| 摘要: |
| 目的 基于Logistics回归分析建立产后压力性尿失禁(PSUI)的盆底四维超声诊断模型。方法 选取2018年7月至2020年8月于本院生产的236例产妇,均予以四维盆底超声动态检查。根据产妇产后3个月内是否出现PSUI将其分为PSUI组和非PSUI组。比较两组年龄、孕前体重指数、孕次、产次、孕周、分娩方式、阴道分娩者会阴侧切或裂伤、新生儿体重、盆底四维超声参数(BND、URA、PUVA、LHA)。采用多因素Logistic回归分析确定PSUI的影响因素。根据危险因素的回归系数构建盆底四维超声诊断模型,绘制ROC曲线分析盆底四维超声诊断模型对PSUI的诊断价值。结果 236例产妇中共38例出现PSUI,发生率16.10%。PSUI组和非PSUI组年龄、孕次、产次、孕周、会阴侧切情况的差异无统计学意义(P>0.05)。PSUI组孕前BMI、阴道分娩占比、新生儿体重、BUD、URA、LHA及Valsalva状态下PUVA明显高于非PSUI组(P<0.05)。多因素Logistic回归分析显示,BUD(OR=4.412,95%CI：1.794～10.850)、URA(OR=4.176,95%CI：1.809～9.640)、Valsalva状态LHA(OR=2.379,95%CI：1.217～4.650)、Valsalva状态PUVA(OR=2.841,95%CI：1.117～7.226)是PSUI的影响因素(P<0.05)。将多因素Logistic回归分析筛选出的影响因素中最小的标准化β值所对应的变量(即变量边界的β值0.867)赋分为1,其它变量则用β值除以0.867后四舍五入取整数,得到相应的整数分值后构建诊断模型,总分0～6分。模型评分诊断PSUI的ROC曲线下面积0.903(95% CI：0.780~0.953),以4.89分为截断值的敏感度和特异度分别为87.7%和78.8%。BND诊断PSUI的ROC曲线下面积0.843(95% CI：0.725~0.903),以12.2 mm为截断值的敏感度和特异度分别为85.6%和79.3%。结论 BND、URA、Valsalva状态下LHA及PUVA是PSUI的影响因素,基于上述因素构建的盆底四维超声诊断模型对PSUI的诊断价值优于单独BND。 |
| 关键词: Logistics回归分析 产后压力性尿失禁 盆底四维超声 诊断模型 |
| DOI： |
| 投稿时间：2021-01-14修订日期：2021-05-07 |
| 基金项目: |
|
| Establishment of four-dimensional ultrasound diagnosis model of pelvic floor for postpartum SUI based on Logistics regression |
| liqian |
| () |
| Abstract: |
| OBJECTIVE Based on Logistics regression analysis to establish a four-dimensional ultrasound diagnosis model of postpartum stress urinary incontinence (PSUI). METHODS A total of 236 parturients who gave birth in our hospital from July 2017 to August 2020 were selected, and all were examined by four-dimensional pelvic floor ultrasound. According to whether there was PSUI within 3 months after delivery, women were divided into PSUI group and non-PSUI group. The age, pre-pregnancy body mass index, gestation times, parity, gestational age, mode of delivery, vaginal delivery, perineal lateral incision or laceration, newborn weight, and four-dimensional ultrasound parameters of the pelvic floor (BND, URA, PUVA, LHA) were compared between the two groups. Multivariate Logistic regression analysis was used to determine the influencing factors of PSUI. According to the regression coefficients of risk factors, a four-dimensional ultrasound diagnosis model of the pelvic floor was constructed, and ROC curves were drawn to analyze the diagnostic value of the four-dimensional ultrasound diagnosis model of the pelvic floor for PSUI. RESULTS PSUI occurred in 38 of 236 parturients, the incidence was 16.10%. There was no significant difference between PSUI group and non-PSUI group in age, pregnancy times, parity, gestational age, and perineal lateral resection (P>0.05). The pre-pregnancy BMI, percentage of vaginal delivery, newborn weight, BUD, URA, LHA, and Valsalva PUVA in the PSUI group were significantly higher than those in the non-PSUI group (P<0.05). Multivariate logistic regression analysis showed that pre-pregnancy BMI (OR=3.112, 95%CI: 1.478～6.552), delivery method (OR=2.776, 95%CI: 1.252～6.155), newborn weight (OR=3.721, 95%CI: 1.259～10.997), BUD (OR=4.412, 95%CI: 1.794～10.850), URA (OR=4.176, 95%CI: 1.809～9.640), Valsalva state LHA (OR=2.379, 95%CI: 1.217～4.650) , Valsalva state PUVA (OR=2.841, 95%CI: 1.117～7.226) is the influencing factor of PSUI (P<0.05). The variable corresponding to the smallest standardized β value (ie, the β value of the variable boundary of 0.867) among the influencing factors screened out by the multivariate logistic regression analysis is assigned to 1, and the other variables are divided by the β value by 0.867 and rounded to the nearest integer. The diagnosis model is constructed after the corresponding integer scores, with a total score of 0-6. The area under the ROC curve of model score diagnosis of PSUI was 0.903 (95% CI: 0.780~0.953), and the sensitivity and specificity of the cut-off value of 4.89 were 87.7% and 78.8%, respectively. The area under the ROC curve of BND in the diagnosis of PSUI was 0.843 (95% CI: 0.725 to 0.903), and the sensitivity and specificity with a cut-off value of 12.2 mm were 85.6% and 79.3%. CONCLUSION LHA and PUVA in BND, URA, Valsalva state are the influencing factors of PSUI, and the four-dimensional ultrasound diagnosis model of pelvic floor based on the above factors is better than BND alone in the diagnosis of PSUI. |
| Key words: Logistics regression analysis postpartum stress urinary incontinence pelvic floor four-dimensional ultrasound diagnostic model |
