Reversal of association between prenatal phthalate levels and birth size when measuring phthalates in placental tissue versus maternal urine

Abstract

Di (2-Ethylhexyl) phthalate (DEHP) is a widely used plasticizer worldwide and is a potent endocrine disruptor. It is unclear which biospecimen from pregnancy best represents fetal exposure when estimating the relationship between maternal DEHP exposure and child health. Therefore, we compared the association between DEHP metabolites and birth size when biomarkers of exposure were measured in placental tissue versus maternal urine.This study included fifty mother-infant dyads from the CANDLE study. Samples were analyzed for DEHP metabolites (MEHP, MEOHP, MEHHP, MECPP), summarized as a molar sum (DEHP-oxo). Associations of placental and urinary DEHP metabolites were calculated, considering fetal sex, maternal race, and BMI as confounders and effect modifiers of maternal-placental transfer and metabolism. Fetal sex differences in DEHP pharmacokinetics are conceptualized as genetic/epigenetic in origin; whereas race and BMI could represent cultural, dietary, or other sources of variation in exposures and/or pharmacokinetics. In the second aim, multiple regression models were used to assess associations of DEHP oxidized metabolites and birth length and weight z scores, and head circumference (HC). DEHP metabolites in the two tissues were uncorrelated in univariate models. A negative effect was found between MEHP in placenta and MECPP in urine, interacted on race (β=-0.36 log units MEHP [95% CI -0.67, -0.05] in Black vs. White women per log MECPP). The association was positive for the effect of urinary DEHP-oxo on birth size; and negative for the effect of placental tissue DEHP-oxo. The highest precision was observed in the association between placental DEHP-oxo and HC (β= -0.94 cm per log DEHP-oxo; 95% CI=-2.03, 0.15). Maternal urine might provide different information than the placenta on prenatal exposure to phthalates. Future work will compare MEHP in these two tissue types, to better understand the pharmacokinetics and to clarify if these results indicate a measurement error or a biologic inference problem.