Anatomical Cystocele Recurrence Development and Internal Validation of a Prediction Model
Anatomical Cystocele Recurrence Development and Internal Validation of a Prediction Model
复制标题
DOI：
10.1097/aog.0000000000001272
复制
发表时间：
2016-02-01
影响因子：
7.2
通讯作者：
Weemhoff, Mirjam
中科院分区：
文献类型：
--
作者：
Vergeldt, Tineke F. M.;van Kuijk, Sander M. J.;Weemhoff, Mirjam
目的：建立一种预测解剖性膀胱膨出术后复发风险的预测模型。方法：合并两项多中心前瞻性队列研究的数据库，并对这些数据进行回顾性二次分析。未使用补片材料且未做过盆腔器官脱垂（POP）手术的前阴道破裂术患者填写问卷，行经唇三维超声检查，并进行术前和术后POP分期。我们使用多变量逻辑回归开发了一个预测模型，并使用标准自举技术对其进行了内部验证。通过综合性能、判别能力、校准等计算指标评价预测模型的性能，通过计算试验特征评价预测模型的临床实用性。结果：287例女性中，149例（51.9%）有解剖性膀胱膨出复发。预测模型的因素包括辅助分娩、术前膀胱膨出期、累及的腔室数、主要提肛肌缺损和Valsalva期间的提肛裂孔面积。由于高P值，在反向排除后排除的潜在预测因素包括年龄、体重指数、阴道分娩次数和POP家族史。自举过程产生的收缩系数为0.91。乐观校正后，Nagelkerke's R-2和Brier评分分别为0.15和0.22。这表明模型拟合令人满意。预测模型的受试者工作特征曲线下面积为71.6%（95%置信区间为65.7 ~ 77.5）。经乐观校正后，受试者工作特征曲线下面积为69.7%。结论：该预测模型包括辅助分娩史、术前分期、间室数、提上睑肌缺损和提上睑肌裂孔，可估计解剖性膀胱膨出复发的风险。
OBJECTIVE:To develop a prediction model that estimates the risk of anatomical cystocele recurrence after surgery.METHODS:The databases of two multicenter prospective cohort studies were combined, and we performed a retrospective secondary analysis of these data. Women undergoing an anterior colporrhaphy without mesh materials and without previous pelvic organ prolapse (POP) surgery filled in a questionnaire, underwent translabial three-dimensional ultrasonography, and underwent staging of POP preoperatively and postoperatively. We developed a prediction model using multivariable logistic regression and internally validated it using standard bootstrapping techniques. The performance of the prediction model was assessed by computing indices of overall performance, discriminative ability, calibration, and its clinical utility by computing test characteristics.RESULTS:Of 287 included women, 149 (51.9%) had anatomical cystocele recurrence. Factors included in the prediction model were assisted delivery, preoperative cystocele stage, number of compartments involved, major levator ani muscle defects, and levator hiatal area during Valsalva. Potential predictors that were excluded after backward elimination because of high P values were age, body mass index, number of vaginal deliveries, and family history of POP. The shrinkage factor resulting from the bootstrap procedure was 0.91. After correction for optimism, Nagelkerke's R-2 and the Brier score were 0.15 and 0.22, respectively. This indicates satisfactory model fit. The area under the receiver operating characteristic curve of the prediction model was 71.6% (95% confidence interval 65.7-77.5). After correction for optimism, the area under the receiver operating characteristic curve was 69.7%.CONCLUSION:This prediction model, including history of assisted delivery, preoperative stage, number of compartments, levator defects, and levator hiatus, estimates the risk of anatomical cystocele recurrence.