用于类风湿性关节炎患者脆性骨折发生风险列线图预测模型的建立与验证

Development and Validation of a Nomogram Prediction Model for the Risk of Fragility Fracture in Patients with Rheumatoid Arthritis

目的建立预测类风湿性关节炎(RA)患者发生脆性骨折的列线图模型并对其进行验证。方法回顾性分析2019年1月至2022年1月我院收治的205例RA患者的临床资料。根据是否发生脆性骨折分为RA合并骨折组(n=69)和RA组(n=136),统计分析两组患者的一般资料、实验室指标及L1椎体HU值。利用单因素、多因素Logistic回归分析相关危险因素,使用R语言软件建立预测脆性骨折发生风险的列线图模型,采用Bootstrap法进行验证。以ROC曲线下面积大小评估模型的预测效能,运用校准曲线、决策曲线分析(DCA)评估模型的准确度及实用性。结果多因素Logistic回归分析表明:钙离子水平、C-反应蛋白、体质指数(BMI)、糖皮质激素使用>3月和HU值是脆性骨折的独立危险因素(P<0.05)。列线图显示出较好的辨别力,训练集ROC曲线下面积为0.944(95%CI 0.901~0.987),灵敏度为85.7%,特异度为92.6%。验证集ROC曲线下面积为0.939(95%CI 0.895~0.984),灵敏度为85.3%,特异度为91.2%。校准曲线结果表明RA患者实际发生脆性骨折概率与预测概率之间有良好的一致性。结论RA患者发生脆性骨折的列线图模型预测能力和区分能力较好,对RA患者脆性骨折的发生具有较高的预测效能,有一定的临床转化应用价值。

Objective The purpose of this paper is to develop and validate a nomogram model for predicting the risk of fragility fracture in patients with rheumatoid arthritis(RA).Methods The clinical data of 205 patients with RA admitted to our hospital from January 2019 to January 2022 were retrospectively analyzed.The patients were divided into RA combined with the fragility fracture group(n=69)and RA group(n=136)according to whether a fragility fracture occurred,and the general data,laboratory indexes,and L1 vertebral HU values of the two groups were statistically analyzed.The relevant risk factors were analyzed using univariate and multifactor Logistic regression,and the nomogram model for predicting the risk of fragility fracture was established using R language software and validated by the Bootstrap method.The area under the ROC curve is used to assess the predictive efficacy of the model,and the accuracy and usefulness of the model are assessed using calibration curves and decision analysis curves(DCA).Results Multifactorial Logistic regression analysis showed that calcium levels,C-reactive protein,body mass index(BMI),glucocorticoid use>3 months,and HU values were independent risk factors for fragility fracture(P<0.05).The nomogram showed good discrimination,with an area under the ROC curve of 0.944(95%CI 0.901~0.987)for the training set,with a sensitivity of 85.7%and a specificity of 92.6%.The area under the validation set ROC curve was 0.939(95%CI 0.895~0.984),with a sensitivity of 85.3%and a specificity of 91.2%.The calibration curve results showed good agreement between the actual probability of fragility fracture and the predicted probability in RA patients.Conclusion The nomogram model of the occurrence of fragility fracture in RA patients has a good predictive,discriminatory ability and high predictive value for the occurrence of fragility fracture in RA,and has certain clinical transformation application value.