LAUSR

BACKGROUND AND OBJECTIVE Despite plausible laboratory evidence about harms of endocrine-disrupting chemicals (EDCs), systematic reviews and meta-analyses of epidemiological studies do not consistently find such harms. The purpose of this article was to discuss why solid human evidence on EDCs was unlikely to be obtainable even though they are truely harmful to humans. RESULTS Indeed, there are insurmountable methodological limitations in human studies. They include unpredictable net effects of diverse EDC mixtures, low reliability of exposure assessment, nonmonotonic dose-response relationships, nonexistence of an unexposed group, and complicated interactions with diet and obesity. The exposome is never free from these methodological issues. Therefore, the most persuasive evidence about human harms of EDCs may be increasing EDC-linked diseases at population levels, but traditional epidemiological studies of EDCs, especially short-lived and widely used EDCs, fail to provide consistent results. Nevertheless, human studies of EDCs with long half-lives, for example, persistent organic pollutants (POPs), are still worthwhile because there are fewer methodological issues. Also, they can play a role as surrogate markers of lipophilic chemical mixtures. Notably, although POPs are well-known EDCs, human findings on POPs cannot be attributed to common hormone-disrupting properties because the net effect of diverse EDC mixtures cannot be reliably predicted even with POPs. Homeostasis disruption through mitochondrial dysfunction may be more relevant to such effects. CONCLUSION Logical inference should play a primary role in judging harms of EDCs in humans.