Studies supposedly linking the plastic additive to diabetes, heart disease and coronary artery disease have been called a “bombshell” by anti-BPA NGOs and many journalists. Now those conclusions, and a central contention of
campaigners, is in doubt.
The most explosive claim of anti-BPA campaigners—that the plastic additive BPA causes an array of heart-related diseases—is in question, according to a
peer reviewed paper on the science website PLOS One.
Environmental health scientist Judy LaKind from Penn State University and the University of Maryland and epidemiologist Michael Goodman from
Emory University reviewed data from the National
Health and Nutrition Survey (NHANES) that previous researchers concluded linked BPA to chronic diseases. Johns Hopkins mathematician Daniel Naiman did the analysis.
In contrast to those previous studies, which looked at only one, two or three datasets, these researchers found no associations between urinary BPA and heart disease or diabetes across four NHANES datasets. Their conclusions challenge one of the central contentions of researchers who believe that BPA is harmful.
The influence of the NHANES data in creating the popular belief that BPA is harmful cannot be overstated. The controversy originated just a few years ago, when bisphenol A was still a relatively obscure plastic additive that a group of obscure scientists had targeted as dangerously toxic.
Based on controversial studies of rodents injected with the chemical, they had come to believe that BPA was what they called an “endocrine disruptor” that did its dirty work at low doses. It distorted hormonal functions, they claimed, and could be blamed for a host of problems from cancer to reproductive and metabolic issues to heart disease. It was a controversial contention, as toxicity has traditionally been linked to exposure—the dose makes the poison, in Paracelsus’ famous phrase.
Heart disease theory rests on questionable data?
A key turning point in the debate came in 2008 with the release of a
study based on the NHANES data covering 2003/4 of nearly 1500 adults in the
Journal of the American Medical Association. A team of researchers led by David Melzer, an epidemiologist at the Peninsula College of Medicine and Dentistry at the University of Exeter in the
United Kingdom concluded that respondents with higher amounts of BPA in their urine were more likely to report having heart disease and diabetes.
“This is a big deal,”
said University of Missouri biologist Frederick vom Saal, the chief proponent of the “endocrine disruptor” hypothesis, who
co-authored an
opinion piece that accompanied the study in
JAMA. He and John Peterson Myers, a biologist and longtime collaborator, demanded immediate regulatory restrictions on BPA and phthalates, another class of chemicals they contend is dangerous.
The associations were modest, which led the Food and Drug Administration to immediately reaffirm its belief that BPA was safe. But that’s not how it was played in the media and by advocacy NGOs, which flooded the Internet with hundreds of stories “linking” BPA to heart disease. Thousands of articles have since cited the NHANES study as “proof” of BPA’s harmful effects or otherwise casually asserted that BPA is “linked to” or “associated with” chronic heart problems.
After the release of yet another Melzer study based on more recent NHANES data, in 2010, the Natural Resources Defense Council hyperbolically
characterized the findings as a “bombshell” as part of its campaign to connect common exposure to everyday chemicals to serious diseases, such as cancer—claims that are not supported by the evidence.
“Health care reform should be linked directly to toxic chemical reform,”
wrote Gina Solomon, a scientist and former blogger for the NRDC. “Chemicals such as BPA are a potentially preventable cause of serious illness, and prevention saves lives and dollars.”
Cherry picking data?
The LaKind-Goodman study identified what appear to be two anomalies in the analysis by Melzer and two other related papers released in
2010 and
2011. A diabetes study included as diabetic people who did not have diabetes but had borderline symptoms—a non-standard definition of diabetes. Without those people included, the BPA-diabetes link disappeared.
The heart disease
study found a weak association between BPA and heart disease—but it excluded six people who had the highest BPA concentrations. It turns out that none of those left out had heart disease. The inclusion of those respondents would have led to a finding of no association between BPA and serious heart problems. This contentious evidence led to the “bombshell” finding the NRDC crowed about.
The only explanation for leaving out the healthy respondents provided in the Melzer paper is that those excluded were “outside the range of BPA in the original 2003/04 sample,” which topped out at 80.1 ng/mL. According to
epidemiologists I spoke with, they made an odd and arbitrary choice. In a blistering
online response to the LaKind-Goodman study they now maintain that the excluded samples, which range from 83.6 to 150, and one outlier at 383, might have been “contaminated.”
In their response, Melzer et al. sharply challenged the overall thrust of the new study, calling it “unfocused” and “poorly documented, and noted that the LaKind-Goodman research was supported by the Polycarbonate/BPA Global Group of the American Chemistry Council. According to the paper, and under the rules of the peer review process, “the ACC was not involved in the design, collection, management, analysis, or interpretation of the data; or in the preparation or approval of the manuscript.”
In more substantive criticism, Melzer said that the new study had left out more than 400 survey respondents, implying those excluded could have skewed the results. LaKind and Goodman wrote they excluded survey respondents who omitted their age, body mass, smoking behavior or other variables to keep the data consistent. Melzer and the primary author of the diabetes study, University of Michigan doctoral pre-candidate Monica Silver, also pointed out that the new study included children in the diabetes assessment, which could also account for the different conclusions. Regardless, responded Lakind and Goodman, they claim they consistently found no associations between urinary BPA and heart disease or diabetes across four NHANES datasets.
Melzer pointedly noted that in a more recent study,
published earlier this year, his team found that those who developed coronary artery disease tended to have higher urine BPA concentrations up to ten years earlier than those who did not develop heart disease.
The dispute over the data threatens to obscure the LaKind and Goodman’s most salient conclusion. NHANES is a robust and critically important public health database, they maintain. However, it only measures concurrent exposure to chemicals as reflected in urine, and not long-term impacts.
Limitations of NHANES survey to analyze BPA
“Our results don’t shed light on whether BPA is or isn’t a risk factor for diabetes or heart disease,” said LaKind. “Rather, the point we are making is that using data from cross-sectional studies like NHANES surveys to draw such conclusions about relations between short-lived environmental chemicals and chronic diseases is inappropriate.”
Melzer brushed off that point completely in his response. But Monica Silver,
who headed the diabetes study using the NHANES data, emailed me: “I completely agree [with LaKind and Goodman on this point] and make similar conclusions in our paper. NHANES’ utility is not in making broad statements of causation of a given disease by a given exposure, but rather in providing preliminary, hypothesis building evidence that can inform future work.”
Many science-challenged journalists and activist NGOs, like the NRDC and Environmental Working Group that put advocacy ahead of science, consistently misrepresent and hype studies that show the presence of chemicals in urine, as if that signals likely toxic effects. The use of biomonitoring data is problematic, say scientists, particularly as it pertains to BPA. According to the
FDA reflecting the emerging scientific consensus, “[O]ral BPA administration [of BPA] results in rapid metabolism of BPA to an inactive [and therefore harmless] form.” In other words, BPA is detoxified and excreted.
That was confirmed in what is considered the state-of-the-art,
independent study financed by the Environmental Protection Agency on the potential harm of BPA—headed 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. 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.
Although low doses of certain chemicals
can induce non-monotonic effects, scientists who have reviewed these studies, time and again, have come away unconvinced these effects consistently or even generally suggest harm. Since 2007, there have been more than a dozen comprehensive reviews of BPA studies by independent government scientists around the world, including in
Canada, Europe,
Japan, Australia and the United States, and each has concluded that current uses of the chemical are safe.
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 have all 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.
Most recently, in October, 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.”
More on science literacy at the Genetic Literacy Project
Follow Jon on Twitter
Jon Entine is senior fellow at the Center for Health & Risk Management and STATS at George Mason University.
