构建Logistic回归预测模型分析颅脑损伤患者并发凝血功能障碍的影响因素

Construction of Logistic regression prediction model to analyze the influencing factors of coagulation dysfunction in patients with head injury

目的构建Logistic回归预测模型分析颅脑损伤(TBI)患者并发凝血功能障碍的影响因素,以优化颅脑损伤患者管理,降低凝血功能障碍风险。方法收集2019年5月—2021年5月南方医科大学南方医院收治的TBI患者的临床资料,对符合纳入和排除标准的172例患者进行回顾性病例研究。观察患者凝血功能情况,根据是否发生凝血功能异常分为凝血功能障碍组、正常组,监测患者伤后7d内的凝血功能,采用Logistic回归方程分析颅脑损伤患者并发凝血功能障碍的影响因素,运用Wald卡方、似然比卡方、拟合优度检验评估Logistic回归模型,利用受试者工作特征(ROC)曲线检验Logistic回归模型预测价值。结果172例TBI患者凝血功能障碍发生率29.65%;凝血功能障碍组活化部分凝血活酶时间(APTT)、血浆凝血酶原时间(PT)、国际标准化比值(INR)、血小板(PLT)等凝血功能指标在伤后7d内持续存在异常,但整体趋于恢复正常。而正常组的APTT、PT、INR、INR等指标在伤后7 d内均在正常范围波动,且两组凝血功能指标差异存在统计学意义(P<0.05);Logistic回归方程分析显示,动脉血pH值、抗凝药物长期使用史、手术治疗、弥漫性轴索损伤、病情程度是TBI患者并发凝血功能障碍的影响因素(P<0.05);Logistic回归模型对TBI患者并发凝血功能障碍的预测敏感度为82.35%,特异度为91.74%。结论①TBI患者具有凝血功能障碍风险;②动脉血pH值、抗凝药物长期使用史、手术治疗、弥漫性轴索损伤、病情程度是TBI患者并发凝血功能障碍的影响因素;③Logistic回归模型对TBI患者并发凝血功能障碍具有一定预测价值。

Objective:To construct a logistic regression prediction model to analyze the influencing factors of coagulation dysfunction in patients with head injury(TBI),so as to optimize the management of patients with head injury and reduce the risk of coagulation dysfunction.Methods Collect the clinical data of TBI patients admitted to our hospital from May 2019 to May 2021,and conduct a retrospective case study on 172 patients who met the inclusion and exclusion criteria.Observe the coagulation function of patients,divide them into coagulation dysfunction group and normal group according to whether abnormal coagulation function occurs,monitor the coagulation function of patients within 7 days after injury,and use Logistic regression equation to analyze the influencing factors of coagulation dysfunction in patients with brain injury.Wald chi-square,likelihood ratio chi-square,and goodness-of-fit test evaluate Logistic regression model,and use receiver operating characteristic(ROC)curve to test the predictive value of Logistic regression model.Results The incidence of coagulation dysfunction in 172 TBI patients was 29.65%;the coagulation dysfunction group activated partial thromboplastin time(activated partial thromboplastin time,APTT),plasma prothrombin time(prothrombin time,PT),international normalized ratio(international normalized ratio,INR),platelet(blood platelet,PLT)and other blood coagulation function indicators continue to be abnormal within 7 days after injury,but the overall trend tends to return to normal.The APTT,PT,INR,INR and other indicators of the normal group fluctuated within the normal range within 7 days after injury,and there were significant differences in the coagulation function indicators between the two groups(P<0.05);Logistic regression analysis showed that the arterial blood pH value,Long-term use history of anticoagulant drugs,surgical treatment,diffuse axonal injury,and disease severity were the influencing factors of TBI patients with coagulation dysfunction(P<0.05);the logistic regression model had a predictive sensitivity of 82.35 for TBI patients with coagulation dysfunction,the specificity was 91.74%.Conclusion①TBI patients are at risk of coagulation dysfunction;②Arterial blood pH,long-term use history of anticoagulant drugs,surgical treatment,diffuse axonal injury,and disease severity are the influencing factors of TBI patients with coagulation dysfunction;③Logistic regression model for TBI Patients with coagulation dysfunction have certain predictive value.