骨质疏松症患者腰椎骨折预测模型的构建

Construction of a prediction model for lumbar fracture in patients with osteoporosis

目的基于腰椎定量CT(QCT)骨密度值和骨代谢指标构建骨质疏松症(OP)患者发生腰椎骨折的预测模型。方法以骨质疏松性腰椎骨折患者52例(骨折组)和OP患者104例(未骨折组)为研究对象,分析骨折组腰椎QCT骨密度值与血清Ⅰ型原胶原C-端前肽(P1CP)、骨钙素N端中分子片段(N-MID)及β-胶联降解产物(β-CTX)指标间的相关性,Logistic回归分析患者腰椎骨折发生的影响因素,构建预测模型并分析其预测价值。结果与未骨折组比较,骨折组的女性、跌倒史占比高,β-CTX水平高,维生素D/钙剂用药史占比、腰椎QCT骨密度值、P1CP、N-MID水平低(P<0.05)。Pearson相关分析显示,骨折患者腰椎QCT骨密度值与血清P1CP、N-MID呈正相关,与β-CTX呈负相关(r分别为0.523、0.506、-0.536,P<0.05)。多因素Logistic回归分析显示,较高水平的腰椎QCT骨密度值、P1CP、N-MID及维生素D/钙剂用药史是腰椎骨折发生的保护因素,较高水平β-CTX是危险因素(P<0.05)。构建的预测模型预测腰椎骨折的敏感度为84.62%,特异度为97.12%,曲线下面积为0.926(95%CI:0.873~0.962)。预测模型的校准度良好,一致率为0.910。结论较高水平的腰椎QCT骨密度值、P1CP、N-MID及低水平β-CTX的OP患者腰椎骨折风险较低,使用维生素D及钙剂可有效预防OP患者,尤其女性腰椎骨折的发生。

Objective To construct a prediction model for lumbar fracture in patients with osteoporosis(OP)based on lumbar bone mineral density measured by quantitative computed tomography(QCT)and bone metabolism indicators.Methods A total of 52 patients with OP and lumbar fracture(the fracture group)and 104 patients with OP without lumbar fracture(the non-fracture group)were selected as study subjects.The correlation between lumbar bone mineral density measured by QCT and serum procollagen type I C-propeptide(P1CP),N-terminal mid-fragment of osteocalcin(N-MID)and beta collagen degradation products(β-CTX)was analyzed.Logistic regression analysis was conducted to identify influencing factors of lumbar fracture.A prediction model was constructed and its predictive value was analyzed.Results Compared with the non-fracture group,the proportions of females and patients with fall history,β-CTX level and vitamin D/calcium medication history were higher in the fracture group,and lumbar bone mineral density measured by lumbar QCT,P1CP and N-MID levels were lower(P<0.05).Pearson correlation analysis found that lumbar bone mineral density measured by QCT was positively correlated with serum levels of P1CP and N-MID,and negatively correlated withβ-CTX(r=0.523,0.506,-0.536,P<0.05).Multivariate Logistic regression analysis found that higher lumbar bone mineral density measured by QCT,P1CP,N-MID and history of taking vitamin D/calcium were protective factors for lumbar fracture,and igherβ-CTX level was a risk factor(P<0.05).The sensitivity and specificity of the prediction model constructed based on Logistic regression analysis results for predicting lumbar fracture were 84.62%and 97.12%.The area under the curve was 0.926(95%CI:0.873-0.962).The prediction model had good calibration,and the agreement rate was 0.910.Conclusion OP patients with higher levels of QCT,P1CP and N-MID,and lower level ofβ-CTX have lower risk of lumbar fracture.Vitamin D and calcium can effectively prevent the occurrence of lumbar fracture in OP patients,especially in women.