个体化预测高龄腰椎间盘突出症患者显微内窥镜腰椎间盘切除术后深静脉血栓形成列线图模型的构建与验证

A nomogram model for individualized prediction of deep vein thrombosis in elderly patients with lumbar disc herniation after micro endoscopic discectomy surgery

目的构建个体化预测高龄(年龄≥80岁)腰椎间盘突出症患者显微内窥镜腰椎间盘切除术(micro endoscopic discectomy,MED)后发生深静脉血栓(deep vein thrombosis,DVT)的列线图模型,并验证该模型的准确性和临床适用性。方法选取2013年1月1日—2022年1月1日于西部战区空军医院行MED手术的912例高龄腰椎间盘突出症患者的临床资料,根据术后是否发生DVT,将其分为DVT组和非DVT组,采用单因素分析和多因素Logistic分析找出术后形成DVT的独立危险因素,基于独立危险因素利用R语言包构建列线图模型,利用ROC曲线、Bootstrap方法以及临床决策曲线验证该模型的准确性和临床决策的获益性。结果非DVT组患者834例,DVT组患者78例,多因素Logistic分析显示影响高龄腰椎间盘突出症患者MED术后DVT形成的独立危险因素为:D-二聚体较高、脑血管疾病病史、糖尿病病史和DVT病史。ROC曲线验证列线图模型显示:AUC=0.856,95%CI:0.72~0.93,上述研究结果显示构建的列线图模型预测高龄腰椎间盘突出症患者MED后发生DVT的预测能力较强,随后采用Bootstrap方法重复抽样1000次验证列线图,发现校准曲线的平均绝对误差为0.016,说明校准曲线与理想曲线贴合良好。临床决策曲线显示,列线图模型预测高龄腰椎间盘突出症患者MED术后发生DVT的发生阈值为0.03~0.93之间时该模型图的适用性最佳。结论影响高龄腰椎间盘突出症患者MED术后DVT形成的独立危险因素为:D-二聚体较高、脑血管疾病病史、糖尿病病史和DVT病史,基于上述危险因素构建的列线图模型准确度高,且具有较高的临床应用价值。

Objective To establish a nomogram model for predicting the occurrence of deep vein thrombosis(DVT)after micro endoscopic discectomy(MED)in elderly patients(ages 80 and older)with lumbar disc herniation,and to verify the accuracy and clinical applicability of the model.Methods The clinical data of 912 elderly patients with lumbar disc herniation who underwent MED surgery at our hospital between 2013 and 2022 was retrospectively analyzed.These patients were divided into a DVT group and a non-DVT group based on the postoperative occurrence of DVT.Univariate and multivariate logistic regression analysis was conducted to identify independent risk factors for the development of DVT after surgery.A nomogram model was constructed based on these independent risk factors using the R language package.The accuracy and clinical decision-making benefits of the model were validated using the ROC curve,the Bootstrap method,and clinical decision curves.Results In the non-DVT group,there were 834 patients,compared with 78 patients in the DVT group.Multivariate logistic regression analysis identified the following independent risk factors for DVT after MED surgery in elderly patients with lumbar disc herniation:elevated D-dimer levels,history of cerebrovascular disease,diabetes,and history of deep vein thrombosis.The ROC curve validation of the nomogram model showed an AUC of 0.856 with a 95%CI of(0.72~0.93),indicating a strong predictive capability of the constructed nomogram model for the occurrence of DVT after MED.Bootstrap resampling(1000 times)validated the nomogram,revealing an average absolute error of 0.016 in the calibration curve and suggesting a good fit with the ideal curve.Clinical decision curves showed that the nomogram model's applicability was optimal for predicting the occurrence of DVT after MED surgery in elderly patients with lumbar disc herniation when the decision threshold ranged from 0.03 to 0.93.Conclusion The independent risk factors affecting the formation of DVT after MED surgery in elderly patients with lumbar disc herniation include elevated D-dimer levels,history of cerebrovascular disease,diabetes,and history of deep vein thrombosis.The nomogram model established based on these risk factors is highly accurate and of high clinical value.