Methodological issues in human studies of endocrine disrupting chemicals

Possible harm from endocrine disrupting chemicals (EDCs) in humans is speculated based on two types of evidence; 1) increasing trends of suspected diseases in ecological studies of populations and 2) findings from traditional epidemiological studies of individuals. However, ecological findings are not regarded as direct human evidence of the relation between EDCs and disease, while the evidence among epidemiological studies of individuals is often inconsistent. Thus, a criticism is that linking EDCs and health in human is naively presumed without solid evidence. However, human studies of EDCs are methodologically complex and understanding methodological issues will help to interpret findings from existing human studies and to properly design optimal human studies. The key issues are low reliability of exposure assessment of EDCs with short half-lives, EDC mixtures, possibility of non-monotonic dose–response relationships, non-existence of an unexposed group, difficulties in measuring exposure during critical periods, and interactions with established risk factors. Furthermore, EDC mixtures may affect human health through other mechanisms than traditional endocrine disruption, for example glutathione depletion or mitochondrial dysfunction. Given this complexity, the most plausible scenario in humans is that exposure to EDC mixtures leads to increasing risk of related diseases at the ecological level, but inconsistent associations would be expected in traditional epidemiological studies. Although epidemiologists have long relied on Bradford Hill’s criteria to objectively evaluate whether associations observed in epidemiology can be interpreted as causal, there are challenges to use these criteria for EDCs, particularly concerning consistency across studies and the findings of linear dose–response relationships. At the individual level, compared to EDCs with short half-lives, epidemiological studies of EDCs with long half-lives among populations with a relatively low exposure dose range of exposure can likely produce relatively more reliable results, because the measurement of EDCs with long half-lives likely represents typical long-term exposure and populations with exposure in the low range of doses are likely to have a reference group closer to non-exposure.