehensive reviews of BPA by independent government scientists around the world, including in Canada, Europe, Japan, Australia and the United States, and each has concluded that the current uses of the chemical are safe.
Among the more high profile reports, the European Food Safety Authority in summer 2010, a joint UN Food and Agriculture Organization/WHO expert panel on BPA in November 2010, and a special Advisory Committee of the German Society of Toxicology in spring 2011 independently concluded that the collective body of evidence demonstrates that BPA does not pose serious neurological dangers or cause cancer in humans, and has not even been shown to be an “endocrine disruptor,” although it does have modest but not necessarily harmful endocrine effects.
In fact, the same week the Baker study was released—and ballyhooed by activist bloggers and journalists as decisive new evidence—Health Canada and that country’s Bureau of Chemical Safety upheld its prior scientific finding that found BPA poses no serious threat. “Based on the overall weight of evidence,” reads the report, “the findings of the previous assessment remain unchanged and Health Canada’s Food Directorate continues to conclude that current dietary exposure to BPA through food packaging uses is not expected to pose a health risk to the general population, including newborns and young children.”
Is BPA bioactive?
The most common and seemingly damning allegation against BPA, and one that shows up repeatedly in media reports, is that the Centers for Disease Control and Prevention (CDC) has found traces of the chemical in the urine of more than 90% of adults and children. That sounds frightening. But is it cause for concern? How scientists and journalists frame this often-stated fact is a good barometer of their understanding of risk analysis—whether they genuinely wrestle with complex science or are mouthpieces, intentionally or not, for a predetermined, chemophobic perspective.
The CDC has repeatedly stated that while biomonitoring “can … help scientists plan and conduct research on exposure and health effects,” the presence of a chemical in urine—whether BPA, atrazine, a phthalate or some other chemical targeted by advocacy groups—is not necessarily a sign that it’s harmful. Because advanced biomonitoring technology can detect vanishingly low concentrations of chemicals, “[f]inding a measurable amount of BPA in the urine does not mean that the levels of BPA cause an adverse health effect,” the CDC has written.
Regulators and most scientists believe the central focus of research should be to determine whether a bioactive form of a chemical makes it through the liver into our bloodstream, where it could effect our health. Years ago, that issue had been in doubt, which is why many scientists were initially agnostic over the question of BPA’s potential health impact. But by 2010 the FDA, reflecting the emerging scientific consensus, concluded, “[O]ral BPA administration results in rapid metabolism of BPA to an inactive [and therefore harmless] form.”
Such declaratory statements, echoed by every major international agency before and since, have not mollified the determined proponents of the endocrine disruption hypothesis. They remain convinced that Paracelsus’ famous declaration—“the dose makes the poison”—which has been the defining standard in toxicology, risk assessment and regulatory oversight—is simplistic and outdated. They believe in what’s called the non-monotonic response: A class of substances, BPA included, can interact with cellular hormone receptors in such a way that lower doses can pose higher risks.
Indeed, studies have shown that low doses of certain chemicals can induce non-monotonic effects. There are more than 600 studies to date according to a survey published last March in Endocrine Reviews. But independent scientists who have reviewed these studies, time and again, have come away unconvinced these effects consistently or even generally suggest harm. For the most part, they see biological activity—lots of statistical noise with contradictory indications of harm. And in the overwhelming majority of cases in which effects have been seen, rodents were administered the chemical by injection, bypassing the oral route, which is how humans typically encounter these substances. Regulators view these studies as marginally relevant to assess human health.
What’s an “endocrine disrupting” chemical?
Four years ago, in light of their unwavering and highly publicized concerns over BPA, the organized group of scientists who embrace the endocrine disruption hypothesis demanded, and were granted, yet another round of independent, government-financed peer review studies. The Obama Administration authorized under the stimulus package more than 30 million dollars for new research. The first set of studies began to appear in 2010. As the evidence began to come in and much to the discouragement of a shrinking group of true believers, the mainstream scientific consensus began to shift away from their determined position.
Originally open to considering that BPA might be harmful, L. Earl Gray Jr., senior scientist at the Environmental Protection Agency’s Research Triangle center in North Carolina, and his colleagues addressed whether endocrine effects translate into “endocrine disruption.” Their government-funded study, published in 2010, covered a wide range of BPA doses. It found extremely weak effects at low and high exposure levels. Even when four thousand times higher than the maximum exposure of humans in the general population, there were no discernible effects. Gray and his co-authors concluded that BPA has an extremely mild estrogenic impact but should not be called an “endocrine disruptor.”
This comprehensive EPA study—which should have considerable doubt over the low-dose, hypothesis—set off a new round of contretemps. Egged on by the university researchers behind the endocrine disruption hypothesis, Consumer Reports, the Milwaukee Journal Sentinel other nonscientific media, and activist blogs led by the NRDC and EWG, dismissed these results, accusing the EPA of using a strain of rat that is extremely insensitive to estrogen. Articles abounded ringing the “endocrine disruption” alarm bell.
The government responded by funding yet more studies to address these concerns. To specifically assess whether BPA could be transmitted to a fetus following its absorption in a mother, a research team led by Daniel Doerge, a chemist with the Division of Biochemical Toxicology at the Food and Drug Administration’s National Center for Toxicological Research in Jefferson, Arkansas, performed a series of studies on rats supposedly more sensitive to estrogen. Even after rodents were fed 100 to 1000 times more BPA than people are exposed to in their daily diets, the scientists could not detect the active form of BPA in the fetus 8 hours after the mother’s exposure. They concluded that the potentially toxic chemical is rendered inactive as it passes through the liver.
Doerge and his team followed with studies of primates, and concluded the same thing: Oral BPA administration results in rapid metabolism of BPA to an inactive form. There is complete glucuronidation of the BPA.
But what about in humans? Many mainstream scientists believe the final nail in the coffin of the BPA scare saga came in 2011 with what is now known as the Teeguarden study. The FDA’s Doerge joined a team of researchers led by Justin Teeguarden, a senior scientist at Battelle’s Pacific Northwest National Laboratory, one of the nation’s premier research centers, to assess how humans process BPA. The study was entirely funded by the Environmental Protection Agency and conducted in collaboration with CDC and FDA researchers. Rather than using animals and injecting them with BPA—humans generally come in contact with the chemical through food contact packaging—the researchers focused on human volunteers exposed to high levels of the substance by having them eat canned food and beverages from plastic bottles.
Their conclusion: Despite the presence of the chemical in urine, human blood concentrations of BPA are infinitesimally low—undetectable in most cases and thousands of times lower than any level that is likely to cause harm to humans. As Wolfgang Dekant, a professor of toxicology at the University of Wurzburg, who co-authored the European Union’s 2006 risk assessment of BPA (which found that the chemical did not pose a health hazard) has noted, Teeguarden’s study confirmed the earlier work undertaken for the European Food Safety Authority and FDA, which concluded that the human digestive system efficiently metabolizes the chemical after which its detoxified form is harmlessly excreted.
The most definitive rebuke by the government of the endocrine disruptor hypothesis—the persistent and central argument pushed by BPA’s determined band of critics—came last March in a dispassionately written but scathing 15-page response by the FDA rejecting a petition by the NRDC demanding that BPA be banned. On its Consumer Update website, the FDA also announced that its scientists had further determined that exposure for infants to BPA through foods is much less than had been previously believed, reaffirming that the trace amounts of the chemical that enter the body are rapidly metabolized and eliminated.
Unbowed, the NRDC and like-minded anti-BPA activists dismissed the findings and renewed what appears more and more like a desperate attempt to salvage a chemical witch-hunt.
The Baker study
Why is there such a disconnect between what activists and even many mainstream journalists choose to believe versus what independent scientists and regulators from around the world with no horse in this race have concluded based on reviews of literally thousands of studies?
The response to the recent Baker study is illustrative of what might be called an ‘evidence denial complex’. The overblown reaction can be traced to the initial news release. Its hyperbolic headline and text, approved by Baker, detailed the results of his computer modeling study of what happens when the liver breaks down BPA into metabolites. He reaffirmed findings that a BPA metabolite binds to a hormone receptor. “MBP is one [of perhaps several] metabolite that causes disruption of estrogen signaling in humans and other animals,” Baker is quoted as saying.
The news release suggested that his study is one more critical piece of evidence confirming “alarming associations between BPA and myriad adverse health and developmental effects, from cancer and neurological disorders to physiological defects and, perhaps, cases of childhood obesity.” In other words, as the release would have it, BPA is extremely dangerous but BPA metabolites are even worse!
Bingo, the UC-San Diego Health System press office and perhaps professor Baker got just what they wanted. The release ricocheted through cyberspace accompanied by alarmist headlines. Intriguingly not one media report mentioned the concurrently released Canadian government study contradicting Baker’s conclusions or included any reference to the spate of recent independent reviews that have found BPA safe as used.
Rather, one story after another hyped the Baker research and tied it to unrelated studies of BPA that also made questionable claims. The reporting by California Watch, which promotes itself as an independent investigative news organization, offers a distressing example of this botched coverage. Its environmental reporting and its specific focus on BPA is supervised by a former reporter from the Milwaukee Journal Sentinel, which established itself years ago as a echo chamber for the coalition of scientists who uncritically embrace the endocrine-disruption hypothesis.
In its article, California Watch quoted Baker as saying that BPA does “nasty stuff” and that he believes researchers should be studying how its metabolite works in the body to determine if it is a health threat. The article then referenced a separate, unrelated study of Latino women who work and live in the agricultural Salinas Valley, where they are exposed to pesticides and other chemicals as a daily work hazard. California Watch asserted that the Latino study “also ties BPA to disrupted hormone function.”
In fact, the Latino study provided no support of the endocrine disruption theory: “The association between the average of the two BPA measurements and maternal thyroid hormone levels was not statistically significant,” it concluded. In other words, California Watch linked one poorly explained study—the Baker research—to a misleadingly reported study in another journal in an attempt to hype its predetermined belief that BPA and other chemicals are dangerous. That’s shoddy journalism.
I contacted Baker to get his take on the brouhaha. He confessed he had no prior expertise in studying BPA and had been unaware of the groundbreaking work by Teeguarden, Doerge, Dekant and others. However, he said that he and the University of San Diego Health Systems writer were aware of the slew of popular and research reports damning BPA as dangerous, which shaped the tone and wording of the release.
Is the Baker study a new piece of the puzzle that could in fact rattle the new scientific consensus that BPA is biologically inactive as humans encounter it? Did the activists get this right? Or was the public yet again a victim of ‘single study syndrome’? After all, if Baker is right and if BPA metabolites and the “culprit,” then the Teeguarden study might be beside the point. Is this research the anti-BPA movement’s smoking gun?
I posed the question directly to Dr. Teeguarden, as California Watch or any reputable journalist covering this story should have done before writing an article on endocrine disruption or BPA. Like other scientists I talked with, he was complimentary of the basic design of the Baker study, calling it a “good example of structural modeling of receptors for screening for biological activity.”
But that’s where the praise ended. In fact, as Teeguarden and Dr. Gray both noted, it broke no new scientific ground. In 2004, Japanese scientist Shin’ichi Yoshihara had found that a BPA metabolite has a higher likelihood to bind with the estrogen receptor. Using computer simulation, Baker simply and elegantly offered a model of how that might work.
But Baker’s study did not touch on the central question: Could BPA or its metabolite effect humans? What’s the implication for human health? On this point, all of the half dozen scientists I talked with, including Baker, were unanimous. There’s nothing new here to suggest that BPA is active or effects humans.
In his study, Yoshihara had pointed out that the BPA metabolite MBP has never been detected in the human body or in active cells. “[M]etabolic activation to MBP may not be significant, at least in usual circumstances,” he and his colleagues had concluded—which is exactly what Doerge, Teeguarden and other scientists have since confirmed.
What about the tiny fraction of BPA that is not rendered harmless and therefore could, theoretically, be turned into a metabolite that might be highly estrogenic? It’s just too small a quantity to have any effect, all the scientists told me.
Baker speculated in the article and to me that perhaps MBP “could become a problem” if it is found at 0.5 nM levels in blood. However, as Teeguarden has noted, extensive blood data repeatedly demonstrates that BPA concentrations are far tinier. Teeguarden got technical in an email, but his comments are worth quoting—maybe serious science journalists will actually take this information to heart:
"The extensively documented, rapid, virtually complete metabolism of BPA in humans to compounds without estrogenic activity would seem to preclude the presence of MBP in sufficient amounts for biological activity. If we accept the speculation that MBP is formed during metabolism in humans, we might look to the elimination data to bound how much that could be. That number is likely to be 0-13%. Call it 10% so the math is easy. If blood BPA is 10 pM, and MBP is 10%…MBP is on the order of 1 pM at peak. Flip it around. If MBP was present at 0.5 nM [as Baker hypothesized], it would be 100x the mass of BPA in the system, which cannot be."
“There is just no evidence that it’s a relevant metabolite in vivo,” the EPA’s Gray added in a conversation. “To make any claims about its potential harm is just pure speculation. There is no evidence.”
“I think the University [of California-San Diego Health System] communications group did what communication groups often do: they looked to sell some press,” added Teeguarden. “Baker’s work is a solid example of structural modeling. The BPA stuff is unnecessary and secondary at best.”
Baker, who reviewed this story before publication, now says that the California Watch report that quoted him as suggesting BPA is harmful is wrong and out of context. “That’s not what I believe,” he told me. “I’m not an expert on BPA, by any means. It’s possible, that the metabolite that I studied would not have any effect on humans and I have no evidence, none at all, that BPA causes any problems in humans. This was a theoretical exercise, and it would be trumped by what actually happens in the real world. Based on what I know now, neither BPA nor its metabolites are harmful. I am upset that my structural study is misused by some.”
What will regulators conclude and how will reporters report
There remains a very large divide between the views of regulators who rely on risk assessment and the academic scientists who look for endocrine effects regardless of whether those effects can cause harm. University researchers have one high-profile supporter in Linda Birnbaum, director of the National Institute of Environmental Sciences. In an April editorial in the NIEHS-published journal Environmental Health Perspectives, Birnbaum argued that “it is time to start the conversation” about incorporating low doses and non-monotonic relationships into regulatory decisions.”
No prominent regulatory scientist has followed her lead. Why? Because no evidence has yet been presented that suggests that BPA, atrazine, phthalates or related chemicals are biologically active enough to harm the human endocrine system.
In making a decision whether to ban or restrict a chemical, regulators do not simply count positive and negative studies as if they are votes. In their assessment of the safety of BPA and other “endocrine disrupting” chemicals, the FDA and other oversight agencies review all the studies to evaluate what the data demonstrates with respect to safety. They look for meaningful patterns; otherwise they could blunder and react to statistical noise, and potentially ban a relatively safe chemical that would be replaced by something less well tested—one that could create actual harm. They then make weight-of-evidence determinations. They have consistently concluded that the data generated so far by university scientists who embrace the endocrine disruption hypothesis are not as reliable, relevant or consistent as the data produced by independent government ingestion studies.
“There is non-monotonicity, but the question is, is it toxicologically relevant?” asked Jason Aungst, a supervisory toxicologist with the FDA’s Center for Food Safety and Applied Nutrition in Silver Spring, Maryland, in a recent article in Nature. He and Earl Gray have argued that the low-dose health effects supposedly identified in studies contradict one another and have not been conclusively linked to health problems.
Some chemicals do cause serious effects, such as organ deformities, at low-dosages, but they are uniformly monotonic and can be identified under current regulatory testing protocols, Gray told Nature in a recent article. “You could never say that non-monotonicity doesn’t happen, but as far as its relevance to risk assessment, we really haven’t seen it in the high-quality studies,” he said.
The NIEHS and the National Center for Toxicological Research will be releasing a slew of state-of-the-art oral ingestion studies over the next 18 months testing, yet again, the non-monotonic thesis. The preliminary results began trickling out at toxicology meetings over the summer. To date, they consistently reaffirm prior studies that BPA has no harmful estrogenic effect except at high doses—thousands of times beyond what humans consume.
Richard Sharpe, head of the Centre for Reproductive Biology at the Medical Research Institute in Edinburgh and an internationally recognized pioneer on endocrine disruption, was originally drawn to the thesis but changed his views in the face of the accumulating evidence. Writing in Toxicological Sciences, in an article titled “Is It Time to End Concerns over the Estrogenic Effects of Bisphenol A?” Sharpe wrote:
"Fundamental, repetitive work on Bisphenol A has sucked in tens, probably hundreds of millions of dollars from government bodies and industry which, at a time when research money is thin on the ground, looks increasingly like an investment with a nil return. . . . If this opinion piece [in Toxicological Sciences] does nothing else, I hope that it will remind us all of the central importance to be attached to the repeatability of experiments and how we should react when a study proves to be unrepeatable"
In theory, this latest round of studies demanded by scientists supporting the endocrine disruption theory should settle the science once and for all. But that’s unlikely. Sensing the regulatory winds are not blowing their way, the increasingly isolated group of scientists who have built their reputations on the dubious endocrine disruption theory are taking their case out of the lab and into the streets.
The undisputed leader of the anti-BPA faction is Frederick vom Saal, an endocrinologist at the University of Missouri. In a paper to be published in the journal Green Chemistry in January, vom Saal and his supporters argue that there is little more that can be determined from new studies. Rather, legislators should bypass the international regulatory system altogether and just ban chemicals based on concern alone. In other words, they argue, the US should abandon classic risk analysis, which is the basis of regulatory science around the world, and adopt the slippery slope of the precautionary principle—regulation based on hazard and fear rather than on empirical evidence.
Their appeal is brazenly political, and some politicians are listening and responsive. Just last March, Rep. Edward Markey (D-Mass.) filed three separate petitions asking FDA to ban the use of BPA in infant formula and toddler food packaging (where it’s no longer used), in reusable food and beverage containers and in canned food, where its use is indispensible and there are no substitutes. Sen. Dianne Feinstein (D-CA) has a similar measure before the upper chamber.
Why, after such a massive accumulation of evidence do activist NGOs, many reporters and even some scientists continue to reject the emerging, evidence-based scientific consensus? “At this point,” one prominent government scientist told me, “with all the recent studies in hand, calling BPA harmful and dangerous, that’s just mass hysteria.” Clinicians call this apophenia—a neurological condition in which people insist they see meaningful patterns or connections in contradictory or random data.
“There are people who believe in the low dose effect of BPA and will always believe its disruptive effect, regardless of the evidence,” Dr. Gray told me. The ‘BPA-is-harmful’ crowd appears impervious to peer reviewed studies,” he added. “That’s a religious position. But as scientists and regulators, we have to go where the data takes us.”
More on science literacy at the Genetic Literacy Project.
Jon Entine, founding director of the Genetic Literacy Project, is senior fellow at the Center for Health & Risk Communication at George Mason University, and a senior fellow at the Statistical Assessment Service (STATS).