Published online 16 September 2008 | Nature | doi:10.1038/news.2008.1110
Updated online: 16 September 2008


Bisphenol A linked to disease in humans

More studies of the controversial chemical are on the way.

Energy drinksFood packaging has come under suspicion in worries about BPAWestend61

High levels of bisphenol A (BPA) — a chemical used in some containers for food and drink — may be associated with an increased risk of diabetes and cardiovascular disease in humans, a new study has found.

The study, published this week by the Journal of the American Medical Association1, is the first large-scale investigation of the controversial chemical's effect on human disease. It is a welcome addition to the body of BPA research, the vast majority of which has been conducted in animals.

But the results do not establish a causal link between BPA and disease, and the study design does not allow researchers to determine which came first: higher exposure to BPA or illness. "I think our study definitely puts a scientific question mark over this compound," says epidemiologist David Melzer of the Peninsula Medical School in Exeter, UK, who led the research. "Still, this is the first study. It has to be repeated."

Regulatory morass

Work in animals has suggested that BPA has the potential to disrupt normal hormone signalling by mimicking the natural hormone, oestrogen. Such studies have linked the chemical to a wide range of conditions, including low sperm count, altered fetal development, behavioural disorders in children and prostate cancer.

“I think the data five years ago warranted this concern.”

Frederick vom Saal
University of Missouri-Columbia

Governments around the globe have struggled to evaluate the risk of BPA. In April this year, Canadian officials called the chemical "potentially harmful", and have proposed a ban on baby bottles made with BPA. In January 2007, the European Food Safety Authority said that dietary exposure to BPA was well below what it defined as the tolerable daily intake — but it has since pledged to re-evaluate the compound in light of decisions in the United States and Canada. Meanwhile, US officials have convened multiple panels to evaluate the risk posed by BPA. Each reached a different conclusion.

Some of the discrepancy stems from differences in how each panel values the previous research in animals, says epidemiologist Russ Hauser of the Harvard School of Public Health in Boston, Massachusetts. One panel might give less weight to studies that injected animals with BPA rather than feeding them the compound; another might view industry-sponsored research with less confidence than academic research.


How the new data from humans will affect regulators remains to be seen: Melzer and his colleagues plan to deliver their results today to a public hearing on the US Food and Drug Administration's draft BPA report. Released in August this year, the draft report declared BPA-containing food containers to be safe. Meanwhile, more studies in humans are likely to be published in the near future, Hauser says.

Risk assessment

As BPA is believed to be a hormone disrupter, most of the discussion about its dangers has revolved around possible effects on reproduction and development. But Melzer, who studies ageing, wondered if BPA might affect diseases commonly associated with that area as well.

Melzer and his colleagues used data from 1,455 adults who participated in the US National Health and Nutrition Examination Survey, conducted by the US Centers for Disease Control and Prevention. Participants donated urine samples and filled out questionnaires about their health. Previous analysis of BPA levels in these urine samples had shown that 92.6% of people over the age of six had detectable levels of BPA in their urine2.

“We don't know if exposure to BPA happened first, or if the health problem happened first”

De-Kun Li
Kaiser Permanente

The researchers compared BPA levels to responses on the health questionnaire. When the data were divided into four equally sized groups based on BPA concentration, those in the group with the highest concentration were three times more likely to have cardiovascular disease and 2.4 times more likely to have diabetes than those in the group with the lowest concentration. Even when the analysis accounted for differences in nine parameters — including age, sex, race, education, income, smoking and body mass index — the pattern was the same.

There was also a link between increasing BPA concentrations and abnormal levels of two liver enzymes, suggesting possible liver damage. But no significant link was found between the compound and other conditions including stroke, cancer and arthritis.

Snapshots of exposure

The data analysis was thorough and sophisticated, says Hauser, but the study might capture only a snapshot of BPA exposure. In adults, most BPA is typically excreted from the body within a day of ingestion, and it can be difficult to correlate such short-term data with diseases that take years to develop. That uncertainty could cause such a study to underestimate the effects of BPA, he says.

Then there is the chicken-and-egg question. "The biggest problem here is that we don't know if exposure to BPA happened first, or if the health problem happened first," says De-Kun Li, a reproductive and perinatal epidemiologist who works for the division of research at Kaiser Permanente, a health-care organization in Oakland, California. Conditions such as obesity or altered liver function could affect the rate at which BPA is metabolized by the body. Given such limitations, Li urges caution before making a causal connection between BPA and disease: "At most, it's very preliminary," he says.

But for Frederick vom Saal, a toxicologist at the University of Missouri-Columbia who has studied BPA in animal models, the results are entirely expected. "These findings should not be a surprise to anybody," he says. "I think the data five years ago warranted this concern." 


Senator Chuck Grassley (Republican, Iowa) responded quickly to the study's publication today by launching an investigation of the FDA panel's decision.

  • References

    1. Lang, I. A. et al. J. Am. Med. Assoc. 300, 1303-1310 (2008). | Article |
    2. Calafat, A. M. et al. Environ. Health Perspect. 116, 39-44 (2008). | PubMed | ChemPort |
Commenting is now closed.