| Objective:Based on the maternal health registration system,we established a prospective cohort to explore the incidence of macrosomia in pregnant women,analyze its influencing factors,and establish a risk prediction model for macrosomia,thus providing scientific basis for improving newborn birth quality.Methods:Pregnant women registered in the early stage of pregnancy were recruited from June 2013 to December 2015 in a county of Hunan Province,and the pregnancy outcome of pregnant women was followed up and tracked by maternal and child health professionals.The data of socio-demography,pregnancy history,pre-pregnency exposure to environmental factors,family history for pregnent woman,husband’s situation,pregnancy complications and childbirth outcome were obtained through questionnaire survey,hospital visit and hospitalization records.The basic characteristics of pregnant women were determined by descriptive analysis.Single factor analysis of macrosomia influencing factors,continuous variables were analyzed by t-test or rank sum test,the enumeration data was analyzed by Chi-square test.Combing with the results of single factor analysis and professional knowledge to select variables,COX ratio risk model was used to Build prediction model for macrosomia.ROC curve was used to evaluate the fitting degree and forecasting efficiency of the prediction model.Results:(1)Incidence of macrosomia:among 9605 pregnant women with single term live birth,640 cases were macrosomia,accounting for 6.7%.(2)Univariate analysis:The age of pregnant women,pre-pregnancy BMI,history of birth,premature delivery,induced labor,and macrosomia,frequency of eating meat,fish,and shrimp,engaging in the production of fireworks,pollution-free environment around the residence,pregnancy anemia and gestational diabetes were associated with macrosomia.The age of pregnant women(χ~2=22.274,P<0.001),pre-pregnancy BMI(χ~2=63.344,P<0.001),production history(χ~2=8.151,P=0.004),history of preterm birth(χ~2=5.637,P=0.018),indu ced labor(χ~2=8.663,P=0.003),frequency of eating meat,fish,and shrimp(χ~2=7.432,P=0.024),macrosomia(χ~2=1442.571,P<0.001),production and processing of fireworks(χ~2=6.372,P=0.012),environmental pollution around residence(χ~2=5.647,P=0.017),anemia(χ~2=5.366,P=0.021)and gestational diabetes mellitus(χ~2=23.693,P<0.001).(3)COX regression analysis:male infants(HR=1.556,95%CI:1.279～1.893),maternal age(HR=1.051,95%CI:1.030～1.074),pre-pregnancy BMI(HR=1.101,95%CI:1.069～1.134),maternal height(HR=1.075,95%CI:1.053～1.098),history of macrosomia(HR=13.712,95%CI:10.622～17.700),gestational diabetes mellitus(HR=1.895,95%CI:1.161～3.092),and gestational week of termination(HR=1.349,95%CI:1.226～1.484)were associated with macrosomia.(4)Risk prediction of macrosomia:ROC curve analysis demonstrated that the optimal critical value of the prediction model is 0.0916,the sensitivity is 52.54%,the specificity is 82.19%,and the area under the ROC curve(AUC)is 0.748(0.720～0.769).The pregnancy outcomes of 2884 pregnant women were predicted,showing that 81 out of 184 macrosomia were predicted to be macrosomia,and 2276 out of2700 non-macrosomia were predicted to be non-macrosomia,with sensitivity of 44.00%,specificity of 84.30%and accuraty of 81.73%.Conclusions:(1)Due to the high incidence of macrosomia in singleton term live births,Pregnancy health care should be taken to prevent the occurrence of macrosomia.(2)Male infants,maternal age,pre-pregenency BMI,maternal height,history of macrosomia,gestational diabetes mellitus and longer gestational age at termination of pregnancy all increased the risk of macrosomia.(3)The prediction model of macrosomia risk established based on influencing factors has high specificity and ideal prediction efficiency,but with low sensitivity,thus the prediction model needs to be improved. |
PDF Full Text Request