Role of epidemiology in risk assessment: a case study of five ortho-phthalates.

BACKGROUND: The association between environmental chemical exposures and chronic diseases is of increasing concern. Chemical risk assessment relies heavily on pre-market toxicity testing to identify safe levels of exposure, often known as reference doses (RfD), expected to be protective of human health. Although some RfDs have been reassessed in light of new hazard information, it is not a common practice. Continuous surveillance of animal and human data, both in terms of exposures and associated health outcomes, could provide valuable information to risk assessors and regulators. Using ortho-phthalates as case study, we asked whether RfDs deduced from male reproductive toxicity studies and set by traditional regulatory toxicology approaches sufficiently protect the population for other health outcomes. METHODS: We searched for epidemiological studies on benzyl butyl phthalate (BBP), diisobutyl phthalate (DIBP), dibutyl phthalate (DBP), dicyclohexyl phthalate (DCHP), and bis(2-ethylhexyl) phthalate (DEHP). Data were extracted from studies where any of the five chemicals or their metabolites were measured and showed a statistically significant association with a health outcome; 38 studies met the criteria. We estimated intake for each phthalate from urinary metabolite concentration and compared estimated intake ranges associated with health endpoints to each phthalate's RfD. RESULT: For DBP, DIBP, and BBP, the estimated intake ranges significantly associated with health endpoints were all below their individual RfDs. For DEHP, the intake range included associations at levels both below and above its RfD. For DCHP, no relevant studies could be identified. The significantly affected endpoints revealed by our analysis include metabolic, neurodevelopmental and behavioral disorders, obesity, and changes in hormone levels. Most of these conditions are not routinely evaluated in animal testing employed in regulatory toxicology. CONCLUSION: We conclude that for DBP, DIBP, BBP, and DEHP current RfDs estimated based on male reproductive toxicity may not be sufficiently protective of other health effects. Thus, a new approach is needed where post-market exposures, epidemiological and clinical data are systematically reviewed to ensure adequate health protection.