髋关节置换术后股骨假体周围再次骨折的预测模型建立及手术治疗效果分析  

Construction of prediction model for periprosthetic femoral fractureafter hip arthroplasty and analysis of surgical effect

作　　者：姜瑞 侯志伟 JIANG Rui;HOU Zhi-wei(The Third Ward,Department of Orthopedics,Xigong Branch of Luoyang Hospital of TCM,Luoyang Branch of Dongzhimen Hospital Affiliated to Beijing University of Chinese Medicine,Luoyang,Henan 471000,China)

机构地区：[1]北京中医药大学东直门医院洛阳医院西工院区骨科三病区,河南洛阳471000

出　　处：《医药论坛杂志》2025年第3期269-274,共6页Journal of Medical Forum

摘　　要：目的Logistic回归分析影响髋关节置换术(hip arthroplasty,HA)后股骨假体周围骨折(periprosthetic femoral fractures,PFF)的相关因素,建立相关风险预测模型,分析HA术后发生PFF的手术治疗效果,为临床防治决策提供参考建议。方法选取2019年1月—2021年12月北京中医药大学东直门医院洛阳医院76例HA术后发生PFF的患者纳入PFF组,依据Vancouver分型为B2、B3型,均接受翻修骨折内固定+植骨手术治疗;另76例同期行HA术后未发生PFF的患者纳入非PFF组。单因素和logistic回归分析影响HA术后PFF发生的危险因素;建立HA术后PFF发生的预测模型,ROC评估预测模型的区分度,Hosmer-Lemeshow检验评估校正能力。比较HA术后PFF患者翻修骨折内固定+植骨手术治疗前和治疗后末次随访的患肢较健侧缩短长度、视觉模拟(VAS)评分、髋关节Harris评分和膝关节Hiss评分。结果HA术后PFF组与非PFF组在年龄、性别、病因、手术入路、假体固定方式和骨质疏松发生情况等方面差异有统计学意义(P<0.05);logistic回归分析显示,年龄、假体固定方式和骨质疏松均为影响HA术后发生PFF的危险因素(P<0.05);建立logistic回归风险预测模型Logit(P)=15.242×年龄-21.188×假体固定方式-0.176×骨质疏松-17.890。ROC曲线结果显示风险预测模型数据的AUC为0.789(95%CI:0.718~0.856),预测模型具有较好区分度,Hosmer-Lemeshowχ^(2)=7.169,P=0.923,模型具有较好校准度。HA术后PFF患者翻修骨折内固定+植骨手术后末次随访的患肢较健侧缩短长度明显短于术前,VAS评分、髋关节Harris评分和膝关节Hiss评分均低于术前,差异有统计学意义(P<0.05)。结论HA术后PFF预测模型Logit(P)=15.242×年龄-21.188×假体固定方式-0.176×骨质疏松-17.890,可通过简便数据收集预测HA术后PFF发生风险,HA术后Vancouver分型B2型、B3型PFF患者应用翻修骨折内固定+植骨治疗方案治疗效果良好。Objective To analyze influencing effects of periprosthetic femoral fractures(PFF)after hip arthroplasty(HA)by Logistic regression analysis,construct the related risk prediction model and analyze surgical effect in PFF after HA so as to provide reference for clinical prevention and treatment decision-making.Methods A total of 76 patients with PFF after HA in the hospital were enrolled as PFF group between January 2019 and December 2021.According to Vancouver classifications,there were types B2 and B3.All patients underwent revision fracture internal fixation and bone graft.A total of 76 patients without PFF after HA during the same period were enrolled as non-PFF group.The risk factors of PFF after HA were analyzed by univariate and logistic regression analysis.The prediction model for PFF was constructed,the discrimination of prediction model was evaluated by ROC curves,and its correction ability was evaluated by Hosmer-Lemeshow test.Before revision fracture internal fixation and bone graft and at the last follow-up,shortening length of affected limb,scores of visual analogue scale(VAS),hip Harris and knee Hiss were compared.Results After HA,there were differences in age,gender,etiology,surgical approaches,prosthesisfixation methodsand osteoporosis between PFF group and non-PFF group(P<0.05).Logistic regression analysis showed that age,prosthesis fixation methodsand osteoporosis were all risk factors of PFF after HA(P<0.05).Logistic regression risk prediction model was as follow:Logit(P)=15.242×age-21.188×prosthesis fixation method-0.176×osteoporosis-17.890.The results of ROC curves analysis showed that AUC of the risk prediction model was 0.789(95%CI:0.718-0.856),showing good discrimination,Hosmer-Lemeshowχ^(2)=7.169,P=0.923,and the model was of good calibration degree.After revision fracture internal fixation and bone graft,shortening length of affected limb was significantly shortened,scores of VAS,hip Harris and knee Hiss were decreased(P<0.05).Conclusion The prediction model for PFF after HA is as follow

关 键 词：髋关节置换术 假体周围再次骨折 危险因素 预测模型 手术效果 

分 类 号：R684[医药卫生—骨科学]