妊娠期糖尿病孕妇分娩巨大儿风险预测模型的构建

Construction of a risk prediction model for delivering macrosomic infants in pregnant women with gestational diabetes mellitus

目的探讨妊娠期糖尿病孕妇分娩巨大儿的影响因素,利用数据构建列线图模型并实施验证。方法选取陕西省人民医院2023年1月至2024年1月收治的妊娠期糖尿病孕妇1493例为研究对象,采用回顾性研究方法,分析妊娠期糖尿病孕妇分娩出现巨大儿的比例,对妊娠期糖尿病孕妇分娩巨大儿的相关影响因素进行单因素分析、多因素分析,建立妊娠期糖尿病孕妇分娩巨大儿的预测模型并分析其预测效能。结果在该次调查的妊娠期糖尿病孕妇1493例中,共有51例分娩巨大儿,占3.42%。单因素分析结果显示,孕妇妊娠期增重[(14.11±3.25)kg比(10.62±2.72)kg,t=8.94]、妊娠前身体质量指数[(24.31±2.51)kg/m^(2)比(23.25±2.13)kg/m^(2),t=8.94]、分娩孕周[(39.14±0.42)周比(38.92±0.51)周,t=3.04]、空腹血糖[(5.15±0.41)mmol/L比(4.75±0.35)mmol/L,t=7.97]、口服糖耐量试验0 h的血糖水平[(5.71±0.42)mmol/L比(5.49±0.41)mmol/L,t=3.76]、胰岛素抵抗指数[(0.54±0.13)比(0.41±0.10),t=9.02]等对分娩巨大儿有一定的影响(均P<0.05)。logistic多因素分析结果显示,妊娠期增重、空腹血糖、胰岛素抵抗指数是妊娠期糖尿病孕妇分娩巨大儿的独立影响因素(OR=1.685、27.113、25.816,均P<0.05)。经过验证,logistic多因素回归模型具有良好的拟合优度(Hosmer-Lemeshowχ^(2)=10.34,P>0.05)。依据多因素分析所筛选出来的因素构建列线图风险模型,C-index为0.742。利用logistic回归模型的独立影响因素及P值预测概率对妊娠期糖尿病孕妇分娩巨大儿的情况进行受试者工作特征曲线的预测,曲线下面积分别为0.815、0.779、0.795、0.938。结论以妊娠期增重、空腹血糖、胰岛素抵抗指数建立的综合预测模型对妊娠期糖尿病孕妇分娩巨大儿有显著的预测价值。

Objective To investigate the factors affecting the delivery of macrosomic infants in pregnant women with gestational diabetes mellitus,and utilize data to construct a nomogram model and validate it.Methods A total of 1493 pregnant women with gestational diabetes mellitus who received treatment at Shaanxi Provincial People's Hospital from January 2023 to January 2024 were included in this study.A retrospective study was conducted to analyze the proportion of pregnant women with gestational diabetes mellitus who delivered macrosomic infants.The influential factors in pregnant women with gestational diabetes who delivered macrosomic infants were analyzed using both univariate and multivariate analyses.A predictive model for macrosomia in pregnant women with gestational diabetes mellitus was established,and its predictive efficiency was evaluated.Results Among the 1493 pregnant women with gestational diabetes mellitus included in this survey,51 delivered macrosomic infants,accounting for 3.42%.Univariate analysis revealed that weight gain during pregnancy[(14.11±3.25)kg vs.(10.62±2.72)kg,t=8.94],pre-pregnancy body mass index[(24.31±2.51)kg/m^(2)vs.(23.25±2.13)kg/m^(2),t=8.94],gestational week[(39.14±0.42)weeks vs.(38.92±0.51)weeks,t=3.04],fasting blood glucose[(5.15±0.41)mmol/L vs.(4.75±0.35)mmol/L,t=7.97],blood glucose level during the oral glucose tolerance test at 0 hours[(5.71±0.42)mmol/L vs.(5.49±0.41)mmol/L,t=3.76],and insulin resistance index[(0.54±0.13)vs.(0.41±0.10),t=9.02]had a significant impact on the delivery of macrosomic infants(all P<0.05).Multivariate logistic analysis indicated that weight gain during pregnancy,fasting blood glucose,and insulin resistance index were independent risk factors for delivering macrosomic infants in pregnant women with gestational diabetes mellitus(OR=1.685,27.113,25.816,all P<0.05).The multivariate logistic regression model demonstrated good goodness of fit(Hosmer-Lemeshowχ^(2)=10.34,P>0.05).Based on the factors identified through multivariate analysis,a nomogram risk model was constructed,yielding a C-index of 0.742.The independent risk factors from the logistic regression model and their prediction probabilities were utilized to generate the receiver operating characteristic curve for predicting the likelihood of delivering macrosomic infants among pregnant women with gestational diabetes mellitus.The areas under the curve were 0.815,0.779,0.795,and 0.938,respectively.Conclusion The predictive model established based on weight gain during pregnancy,fasting blood glucose levels,and the insulin resistance index demonstrates significant predictive value for the delivery of macrosomic infants in pregnant women with gestational diabetes mellitus.