子宫全切术后盆底功能障碍性疾病发生危险因素Logistic回归分析及预测模型建立  

Logistic regression analysis and prediction model establishment of risk factors for PFD after total hysterectomy

作　　者：刘艳娟[1] 张斌[1] 王慧敏 李晶[1] 李英惠[2] 熊俊坤 Liu Yanjuan;Zhang Bin;Wang Huimin;Li Jing;Li Yinghui;Xiong Junkun(Department of Obstetrics and Gynecology,Baoding Third Central Hospital,Baoding Hebei 071051;Department of Obstetrics and Gynecology,Baoding First Central Hospital,Baoding Hebei 071000,P.R.China)

机构地区：[1]保定市第三中心医院妇产科,河北保定071051 [2]保定市第一中心医院妇产科,河北保定071000

出　　处：《中国计划生育和妇产科》2021年第8期52-55,共4页Chinese Journal of Family Planning & Gynecotokology

基　　金：保定市科学技术研究与发展指导计划项目(项目编号:18ZF144);河北省卫生和计划生育委员会科研基金项目(项目编号:20191228)。

摘　　要：目的探讨子宫全切术后盆底功能障碍性疾病(pelvic floor dysfunction,PFD)发生的危险因素并基于Logistic回归分析结果建立预测模型。方法选取2017年9月至2019年9月保定市第三中心医院妇产科收治的90例预行子宫全切术患者,根据术后PFD发生情况,将患者分为无PFD组和PFD组。对子宫全切术后PFD发生相关因素进行单因素和多因素Logistic回归分析,对筛选出的预测因子根据回归系数构建合适的预测模型评价系统,并绘制受试者工作特征曲线(receiver operating characteristic curve,ROC),评估其对预测子宫全切术后PFD发生的预测价值。结果子宫全切术后发生PFD 52例,占比57.78%,无PFD 38例,占比42.22%。两组年龄、超重或肥胖、孕次、产次、吸烟或饮酒比较,差异有统计学意义(P<0.05)。Logistic回归结果显示,年龄>45岁(OR=4.665,95%CI:1.411-15.419)、超重或肥胖(OR=8.415,95%CI:1.789-39.583)、孕次>2次(OR=3.525,95%CI:1.131-10.988)、产次>2次(OR=4.904,95%CI:1.399-17.191)、吸烟或饮酒(OR=7.846,95%CI:1.542-39.917)是子宫全切术后PFD发生的影响因素(OR>1,P<0.05)。根据以上因素建立危险评分系统,其中年龄>45岁、孕次>2次、产次>2次为1分,超重或肥胖、吸烟或饮酒为2分。评分系统对预测子宫全切术后PFD发生的ROC曲线下面积为0.848(95%CI:0.772-0.934),以8.94分为最佳临界值,此时的敏感度和特异度分别为85.3%和80.2%,Hosmer-Lemeshow拟合优度检验r=4.4,P=0.479。结论年龄>45岁、存在超重或肥胖、孕次>2次、产次>2次、存在吸烟或饮酒是子宫全切术后PFD发生的危险因素,评估预测模型对于子宫全切术后PFD发生具有较好预测价值。Objective To explore the risk factors of pelvic floor dysfunction(PFD)after total hysterectomy and establish a predictive model based on the results of Logistic regression analysis.Methods A total of 90 patients undergoing total hysterectomy admitted to Department of Obstetrics and Gynecology of Baoding Third Central Hospital from September 2017 to September 2019 were selected.According to the occurrence of postoperative PFD,the patients were divided into non-PFD group and PFD group.Performed single-factor and multi-factor Logistic regression analysis on the factors related to PFD after total hysterectomy,constructed an appropriate predictive model evaluation system based on the regression coefficients for the selected predictors,and drawed ROC curves to evaluate their predictive value in predicting the occurrence of PFD after total hysterectomy.Results PFD occurred in 52 cases after total hysterectomy,accounting for 57.78%,and 38 cases without PFD,accounting for 42.22%.There were significant differences between the two groups of age,overweight or obesity,pregnancy times,parity,smoking or drinking(P<0.05).Logistic regression results showed that age>45 years old(OR=4.665,95%CI:1.411-15.419),overweight or obese(OR=8.415,95%CI:1.789-39.583),pregnancy times>twice(OR=3.525,95%CI:1.131-10.988),parity>twice(OR=4.904,95%CI:1.399-17.191),smoking or drinking(OR=7.846,95%CI:1.542-39.917)after total hysterectomy were factors affecting the occurrence of PFD(OR>1,P<0.05).A risk scoring system was established based on the above factors.Among them,age>45 years old,pregnancy times>twice,parity>twice was 1 point,and overweight or obesity,smoking or drinking was 2 points.The area under the ROC curve of the scoring system to predict the occurrence of PFD after total hysterectomy was 0.848(95%CI:0.772-0.934).Taking 8.94 as the best critical value,the sensitivity and specificity at this time were 85.3%and 80.2%,respectively.The Hosmer-Lemeshow goodness of fit test r=4.4,P=0.479.Conclusion Age>45 years old,overweight or obesity,pregnan

关 键 词：子宫全切术 盆底功能障碍性疾病 危险因素 Logistic回归分析 预测模型 ROC曲线 

分 类 号：R713[医药卫生—妇产科学]