建立经皮椎体成形术后椎体再骨折风险的预测模型

Development of a Predictive Model for the Risk of Vertebral Refracture after Percutaneous Vertebroplasty

目的:探讨老年患者椎体成形术后椎体再骨折的危险因素,构建风险预测模型。方法:对2018年1月—2020年6月在马鞍山市人民医院接受经皮椎体成形术的370例患者进行回顾性分析。根据术后2年随访期内有无椎体骨折,将患者分为骨折组和对照组,采用Cox回归分析构建预测模型列线图。结果:370例患者(骨折组55例,对照组315例)纳入本研究。通过单因素及多因素Cox风险回归分析发现CT值、胸腰段、骨水泥的渗漏、抗骨质疏松治疗差异具有统计学意义(P<0.05)。列线图AUC值分别为0.827,说明模型具有良好的区分度。最优校准曲线显示预测值和实际值具有良好的一致性,决策曲线显示模型在整个阈值概率范围内具有较高的临床净获益。结论:患者在椎体成形术后,CT值、胸腰段、骨水泥的渗漏、抗骨质疏松治疗是术后新发椎体骨折的危险因素。建立预测模型,使脊柱科医师能采取有效预防治疗措施,防止并发症的发生。

Objective:To explore the risk factors of vertebral refracture in elderly patients after percutaneous vertebroplasty,and build a nomogram to provide reference for clinical prevention.Methods:A retrospective analysis was performed on 370 patients who underwent percutaneous vertebroplasty at Maanshan people’s hospital from January 2018 to June 2020.Patients were divided into a fracture group and a control group according to the presence or absence of vertebral fractures during the 2-year postoperative follow-up period,and Cox regression analysis was used to construct a predictive model column line graph.Results:370 patients(fracture group 55 patients,control group 315 patients)were included in this study.Univariate and multifactorial Cox analyses revealed statistically significant differences in CT values,thoracolumbar segment,leakage of bone cement,and anti-osteoporotic treatment(P<0.05).The AUC value was 0.827 respectively,indicating that the model was well differentiated.Optimal calibration curves showed good agreement between predicted and actual values,and decision curves showed that the model had a high net clinical benefit across the range of threshold probabilities.Conclusion:After vertebroplasty,HU values,thoracolumbar segment,cement leakage and anti-osteoporosis treatment are risk factors for new vertebral fractures after surgery.A predictive model was developed to enable spine surgeons to take effective preventive measures to prevent complications.