Analysis Of Influential Factors Related To Neonatal Birth Weight Based On Gauangxi Birth Cohort

Objective1.To investigate the epidemiologic feature of low birth weight(LBW)and macrosomia and establish birth cohort through surveying the pregnant women,who were permanent residents in Guangxi and considered to give birth in the local Maternal & Child Health Hospitals.2.To analyze the influential factors related to neonatal birth weight,and evaluate the effect from those factors by Logistic models,which provided theoretical basis for the early prediction and intervention of abnormal birth weight.MethodsThis study was based on an ongoing collaborate prospective of Guangxi Birth Cohort Study(GXBCS).Between Jul.2015 and Sept.2015,the pregnant women were recruited in eight Maternal & Child Health Hospitals in Guangxi Zhuang Autonomous Region.We collected the data of pregnant women and neonates.The detailed questionnaire collected information regarding demographic factors,reproductive history,personal habits,psychological factors,disease,pregnancy complications,medication,etc.Meanwhile,tracking the delivery information about infants’ gender and birth weight,etc.by followed-up.Until Jun.2016,after exclusions including missing birth outcome,multiple birth,unknown infant gender,abortion,stillbirth and birth defect,5251 single,live,full-term infant-mother pairs were selected for this study.Low birth weight and macrosomia were selected as research endpoints.Epidata 3.1 was used to record data.SPSS 21.0 was responsible for analyzing the data of pregnant women including demographic characteristics,customs,health conditions,history of pregnancy,history of diseases and medication history before pregnancy,etc.Firstly,single factor analyses were conducted by t-test,Chi-square test and Fisher exact test.Two-side alpha level of 0.05 was considered statistically significant for all statistical tests.The forecast models for low birth weight and macrosomia were established by Logistic regression analyses.Homer-Lemeshow goodness-of-fit test and Receiver Operating Characteristic Curve(ROC)were performed to evaluate the Logistic regression models.Results1.In the newborns,the average gestational age was 38.68±1.46weeks(28~42 weeks).The average birth weight was 3173.24±433.26 g,the min birth weight was 1000 g and the max birth weight was 4910 g.According to the category of birth weight,there were 234 infants with low birth weight(4.46%),127 infants diagnosed macrosomia(2.42%)and 4890 infants with normal birth weight(93.12%).2.Logistic regression model for prediction of low birth weight showed that the influential factors of LBW including(P<0.05): pre-pregnant BMI<18.5kg/m2(OR=1.482,95%CI:1.482-1.043),gestational age(OR=0.319,95%CI:0.286-0.570),first pregnancy(OR=1.622,95%CI:1.155-2.279),nephritis diagnosed before pregnancy(OR=4.914,95%CI:1.531-15.777),irregular menstruation(OR=1.485,95%CI:1.039-2.121)and smoking(OR=3.430,95%CI:1.166-10.086).3.Logistic regression model for prediction of macrosomia showed that the influential factors of macrosomia including(P < 0.05): gestational age of pregnant women between 30 and 34 years old(OR=1.591,95%CI:1.015-2.496),gestational age of pregnant women ≥35 years old(OR=2.089,95%CI:1.212-3.601),pre-pregnant BMI < 18.5 kg/m2(OR=0.283,95%CI:0.135-0.592),pre-pregnant BMI between 24.0 and 27.9 kg/m2(OR=1.738,95%CI:1.065-2.835),pre-pregnant BMI ≥28 kg/m2(OR=5.266,95%CI:2.795-9.922),maternal educational degree of junior high school level and below(OR=1.764,95%CI:1.113-2.797),gestational age(OR=1.773,95%CI:1.501-2.095),previous pregnancy with gestational diabetes mellitus(OR=6.889,95%CI:1.912-24.818).4.The logistic regression model for prediction of LBW was Logit P=39.496+0.393×pre-pregnant BMI(<18.5kg/m2)﹣1.142×gestational age+0.484×first pregnancy + 1.592×nephritis diagnosed before pregnancy +0.395×irregular menstruation + 1.233×smoking.P-value of Homer-Lemeshow goodness-of-fit test was 0.500.The result of AUC was 0.886(95%CI:0.860,0.911),the sensitivity and specificity of this prediction model was 71.4% and90.7%,respectively.5.The logistic regression model for prediction of macrosomia was Logit P= ﹣ 26.590 + 0.465×age of pregnant women(30 ~ 34 age)+ 0.737×age of pregnant women(≥35 age)﹣ 1.262×pre-pregnant BMI(< 18.5 kg/m2)+0.553×pre-pregnant BMI(24.0~27.9 kg/m2)+1.661×pre-pregnant BMI(≥28.0kg/m2)-1.142×gestational age + 0.568×maternal educational degree of junior high school level and below﹣1.022×thalassemia diagnosed before pregnancy+1.93×previous pregnancy with gestational diabetes mellitus.P-value of Homer-Lemeshow goodness-of-fit test was 0.408.The result of AUC was 0.775(95%CI:0.735,0.815),the sensitivity and specificity of this prediction model was 73.5% and 81.5%,respectively.ConclusionsBased on Guangxi birth cohort,the analysis of influential factors related to neonatal birth weight indicated that pregnant women with low pre-pregnant BMI,first pregnancy,nephritis diagnosed before pregnancy,irregular menstruation and smoking were more likely to give birth to LBW infants,and pregnant women with high pre-pregnant BMI,advanced gestational age,low maternal educational level,and previous pregnancy with gestational diabetes mellitus were more likely to produce macrosomia.The prediction model for low birth weight was build by Logistic regression,which can predict the incident of low birth weight and offer guidance for screening and health care before and during early pregnancy.In order to provide theoretical basis for reduce the incident of LBW in the future.