| Per-/polyfluoroalkyl Substances?PFASs?are a class of synthetic fluorinated compounds with the properties of surface activity,water/oil repellency,and thermal stability.PFASs were commonly used in commercial and consumer products,such as stain-resistant and non-stick coatings,food packaging,and fire-fighting foams.Humans are widely exposed to PFASs via contaminated food,drinking water and air and through dermal contact.Major manufacturers have voluntarily phased out perfluorooctane sulfonate?PFOS?production since 2000.Studies reported PFOS levels in Americans and Europeans had been decreseing in recent years.However,other PFASs,such as perfluorononanoic acid?PFNA?,are still widely and even increasingly detected in humans.Chlorinated polyfluorinated ether sulfonate?Cl-PFESA?,known by the trade name F-53B,is an alternative to PFOS that has been used in the Chinese electroplating industry for near 40 years.As PFOS is gradually phased out,F-53B will occupy a larger market in future.In China,about 10-14 tons of F-53B was released to the environment every year from 2006 to 2015,and the volum is increasing.The main component of F-53B,6:2 Chlorinated polyfluorinated ether sulfonate?6:2 Cl-PFESA?,has ranked as the top third PFASs in Chinese body burden after PFOA and PFOS in recent studies.The extensive and high exposure of F-53B is one of the characteristics of PFASs exposure in China,thus the health risk of F-53B deserves attention.Prior evidence has proven that exposure to PFOS and PFOA can disrupt endocrine system.However,there are few studies on the endocrine disruption of PFOS alternatives,and the results are inconsistent.Especially F-53B,which is widely used in China,has not been associated with endocrine system in epidemiological data to our knowledge.In addition,animal studies have proven that PFOS and PFOS showed developmental toxicity.Many epidemiological researches studied associations of PFOS and PFOS with adverse birth outcomes,but the results are inconsistent.Previous studies focused on PFOA and PFOS,few of them study on other long-chain PFASs.To our knowledge,no epidemiological data studied on the association of F-53B with fetal and infant growth.Therefore,studying the relationship of PFOS,PFOA and other PFASs with fetal and infant growth in China is of great significance.Taking advantage of a prospective birth cohort,we determined 12 PFASs levels in cord serum of 1048 newborns,and analyzed the associations between PFASs exposure and thyroids hormones in pregnancy,and three estrogens in newborns.We also repeatedly collected prenatal ultrasound data during pregnancy,and repeatedly measured the weight and height of newborns and infants aged 1-24 months.To explore the possible effects of PFASs on growth,we analyzed the associations between PFASs exposure and the growth of fetuses,newborns and infants.Thyroid hormone and estrogen play critical roles in the the growth of fetuses and infants.Given the background of the disrupting effects of PFOS and PFOA on thyroid hormone and estrogen,and the developmental toxicity of PFOS and PFOA,we raised a hypothesize:PFASs may affect the fetal and infant growth by interfering with thyroid hormone and estrogen.Thus,we conducted mediation analysis to analyze whether thyroid hormone and estrogen were intermediate variables among the associations between PFASs and fetal and infant growth.Part 1:Associations of per-/polyfluoroalkyl substances exposure and maternal thyroid function in early pregnancyObjective:We examined associations of cord serum PFAS with thyroid hormone levels in early pregnancy.Methods:We measured 12 PFASs in cord serum of 1048 newborns in the birth cohort,and obtained maternal thyroid hormones[Free Thyroxine?FT4?,Free Triiodothyronine?FT3?,Thyroid-stimulating hormone?TSH?]in their early pregnancy?5-16 gestational weeks?from the electronic medical records in the study hospital.After excluding 322women without thyroid hormones data during 5-16 weeks,726 pregnant women were included for further analysis.The basic charactoristics of women and children were obtained by face-to-face questionnaire and electronic medical records.We tested the linearity of the associations between PFASs and thyroid hormones via Restricted Cubic Spline?RCS?,and found most were linear(P-nonlinear>0.05),except relationships between PFUnDA and FT4,PFTriDA and FT4(P-nonlinear<0.05).Thus,we used multiple linear regression models to examine each PFAS and hormone.PFASs were analyzed as both continuous variables and categorical variables cutted by quartiles,with the first quartile as the reference.We controlled the following covariates in the models:age,pre-pregnancy BMI,parity,fetal sex,education status of pregnant women and gestational week of thyroid hormone test.Further,we assessed effect measure modification by fetal sex by stratification and interaction testing using PFASs×sex as covariate in the models.Finally,in order to avoid the bias caused by multiple comparisons,we corrected P values by False Discovery Rate?FDR?.Result:Among the twelve PFASs,nine of them were detected in more than 95%samples,indicating the pregnant women and newborns in Wuhan were widely exposed to those PFASs.As the main component of F-53B,the detection rate of 6:2 Cl-PFESA is 100%,and the cord serum concentration?median:0.76 ng/mL?ranked as top 3 in all PFASs,less than PFOS?median:4.06 ng/mL?and PFOA?median:1.60 ng/mL?,indicating that F-53B is a widely and highly exposed PFASs in newborns in Wuhan.In the model of PFASs as a continuous variable,we found that FT4 decreased1.19%?95%CI:-2.07%,-0.29%,P-FDR=0.04?for each interquartile range?IQR?level increase in PFOS.There was no significant association between other PFASs and thyroid hormones.In the categorized PFASs models,except for the significant relationship between PFOS and FT4,PFDA concentrations were inversely associated with the FT3 levels and TSH levels(FT3,quartile 4 vs.1:-3.12%,95%CI:5.42%,0.76%,P-FDR=0.03;TSH,quartile 3 vs.1:-21.85%,95%CI:-37.68%,-1.99%,P-FDR<0.01).Although the association between PFUnDA,PFTriDA and FT4 has a non-linear trend,their cofficients have not reached the level of statistical significance.The sex-stratified analyses indicated there was no fetal sex-specific associations between PFASs and thyroid hormones in pregnant women(P-interaction>0.05).Conclusions:In this study,prenatal exposure to some PFASs was inversely associated with maternal FT3,FT4 and TSH in early pregnancy.These results support the hypothesis that prenatal exposure to PFOS,PFDA influences thyroid function in pregnant women.Part 2:Associations of cord serum per-/polyfluoroalkyl substances and estrogen in neonatesObjective:To explore the associations of prenatal exposure to PFASs with estrogen in neonates.Methods:Among the 1048 neonates with PFASs measurement,981 of them had sufficient cord serum sample for estrogen determination.We detected estrone,estradiol,and estriol by high performance liquid chromatography tandem Xevo TQ-S triple quadruple-bar mass spectrometry.Because fetal and maternal health can affect umbilical cord blood estrogen,we excluded neonates with birth defects,preterm birth,pregnant women with gestational hypertension and diabetes mellitus,pregnant women who had smoking or drinking during pregnancy,finally,a total of 828 people were included in the analysis.RCS results indicated that there was no non-linear relationship between any PFASs and estrogen?P-nonlinear>0.05?.Therefore,we used multiple linear regression to analyze the associations between PFASs and estrogen,adjusted for the age of pregnant women,pre-pregnancy BMI,parity,passive smoking during pregnancy,maternal education status,neonatal sex,gestational weeks of delivery and delivery mode.To investigate whether there is a sex-different manner in the association between PFASs and estrogen in cord blood,we performed sex-stratified analysis,adjusted for same covariates except neonatal sex.Result:After adjustment of the possible confounders and FDR correction of P value,most of the PFASs were positively associated with at least one estrogen.Each IQR increase in the level of PFOA,PFNA,PFDA,PFUnDA,PFHxS,PFOS,PFOS,6:2 Cl-PFESA,and?PFASs,estrone increased by 4%-9%?P-FDR<0.05?;An IQR increase in PFOA,PFNA,PFDA,PFUnDA,PFTriDA,PFHxS,PFOS,6:2 CI-PFESA,and?PFASs,the level of estradiol increased 4%-10%?P-FDR<0.05?.As for the estriol,about4%-10%change in estriol associated with an IQR increase in PFOA,PFNA,PFDA,PFUnDA,PFHxS and 6:2 CI-PFESA levels?P-FDR<0.05?.In the sex-stratified analyses,only the association between PFOS and estrone was sex-specific(P-interaction<0.01),which was statistically significant in boys but not girls.Conclusions:The above results suggested that not only PFOS and PFOA,but also other PFASs,such as F-53B exposures may have impacts on the fetal synthesis and secretion of estrogens.Part 3:Associations of per-/polyfluoroalkyl substances with Fetal,Neonatal and Infant Growth:Potential Mediating Roles of Thyroids Hormones and EstrogenObjective:To investigate whether PFASs are associated with fetal,neonatal and infant growth,and explore the potential mediating roles of thyroids hormones and estrogen.Methods:We investigated whether PFASs exposure are associated with fetal,neonatal and early-childhood growth?1-24 months?.Because of the discrepancy in indicators and measure methods,the growth was divided into three periods:prenatal period?fetuses?,at birth?neonates?,and postnatal period?infants aged 1-24 months?.Fetal growth data were obtained by prenatal ultrasonic examination,including repeated measured biparietal diameter?BPD?,abdominal circumference?AC?,femoral length?FL?and estimated body weight?EBW?in the early,middle and late trimesters of pregnancy;anthropometry data at birth were collected from electronic medical records,including birth length,birth weight and ponderal index?NBPI?;postnatal growth data were obtained during follow-up physical examination including body length,weight and BMI at 1 month,12 months and 24 months of age.All fetal growth indicators and neonatal physical indicators were standardized using the generalized additive method for Location?Scale and Shape,GAMLSS?based on position,scale and shape parameters.Similarly,the weight,length and BMI of infants were standardized by Z-score conversion method issued by the World Health Organization?WHO?.Because the fetal and postnatal growth data are repeated measurements,mixed linear regression models were used to examine the relationship between PFASs and fetal and postnatal growth.Multiple linear regression was performed to analyze the associations between PFASs and neonatal physical indicators.In order to investigate the possible influence of PFASs on the long-term weight growth pattern of infants,the weight gain pattern of infants aged 1-24 months was divided into rapid weight gain and non-rapid weight gain patterns,which were included in logistic regression analysis as a binary dependent variable.Finally,based on the above results,PFHxS-estradiol-birth weight,PFOS-estradiol-infant weight,these two groups met the basic conditions of mediation analysis,and we performed the mediation analysis in these two groups using estrodial as mediating role.Result:After adjustment for potential confounders and FDR correction of P value,we found that only 8:2 CI-PFESA had a significant association with BPD of female fetuses.For every IQR increase in 8:2 CI-PFESA level,BPD SD score of female fetuses decreased by 15.45%(95%CI:-26.69%,-4.21%,P-FDR=0.03),but the sex-dieffence was not significant(P-interaction=0.43);An IQR increase in PFHxS was associated with57.16g?95%CI:-96.54g,-15.78g,P-FDR=0.01?decrease in the birth weight of female neonates,which was not observed in male newborns(P-interaction=0.04).For each IQR increase in PFOA level,length z score of infants decreased on average by 9.80%?95%CI:-15.85%,-3.75%,P-FDR=0.02?;additionally,for each IQR increase in PFOS level,the infants weight z score decreased on average by 8.81%?95%CI:-14.35%,-3.27%,P-FDR=0.02?and the infants length z score decreased by 8.55%?95%CI:-14.67%,-2.44%,P-FDR=0.02?.PFHxS and 6:2 Cl-PFESA were positively associated with the rapid weight gain pattern of boys aged 1-24 months?P-FDR<0.05?,but there was no significant gender difference(P-interaction>0.05).The results of mediation analysis indicated that PFHxS reduced the birth weight of female infants by 50.21g,of which 15.98g was reduced via the PFHxS-estradiol-birth weight pathway?indirect effect P<0.01?.The results suggested the indirect effect of PFHxS on birth weight through estradiol accounted for 32%of the total effect of PFHxS on birth weight;the indirect effect of PFOS on weight z score of infant aged 1 month through estradiol was-0.015?95%CI:-0.029,-0.010,P<0.01?,accounting for 25%of the total effect of PFOS on the weight z score of infant aged 1 month.Conclusions:We observed positive significant associations between 8:2 CI-PFESA and fetal BPD,PFHxS and birth weight in girls,PFOA,PFOS and infant growth;PFHxS,6:2 Cl-PFESA were positively associated with the rapid weight growth pattern of male infants.These results suggested that selected PFASs may have impacts on fetal and early-childhood growth.In addition,the results of mediation analysis suggested that the disrupting effect of PFHxS and PFOS on estrodial is an important pathway link PFASs and neonatal and infant weight.However,this study can only provide epidemiological evidences and clues,in vivo and in vitro studies are needed in the future to clarify the mechanisms. |
PDF Full Text Request