Association Of Prenatal Phytoestrogen Exposures And Maternal Estrogen Metabolite Levels With Birth Outcomes

Part I Relationship between prenatal exposure tophytoestrogens and birth outcomesIn the present study, we investigated levels of phytoestrogen exposure in pregnantwomen and the association between phytoestrogen exposure and birth outcomes,including birth weight, length and gestational age. Study subjects (n=328) werepregnant women from the affiliated hospitals of Nanjing Medical University. Byusing UPLC-MS/MS, we measured the urinary levels of secoisolariciresinol (SEC),pinoresinol (PNR), matairesinol (MAT), enterodiol (END), enterolactone (ENT),genistin (GEN), dadzein (DAD), equol (EQU) and Naringenin (NAR) in each subject.After being adjusted by urinary creatine (CR), the geometric means (μg/g CR) ofthese compounds were12.6(SEC),5.39(PNR),3.11(MAT),24.1(END),56.18(ENT),110.35(GEN),115.18(DAD),9.27(EQU) and40.83(NAR), respectively.There were significantly positive association between gestational age and urinary MAT [adjusted=0.69week per log10unit increase;95%confidence interval (CI):0.22to1.16], END [adjusted=0.64week;95%CI:0.27to1.01] and ENT [adjusted=0.54week;95%CI:0.25to0.84]. Additionally, we also found that the metabolite ratio oflignans [(END+ENT)/(SEC+PNR+MAT+END+ENT)] was more correlated withincreased gestational age [adjusted=0.87week;95%CI:0.19to1.55] than any of theMAT, END and ENT. No associations for any of other phytoestrogen exposure werepresent for birth outcomes. These data suggested that MAT, END and ENT exposuresduring pregnancy might have protective association with length of gestation, and thisassociation was more pronounced among women with higher metabolite ratio oflignans. Part II Relationship between maternal levels of estrogenmetabolites and birth outcomesIn this study, we estimated the levels of endogenous estrogens and estrogenmetabolites (EMs) in pregnant women and investigated their association with birthoutcomes.328pregnant women were recruited from the affiliated hospitals ofNanjing Medical University. By using UPLC-MS/MS, we analyzed13estrogens andEMs: estrone (E1) and its2-,4-,16α-hydroxy derivatives (2-OHE1,4-OHE1,16α-OHE1) and its2-,4-methoxy derivatives (2-MeOE1,4-MeOE1), and2-hydroxyestrone-3-methyl ether (3-MeOE);17β-estradiol (E2) and its2-hydroxyderivative (2-OHE2) and2-methoxy derivative (2-MeOE2); and estriol (E3),16-epiestriol (16-epiE3) and17-epiestriol (17-epiE3). Linear regression models wereused to estimate regression coefficients and95%CIs after adjusting for potential confounding factors. E3and its precursor16α-OHE1, and their metabolites16-epiE3and17-epiE3showed obviously higher levels than E1, E2and other EMs. There weresignificantly positive associations of E3,16α-OHE1,16-epiE3and17-epiE3withgestational age. And these four EMs were also associated with increased birth weightand length after adjusting for gestational age. These data suggested that E3,16α-OHE1,16-epiE3and17-epiE3were the main EMs during pregnancy andshowed protective association with both gestation duration and fetal growth. Part III Relationship of the interaction between prenatalphytoestrogen exposures and maternal estrogen metabolitelevels with birth outcomesIn the present study, we investigated the relationship of the interaction betweenprenatal phytoestrogen exposures and major estrogen levels during pregnancy withbirth outcomes. Study subjects (n=328) were pregnant women from the affiliatedhospitals of Nanjing Medical University. We analyzed the urinary levels of9PEs(DAD, GEN, EQU, NAR, SEC, PNR, MAT, END and ENT) and4EMs (E3,16α-OHE1,16-epiE3and17-epiE3) in each subject by using UPLC-MS/MS.Through the Pearson Correlation analysis, we estimated the correlation betweenexposures to PEs and levels of the four EMs. And we found that END and ENT werepositively related to all of the four EMs and two of them, respectively. In the first partof our study, we found that END and ENT exposures were associated with increasedgestational age. It can be concluded that maternal exposures to END and ENT mightincrease endogenous estrogen levels during pregnancy, which should lead to their protective associations with gestational age. Additionally, in order to explore theeffects of the four EMs on the association between the PEs and birth outcomes, thesubjects were categorized as low (T1), middle (T2) and high (T3) groups according tothe tertile distribution of the four EMs, and the relationship between phytoestrogenexposures and birth outcomes were estimated for all the three tertile groups. After thestratified analyses, we found that MAT, END and ENT exposures were positivelyrelated to gestational age mainly in the T1, T2strata; while GEN and DAD exposureswere inversely associated with gestational age mainly in the T1strata. These resultsindicated that MAT, END and ENT might exert estrogenic effects when theendogenous estrogen levels were relatively low during pregnancy, which should leadto their protective associations with gestational age. On the other hand, the adverseassociations of GEN and DAD with gestational age might due to their antiestrogenicactivities, exerted when the endogenous estrogen levels were relatively low duringpregnancy. In summary, we found that the protective associations of some PEs withgestational age might due to their positive correlation with endogenous estrogenlevels during pregnancy. And the relationship between PEs and gestational age mightbe modified by levels of endogenous estrogen. Our findings, if confirmed in futurelarger studies, could have significant public health implications, as women with a lowestrogen levels during pregnancy could be specifically targeted for increasing lignansintake.