经皮椎体后凸成形术后骨水泥渗漏的相关危险因素分析

Analysis of risk factors causing cement leakages after percutaneous kyphoplasty for osteoporotic vertebral fractures

目的:探讨经皮椎体后凸成形术(percutaneous kyphoplasty,PKP)治疗骨质疏松性椎体骨折(osteoporotic vertebral body fracture,OVF)后骨水泥渗漏的相关危险因素。方法:回顾性搜集252例经PKP治疗且确诊为OVF的患者临床及影像学特征,包括年龄、性别等17个因子。利用单因素分析将上述与骨水泥渗漏有关的因子纳入多因素Logistic回归分析,并建立预测模型。应用受试者工作特征曲线的曲线下面积(area under curve,AUC)评估模型的诊断价值。结果:单因素分析结果显示,骨密度、骨水泥量、皮质缺损以及椎体内裂隙征与术后骨水泥渗漏有关,多因素Logistic回归分析示皮质缺损(P=0.001)、大剂量骨水泥量(P=0.047)和较低骨密度值(P=0.002)为术后骨水泥渗漏的独立预测因素,这三者建立的骨水泥渗漏预测模型,AUC为0.751,灵敏度为61.0%,特异度为82.5%。结论:PKP治疗OVF术后,椎体出现皮质缺损、大剂量骨水泥量和较低骨密度值为骨水泥渗漏的独立危险因素,其构成的联合模型具有较好的预测效能。

Objective:To evaluate risk factors for the occurrence of cement leakages after percutaneous kyphoplasty for osteoporotic vertebral fractures. Methods:A total of 252 vertebrae with percutaneous kyphoplasty were investigated. The following 17 parameters were evaluated:patient age,sex,and so on. The influence of possible risk factors affecting cement leakage was assessed using univariate analysis,and the parameters with statistical significance were incorporated into the multivariate logistic regression model,and the prediction model was established. The AUC value of ROC was used to evaluate the value of diagnostic tests. Results:The presence of intravertebral cleft,cortical disruption,large volume of cement and low bone mineral density were selected with statistical significance by univariate analysis for cement leakage in general. Multivariate logistic regression showed that the presence of cortical disruption(P=0.001),large volume of cement(P=0.047)and low bone mineral density(P=0.002)were independent predictors for cement leakage. The combined prediction model of cement leakage was established. The AUC value was 0.751,the sensitivity was61.0%,and the specificity was 82.5%. Conclusion:The presence of cortical disruption,large volume of cement and low bone mineral density of treated level are three independent risk factors for cement leakage after percutaneous kyphoplasty for osteoporotic vertebral fractures,and the combined prediction model has excellent prediction efficiency.