Résumé
The Scientific Panel on Food Additives, Flavourings, Processing Aids and Materials in Contact with Food has been asked to re-evaluate the use of bisphenol A (BPA) in plastic materials and articles intended to come into contact with food, giving particular attention to the exposure of infants.
Uses in food contact materials
BPA is present in certain food contact materials because it is used in the production of polycarbonate and epoxy-phenolic resins. Polycarbonate (PC) is a plastic widely used in articles such as infant feeding bottles, tableware (plates, mugs, jugs, beakers), microwave ovenware, storage containers, returnable water and milk bottles, and refillable water containers. PC is also used for water pipes. Epoxy-phenolic resins are used as an internal protective lining for food and beverage cans and as a coating on metal lids for glass jars and bottles. Epoxy-phenolic resins are also used as a surface-coating on residential drinking water storage tanks and vine vats.
Dietary exposure to BPA
Only consumer exposure via the diet has been considered in the present assessment. Conservative dietary exposure assessments on BPA have been made by the Panel for adults, infants and children. The estimates of total dietary exposure are as shown in Table 1.
[GRAPHIC]
Assessment of human bisphenol A exposure by biomonitoring of urinary excretion of bisphenol A metabolites in the general population gives an estimated average daily total exposure to BPA of up to 7 microg BPA/adult/day and upper range exposures up to 10 microg BPA/adult/day (0.16 microg BPA/kg bw/day for a 60 kg person) in the USA and 0.04 to 0.08 microg/kg bw/day in Japan (95 % confidence interval). The discrepancy between the levels of exposure estimated through urinary biomarkers and the levels of exposure estimated above by combining food consumption data with BPA concentrations in the diet is likely to be due to the conservative assumptions made.
Previous evaluations
The European Union in 2003 published a comprehensive Risk Assessment Report (RAR) for BPA from all sources in the context of the legislation on existing substances. BPA was also evaluated for use in plastic materials and articles intended to come into contact with foods by the Scientific Committee on Food (SCF) in 2002. The SCF considered that the overall oral No-Observed-Adverse-Effect Level ( NOAEL) for BPA, from the experimental animal data then available, was 5 milligrams/kg bw/day. The NOAEL was taken from a comprehensive 3-generation reproduction study in the rat that used dose levels ranging down to 1 microgram/kg bw/day. Applying an uncertainty factor of 500 to the NOAEL, the SCF derived a temporary Tolerable Daily Intake (TDI) of 0.01 mg BPA/kg bw/day. Since then, a considerable number of new papers have been published in the scientific literature, addressing various aspects of the toxicity of BPA, including its fate in the body and the issue of possible low-dose effects. BPA is known to have weak oestrogenic activity. The present re-evaluation builds on the previous evaluation by the SCF and focuses on the effects of BPA on reproduction and the endocrine (hormonal) system, about which there has been much scientific debate.
Due to the much lower bioavailability of BPA from oral administration compared to that from other routes of exposure and the relevance of the oral route for human exposure from food, effects seen in experimental animals following oral administration were considered the most pertinent data for the risk assessment.
Toxicokinetics and toxicodynamics
The Panel noted that new comparative data on toxicokinetics of BPA show that there are major species differences between rodents and humans in the way that BPA is handled in the body. For example, there are major differences in disposition of BPA-glucuronide due to different pathways of elimination from the liver in rodents and primates. This has important implications for the relevance of observations on the effects of BPA in sensitive strains of rodents, including low-dose effects, for human health risk assessment.
In humans and other primates, BPA given orally is rapidly transformed to BPA-glucuronide during first pass metabolism in the gut wall and the liver. The BPA-glucuronide formed, which is devoid of endocrine activity, is rapidly excreted in the urine, with an elimination half-life of less than 6 hours. Thus, there is very low oral bioavailability of the parent substance, BPA, in humans and other primates. Due to this rapid biotransformation and excretion and plasma protein binding in humans, peak BPA-concentrations after dietary exposures to BPA available for receptor binding are predicted to be very low even in worst case exposure scenarios. In rats, orally administered BPA also predominantly undergoes glucuronidation, but the BPA-glucuronide formed is excreted from the liver into the gut in the bile. In the gut, BPA-glucuronide is then cleaved into BPA and glucuronic acid and BPA is reabsorbed as such into the blood stream. This enterohepatic recirculation results in slow elimination of BPA in rodents. The Panel noted that while glucuronidation of BPA seems to be the major pathway of BPA biotransformation in rats, in mice oxidation products of BPA have been identified after low-dose administration, suggesting possible formation of metabolites with higher oestrogenic potency. The Panel also noted that there are major species differences between the mouse and the human, both in the physiology of gestation and in their toxicodynamic sensitivity to oestrogens, the mouse being particularly sensitive to oestrogens, which could predispose that species to sensitivity to weak oestrogens such as BPA.
Toxicity studies
In reviewing the earlier and the recently published studies on BPA, the Panel noted that some studies indicated differences in behaviour or reproductive parameters between control and treated animals at dose levels lower than the previously accepted overall NOAEL of 5 mg/kg bw/day. However, the Panel had considerable reservations both about the biological significance of the reported observations and the robustness of the studies.
The effects of BPA reported in some studies at low doses in sensitive animal systems were small changes in organ weight or changes in tissue architecture, increased or decreased receptor expression, changes in hormone concentrations in plasma or tissues, small changes in the time required to attain puberty landmarks, and behavioural effects. The Panel noted that the changes observed were often not sustained through adulthood. The biological consequences of many of the changes in the affected animals are unknown and some, such as small increases in prostate weight, are not considered as precursors of pathological change. While some of the changes may be indicative of biomarkers of effect in very sensitive species and strains, in the light of present knowledge, they cannot be readily interpreted as adverse effects.
The Panel also noted that in some studies reporting low-dose effects, only a single dose level was investigated, or there was absence of a dose-response relationship where several dose levels had been used. Many studies also used only small numbers of animals per dose group. There are also a number of other potential confounding factors in these types of study that may contribute to the lack of consistency in the database and this is discussed in more detail in the opinion.
For studies to be used for risk assessment purposes it is important that adequate numbers of animals are tested to control for individual variability of responses and that an adequate range of doses is tested to show a dose-response relationship. With regard to the claimed nonmonotonic dose-responses for BPA, the Panel notes that toxicologists are familiar with Ushaped and inverted U-shaped dose-response curves for hormonal activities, but the presence of a response at one dose level only does not necessarily indicate a causal relationship between the administration of a substance and an observed change. To demonstrate U-shaped dose responses in a robust way, it is necessary to have reasonably spaced dose intervals, usually of less than 10-fold, and not steps of 1000-fold as in some recent studies.
The Panel also noted that the results of the studies reporting low-dose effects are in contrast to the results of other studies using comprehensive protocols developed for testing both structure and reproductive function in parents and offspring and performed following internationally recognised guidelines with regard to study design and animal model selection. A two-generation reproductive toxicity study has recently been reported in an oestrogensensitive strain of mice administered a wide range of BPA doses in the diet. BPA administration in the low-dose range did not result in changes in reproductive organs or performance and gave an overall NOAEL of 5 mg/kg bw/day, with liver toxicity as the most sensitive endpoint. The positive control substance, 17 beta-oestradiol, resulted in reproductive and developmental toxicity.
Moreover, a number of other studies applying low doses of BPA were also unable to demonstrate low-dose effects on reproduction or development. Thus, the literature remains inconsistent with regard to strain and species sensitivity to low-dose effects of BPA.
Conclusions
The Panel considered that low-dose effects of BPA in rodents have not been demonstrated in a robust and reproducible way, such that they could be used as pivotal studies for risk assessment. Moreover, the species differences in toxicokinetics, whereby BPA as parent compound is less bioavailable in humans than in rodents, raise considerable doubts about the relevance of any low-dose observations in rodents for humans. The likely high sensitivity of the mouse to oestrogens raises further doubts about the value of that particular species as a model for risk assessment of BPA in humans.
For these reasons, the Panel c