股骨髁间骨折术后DVT形成的相关危险因素分析以及预测模型的建立  

Relevant risk factors of DVT after operation for intercondylar fracture of femur and the establishment of prediction model

作　　者：古扎丽阿衣·努尔买买提 谢增如[1] GU Zaliayi;XIE Zeng-ru(Emergency Department,The First Teaching Hospital of Xinjiang Medical University,Urumqi 830054,China)

机构地区：[1]新疆医科大学第一附属医院急诊创伤科

出　　处：《创伤外科杂志》2020年第2期116-120,共5页Journal of Traumatic Surgery

基　　金：国家自然科学基金资助项目(81660368)

摘　　要：目的探讨股骨髁间骨折术后深静脉血栓(DVT)形成相关危险因素以及建立其预测模型。方法采用回顾性病例描述研究分析。2014年1月-2017年12月新疆医科大学第一附属医院急诊创伤科收治248例股骨髁间创伤骨折患者,其中男性145例,女性103例;年龄31~67岁,平均52.2岁;骨折类型:C1型95例,C2型86例,C3型67例;轻伤104例(ISS<16分),重伤75例(16≤ISS<25分),严重伤69例(ISS≥25分)。所有患者术前下肢血管B超检查未发现DVT,术后7~14d常规进行下肢血管B超明确有无DVT形成分为血栓组98例和非血栓组150例。对各相关影响因素进行单因素和多因素Logistic回归分析,得出影响术后DVT形成的独立危险因素。根据多因素分析结果建立Logistic回归预测模型,绘制Logit P值预测股骨髁间骨折术后DVT形成的ROC曲线,评价该模型对于判断术后形成DVT风险的预测价值。结果多因素Logistic回归分析显示,两组烟酒嗜好(SDH)、体重指数(BMI)、糖尿病(DM)、骨折类型(CF)、术前ISS评分、术后2~5d D-二聚体(D-Dimer),差异有统计学意义,为术后7~14d内发生DVT的独立危险因素,其OR值分别为1.35、1.46、1.85、2.56、2.34、2.25。根据上述指标建立Logistic回归预测模型:Logit P值=-2.715+0.364×SDH(是=1,否=0)+0.562×BMI(kg/m2)+0.647×D-Dimer(mg/L)+0.612×DM(是=1,否=0)+0.671×ISS(轻伤=1,重伤=2,严重伤=3)+0.754×CF(C1=1,C2=2,C3=3),Logit P值的临界值为15.96,ROC曲线下面积为0.865,准确度为83.1%。结论通过研究首次提出了股骨髁间骨折术后DVT形成的预测模型以及预测值,并评价了该模型的预测价值,判断此类患者术后DVT形成的准确度高,有利于指导临床针对不同患者制定个体化的干预措施预防术后DVT形成。Objective To discuss the relevant risk factors and establish the Logistic regression prediction model of deep vein thrombosis(DVT) after operation for intercondylar fracture of femur.Methods Clinical data on perioperative 248 patients with intercondylar fracture of femur who received surgery in our hospital from Jan.2014 to Dec.2017 were enrolled. Among them,145 cases were male,103 cases were female;the age range was between 31 and 67, with an average age of 52.2 years;fracture types:95 cases of C1 type,86 cases of C2 type,67 cases of C3 type;104 cases of serere injury(ISS < 16),75 cases of serious injury(16 ≤ ISS < 25),69 cases of serious injury(ISS ≥ 25).There was non DVT in the B-ultrasound examination of the lower extremity blood vessels of all patients before operation.The routine B-ultrasound examination of the lower extremity blood vessels was performed 7-14 days after operation to determine whether there was DVT,which was divided into DVT group of 98 cases and non-DVT group of 150 cases. Then univariate and multiple Logistic regression analysis on the relevant risk factors was used to sort out the independent risk factors,which influenced the incidence of DVT after surgery.Based on the multivariate analysis results,the Logistic regression prediction model was established,then the relevant ROC curve predicting DVT after operation for intercondylar fracture of femur was drawn,and the value of the model would be evaluated on predicting DVT of these patients.Results Multiple Logistic regression analysis showed that smoking and drinking habits(SDH),body mass index(BMI),diabetes(DM),classification of fracture(CF), preoperative injury severity score(ISS),D-Dimer(2-5 days after surgery) were independent risk factors to affect the incidence of DVT witin 7-14 days after surgery,and their ORs value were 1.35,1.46,1.85,2.56,2.34,2.25,respectively.Logistic regression prediction model established according to the above factors was expressed as Logit P =-2.715+0.364×SDH(Yes=1,No=0)+0.562×BMI(kg/m^2)+0.647×D-Dimer

关 键 词：股骨髁间骨折 深静脉血栓 预测模型 危险因素 

分 类 号：R683.42[医药卫生—骨科学]