Prenatal exposure to phthalates and phenols and infant endocrine-sensitive outcomes: The MIREC study

BACKGROUND: Anogenital distance (AGD) and the second to fourth finger (2D:4D) digit ratio may be early markers of in utero androgen exposure for the infant. Phthalates and phenols have been identified as endocrine disrupting chemicals. OBJECTIVES: To study the association between prenatal exposure to phthalates, bisphenol A (BPA) and triclosan (TCS) and AGD and the 2D:4D digit ratios. METHODS: Single spot urine samples were collected in the first trimester from the MIREC Study and analyzed for phthalates and phenols. Anogenital distance (n = 394) at birth and 2D:4D digit ratios (n = 420) at 6 months were measured in male and female infants. Associations between maternal concentrations of phenols and phthalate metabolites and these outcomes were estimated using multiple linear regression models. RESULTS: In females, the anoclitoris distance (ACD) was negatively associated with mono-benzyl phthalate (MBzP) (β = -1.24; 95% CI -1.91, -0.57) and positively associated with mono-ethyl phthalate (MEP) (β = 0.65; 95% CI 0.12, 1.18) (masculinizing). In males, anopenile distance (APD) was positively associated with mono-n-butyl phthalate (MnBP) (β = 1.17; 95% CI 0.02, 2.32) and the molar sum of low molecular weight phthalates (ΣLMW). Female 2D:4D of the right hand was positively associated with MnBP and negatively with total BPA (masculinizing). CONCLUSIONS: Significant associations were only observed for the long AGD metrics. Positive associations were observed between MnBP or LMW phthalates and APD in males. In females, prenatal MEP was associated with a masculinizing effect on ACD, while MBzP was associated with a feminizing effect. No significant associations were observed between prenatal phenols and AGD. Given the paucity of research on digit ratios and prenatal chemical exposures, it is difficult to say whether this metric will be a useful marker of prenatal androgen or anti-androgen exposure. Given the large number of associations examined, the statistical associations observed may have been due to Type 1 error. The inconsistencies in results between studies suggest that this issue is yet to be resolved.