Trimester-specific, gender-specific, and low-dose effects associated with non-monotonic relationships of bisphenol A on estrone, 17β-estradiol and estriol

Background: Bisphenol A (BPA) may cause some adverse effects on human health by mimicking estrogen activities. In vitro and animal studies have observed the non-monotonic associations between BPA and natural estrogens, but the evidence in human study is lacking, particularly at multiple points in time during pregnancy. Objective: We aimed to examine the relationships between BPA and estrogens in the three trimesters among Chinese pregnant women and their gender variations. Methods: This study included 851 participants from a birth cohort conducted in Wuhan, China between 2014 and 2015. We measured concentrations of BPA and three estrogens (estrone (E1), 17β-estradiol (E2) and estriol (E3)) in urine samples collected in the three trimesters of pregnancy (mean for each visit: 13.0, 23.6, and 35.9 weeks’ gestation). We calculated the estimated daily intakes using urinary BPA concentrations and compared them with the tolerable intake value to assess potential health risks. We used multivariate linear regression models stratified by trimester and gender to explore trimester-specific and gender-specific associations of BPA with E1, E2, and E3. Results: We found the decreased levels of estrogens (β < 0, P < 0.05) in the upper BPA quartiles over three trimesters, except for the elevated levels of E3 (β = 0.20, 95% CI: 0.02, 0.38) in the highest BPA quartile in the 2nd trimester. There were significant non-linear associations (overall associations P < 0.05, non-linear associations P < 0.05) between BPA and E3 in the three trimesters. In the gender-stratified analysis, we observed significant negative relationships (β < 0, P < 0.05) between BPA and E2 among mothers carrying male fetuses in the 1st trimester and significant associations between BPA and E3 among mothers carrying female fetuses in the 2nd trimester. However, we found no significant relationship between BPA and E2 among mothers carrying female fetuses over three trimesters. Conclusions: Our findings support experimental evidence of non-monotonic relationships between BPA and three major estrogens, even at low doses of BPA. Mothers delivering male fetuses may be more sensitive to E2 at early pregnancy, and those delivering female fetuses may be more susceptive to E3 at mid-pregnancy.