The toxicity of bisphenol A (BPA) has been a fairly controversial subject for a while. Industry groups have been fairly adamant about its safety, while many environmental groups suggest it is causing adverse health effects in humans. Messages from regulatory agencies have been fairly mixed, or even confusing. While Canada declared the substance “toxic” under the Canadian Environmental Protection Act, they did not actually conclude that it was causing a risk to human health. The US National Toxicology Program found “some concern” for infants and young children, based on insufficient data in humans and limited evidence of developmental effects in animals. The World Health Organization concluded that further regulation was not warranted based on current information, but that there were data gaps where further research is needed.
A couple of years ago, a paper was published by Myers et al. (2009) (a paper with 36!!! authors from 30 different institutions) that was highly critical of the methods used by public health agencies for the evaluation of BPA toxicity, and in particular their reliance on studies conducted using “Good Laboratory Practice” (GLP) standards over academic studies. Now a new critical review has been published by the Advisory Committee of the German Society of Toxicology (Hengslter et al., 2011) that concludes the European tolerable daily intake (or “safe dose”) of 0.05 mg/kg body weight/day is justified (for comparison, the US also uses 0.05 mg/kg/d, while Canada uses 0.025 mg/kg/d). This review also takes specific aim at the Myers et al. paper and rejects its conclusions.
To keep things relatively brief, I’m just going to focus on a few of the more controversial or important aspects of BPA toxicity. One of the key questions is whether the multi-generation rodent studies used to derive the tolerable daily intake are valid. Myers et al. criticized these studies, based on apparently insensitive species used, insensitive protocols, and issues with how prostates were examined, among other issues. The Hengstler et al. review systematically addresses these arguments based on additional data provided by the study authors, as well as subsequent studies that have been undertaken and published in 2010. Overall I think most of the criticisms were adequately addressed, except perhaps some of the comments relating to the study methods, and in particular that there are more sophisticated techniques that are more likely to find anomalies in tissues such as mammary glands than the histopathological analyses used in GLP studies (though on the other hand, the toxicological significance of some of these anomalies is often unknown).
Hengstler et al. also looked at why industry-sponsored and publicly sponsored studies often show different results. They concluded that the academic studies are often exploratory studies with low numbers of animals, looking at a large number of potential effects, generally without accounting for multiple comparisons in the statistical evaluation, resulting in a large number of “false positives” (apparent effects that are actually just random variation in the population), combined with publication bias (only publishing studies that generate positive results, while negative studies remain unpublished). I agree these are valid concerns, although if a specific effect occurs repeatedly it shouldn’t be ignored.
Another key issue is the exposure route. Many of the academic studies that have found effects from BPA at low doses involved injecting BPA into animals. However, human exposures are usually through ingestion in food and water. While the ingested BPA is almost completely absorbed by humans, it is passed through the liver and metabolized before going to other tissues. This basically means that most of these studies are not useful for determining what doses of BPA are toxic or non-toxic.
Sensitive populations, and in particular whether infants and young children are more sensitive to BPA, is another key issue, and one where I feel Hengstler et al. might not have made their case as strongly. The metabolism of infants and young children is not as developed as that in adults. Information in the Hengstler et al paper suggests that the activity of a key enzyme in BPA metabolism is probably about 5% of the adult level in newborns, 30% of the adult level at 3 months, 80% at 6 months, and 100% by 1 year. However, they then suggest that the standard safety factor of 10 used when determining safe doses for sensitive humans (compared to the general population), which is generally considered to be a factor of 3.3 for differences in absorption, distribution in the body, metabolism and excretion, and a factor of 3.3 for differences in how the toxin acts at the molecular level to cause an effect. In this case, I would argue that the much lower metabolism in infants indicates that a larger safety factor would be required for the protection of infants – perhaps a factor on the order of 60 (20 for differences in metabolism and 3.3, rounded to 3, for differences in toxic effect) or more.’
The review also briefly considers the available epidemiological studies in humans, including the highly publicized Lang et al. (2008) study that associated high levels of BPA in urine with effects such as cardiovascular disease and diabetes. However, the design of these studies, and in particular the use of a single measurement of BPA, which has a very short half-life in the body, to long-term health effects, and concluded that the studies cannot establish a cause-effect relationship and can at best generate hypotheses. I agree this is true; furthermore, given some of the common sources of BPA exposure (e.g. canned food), I would not be surprised to find a relationship between BPA levels in urine and poor diet, which can certainly be associated with cardiovascular disease and diabetes.
So what can we conclude? I’m inclined to agree with the authors that there is still no compelling evidence that exposures below the tolerable daily intake are harmful to adults or older children. Personally I think a lower tolerable daily intake for infants may be warranted though, and banning BPA in baby bottles and other infant products is probably not a bad idea, given both the differences in metabolism at this stage and the potential for developmental effects being highest at this age. Since typical adult exposures are well below these tolerable daily intakes, there isn’t really a panic for other uses. However, there are still some uncertainties and data gaps, and more extensive testing to further evaluate some of the suggested low-dose effects is still warranted.
Hengstler, J., Foth, H., Gebel, T., Kramer, P., Lilienblum, W., Schweinfurth, H., Völkel, W., Wollin, K., & Gundert-Remy, U. (2011). Critical evaluation of key evidence on the human health hazards of exposure to bisphenol A Critical Reviews in Toxicology, 41 (4), 263-291 DOI: 10.3109/10408444.2011.558487
Lang, I., Galloway, T., Scarlett, A., Henley, W., Depledge, M., Wallace, R., & Melzer, D. (2008). Association of Urinary Bisphenol A Concentration With Medical Disorders and Laboratory Abnormalities in Adults JAMA: The Journal of the American Medical Association, 300 (11), 1303-1310 DOI: 10.1001/jama.300.11.1303
Myers, P.J., vom Saal, F., Akingbemi, B., Arizono, K., Belcher, S., Colborn, T., Chahoud, I., Crain, D., Farabollini, F., Guillette, L., Hassold, T., Ho, S., Hunt, P., Iguchi, T., Jobling, S., Kanno, J., Laufer, H., Marcus, M., McLachlan, J., Nadal, A., Oehlmann, J., Olea, N., Palanza, P., Parmigiani, S., Rubin, B., Schoenfelder, G., Sonnenschein, C., Soto, A., Talsness, C., Taylor, J., Vandenberg, L., Vandenbergh, J., Vogel, S., Watson, C., Welshons, W., & Zoeller, R. (2008). Why Public Health Agencies Cannot Depend upon ‘Good Laboratory Practices’ as a Criterion for Selecting Data: The Case of Bisphenol A Environmental Health Perspectives DOI: 10.1289/ehp.0800173