老年髋部骨折术后二次骨折风险预测模型的构建与验证  

Construction and validation of a nomogram prediction model for secondary fracture risk after hip fracture in elderly

作　　者：王礼宁 李新艳 朱弈桦 王远智 涂鹏程 郭杨[2] 沈计荣 马勇[1,2] 张春雷 WANG Li-ning;LI Xin-yan;ZHU Yi-hua;WANG Yuan-zhi;TU Peng-cheng;GUO Yang;SHEN Ji-rong;MA Yong;ZHANG Chun-lei(School of Integrated Medicine,Nanjing University of Chinese Medicine,Nanjing 210023,China;Laboratory of New Techniques of Restoration and Reconstruction of Orthopedics and Traumatology,Nanjing University of Chinese Medicine,Nanjing 210023,China;Department of Orthopedic Surgery,Nanjing Hospital of Chinese Medicine Affiliated to Nanjing University of Chinese Medicine,Nanjing 210022,China;Jiangsu Health Development Research Center/NHC Contraceptives Adverse Reaction Surveillance Center,Nanjing 210036,China;Department of Orthopedic Surgery,Affiliated Hospital of Nanjing University of Chinese Medicine,Nanjing 210004,China)

机构地区：[1]南京中医药大学中西医结合学院,南京210023 [2]南京中医药大学骨伤修复与重建新技术实验室,南京210023 [3]南京中医药大学附属南京中医院骨伤科,南京210022 [4]江苏省卫生健康发展研究中心/国家卫生健康委计划生育药具不良反应监测中心,南京210036 [5]南京中医药大学附属医院骨伤科,南京210004

出　　处：《中华骨质疏松和骨矿盐疾病杂志》2025年第1期49-59,共11页Chinese Journal Of Osteoporosis And Bone Mineral Research

基　　金：国家自然科学基金面上项目(82074458);江苏省自然科学基金面上项目(BK20221351);江苏省自然科学基金青年项目(BK20220470);2021年度江苏省卫生发展研究中心开放课题(JSHD2021026)。

摘　　要：目的构建老年髋部骨折患者术后发生二次骨折风险预测模型。方法采用回顾性研究纳入2018年1月至2023年2月在南京中医药大学附属医院骨伤科就诊年龄≥60岁的髋部骨折术后患者的临床数据,使用单因素Logistic回归、多因素Logistic回归筛选最佳预测变量,对预测变量进行多重共线性检验,基于多因素Logistic分析结果利用R软件构建术后二次骨折列线图预测模型。采用受试者工作特征曲线(receiver operating characteristic curve,ROC)评估模型的区分度,建立校准曲线评估模型的校准度,绘制临床决策曲线评价模型的临床实用性。并采用内部验证Boostrap和十折交叉验证法评估模型的可重复性。结果本研究共纳入629例老年髋部骨折术后患者,其中初次骨折576例,二次骨折53例,二次骨折发生率为8.4%。通过单因素及多因素Logistic分析筛选出5个危险因素:Harris评分(OR=0.916,95%CI:0.872~0.961)、日照时间(OR=0.397,95%CI:0.217~0.728)、功能锻炼(OR=0.400,95%CI:0.215~0.744)、阿尔茨海默病(OR=3.069,95%CI:1.034~9.110)、住院时长(OR=1.033,95%CI:1.004~1.063)。基于回归结果利用R软件构建髋部骨折术后二次骨折列线图。预测模型得出C指数与ROC曲线下面积为0.778,表明该模型有良好的区分度。校准曲线与理想曲线重复度较高,表明该模型校准度较好,预测二次骨折的能力与实际误差小,一致性高。临床决策曲线结果显示预测模型设定风险阈值为5%~55%时净获益大于0。结论本研究初步构建老年髋部骨折术后二次骨折的风险预测模型,该模型区分度较高、一致性良好,有一定的临床应用潜能。Objective To construct a nomogram prediction model for the risk of the secondary fracture in elderly patients with hip fracture after surgery.Methods This retrospective study included the clinical data of patients aged≥60 years who underwent hip fracture surgery in the Department of Orthopedics,Affiliated Hospital of Nanjing University of Chinese Medicine from January 2018 to February 2023.Univariate Logistic regression and multivariate Logistic regression were used to screen the best predictive variables,and multiple collinearity test was performed on the predictive variables.Based on the results of multivariate Logistic analysis,R software was used to construct a nomogram prediction model for postoperative secondary fractures.Receiver operating characteristic(ROC)curve was used to evaluate the discrimination of the model,a calibration curve was established to evaluate the calibration of the model,and a clinical decision curve was drawn to evaluate the clinical practicability of the model.Internal validation Boostrap and ten-fold cross validation were used to evaluate the reproducibility of the model.Results A total of 629 elderly patients after hip fracture were included in this study,including 576 primary fractures and 53 secondary fractures,with the incidence of secondary fracture as 8.4%.Five risk factors were screened by univariate and multivariate Logistic analysis as Harris score(OR=0.916,95%CI:0.872-0.961),sunshine time(OR=0.397,95%CI:0.217-0.728),functional exercise(OR=0.400,95%CI:0.215-0.744),Alzheimer s disease(OR=3.069,95%CI:1.034-9.110),and length of hospital stay(OR=1.033,95%CI:1.004-1.063).Based on the results of multivariate Logistic regression,R software was used to construct the nomogram for secondary fracture risk after hip fracture surgery.The area under the C index and ROC curve obtained by the prediction model was 0.778,indicating that the model had good discrimination.The repeatability between the calibration curve and the ideal curve was high,indicating that the model had good calibrati

关 键 词：髋部骨折 术后 二次骨折 LOGISTIC回归 列线图 

分 类 号：R687.3[医药卫生—骨科学]