基于列线图预测模型分析早期稽留流产的影响因素

Analysis of influencing factors of early missed abortion based on nomogram prediction model

摘要:

目的 基于列线图预测模型分析早期稽留流产的影响因素。方法 前瞻性选取2021年11月至2023年7月重庆市璧山区妇幼保健院2 550名孕妇作为建模人群，根据是否发生早期稽留流产分为早期稽留流产组和健康对照组。比较孕妇一般资料，以Logistic多因素回归分析早期稽留流产影响因素。采用R语言绘制预测早期稽留流产的列线图预测模型，以受试者工作特征曲线(ROC)评价列线图预测模型的预测效能。另随机选取2023年8月至2024年5月356名符合纳入、排除标准的孕妇为验证组，在验证人群中对列线图预测模型进行外部验证。结果 本研究共发放问卷及量表2 550份，有效回收2 537份，有效回收率为99.49%(2 537/2 550);早期稽留流产发生率为10.13%(257/2 537);年龄、妊娠压力量表(PPS)评分、合并慢性高血压、吸烟史、是否开展染色体检查、同型半胱氨酸(Hcy)水平、促甲状腺激素(TSH)水平、全氟辛酸(PFOA)水平、D-二聚体(D-D)水平是早期稽留流产的影响因素。内部、外部验证结果显示，列线图模型具有良好预测效能。结论 年龄、合并慢性高血压、吸烟史、妊娠压力、是否开展染色体检查、Hcy、TSH、PFOA、D-D是早期稽留流产发生的影响因素，基于影响因素构建早期稽留流产列线图预测模型具有一定应用价值，可指导临床制定针对性防治措施，预防早期稽留流产发生。

Abstract:

Objective Toanalyzetheinfluencingfactorsofearly missedabortionbasedonanomogram prediction model.Methods Aprospectivestudywasconductedtoselect2550pregnantwomenfromBishanMaternalandChildHealth CareHospitalinChongqingfrom November2021toJuly2023asthemodelingpopulation.Theparticipantsweredividedinto anearlymissedabortiongroupandahealthycontrolgroupaccordingtowhetherearlymissedabortionoccurred.Thegeneral informationofpregnantwomenwascompared.Andtheindependentriskfactorsforearlymissedabortionwereanalyzedby multivariatelogisticregressionanalysis.AnomogrampredictionmodelwasconstructedusingRlanguagesoftware,withits predictiveperformanceevaluatedbyreceiveroperatingcharacteristic (ROC)curveanalysis.Externalvalidationofthe predictionmodelwasperformedonthevalidationgroupof356eligiblepregnantwomenrandomlyselectedfrom August2023 toMay2024.Results Atotalof2,537validresponseswererecoveredamongthe2550distributedquestionnairesandscales, withaneffectiverecoveryrateof99.49% (2537/2550).Theincidenceofearlymissedabortionwas10.13% (257/2537). Maternalage,PregnancyPressureScale(PPS)score,co-existingchronichypertension,smokinghistory,andchromosome screening,aswellasserumlevelsofhomocysteine (Hcy),thyroid-stimulatinghormone (TSH),perfluorooctanoicacid (PFOA),andD-dimer(D-D)weretheinfluencingfactorsforearlymissedabortion(P<0.05).Thenomogramprediction modeldemonstratedrobustpredictiveperformanceinboththeinternalvalidationandtheexternalvalidation.Conclusion Maternal age,chronichypertension,smokinghistory,pregnancy-relatedpressure,chromosomescreening,Hcy,TSH,PFOA,D-D aretheinfluencingfactorsforearly missedabortion.Thenomogram prediction modeldeveloped based ontheabove influencingfactorsexhibitscertainclinicalutilityinguidingtargetedpreventiveinterventionstopreventtheoccurrenceofearly missedabortion.