基于多指标建立妊娠期糖尿病列线图模型

Development of a nomogram model for gestational diabetes mellitus based on multi-indexes

目的:建立基于多指标的妊娠期糖尿病(GDM)风险预测模型并予验证。方法:纳入2020年4月至2023年2月在南京医科大学附属南京医院(南京市第一医院)产检并分娩的484例孕妇,其中231例GDM孕妇,253例正常孕妇,按7∶3比例分为训练集与验证集。采用单因素分析两组数据的差异性,将有统计学意义的变量纳入多因素logistic回归模型,分析GDM相关危险因素。筛选出的危险因素使用R软件建立列线图预测模型,并以一致性指数(C-index)、标准校正曲线评价模型的效能,验证其准确性。结果:血红蛋白(Hb)、平均血小板体积(MPV)、血清肌酐、碱性磷酸酶(ALP)和空腹血糖是GDM的独立危险因素。基于筛选出的5项独立危险因素,建立GDM的预测模型。列线图模型预测效果较好,具有良好的精准度及区分度。C-index在训练集为0.829(95%CI为0.785~0.871),在验证集为0.817(95%CI为0.745~0.888)。校正曲线显示预测结果与实际结果的相关性良好。决策曲线(DCA)分析显示该列线图模型具有良好的临床实用性。结论:基于多指标构建的列线图,能较准确地预测GDM发生,可作为临床早期GDM的筛查及评估工具。

Objective:To develop and validate a nomogram to predict gestational diabetes mellitus(GDM)based on multi-indexes.Methods:A total of 484 pregnant women in our hospital from April 2020 to February 2023 were collected,including 231 individuals with GDM and 253 normal pregnant women.For nomogram construction and validation,all the pregnant women were divided into two cohorts(training and testing)at a 7∶3 ratio.Then we identified independent prognostic factors using univariate analysis and multivariate logistic regression models.Based on independent risk factors,a nomogram was established with R software.The predictive accuracy and discriminative ability were evaluated using the concordance index(C-index)and calibration curves.Results:Hemoglobin(Hb),mean platelet volume(MPV),serum creatinine(Cr),alkaline phosphatase(ALP)and blood glucose were independent risk factors for GDM.Finally,a nomogram was constructed by incorporating these five independent predictors.The C-indexes for gestational diabetes mellitus prediction in the nomogram were 0.829(95%CI:0.785~0.871)and 0.817(95%CI:0.745~0.888)for the training set and validation set,respectively.The calibration plot showed good agreement between the actual observations and the predicted outcomes both in the training set and validation set.Decision curve analysis(DCA)demonstrated that the nomogram was clinically useful.Conclusion:The nomogram based on multi-indexes could effectively predict gestational diabetes mellitus,which may provide a useful tool for screening and evaluating gestational diabetes mellitus.