Cardiovascular disease, diabetes linked to bisphenol A levels in NHANES population

Despite reported adverse effects associated with bisphenol A exposure, the FDA said that current exposure levels are safe.

Researchers have discovered a possible link between exposure to the chemical bisphenol A and cardiovascular disease, diabetes and abnormal concentrations of liver enzymes. The study was published in JAMA on Tuesday, the same day the FDA held a hearing in Washington to assess the use of bisphenol A in food contact applications.

Using data from the NHANES study, researchers from Peninsula Medical School and other sites in the United Kingdom, along with the University of Iowa College of Public Health in Iowa City, conducted a cross-sectional analysis to determine the link between bisphenol A concentrations and health status in U.S. adults.

The researchers analyzed measurements of urine bisphenol A and urine creatinine concentrations in 1,455 adults aged 18 to 74 years. They adjusted for sex, age, BMI and others factors. The primary outcome was chronic disease diagnosis plus blood markers of liver function, glucose homeostasis, inflammation and lipid changes, according to the study.

Health effects of bisphenol A

In age-, sex- and fully-adjusted models, cardiovascular diagnoses were associated with higher concentrations of urinary bisphenol A (OR per 1-standard deviation increase in bisphenol A concentration=1.39; P=.001). Similarly, diabetes was also associated with high concentrations of bisphenol A (OR per 1-standard deviation increase in bisphenol A concentration=1.39; P<.001). However, no correlations were found between high levels of bisphenol A and other common diseases studied.

The researchers reported that higher concentrations of bisphenol A were associated with the presence of clinically abnormal concentrations of liver enzymes y-glutamyltransferase (OR per 1-standard deviation increase in bisphenol A concentration=1.29; P<.001) and alkaline phosphatase (OR per 1-standard deviation increase in bisphenol A concentration=1.48; P=.002).

“Until we are able to repeat these results and clarify that the effects are definitely due to bisphenol A itself, we cannot say for certain that bisphenol A causes disease in humans,” the researchers wrote in a briefing document that was sent to Endocrine Today.

Safety concerns and a need for data

The FDA concluded in its draft assessment for Tuesday’s hearing that “an adequate margin of safety exists for bisphenol A at current levels of exposure from food contact uses for infants and adults.”

In an accompanying editorial, however, two experts said that these findings, combined with previous data on the adverse effects of low-dose bisphenol A in animal models, should be evidence enough for government agencies to reevaluate the Environmental Protection Agency’s recommended “safe” dose of bisphenol A in humans — currently at 50 mcg/kg per day — and reduce exposure to bisphenol A.

Frederick S. vom Saal, PhD, in the division of biological sciences at the University of Missouri, and John Peterson Meyers, PhD, from Environmental Health Sciences in Charlottesville, Va., wrote that the data “should spur U.S. regulatory agencies to follow the recent action taken by Canadian regulatory agencies, which have declared bisphenol A a ‘toxic chemical’ requiring aggressive action to limit human and environmental exposures.”

Similarly, in the briefing document from the JAMA researchers, they write that, “if the findings are replicated and it is clear that bisphenol A is causing the problems we have identified, then safe limits would need to be reviewed. Government agencies need to make this decision based on an overview of all the human and animal evidence, not just our analysis.” – by Stacey L. Adams

JAMA. 2008;300:1303-1310.
JAMA. 2008;300:1353-1355.

This is the first large human study that looked at bisphenol A exposure and human health endpoints. Though the study was well done, it did have limitations. It used cross-sectional data from NHANES and assessed disease status — whether cardiovascular disease, liver enzymes — at the same time they measured bisphenol A with a single urine sample. A limitation would be the exposure assessment for bisphenol A, which is a short-lived chemical that has a half-life of several hours. Therefore a spot urine sample measures someone's exposure to bisphenol A in the previous day or so.

However, these are chronic disease conditions that would take years or decades to develop. Having said that, this study does have many strengths and the findings are provocative. Rather than doing the statistical analyses just one way, they performed sensitivity analyses to test the robustness of the results. When you get the same result in the sensitivity analyses, as they did, this tells you that the results are consistent and you tend to have more confidence in the results. With most epidemiologic studies, we need replication to confirm these results in other populations. Therefore, this one study is not going to be the final answer; there needs to be more research done. There are a lot of other endpoints that also need to be studied. These include the relationship of bisphenol A with developmental and reproductive endpoints in humans.

– Russ Hauser, MD, MPH, ScD

Professor of Environmental and Occupational Epidemiology
Harvard School of Public Health