Construction of the prediction model of macrosomia in pregnant women with gestational diabetes mellitus

ObjectiveTo construct the prediction model of macrosomia in pregnant women with gestational diabetes mellitus.

MethodsOne hundred and twenty pregnant women with gestational diabetes mellitus were divided into the macrosomia group(n=33) and normal group(n=87) according to the different birth body mass of newborns.The general clinical data of two groups were collected, the logistic regression was used to construct the prediction model of macrosomia in pregnant women with gestational diabetes mellitus, and the receiver operating characteristic(ROC) curve of subjects was drawn.

ResultsThe differences of the weight gain during pregnancy, pre-pregnancy body mass index, gestational age, fasting blood glucose, OGTT 0 h blood glucose and insulin resistance index between two groups were statistically significant(P < 0.01).The results of logistic regression analysis showed that the weight gain during pregnancy, fasting blood glucose, insulin resistance index were the independent risk factors of the delivery of macrosomia(P < 0.05).The weight gain during pregnancy, fasting blood glucose and insulin resistance index were taken as covariates X₁, X₂ and X₃, respectively, and the prediction model expression was obtained as follows: Logit(P)=-10.522+0.521X₁+0.337X₂+0.216X₃.The optimal critical value, sensitivity, specificity and area under the ROC curve of model were 0.522, 73.68%, 93.90% and 0.856(0.614-0.977), respectively.

ConclusionsThe prediction model based on multiple factors has a high value in the prediction of macrosomia of pregnant women with gestational diabetes mellitus.