胸椎黄韧带骨化症术后脑脊液漏危险因素分析及列线图风险预测模型构建  

Analysis of risk factors for cerebrospinal fluid leakage after thoracic ossification of ligamentum flavum surgery and construction of nomogram risk prediction model

作　　者：罗迪 张潇 顾文波 马安宁 吴继德 袁海峰[1] LUO Di;ZHANG Xiao;GU Wenbo;MA Anning;WU Jide;YUAN Haifeng(Spine Surgery,General Hospital,Ningxia Medical University,Yinchuan 750004,China;School of Clinical Medicine,Ningxia Medical University,Yinchuan 750004,China;Wuwei People′s Hospital,Wuwei 733000,China)

机构地区：[1]宁夏医科大学总医院椎柱外科,宁夏银川750004 [2]宁夏医科大学临床医学院,宁夏银川750004 [3]甘肃省武威市人民医院,甘肃武威733000

出　　处：《宁夏医学杂志》2024年第10期874-878,F0003,共6页Ningxia Medical Journal

基　　金：宁夏回族自治区中央引导地方科技发展专项项目(2022FRD05038);宁夏自然科学基金项目(2023AA03543)。

摘　　要：目的分析胸椎黄韧带骨化症(TOLF)患者发生术后脑脊液漏(CSFL)的相关危险因素,构建列线图风险预测模型。方法回顾性分析行手术治疗的111例TOLF患者临床资料,根据术后是否发生脑脊液漏,将其分为脑脊液漏组(n=42)和非脑脊液漏组(n=69),采用LASSO回归及多因素logistic回归分析对其危险因素进行筛选,并构建列线图风险预测模型。结果111例纳入研究的患者中,42例术后出现脑脊液漏,发生率为37.8%。LASSO回归与多因素logistic回归分析结果表明,椎管面积残余率小(OR=0.958,P<0.05)、手术节段数多(OR=2.940,P<0.05)、硬脊膜粘连(OR=9.173,P<0.05)、硬脊膜骨化(OR=69.468,P<0.05)是术后脑脊液漏的独立危险因素。预测模型ROC曲线下面积(AUC)为0.923(95%CI:0.873,0.973),模型的校准曲线与理想曲线接近,决策曲线(DCA)分析显示当该模型预测的概率阈值>5%时,该列线图模型的临床应用价值较高。结论椎管面积残余率低、手术节段数多、硬脊膜粘连、硬脊膜骨化是术后脑脊液漏的独立危险因素,基于此构建的列线图模型可以较好地预测TOLF术后脑脊液漏的风险。Objective Analysis of risk factors related to postoperative cerebrospinal fluid leakage(CSFL)in patients with thoracic ligamentum flavum ossification(TOLF)and construct a nomogram risk prediction model.Methods The clinical data of 111 patients with TOLF who underwent surgical treatment were retrospectively analyzed,including 55 males and 56 females,aged 27~77 years,with an average age of(58.04±9.36)years.According to whether cerebrospinal fluid leakage occurred after surgery,they were divided into a CSFL group(n=42)and a non-CSFL group(n=69).LASSO regression and multifactorial logistic regression analyses were used to screen for their risk factors and to construct a nomogram risk prediction model.Results A total of 111 patients were included in this study,and 42 patients developed CSFL,with an incidence rate of 37.8%.The results of LASSO regression and multivariate logistic regression analyses indicated that less residual rate of spinal canal area(OR=0.958,P<0.05),the more segments involved(OR=2.940,P<0.05),dural adhesion(OR=9.173,P<0.05)and dural ossification(OR=69.468,P<0.05)are independent risk factors for postoperative CSFL.The area under the curve(AUC)of the predicting model analyzed by ROC was 0.923(95%CI:0.873,0.973).The calibration curve of the model is close to the ideal curve.The decision curve analysis(DCA)indicates that when the probability threshold predicted by the model is>5%,this column The line graph model has high clinical application value.Conclusion The lower residual rate of the spinal canal area,the more thoracic segments,dural adhesion and dural ossification are independent risk factors for postoperative CSFL.The Nomogram model based on this can better predict the risk of cerebrospinal fluid leakage after TOLF.

关 键 词：胸椎 黄韧带骨化症 脑脊液漏 危险因素 列线图模型 

分 类 号：R681.5[医药卫生—骨科学]