肩袖修补术后关节僵硬风险评估模型

A novel nomogram for predicting postoperative stiffness after arthroscopic rotator cuff repair

目的探讨肩袖修补术后关节僵硬的危险因素并构建风险评估模型。方法回顾性分析2016年6月至2022年12月于民航总医院因肩袖损伤接受肩袖修补术251例患者的病历资料,男111例、女140例。根据收治时间将纳入的患者分为建模组(2016年6月至2021年6月收治的病例,用于风险评估模型的构建)和验证组(2021年7月至2022年12月收治的病例,用于风险评估模型效能的评价)。建模组统计术后1年时关节僵硬的发生率,收集患者的年龄、性别、体质指数、病程、吸烟史、合并糖尿病、术前肩袖肌肉脂肪浸润程度、肩袖撕裂程度、缝合方式、术前并发僵硬、肩袖愈合情况、术后2和6周疼痛视觉模拟评分(visual analogue scale,VAS)、术后6周肩关节Constant-Murley评分及基于患侧肩关节CT三维重建测量所得的术前和术后第1天临界肩关节角(critical shoulder angle,CSA)、肩峰外侧角(lateral acromion angle,LAA)、肩峰指数(acromial index,AI)。采用单因素分析筛选术后发生关节僵硬的影响因素,将单因素分析中差异统计学意义的指标纳入logistics回归模型进行多因素分析,构建风险评估模型。通过验证组对影响因素再次评估,并对该模型的区分度和校准度进行评价。结果术后1年时建模组176例患者中21例发生关节僵硬。多因素logistic回归分析显示合并糖尿病、脂肪浸润程度、肩袖撕裂程度、术前并发僵硬、术后6周VAS评分偏高及术后6周Constant-Murley评分偏低为术后1年时发生关节僵硬的危险因素,以logistic回归模型为基础绘制列线图(nomogram图)。术后1年时验证组75例中11例发生关节僵硬,受试者工作特征曲线(receiver operating charac-teristic curve,ROC)的曲线下面积(area under curve,AUC)为0.926。拟合优度检验显示H-L检验(χ^(2)=2.215,P=0.947),表明该模型具有较好的校准度。结论合并糖尿病、肩袖肌肉脂肪浸润程度高、大型与巨大肩袖撕裂、术前并发僵硬、术后6周VAS评分偏高及术后6周Constant-Murley评分偏低为肩袖修补术后发生关节僵硬的危险因素,该风险评估模型有良好区分度和校准度,可为肩袖修补术后发生关节僵硬的风险评估提供参考。

ObjectiveTo investigate the risk factors for postoperative stiffness following rotator cuff repair and to develop a predictive risk assessment model.MethodsA retrospective analysis was conducted on 251 patients(111 males and 140 females)who underwent rotator cuff repair at the Department of Orthopedics,Civil Aviation General Hospital,from June 2016 to December 2022.Patients were divided into two groups based on the time of admission:the modeling group,comprising patients treated from June 2016 to June 2021,was used to construct the risk assessment model,while the validation group,including those treated from July 2021 to December 2022,was used to evaluate the model's effectiveness.In the modeling group,the incidence of postoperative stiffness one year after surgery was assessed.The study collected data on age,sex,body mass index,disease duration,smoking history,diabetes history,preoperative fat infiltration of the rotator cuff muscles,tear size,suturing technique,preoperative stiffness,re-tear rate,visual analogue scale(VAS)scores at two and six weeks postoperatively,Constant-Murley scores at six weeks postoperatively,and both preoperative and postoperative critical shoulder angle(CSA),acromial index(AI),and lateral acromion angle(LAA).Univariate analysis was used to identify potential risk factors for postoperative stiffness,followed by multivariate logistic regression to construct the risk assessment model.The validation group was used to reassess the identified risk factors.ResultsPostoperative stiffness occurred in 21 out of 176 patients in the modeling group.Logistic regression analysis revealed that diabetes history,higher fat infiltration of the rotator cuff muscles,larger tear size,preoperative stiffness,higher VAS score at six weeks postoperatively,and lower Constant-Murley score at six weeks postoperatively were significant risk factors for postoperative stiffness.Based on the logistic regression model,a nomogram was created using R software.In the validation group,postoperative stiffness was observed in 11 out of 75 patients.The area under the ROC curve(AUC=0.926)indicated good discriminative ability in predicting postoperative stiffness.The goodness-of-fit test(H-L test:χ^(2)=2.215,P=0.947)demonstrated moderate calibration of the model.ConclusionA history of diabetes,high fat infiltration of the rotator cuff muscles,large or massive rotator cuff tears,preoperative stiffness,higher VAS scores at six weeks postoperatively,and lower Constant-Murley scores at six weeks postoperatively are significant risk factors for postoperative stiffness after rotator cuff repair.The risk assessment model shows good discriminative power and calibration,making it a useful tool for predicting the risk of postoperative stiffness following rotator cuff repair.