Trevor Butterworth, Contributor
If there is one certainty that greeted publication of an Environmental Protection Agency-funded study showing that bisphenol A (BPA) was below the level of detection in people fed a high BPA diet, it was that University of Missouri biologist, Dr. Frederick vom Saal, the scientist who first ignited the global panic on BPA, would, sooner or later, denounce it as astoundingly flawed and influenced by industry research.
Why? This has been his stock response to every major study that has either failed to replicate his original claims or has demonstrated that BPA is not, as he claims, ‘the biological equivalent of global warming.’ In vom Saal’s world of science, everyone who disagrees with him is either guilty of gross incompetence or unethical behavior or both. It would not be unfair to say this attitude of attack the person first and (maybe) the science afterwards has not won him many fans in the scientific fields which deal with the risk from chemicals; indeed, there is a small but telling academic literature dealing with his unscientific behavior (Purchase, 2004; various letters in response to his accusations in journals; pleas for civility).
All of this has been largely ignored by the mainstream media, which has, instead, portrayed vom Saal as either a representative of the movement he, in effect, created – or a heroic, independent researcher battling evil industry-funded scientists.
The second reason to expect a denunciation was that the new study, Teeguarden et al., conducted by Battelle researchers working at the Pacific Northwest National Laboratory (PNNL), the Centers for Disease Control and Prevention (CDC), and the U.S. Food and Drug Administration (FDA), andpublished in the leading toxicology journal, Toxicological Sciences on June 24, effectively demonstrated that the studies driving fears about BPA couldn’t be correct about levels of the chemical in people’s blood. Plus, if BPA was below the level of detection using the most advanced technology for determining the presence of such chemicals, all those animal studies claiming adverse effects based on much, much higher exposures were simply irrelevant to human health.
This also meant that Teeguarden et al. also threw a wrench in the environmental lobbying movement’s campaign to ban BPA, one that has been supported by prominent Democrats such as California’s Senator Diane Feinstein. In the logic of public activism, it is better to destroy obstacles in one’s path than admit that things aren’t quite as simplistic – or as deadly – as one’s campaign slogans.
In fact, Teeguarden et al.’s methodology demonstrated a key analytical flaw in the ‘BPA-is-deadly’ hypothesis which anyone who was truly concerned about scientific integrity should be be concerned about: contamination. The presence of BPA in lab equipment meant to test for BPA has clearly been distorting the data. By contrast, and as monitoring technology got more sophisticated, the evidence was going in the other direction, namely, that the more carefully you looked for BPA, the less of it you found.
Teeguarden et al., as I noted earlier in Forbes, drew praise from leading European scientists for its rigor and clarity. “Beautifully designed and executed” with “fundamentally important implications” said Richard Sharpe, one of the UK’s leading independent researchers on the risks to reproductive health from environmental chemicals. So, the betting was that it would elicit the exact opposite response from vom Saal. He did not disappoint.
Its conclusions are preposterous,” he told the Atlantic. “How could a federal agency be associated with this? It is profoundly bad.”
Actually, three federal agencies and one of the nation’s top research laboratories – so top-of-the-heap it has been the subject of intense cyber attack – were behind Teeguarden et al.
The Atlantic’s writer claimed that the falsity of the Teeguarden study had “outraged prominent members of the scientific community,” without indicating just who, exactly, was in that community – beyond vom Saal, his associates, and the environmental activists who have adopted a halal-like view on chemicals in the environment. While members of this community are certainly prominent in the media coverage of BPA, it is far from clear that they are regarded as prominent within, say, toxicology.Still, rumors began to circulate –the journal Toxicological Sciences was going to pull Teeguarden et al. because it really was as bad as vom Saal claimed – but this proved as false as it was incredible. Could the top researchers at three world-leading institutions (and by implication the academic reviewers for a leading journal in the field of toxicology) be that “profoundly bad” or could it be that the scientist who has repeatedly seen his work rejected by international risk assessments was simply attacking in order to defend his reputation?
Finally, last week, Toxicological Sciences published vom Saal’s formal, if rhetorically extravagant, denunciation, a letter cosigned by longtime collaborators Gail Prins and Wade Welshons: “Report of Very Low Real World Exposure to Bisphenol A is Unwarranted Based on a Lack of Data and Flawed Assumptions.”
But it also published, at the same time, a response from the lead authors of the study, Justin Teeguarden (PNNL), Antonia Calafat (CDC) and Daniel Doerge (FDA): Adhering to Fundamental Principles of Biomonitoring, BPA Pharmacokinetics, and Mass Balance is no “Flaw”
Now, typically, this is the kind of exchange that has allowed many journalists, and not a few scientists to assert that the argument over BPA is deeply divisive – and that there is no agreement among the experts. This is simply because the science is extremely difficult to follow – not least because it is written for other scientists. And so, what this reportage misses is whether one or both sides are in fact, equal; that is, do they both demonstrate methodological rigor and textual fidelity to the basic practices and weight of evidence of the requisite scientific fields?
In other words, if there are established practices in, say, pharmacokinetics, which is the study of the mechanisms of distribution and absorption of a substance in the body, and those practices are ignored in a critique of a study of the distribution and absorption of a substance in the body, then, factually, it is not a simply a matter of he says, she says; the critique is disingenuous.
This is the underlying theme of Teeguarden et al.’s response: vom Saal et al.flaunt fundamental principles while, at the same time, claim flaws in methods that they otherwise endorse in studies that agree with their position.
For example, vom Saal et al. criticize Teeguarden et al. for not measuring the BPA content in the canned food they gave to their subjects, or identifying the brand of food because, they say, BPA content varies widely depending on the manufacturer. What they want to say is that if you don’t know how much BPA was in the food to begin with, you cannot know what daily exposure to BPA was.
But as Teeguarden et al. respond, we already know from previous research that ~100 percent of BPA ingested and absorbed is excreted in urine, so the total amount of BPA in urine has to be equal to or greater than the amount ingested ) and, therefore, equal to the internal dose (the amount inside the body). Given that their blood tests found that BPA was below the level of detection for the majority of their feeding group (and miniscule for the remainder), and given that the total BPA in urine was substantially higher than the CDC biomonitoring data for the general population, their methodology is not only sound, it is based on basic principles of mass balance: knowing the brand of canned food or the amount of BPA in the food is irrelevant. They have shown that high exposure to BPA in the diet does not lead to measureable internal exposure to active BPA and only to very small amounts of inactive BPA in some people.
Vom Saal et al. want to be able to say that food is not the only source of BPA, and that the measurements do not reflect people’s exposure based on CDC biomonitoring data; yet these lack force: the CDC’s leading expert in biomonitoring, Antonia Calafat, is a co-author of the Teeguarden paper and the letter in response to criticism by vom Saal. Second,the idea that we may be surreptitiously exposed to large amounts of BPA from, as vom Saal et al.suggest, through cash receipts, was rejected by the World Health Organization Expert Panel on BPA last year. Diet, the panel concluded, is the primary and overwhelming source of our exposure to BPA.
As you can now grasp, if you’ve made it this far in the article, the dispute is on a level of technical sophistication (of which the above example is a summary) that doesn’t typically make it into news reporting. This is understandable: who would want to read this? At the same time, this is why the public has been treated to spin rather than science in the story of BPA. However, there is a problem with vom Saal et al. that anyone can grasp. As Teeguarden et al. note:
Vom Saal et al. expressed the opinion that our reliance on three papers (Dekant and Volkel 2008; Lakind and Naiman 2008; Lakind and Naiman 2010) to support one of our conclusions was unjustified because the estimates in those studies “have been deemed to be flawed by others,” citing a 2007 publication (vom Saal, Akingbemi et al. 2007) and two others (Vandenberg, Chahoud et al. 2010a; Vandenberg, Chahoud et al. 2010b). We confirmed, by careful reading and text searches, that none of the three papers were analyzed, cited, or mentioned in the vom Saal, Akingbemi et al. 2007 publication. We further confirmed that the two Vandenberg et al. papers neither mention nor cite the two Lakind papers they are alleged by vom Saal et al. to refute.”
Do you see the math problem? Vom Saal claims that he had demonstrated that a study published in 2010was false in 2007 – three years before it was actually published.
Could the LaKind and Naiman paper have languished on the web for at least three years before official publication? I posed the latter question to the University of Maryland’s Dr. Judy Lakind: no, she said; the 2008 paper – “Bisphenol A (BPA) daily intakes in the United States: estimates from the 2003-2004 NHANES urinary BPA data” – first appeared online on 16 April 2008; and the 2010 paper – “Daily intake of bisphenol A and potential sources of exposure: 2005-2006 National Health and Nutrition Examination Survey” was first published online on 17 March 2010. LaKind also says she never shared any of their data with Dr. vom Saal prior to publication.
I checked vom Saal’s citation of himself, and just as Teeguarden et al. point out, there is no mention of the LaKind and Naiman papers, or the 2008 Dekant and Volkel paper. Similarly, a check of Vandenberg finds no mention of the LaKind and Naiman papers.
This is carelessness of a kind that wouldn’t pass muster on the letters page of a newspaper,but which is especially out-of-place when the context is denouncing government scientists for flawed science (and, by implication, Toxicological Sciences, and its academic reviewers for publishing it). Surely, if you are going to take the step of denouncing fellow scientists and a leading journal, you don’t want to make stuff up? Moreover, if you can be so cavalier about making self-evidently implausible claims, what else in your argument suffers from the same defect?
In fact, as Teeguarden et al. note, they are not the first to call out vom Saal for unscientific tactics in academic research. In 2004, the University of Manchester’s Iain Purchase published a paper in Toxicology titled “Fraud, errors and gamesmanship in experimental toxicology,” which drew attention to vom Saal under the category of gamesmanship.
“Gamesmanship,” writes Purchase,“is where the normal paradigm of the self-corrective mechanism in science, that is verification and scientific review, is not followed. Rather, in place of scientific criticism, the focus of the criticism is the scientist or the scientist’s affiliation.”
Vom Saal has spared no vitriol in denouncing studies that failed to reproduce his or his collaborators’ findings as the work of rank incompetence or corrupted by industry-funding, while at the same time announcing in the media that his initial study demanded a revolutionary rethink of the fundamental principles of toxicology. The first two failed attempts to reproduce vom Saal’s seminal paper (Ashby et al. 1999, Cagenet al. 1999) were indeed industry funded; but, as Purchase takes up the story:
First the technical competence of Ashby’s group was questioned, but a member of vom Saal’s group trained Ashby’s group in the technique of prostate dissection. Then high levels of phytoestrogens in Ashby’s laboratory diets were claimed to invalidate the results, but the phytoestrogen levels in the diets employed by vom Saal were higher. Then the environment in the animal rooms was considered an issue and finally it was suggested that the mice used in Ashby’s experiments were too heavy rendering their reproductive tissues resistant to the effects of BPA-mediated effects; however, analysis of all the studies available renders this explanation unlikely…”
More important, vom Saal’s original study was based on only 14 animals and 11 controls – a tiny amount in terms of statistical rigor. Ashby used 66 and 54 controls, Cagen, 87 and 44 controls. Purchase’s point was that you can’t, if you want to have scientific credibility, keep insisting you are right and they are stupid without doing more science which addresses the limitations of the original experiment. But as the controversy progressed over the past decade, vom Saal – aided, by now, by environmental activists such as The Silent Spring Institute – increasingly insisted on an industry conspiracy or that his critics were not “experts” in BPA.
This has become increasingly difficult to sustain in the face of massive, multigenerational toxicity studies funded by government agencies around the world which have either failed to reproduce his low dose theory of BPA or to find the adverse effects found in the small studies he cites, which used methods now generally accepted as inappropriate to determine human risk; and yet the industry conspiracy narrative is still alive and kicking in the media coverage of the topic, as the recent Atlantic piece by Elizabeth Grossman demonstrates. Indeed, it has proven vom Saal’s most potent weapon in driving the media narrative to cover his position as the dominant scientific narrative even though that narrative has been rejected by regulatory science around the world
But if you consistently reduce scientific research to motive and bias in order to attack your opponents, you leave yourself obnoxious to the same attack.
As Teeguarden et al., note, “vom Saal et al. also cast doubt on the validity of studies we cite by implying that funding for those studies from the Polycarbonate/BPA Global Group of the American Chemistry Council indicates a potential for bias. We caution Dr. vom Saal et al. that an objective scientific community understands the potential for bias cuts both ways (see Dr. vom Saal’s conflict of interest statement in (Sieli, Jasarevic et al. 2011)).
That conflict of interest statement reads: “F.S.v.S. consulted for an attorney involved in civil litigation regarding the health effects of BPA, but he has no financial interests related to plastics, products, or compounds that might serve as alternatives to BPA.”
An additional 2008 disclosure in the Journal of the American Medical Association for vom Saal includes, “serving as an expert witness for the defendant in a trial in 2004 regarding the health effects of bisphenol; serving as a consultant for in-preparation litigation regarding BPA; serving as chief executive officer of XenoAnalytical LLC, which uses a variety of analytical techniques to measure estrogenic activity and BPA in tissues and leachates from products.”
If industry funding provides grounds to claim bias, collaborating with trial lawyers and running a private company designed to test for the chemical you have denounced for the past decade is hardly evidence of academic disinterest. There is just as much motive to insist that there’s a problem with BPA as there isn’t, which is why motive is a dead end for determining scientific validity. As Teeguarden et al. argue: “We strongly believe that the merits of research should be judged on the quality of work, through the peer review process, which exists to maintain scientific standards and assure objectivity.”
The final statement by Teeguarden et al. sums up the problem when gamesmanship replaces an ethical commitment science and method:
Looking forward, we propose that accusations of this sort, particularly those that purport to reveal fatal flaws or deceit, be subjected to fact-finding and full peer-review before publication. Publishing organizations such as Toxicological Sciences are the guardians of scientific objectivity through accurate representation of published data and cannot ignore the possibility that letters to the editor can be used to circumvent these standards. High standards of scientific integrity are essential because nothing is more damaging to any scientific discipline than the promotion of opinion as equivalent to fact.”
But as the media continues to report studies claiming a risk from BPA, no matter how flimsy the methodology, and ignoring regulatory studies which contradict these claims, no matter how rigorous the methodology, the incentives for gamesmanship in science have never been greater. The Atlantic didn’t report the Teeguarden study, but it did report supposed outrage among the scientific community about its flaws. That outrage is proven by nothing more than a letter to the editor that wouldn’t have passed the editorial process of a newspaper.