无自发性早产史单胎孕妇早产预测模型的构建  

Development of a nomogram prediction model for preterm birth in singleton pregnancies without history of spontaneous preterm birth

作　　者：汝萍 倪晓田 徐文怡 史玉霞 雷胜瑶 颜妍 苏秀娟[4] 顾颖 刘铭 刘云 RU Ping;NI Xiaotian;XU Wenyi;SHI Yuxia;LEI Shengyao;YAN Yan;SU Xiujuan;GU Ying;LIU Ming;LIU Yun(Department of Obstetrics,Shanghai East Hospital,School of Medicine,Tongji University,Shanghai 200123,China;Department of Obstetrics,Wuxi Maternal and Child Health Care Hospital,School of Medicine,Jiangnan University,Wuxi 214002,Jiangsu Province,China;Department of Obstetrics and Gynecology,Tongren Hospital,Shanghai Jiao Tong University School of Medicine,Shanghai 200336,China;Clinical Research Center,Obstetrics and Gynecology Hospital,School of Medicine,Tongji University,Shanghai 201204,China)

机构地区：[1]同济大学附属东方医院产科,上海200123 [2]江南大学附属无锡市妇幼保健院产科,江苏214002 [3]上海交通大学医学院附属同仁医院妇产科,上海200336 [4]同济大学附属妇产科医院临床研究中心,上海201204

出　　处：《同济大学学报（医学版）》2024年第6期884-890,共7页Journal of Tongji University(Medical Science)

基　　金：上海市浦东新区卫生健康委员会联合攻关项目(PW2021D-10)。

摘　　要：目的建立及评价无自发性早产(spontaneous preterm birth,sPTB)史单胎孕妇的早产预测模型。方法2021年1月—2021年12月在三家医疗中心分娩的无sPTB史单胎孕妇为研究对象,回顾性采集孕妇的临床特征、妊娠期并发症和妊娠结局,采用多元回归模型构建早产预测模型,以线列图形式展示。采用受试者工作特征(receiver operating characteristic,ROC)曲线下面积(area under the curve,AUC)及校准图等对预测模型进行评价。结果共纳入11371例无sPTB史单胎孕妇进行建模,识别出7个早产预测因子,包括孕前BMI、产次、辅助生殖技术妊娠、子宫颈手术史、子痫前期、未足月胎膜早破和妊娠期糖尿病,该模型的AUC为0.693(95%CI:0.6630.722),Hosmer-Lemeshow检验、二分类变量的决策曲线分析(decision curve analysis,DCA)均说明模型校准度良好。临床影响曲线(clinical impact curve,CIC)表明该模型临床预测有效率高。结论本研究尝试构建无sPTB史单胎孕妇的早产预测模型,构建的模型稳定性好,可作为无sPTB史单胎孕妇预测早产的一种工具,但其临床应用与推广需要进一步验证。Objective To develop a prediction model of preterm birth for singleton pregnancies without history of spontaneous preterm birth.Methods The clinical characteristics,pregnancy complications,and pregnancy outcomes of a cohort of singleton pregnancies with no history of spontaneous preterm birth who delivered at three tertiary hospitals from January 2021 to December 2021 were retrospectively analyzed.A multivariate regression model was used to construct a prediction nomogram model.The prediction model was evaluated using receiver operating characteristic(ROC)curve analysis and calibration plots.Results A total of 11371 singleton pregnancies without a history of spontaneous preterm birth were included in the study.Seven predictive factors were identified,including pre-pregnancy BMI,parity,assisted reproductive technology pregnancy,history of cervical surgery,preeclampsia,preterm premature rupture of membranes,and gestational diabetes.The area under the curve(AUC)of the model was 0.693(95%CI:0.6630.722).The Hosmer-Lemeshow test and decision curve analysis(DCA)for binary variables indicated good model calibration.The clinical impact curve(CIC)demonstrated high clinical prediction efficiency of the model.Conclusion A prediction model of preterm birth has been developed in this study,which may serve as a tool for predicting preterm birth in singleton pregnancies with no history of spontaneous preterm birth,however,further validation is needed before clinical application and generalization.

关 键 词：早产 无自发性早产史 单胎妊娠 预测模型 列线图 

分 类 号：R722[医药卫生—儿科]