摘要
目的探讨脓毒症相关性脑病(SAE)发生的影响因素,并构建预测SAE发生的列线图模型。方法选取2021年3月—2023年2月贵州医科大学附属医院重症监护室(ICU)收治脓毒症患者213例,依据是否发生SAE分为SAE组和非SAE组。收集各项资料开展单因素分析,Logistic回归分析SAE发生的相关因素;利用R软件构建SAE发生的风险预测模型,并通过ROC曲线下面积(AUC)与校准曲线对模型的区分度和准确性予以验证。结果脓毒症患者213例中SAE发生84例(39.44%);SAE组APACHEⅡ评分、SOFA评分、大脑中动脉PI、血乳酸(BLA)、血清S100β、ALT、AST、IL-6水平、有创机械通气比例均显著高于非SAE组(t(χ^(2))/P=3.974/<0.001、3.066/0.002、4.610/<0.001、5.528/<0.001、4.750/<0.001、9.024/<0.001、2.810/0.005、5.063/<0.001、7.239/0.007),局部脑组织氧饱和度(rScO 2)、白蛋白水平低于非SAE组(t/P=4.935/<0.001、3.650/<0.001);Logistic回归分析结果显示,SAE发生的危险因素主要有APACHEⅡ评分高、SOFA评分高、大脑中动脉PI高、动脉血乳酸高、ALT高,而rScO 2高、白蛋白高是SAE发生的保护因素[OR(95%CI)=2.347(1.316~4.184)、2.416(1.432~4.077)、2.204(1.234~3.937)、2.311(1.306~4.088)、2.326(1.376~3.933),0.468(0.271~0.809)、0.561(0.342~0.919)];基于7项预测指标构建预测SAE发生的列线图风险模型,结果显示,AUC为0.831(95%CI 0.773~0.889),预测SAE发生的概率和实际概率相近,且拟合优度HL检验提示模型区分度和准确度均较好(χ^(2)=6.282,P=0.616)。结论基于APACHEⅡ评分、SOFA评分、大脑中动脉PI、rScO 2、动脉血乳酸、ALT、白蛋白7项指标构建的列线图风险模型对SAE发生具有较好的预测作用。
Objective To explore the influencing factors of sepsis associated encephalopathy(SAE)and construct a column chart model to predict the occurrence of SAE.Methods Select 213 patients with sepsis admitted to the Intensive Care Unit(ICU)of Guizhou Medical University Affiliated Hospital from March 2021 to February 2023,and divide them into SAE group and non SAE group based on the occurrence of SAE.Collect various data for single factor analysis,and conduct logistic regression analysis on the relevant factors of SAE occurrence;Use R software to construct a risk prediction model for SAE occurrence,and verify the discrimination and accuracy of the model through the area under the ROC curve(AUC)and calibration curve.Results Among 213 patients with sepsis,84(39.44%)had SAE;SAE group APACHE II score,SOFA score,middle cerebral artery PI,arterial blood lactate(BLA),serum S100β、The levels of ALT,AST,IL-6,and the proportion of invasive mechanical ventilation were higher in the non SAE group than in the non SAE group(t(χ^(2))/P=3.974/<0.001,3.066/0.002,4.610/<0.001,5.528/<0.001,4.750/<0.001,9.024/<0.001,2.810/0.005,5.063/<0.001,7.239/0.007),local brain tissue oxygen saturation(rScO 2)and albumin levels were lower than those in the non SAE group(t/P=4.935/<0.001,3.650/<0.001);The logistic regression analysis results showed that the main risk factors for the occurrence of SAE were high APACHE II score,high SOFA score,high middle cerebral artery PI,high arterial blood lactate,and high ALT,while high rScO 2 and albumin were protective factors for the occurrence of SAE[OR(95%CI)=2.347(1.316~4.184),2.416(1.432~4.077),2.204(1.234~3.937),2.311(1.306~4.088),2.326(1.376~3.933),0.468(0.271~0.809),0.561(0.342~0.919)];A column chart risk model was constructed based on 7 prediction indicators to predict the occurrence of SAE.The results showed that the AUC was 0.831(95%CI 0.773~0.889),and the probability of predicting SAE occurrence was similar to the actual probability,with a goodness of fit HL testχ^(2)=6.282,P=0.616.Conclusion The column chart risk model constructed based on seven indicators:APACHE II score,SOFA score,middle cerebral artery PI,rScO 2,arterial blood lactate,ALT,and albumin has a good predictive effect on the occurrence of SAE.
作者
周航向
袁佳
张倩
陶浚泠
刘颖
Zhou Hangxiang;Yuan Jia;Zhang Qian;Tao Junling;Liu Ying(Department of Intensive Care Medicine,Affiliated Hospital of Guizhou Medical University,Guizhou Province,Guiyang 550004,China)
出处
《疑难病杂志》
CAS
2023年第12期1245-1250,共6页
Chinese Journal of Difficult and Complicated Cases
基金
贵州医科大学附属医院2021年度院级临床研究课题项目(2021-GMHCT-015)。
关键词
脓毒症相关性脑病
影响因素
预测
列线图风险模型
Sepsis-associated encephalopathy
Influencing factors
Prediction
Column chart risk model
作者简介
通信作者:刘颖,E-mail:liuying@gmc.edu.cn。
