Drivers of divergent assessments of bisphenol-A hazards to semen quality by various European agencies, regulators and scientists.

The downward revision of the bisphenol A (BPA) Health-based Guidance Value (HBGV) by the European Food Safety Authority (EFSA) has led to disagreements with other regulatory agencies, among them the German Federal Institute for Risk Assessment (BfR). The BfR has recently published an alternative Tolerable Daily Intake (TDI), 1000-times higher than the EFSA HBGV of 0.2 ng/kg/d. While the EFSA value is defined in relation to immunotoxicity, the BfR alternative TDI is based on declines in sperm counts resulting from exposures in adulthood. Earlier, we had used semen quality deteriorations to estimate a BPA Reference Dose (RfD) of 3 ng/kg/d for use in mixture risk assessments of male reproductive health. We derived this estimate from animal studies of gestational BPA exposures which both EFSA and BfR viewed as irrelevant for human hazard characterisations. Here, we identify factors that drive these diverging views. We find that the fragmented, endpoint-oriented study evaluation system used by EFSA and BfR, with its emphasis on data that can support dose-response analyses, has obscured the overall BPA effect pattern relevant to male reproductive effects. This has led to a disregard for the effects of gestational BPA exposures. We also identify problems with the study evaluation schemes used by EFSA and BfR which leads to the omission of entire streams of evidence from consideration. The main driver of the diverging views of EFSA and BfR is the refusal by BfR to accept immunotoxic effects as the basis for establishing an HBGV. We find that switching from immunotoxicity to declines in semen quality as the basis for deriving a BPA TDI by deterministic or probabilistic approaches produces values in the range of 2.4-6.6 ng/kg/d, closer to the present EFSA HBGV of 0.2 ng/kg/d than the BfR TDI of 200 ng/kg/d. The proposed alternative BfR value is the result of value judgements which erred on the side of disregarding evidence that could have supported a lower TDI. The choices made in terms of selecting key studies and methods for dose-response analyses produced a TDI that comes close to doses shown to produce effects on semen quality in animal studies and in human studies of adult BPA exposures.