Abstract

Bisphenol A (BPA) is a key building block of polycarbonate plastic and a precursor for the manufacturing of monomers of epoxy resins. The BPA exposure of the general population is via food as a result of the use of BPA in food packaging and via skin as a result of contact with thermal paper. The vast majority of the population (91–99%) has detectable levels of BPA-conjugates in their urine. BPA is also present in medical devices including implants, catheters, tubing, and some dental materials. This Opinion describes the risk assessment of exposure to BPA via medical devices that are manufactured with materials that potentially leach BPA leading to oral (via dental material), subcutaneous and intravenous (e.g. during heamodialysis) routes of exposure.

After oral exposure BPA is readily absorbed from the gastro-intestinal tract and due to the first pass effect in the liver and the small intestine is rapidly conjugated to non-toxic metabolites. By the oral route BPA has a low systemic bioavailability (1-10% in humans) and has a half life time of a few hours. For parenteral routes of exposure (intravenous, intraperitoneal, subcutaneous), BPA can be considered 100% systemically bioavailable. However, BPA will also be conjugated in the liver and the clearance of free BPA from the circulation appears to be relatively fast.

Toxicity studies indicate that the kidney and the liver are relevant target organs for BPA toxicity. The lowest NOAEL after oral repeated exposure identified in several studies, including multigeneration reproductive toxicity studies, was approximately 5 mg/kg b.w./day. By applying the benchmark dose (BMD) approach, a BMDL10 of 8.96 mg/kg b.w./day was derived (EFSA 2015), based on the alteration in kidney weight. BPA is not likely to pose a genotoxic hazard to humans and has no carcinogenic activity, although there are some effects observed in the mammary gland, which currently are of unknown significance to human health. Neither reproductive nor prenatal developmental toxicity are critical end-points in BPA toxicity, though BPA is associated with reproductive toxicity at doses higher than those causing liver and kidney damage.

There are several indications that BPA might have biological effects below the recently determined BMDL10. However, the evidence is inconsistent, mainly obtained in dedicated studies focussing on different selected health effects and dose-response relationships could not be established. Regarding possible low dose effects, some concern remains for effects on mammary gland, metabolism and adiposity and neurobehaviour.

The SCENIHR adopts the temporary oral TDI (t-TDI) of 4 μg/kg b.w./day derived by EFSA which considers the BMDL10 as point of departure, and the uncertainties related to some effects using a BPA-specific assessment factor of 150. This t-TDI represents a useful base for carrying out a BPA risk assessment for the use of BPA in medical devices.

The BMDL10 dose was translated into a human dose inducing similar effects, the human equivalent dose (HED). The HED of 609 μg/kg b.w./day was determined considering the ratio of internal exposure in mice (the species used in the study from which BMDL10 was derived) versus the internal exposure in humans based on toxicokinetic studies.

For medical devices, several exposure scenarios were evaluated taking into account the material used, information related to BPA leaching, duration of a single treatment and the frequency of treatments, giving rise to toxicologically relevant acute, short and long term exposure. The information available is very limited and in many cases due to the lack of experimental data, only estimations were used. The uncertainties related to the exposure assessment indeed represent the weakest part of this evaluation. For this reason only the highest value obtained in the estimate was indicatively used.

The estimated BPA exposures were: 1) 3000 ng/kg b.w./day (3 μg/kg b.w./day) for prematurely born infants in neonatal intensive care units; 2) 685 ng/kg b.w./day for prolonged medical procedures (such as extracorporeal circulation or transplantation/ implantation of an artificial organ) in infants (b.w. around 5 kg); 3) 57 ng/kg b.w./day for dialysis patients; 4) 0.4 – 12ng/kg b.w./day for long-term exposures to medical devices; 5) 140 to 200 ng/kg b.w./day for, respectively, children and adults due to contact with dental materials (<24 h); 6) 2 to 12 ng/kg b.w./day due to long-term contact with dental materials.

Exposure via medical devices consisting of BPA containing PVC might even be higher with values estimated for adults up to 5000 ng/kg b.w./day and infants up to 12000 ng/kg b.w./day. However, exposure to BPA via BPA-containing PVC has been estimated based on extrapolation from data on phthalate leakage from PVC and has a high degree of uncertainty. Notably, European PVC manufacturers informed the SCENIHR that they no longer use BPA in PVC production, resulting in an expected low level of exposure, BPAcontaining medical devices from extra-EU manufacturers may be available in Europe.

Long-term oral exposure to BPA via dental material is below the t-TDI of 4μg/kg b.w./day, thus posing a negligible risk for human health. Regarding the parenteral exposure via medical devices, the kinetic differences between routes of exposure indicate that the bioavailability after oral route of exposure is significantly lower (around 1%) when compared to the parenteral one (assumed as 100%). Therefore, the SCENIHR applied a margin of safety (MOS) approach using an internal dose (6 μg/kg b.w./day = HED/100) rather than the external t-TDI.

Applying a MOS approach, for the 3 μg/kg b.w./day corresponding to medical devices use in intensive care the MOS is 2; for repeated medical procedures in infants the MOS is 10; for the other short term exposure scenarios estimated for different medical devices, the MOS ranges between 43 and 100; for dialysis treatments the MOS is 105. For the other scenarios the exposure ranges from 0.4 ng/kg b.w./day to 12 ng/kg b.w./day resulting in a MOS range from 500 to 15,000.

The uptake by the oral mucosa can be considerable: assuming a worst case scenario as 100%, for the highest exposures 200 ng/kg b.w./day the MOS would be 30. However, the peak of exposure occurs for <24 hours (representative of acute exposure in a toxicological context), whereas the MOS of 150 is related to chronic exposure. The differences in exposure duration can be larger than 5, therefore a MOS of 30 for acute exposure to dental materials can be considered sufficiently large.

Duration of exposure is indeed very relevant. With the exception of dialysis patients for which treatment is repeated for long time periods and can be considered similar to chronic exposure, the exposure via medical devices is generally of limited duration (representative of acute to subchronic exposure in a toxicological context). The application of correction factors to the MOS accounting for the time of exposure (3 as used by ECHA) might indicate that MOS around 50 could be sufficiently large, although with a high degree of uncertainties. Nevertheless some scenarios still raise concern.

The SCENIHR concludes that risk for adverse effects of BPA may exist when the BPA is directly available for systemic exposure after non-oral exposure routes, especially for neonates in intensive care units, infants undergoing prolonged medical procedures and for dialysis patients. Although the benefit of medical devices has also to be considered, the SCENIHR recommends that, where practicable, medical devices that do not leach BPA should be used. The possibility of replacing BPA in these products should be considered against their efficiency in the treatment, as well as the toxicological profile of the alternative materials.

However, better data on exposure would be beneficial for the refinement of the present risk assessment, to be carried out when new data on exposure via medical devices become available.

Reference
Safety of the use of bisphenol A in medical devices
Testai E, Hartemann P, Rodriguez-Farré E, Rastogi SC, De Jong W, Bustos J, Castle L, Hensten A, Kopperud HM, Olea N, Piersma A.
SCENIHR (Scientific Committee on Emerging and Newly Identified Health Risks) (2015).
ISSN: 1831-4783, ISBN: 978-92-79-30133-9 doi: 10.2772/75546
http://ec.europa.eu/health/scientific_committees/emerging/docs/scenihr_o_040.pdf

Laurence Castle