中华创伤骨科杂志2024,Vol.26Issue(12) :1055-1061.DOI:10.3760/cma.j.cn115530-20240611-00248

老年膝关节骨关节炎患者发生髋部骨折的风险预测模型构建与验证研究

Establishment and validation of a prediction model for hip fracture in the aged patients with knee os-teoarthritis

林钲统 王号 齐瑞龙 徐国红 王利宏
中华创伤骨科杂志2024,Vol.26Issue(12) :1055-1061.DOI:10.3760/cma.j.cn115530-20240611-00248

老年膝关节骨关节炎患者发生髋部骨折的风险预测模型构建与验证研究

Establishment and validation of a prediction model for hip fracture in the aged patients with knee os-teoarthritis

林钲统 1王号 1齐瑞龙 1徐国红 1王利宏1
扫码查看

作者信息

  • 1. 温州医科大学附属东阳医院关节外科,东阳 322100
  • 折叠

摘要

目的 构建并验证老年膝关节骨关节炎(KOA)患者发生髋部骨折的风险预测模型.方法 回顾性分析2013年9月至2023年9月期间温州医科大学附属东阳医院诊断为KOA(Kellgren-Lawrence分级1~4级)的701例老年患者资料.男275例,女426例;年龄(76.5±8.4)岁.根据随访期间是否发生髋部骨折将患者分为两组:骨折组(n=145)和未骨折组(n=556).比较两组患者的年龄、性别、Kellgren-Lawrence分级、合并内科疾病情况及白蛋白值等,将P<0.05的项目纳入多因素logistic回归模型,分析老年KOA患者发生髋部骨折的危险因素,根据以上危险因素构建患者临床预测模型,并进行模型验证.结果 多因素logistic回归分析结果显示:女性(OR=2.009,95%CI:1.280~3.154,P=0.002)、年龄≥75 岁(OR=2.313,95%CI:1.493~3.583,P=0.001)、Kellgren-Lawrence 分级为 3~4 级(OR=2.348,95%CI:1.533~3.596,P=0.001)、白蛋白<35 g/L(OR=0.316,95%CI:O.191~0.522,P=0.001)和淋巴细胞绝对值<0.8 × 109/L(OR=0.133,95%CI:0.069~0.253,P=0.001)为老年KOA患者发生髋部骨折的危险因素.本次构建的髋部骨折风险预测模型在训练集和验证集中受试者工作特征曲线的曲线下面积分别为0.753和0.815(P<0.05).Hosmer-Lemeshow拟合优度检验结果显示:训练集与验证集模型预测风险与实际发生风险的一致性好(P<0.05).训练集和验证集中列线图的校准曲线均与理想曲线相近.临床决策曲线分析结果显示:该列线图模型具有良好的净获益率和预测能力.结论 女性、年龄≥75岁、Kellgren-Lawrence分级为3~4级、白蛋白<35 g/L和淋巴细胞绝对值<0.8 × 109/L是老年KOA患者发生髋部骨折的危险因素.根据风险因素构建的列线图预测模型具有良好的区分度和校准度,且有一定的临床预测能力和应用价值.

Abstract

Objective To develop and verify a predictive model for hip fracture risk in the aged patients with knee osteoarthritis(KOA)on the basis of analysis of the risk factors associated with the hip fracture.Methods A retrospective study was conducted to analyze the 701 patients who had been diag-nosed with KOA(Kellgren-Lawrence grades 1 to 4)at Dongyang Hospital affiliated to Wenzhou Medical Uni-versity from September 2013 to September 2023.The cohort consisted of 275 males and 426 females with an age of(76.5±8.4)years.The patients were divided into a fracture group(n=145)and a fracture-free group(n=556)based on whether a hip fracture occurred during the follow-up period.The 2 groups were compared in terms of age,gender,comorbidities,albumin level,absolute lymphocyte count,and Kellgren-Lawrence grade,etc.The items with P<0.05 were analyzed by a multivariate logistic regression model to identify the risk factors for hip fracture in the aged KOA patients.A clinical prediction model based on the above risk fac-tors was constructed and validated for hip fracture risk in the aged KOA patients.Results Multivariate lo-gistic regression analysis identified the following as independent risk factors for hip fracture in the aged KOA patients:female(OR=2.009,95%CI:1.280 to 3.154,P=0.002),age ≥75 years(OR=2.313,95%CI:1.493 to 3.583,P=0.001),Kellgren-Lawrence grades of 3-4(OR=2.348,95%CI:1.533 to 3.596,P=0.001),an albumin level<35 g/L(OR=0.316,95%CI:0.191 to 0.522,P=0.001),and an absolute lymphocyte count<0.8 × 109/L(OR=0.133,95%CI:0.069 to 0.253,P=0.001).The area under the ROC curve(AUC)for the model developed by this study was 0.753 in the training set and 0.815 in the validation set(P<0.05).The Hosmer-Lemeshow goodness-of-fit test showed that the consistency between the predicted risk and the actual risk was good in the training and validation sets(P<0.05).The calibra-tion curves for both the training and validation sets closely aligned with the ideal curve.The clinical deci-sion curve analysis showed that the nomogram model had a good net benefit rate and a good predictive po-tential.Conclusions Female,age ≥75 years,Kellgren-Lawrence grades of 3-4,an albumin level<35 g/L,and an absolute lymphocyte count<0.8 × 109/L are independent risk factors for hip fracture in the aged KOA patients.Since the nomogram prediction model based on these risk factors is satisfactory in discrimi-nation and calibration,it shows a certain predictive ability and application value in clinic.

关键词

髋骨折/骨关节炎/危险因素/列线图/预测模型

Key words

Hip fractures/Osteoarthritis/Risk factors/Nomogram/Prediction model

引用本文复制引用

出版年

2024
中华创伤骨科杂志
中华医学会

中华创伤骨科杂志

CSTPCDCSCD北大核心
影响因子:1.579
ISSN:1671-7600
段落导航相关论文