探讨骨钙素（bone glaprotein, BGP）及Ⅰ型前胶原N端前肽（procollagenⅠN-terminal peptide, PⅠNP）对骨质疏松性椎体压缩性骨折（osteoporotic vertebral compression fractures, OVCF）患者术后再骨折的预测价值。

选取2013-2016年于浙江医院骨科行经皮椎体后凸成形术的100例OVCF患者为研究对象，术后随访2年以上，以骨折再发为随访终点事件。收集患者的一般资料，包括性别、年龄、BMI等，以及初次骨折术前的骨密度T值、血钙值、BGP、PⅠNP、术中骨水泥注入量。采用t检验、秩和检验、Fisher确切概率法进行单因素分析，并将有意义的指标代入多元Cox回归模型进行骨折再发的多因素分析，并以ROC曲线评估BGP及PⅠNP对骨折再发的预测价值。

100例OVCF患者行椎体后凸成形术后随访3～37个月，其中33例（22.45%）术后再发骨折。将单因素分析有意义的指标（年龄、骨密度值、BGP、PⅠNP）代入多元Cox比例回归模型，结果显示BGP及PⅠNP仍是骨折再发的独立危险因素（HR=2.053，95%CI=1.128-3.734，P < 0.05；HR=1.250，95%CI=1.150-1.360，P < 0.01）。ROC曲线分析结果显示，BGP、PⅠNP及联合预测的曲线下面积分别为0.794、0.911、0.964；BGP预测截点值为0.474，对应灵敏度及特异度分别为72.7%、25.48%；PⅠNP预测截点值为0.685，对应灵敏度和特异度分别为0.879、0.194。

对于骨质疏松骨折术后患者，可通过定期监测血清BGP及PⅠNP水平评估骨代谢的变化，从而早期预测骨折再发风险并及时给予干预措施。

To explore the predictive value of bone glaprotein (BGP) and procollagen typeⅠN-terminal pro peptide (PⅠNP) in patients with osteoporotic vertebral compression fracture (OVCF) after operation.

100 OVCF patients who underwent percutaneous kyphoplasty (PKP) in Zhejiang Hospital from 2013 to 2016 were selected. The patients were followed up for more than 2 years after surgery, and the recurrence of fracture was the end point. The general information of the patients were collected, including gender, age, BMI, bone mineral density (BMD) T value, blood calcium value, BGP, PⅠNP, intraoperative bone cement injection. T test, rank sum test and Fisher exact rate method were used for single factor analysis, and meaningful indexes were substituted into multivariate Cox regression model for multiple factor analysis of fracture recurrence, and ROC curve was used to evaluate the predictive value of BGP and PⅠNP for fracture recurrence.

100 patients with OVCF were followed up for 3~37 months after PKP, and 33 patients (22.45%) of them had recurrent fractures. The significant indexes of single factor analysis (age, BMD, BGP, PⅠNP) were substituted into multiple Cox proportional regression model, and the results showed that BGP and PⅠNP were still independent risk factors of fracture recurrence (HR=2.053, 95%CI=1.128-3.734, P < 0.05; HR=1.250, 95%CI=1.150-1.360, P < 0.01). ROC curve analysis results showed that the area under the curve predicted by BGP, PⅠNP and the combination was 0.794, 0.911 and 0.964 respectively; the predicted cut-off point value of BGP is 0.474, corresponding sensitivity and specificity were 72.7% and 25.48% respectively; the predicted cut-off point value of PⅠNP was 0.685, corresponding sensitivity and specificity are 0.879 and 0.194 respectively.

For patients with osteoporotic fracture after operation, the changes of bone metabolism can be evaluated by regular monitoring of serum BGP and p1np levels, so as to predict the risk of fracture recurrence early and give timely intervention measures.