基于盆底压力评估构建中重度阴道松弛症的预测模型

Constructing a predictive model for moderate to severe Vaginal Laxity syndrome based on pelvic flfloor pressure assessment

目的探讨发生中重度阴道松弛综合症(Vaginal Laxity syndrome,VLS)的危险因素并构建预测模型。方法收集郑州大学第三附属医院2021年5月至2023年5月期间盆底重建门诊及住院部患者的盆底压力评估数据(共948例),采用单因素分析及logistic回归分析发生中重度阴道松弛的影响因素,构建中重度VLS预测模型并进行内部验证。另外收集2023年6月至8月的数据(共143例)进行外部验证。结果纳入发生中度及重度阴道松弛综合征的患者共534例。其中年龄(OR=0.951,95%CI:0.935~0.967,P<0.001)、有一次或多次顺娩史的女性(OR=7.419,95%CI:4.719~11.692,P<0.001;OR=12.329,95%CI:7.324~20.754,P<0.001)、有一次或多次剖娩史的女性(OR=1.697,95%CI:1.083~2.659,P=0.021;OR=3.136,95%CI:1.738~5.659,P<0.001)、POP(OR=2.845,95%CI:2.029~3.987,P<0.001)、子宫切除术(OR=0.094,95%CI:0.025~0.352,P<0.001);后静息阶段平均值(OR=1.135,95%CI:1.085~1.219,P=0.002)与是否发生中重度VLS有关,基于以上因素构建了中度以上VLS的风险预测模型。建模组经内部验证的校正C index为0.729,建模组和验证组的ROC曲线的C index分别为0.788和0.807。两组的DCA曲线分析显示该模型在建模组中的阈值概率20%~90%,在验证组中从40%~100%受益,提示该模型具有良好的区分度和一致性,有较好的临床价值。结论本研究构建的中度及重度VLS发生的风险预测模型,可在一定程度指导临床实际工作,给予患者相应的诊疗建议,为VLS的预防提供了理论支持。

Objective To explore the risk factors for moderate to severe vaginal laxity syndrome(VLS)and build a prediction model.Methods The pelvic floor pressure assessment data of patients in the pelvic floor reconstruction outpatient and inpatient departments of the Third Affiliated Hospital of Zhengzhou University from May 2021 to May 2023 were collected(a total of 948 cases),and univariate analysis and logistic regression were used to analyze the occurrence of moderate to severe cases.Factors affffecting vaginal laxity,a prediction model for moderate to severe VLS was constructed and internally verifified.In addition,data from June to August 2023(143 cases in total)were collected for external verification.Results Moderate and severe vaginal relaxation syndrome occurred in 534 cases.Among them,women with age(OR=0.951,95%CI;0.935~0.967,P<0.001)and women with one or more vaginal births(OR=7.419,95%CI:4.719~11.692,P<0.001;OR=12.329,95%CI:7.324~20.754,P<0.001),women with one or more cesarean deliveries(OR=1.697,95%CI:1.083~2.659,P=0.021;OR=3.136,95%CI:1.738~5.659,P<0.001),POP(OR=2.845,95%CI:2.029~3.987,P<0.001),hysterectomy(OR=0.094,95%CI:0.025~0.352,P<0.001);the average value of the post-resting phase(OR=1.135,95%CI;1.085~1.219,P=0.002)is related to whether moderate to severe VLS occurs.Based on the above factors,a risk prediction model for moderate or above VLS was constructed,and the modeling team was internally verifified The corrected C index is 0.729,and the C index of the ROC curves of the modeled group and the modeled group are 0.788 and 0.807 respectively.DCA curve analysis of the two groups shows that the model has a threshold probability of 20%to 90%in the modeling group and benefifits from 40%to 100%in the validation group,suggesting that the model has good discrimination and consistency.It has good clinical value.Conclusion The risk prediction model for moderate and severe VLS constructed in this study can guide clinical practice to a certain extent,provide patients with corresponding diagnosis and treatment suggestions,and provide theoretical support for the prevention of VLS.