| Object:Abnormal pregnancy outcome is one of the major concerns in obstetrics all over the world.Previous epidemiological studies reported that maternal lifestyle,diet,infectious diseases et al were associated with abnormal birth weight.As the rapid development of industrialization,exposure of metals has been a topic problem to human health.Pregnant women and fetus are sensitivity to the external environment as the immunity of them is deficiency.Previous studies reported that maternal metal exposure was associated with abnormal pregnancy outcomes.Most of them only focused on single or several metals and neglected the effect of multiple metals on adverse pregnancy outcomes.Though lower level metals may have no obvious adverse effect on human health,the multiple metals exposure may have toxic effect on human health.As the development of gene chip technology,the report about genome-wide exon polymorphism analysis and abnormal pregnancy outcomes is increasing in recent years,especially for preterm delivery and gestational diabetes mellitus.However,the report about the relationship of maternal genome-wide exon polymorphisms and single full-term newborn birth weight is limited.Our study based on Guangxi Birth Cohort Study(GBCS),which is a multicenter prospective study.We aimed to investigate the association of prenatal exposure to multiple metals and abnormal pregnancy outcome and the association of maternal gene on newborn birth weight.Method:1.Maternal metals exposure and abnormal pregnancy outcome:(1)Study design:We used nested case-control study based on a prospective birth cohort study and included 5541 pregnant women who had antenatal examination in Guangxi eight hospitals from July in 2015 to June in 2016 and had delivery in these hospitals.After these exclusions(for abortion,stillbirth and cases in which infant gender information was missing),we selected 248 low birth weight(infant birth weight<2500 g)as cases.Controls were selected randomly at 1:2 allocation ratios according to maternal age,infant gender,gestational week at sample collection and the enrolment hospital.Finally,496 normal birth weight(infant birth weight≥ 2500 and<4000 g)as controls were selected from the 5,541 eligible participants.As the same matched method,we selected 131 macrosomia(infant birth weight≥4000 g)and 262 normal birth weight.,206 preterm delivery(delivery week<37 week)and 412 term delivery(delivery week≥37 week).As the sample deficiency,246 low birth weight and 409 normal birth weight,131 macrosomia and 260 normal birth weight,201 preterm delivery and 400 term delivery were included in our study.(2)Measurement of maternal serum metals concentration:We used inductive coupled plasma mass spectrometer to measure maternal 22 serum metal concentrations.(3)The method of statistical analysis:①Single metal model:We used logistic regression analysis to investigate the association of maternal serum metal concentration and the risk of abnormal pregnancy outcome,and used subgroup analysis and the restrict cubic regression to investigate the dose response relationship of serum metal concentration and abnormal birth weight.Metal concentrations were classification variables of quartile concentration in Logistic regression analysis and continuous variable in dose-response relationship.② Multi-metal model:We used elastic network regression to select metals which were associated with the risk of abnormal pregnancy outcomes and build multiple metals exposure model.2.Maternal genes and newborn birth weight:(1)Study design:According the principle of extreme phenotype,we selected the pregnant women who had completed the measurement of 22 serum metals and had full term single live infant.Birth weight was taken as the outcome variable.Singleton full-term newborns had the lowest weight was 103,according to the 1:1 matching of delivery hospital,included gestational age,maternal age and fetal sex,103 normal birth weight babies were screened.A genome-wide exon chip technique based on quantitative trait loci was used to analyze the correlation between maternal genetic polymorphism and newborn birth weight.(2)Measurement of maternal gene:We made genome-wide exon polymorphisms analyses in maternal peripheral blood(3)The method of statistical analysis:①We used liner regression to analyze the maternal gene and infant birth weight.②We used NCBI,Seattle Seq Annotation 151,Ensemble,Genome Brower 94,UniProt,DAVID,KEGG,et al to complete the annotation of gene which was relative with infant birth weight.Moreover,we did GO analysis and pathway analysis by David and KEGG websites.③ We used general linear regression to analyze the association of maternal Co,Ti and Mo and newborn birth weight.④We investigated the interaction effect of maternal serum metals and gene on newborn birth weight.Results:1.Maternal metals exposure and the risk of LBW:(1)The baseline character results of pregnant women and infants:Except pre-pregnancy BMI,newborn birth weight,body length,head and chest circumference had statistical difference during LBW and normal birth weight(P<0.05),other characters had no statistical difference(P>0.05).(2)The logistic regression analysis of maternal serum single metal and the risk of low birth weight:The results of logistic regression analysis showed that lower level serum Co were positively associated with LBW,lower level serum Ti and lower level serum Mo of pregnant women were negatively associated with the risk of low birth weight(Co:Q3 vs Q4’:Adjusted OR=1.80,95%CI:1.14-2.84,P trend=0.033;Ti:Q2 vs Q4:Adjusted OR=0.49,95%CI:0.31-0.78,P trend=0.040;Mo:Q2 vs Q4:Adjusted OR=0.47,95%CI:0.30-0.75,P trend=0.009).Subgroup analysis showed that,when sample collection>13 gestational week,serum Ti,Co and Mo of pregnant women were associated with low birth weight.When maternal age>28 years old,infant gender was girl,lower level serum Ti of pregnant women were negatively associated with low birth weight;when maternal age≤28 years old,infant gender was boy,lower level serum Co of pregnant women were positively associated with the risk of low birth weight.And all P value were<0.05.Lower level serum Mo were negatively associated with the risk of LBW when maternal age≤28 years old and>28 years old,infant gender was girl and boy.(3)The dose response relationship of maternal serum single metal and the risk of low birth weight:RCS results showed that serum Ti and Co of pregnant women have nonlinearity relationship with low birth weight when adjusted by pre-pregnancy BMI,maternal age and gestational age at sample selection(Ti:P for overall=0.015,P for nonlinearity=0.008;Co:P for overall=0.048,P for nonlinearity=0.014).(4)The interaction effect analysis of maternal serum metals and the risk of low birth weight:The interaction effect of metals on the risk of low birth weight:High Mo+Low Ti were negatively associated with the risk of low birth weight.When subgroup by gestational age at sample collection>13 weeks,maternal age>28 years old,infant gender was girl,the interaction effect of High Mo+Low Ti was negatively associated with the risk of low birth weight,respectively.(5)Multi-metal model and the risk of low birth weight:In the multi-metal model of 14 metals selected by elastic regression,lower level serum Ti,Mo and Zn were negatively associated with the risk of LBW(Ti:Q2 vs Q4:OR=0.44,95%CI:0.25-0.78.Mo:Q2 is Q4:OR=0.42,95%CI:0.25-0.72;Zn:Q2 vs Q4:OR=0.50,95%CI:0.29-0.87),and lower level serum Mg and Co were positively associated with the risk of LBW(Mg:Q3 vs Q4:OR=1.11,95%CI:1.15-3.86;Co:Q3 vs Q4:OR=1.52,95%CI:1.44-4.43)2.Maternal metals exposure and the risk of macrosomia:(1)The baseline character of pregnant women and infants:Except pre-pregnancy BMI,education level,parity,newborn birth weight,body length,head and chest circumference had statistical difference during macrosomia and normal birth weight(P<0.05),other characters had no statistical difference during two groups(P>0.05).(2)The Logistic regression analysis of maternal serum single metal and the risk of macrosomia:The results showed that unadjusted higher level serum Ca,higher level serum V and higher level serum Mo of pregnant women were positively associated with the risk of macrosomia(Ca:Q4 vs.QI:Unadjusted OR=2.13,95%CI:1.17-3.89,P trend=0.004;V:Q3 vs.Q1:Adjusted OR=2.95,95%CI:1.50-5.83,P trend=0.082;Mo:Q2 vs Q1:Adjusted OR=2.15,95%CI:1.04-4.46;Q3 vs Q1:Adjusted OR=2.10,95%CI:1.02-4.30,P trend=0.113),and lower level serum Ti of pregnant women was negatively associated with the risk of macrosomia(Q2 vs Q1:Adjusted OR=0.44,95%CI:0.20-0.97,P trend=0.113).The subgroup analysis showed that when maternal age≤28 years old,higher level serum V and higher level serum Mo of pregnant women was positively associated with the risk of macrosomia;when maternal age>28 years old,unadjusted high level serum Ca was positively associated with the risk of macrosomia and lower level serum Ti of pregnant women was negatively associated with the risk of macrosomia.And all P value were less than 0.05.(3)The dose response relationship of maternal serum single metal and the risk of macrosomia:Restrict cubic regression analysis showed that serum V had nonlinearity relationship with the risk of macrosomia(P for overall=0.014,P for nonlinearity=0.047),when serum V>0.70 μg/L,the risk of macrosomia was decrease.(4)The interaction effect analysis of maternal serum metals and the risk of macrosomia:The interaction effect on higher level seram Ca(>114.48 mg/L),Ti(>12.07μg/L),V(>0.65 μg/L)and Mo(>1.08 μg/L)were positively associated with the risk of macrosomia.(5)Multi-metal model and the risk of macrosomia:In the multi-metal model of 8 metals selected by elastic regression,higher level serum V(0.65-0.92μg/L)were negatively associated with the risk of macrosomia(OR=2.79,95%CI:1.29-6.01,P=0.009).3.Maternal metals exposure and the risk of preterm delivery:(1)The baseline character of pregnant women and infants:Except supplement Vitamin,newborn birth weight,body length,head and chest circumference had statistical difference during preterm delivery and term delivery(P<0.05),other characters had no statistical difference during two groups(P>0.05).(2)The Logistic regression analysis of maternal serum single metal and the risk of preterm delivery:The results showed that higher level serum Fe and higher level serum Rb of pregnant women were negatively associated the risk of preterm delivery(Fe:Q3 vs QI:Adjusted OR=0.57,95%CI:0.35-0.93;Q4 vs Ql:Adjusted OR=0.53,95%CI:0.32-0.86,P trend=0.004;Rb:Q4 vs Q1:Adjusted OR=0.36,95%CI:0.59-0.99,P trend=0.057);higher level serum AS was positively associated the risk of preterm delivery(Q2 vs Q1:Adjusted OR=1.62,95%CI:1.02-2.08,P trend=0.576).The subgroup analysis showed that when gestational age at sample collection>13 weeks,maternal age≤28 years old,infant gender was girl,higher level serum Fe was negatively associated with the risk of preterm delivery,respectively;when gestational age at sample collection>13 weeks,maternal age>28 years old,higher level serum Rb was negatively associated with the risk of preterm delivery,respectively;when gestational age at sample collection≤13 weeks,infant gender was boy,adjusted lower level serum AS was positively associated with the risk of preterm delivery respectively.And all P value were less than 0.05.(3)The dose response relationship of maternal serum single metal and the risk of preterm delivery:The restrict cubic analysis showed that serum Fe had linearity relationship with the risk of preterm delivery(P for overall=0.044,P for nonlinearity=0.249).As the increasing of serum Fe concentration,the risk of preterm delivery was decreasing.(4)The interaction effect analysis of maternal serum metals and the risk of preterm delivery:The interaction effect of metals on the risk of preterm delivery:High Fe+Low As,High Fe+Low Rb,High Fe+High Rb were negatively associated with the risk of preterm delivery,respectively.The subgroup analysis showed that when gestational age at sample collection>13 weeks,all of the above interaction effects were negatively associated with the risk of preterm delivery.When maternal age>28 years old,infant gender was boy,the interaction effect of High Fe+Low As were negatively associated with the risk of preterm delivery.When maternal age≤28 years old,infant gender was girl,the interaction effect of High Fe+Low Rb were negatively associated with the risk of preterm delivery.And all P value were less than 0.05.The interaction effect of High Fe+High Rb had no statistical significance when stratified by maternal age and infant gender,all P value>0.05.(5)Multi-metal model and the risk of preterm delivery:In the multi-metal model of 6 metals selected by elastic net regression,higher level serum Fe(1.45-1.78 mg/L)were negatively associated with the risk of preterm d.elivery(OR=0.49,95%CI:0.28,0.83,P=0.008);higher level serum Zn(>0.88)were positively associated with the risk of preterm delivery(OR=1.84,95%CI:1.03-3.29).4.Maternal gene and infant birth weight:(1)Baseline Characters:The average age of pregnant women in this study was 28.29±4.42 years old;the average Pre-pregnant BMI was 20.18±3.14 Kg/m2;the average new born birth weight was 2778.9±556.5 g,from 1800 g to 4000 g;the average length was 48.81±2.50 cm;the average head circumference was 31.78±1.93 cm and chest circumference was 31.35±2.28 cm.(2)The linear regression analysis of genome-wide association study showed that,there are 17 genes(17 SNPS)were associated with full term infant birth weight,all of which were associated with the protein binding.There are ATF6(rs10918270),CAVIN2(rs9789700),GALNT14(rs75732705),MATN3(rs6728440),PLCH1(rs13100610),PLAGL1(rs9373401),NCAPG2(rs78039900),INVS(rs34505550),GLIS3(rsl44576834),PARG(rs3844492),GRID1(rs186666792),TUB(rs143443020),XP04(rs71424096),NDFIP2(rs1417673),TRIP4(rs749504),ARHGAP44(rs2051975),ZNF407(rs373115608).And four of them were associated with the metal ion.There are MATN3,GLIS3,TRIP4 and ZNF407.We did pathway analysis by KEGG web site and found three major pathways related to single term infant birth weight:the phosphoinositol metabolism pathway(Gene:PLCH1;SNP:rs13100610),the endoplasmic reticulum protein synthesis pathway(Gene:ATF6;SNP:rs10918270),and the Wnt signaling pathway(Gene INVS;SNP:rs34505550).(3)Generalized linear regression analysis showed that Serum Co were positively associated with infant birth weight when adjusted by pre-pregnancy BMI,maternal age and gestational age of sample collection.Serum Mo and Ti were not associated with infant birth weight,all P value were more than 0.05.(4)The interaction effect of Maternal TUB gene(SNP:rs143443020)and serum Co were associated with infant birth weight(P interaction<0.05).Conclusion:(1)Serum Ti and Co of pregnant women have nonlinearity relationship with low birth weight when adjusted by pre-pregnancy BMI(Ti:P for overall=0.015,P for nonlinearity=0.008;Co:P for overall=0.048,P for nonlinearity=0.014).(2)Serum V had nonlinearity relationship with the risk of macrosomia(P for overall=0.014,P for nonlinearity=0.047),when serum V>0.70 μg/L,the risk of macrosomia was decrease.(3)Serum Fe had linearity relationship with the risk of preterm delivery(P for overall=0.044,P for nonlinearity=0.249).(4)In the multi-metal compound exposure model,after adjusted by other metals selected by elastic net regression and basic characters,lower level serum Ti,Co and Mo were associated with LBW;higher level serum V were negatively associated with macrosomia;higher level serum Fe were negatively associated with preterm delivery.(5)17 maternal genes(17 SNPS)and serum Co were associated with full term infant birth weight.There are three major signaling pathways associatedwith newborn birth weight:phosphoinositol metabolism,endoplasmic reticulum protein synthesis and Wit signaling pathway.The interaction effect of Maternal TUB gene(SNP:rs 143443020)and serum Co were associated with infant birth weight(P interaction<0.05). |
PDF Full Text Request