[Dilemma of Environmental Health Research].

This article presents not only a brief overview of birth cohort studies focusing on environmental health in which the associations between health and environment were examined, but also a tentative plan to apply epidemiological data to benchmark dose calculation. According to the preceding studies, the checkpoints to be scrutinized when a result is not consistent with those of other researchers are as follows: (1) whether the study included all crucial confounders, (2) whether it included any exposure marker or confounder with a U-shaped dose-response curve, (3) whether the outcome measure was conducted by two or more examiners that might lead to measurement bias, (4) whether such examiners picked up information about exposure levels of the subjects before measuring the endpoints, and (5) whether subjects with different genetic factors were included in the analysis. In addition, (6) researchers conducting a children's study on developmental effects due to toxic substances must keep in mind that the impact of prenatal methylmercury exposure, independent of postnatal exposure, may continue for at least seven years. (7) When an environmental health research emphasizes to be population-based study, the levels of exposure to environmental chemical substances in developed countries with strict environmental regulations may be too low to examine a dose-response relationship for critical dose estimation. Such risk assessment should be carried out among the subjects with a wide range of exposure levels.