脓毒症患者谵妄发生风险预测模型构建  

Construction of a predictive model for the risk of delirium in patients with sepsis

作　　者：杨思敏 卞兴航 邓开基 孙海伟[1] 王少雄[1] 朱建军[1] 周保纯[1] Yang Simin;Bian Xinghang;Deng Kaiji;Sun Haiwei;Wang Shaoxiong;Zhu Jianjun;Zhou Baochun(Department of Emergency and Critical Care Medicine,Second Affi liated Hospital of Soochow University,Suzhou 215004)

机构地区：[1]苏州大学附属第二医院急重症医学科,苏州215004

出　　处：《中华急诊医学杂志》2024年第6期784-791,共8页Chinese Journal of Emergency Medicine

摘　　要：目的通过收集和分析脓毒症患者在重症监护室(intensive care unit,ICU)住院期间谵妄发生的危险因素,构建临床预测模型,为脓毒症患者的谵妄防治提供参考。方法收集2020年9月至2022年8月入住苏州大学附属第二医院ICU的脓毒症患者数据,根据患者是否发生谵妄分为谵妄组和非谵妄组。比较两组一般资料及临床资料的差异,采取向后逐步回归法筛选出谵妄的独立危险因素,构建谵妄风险预测模型,并对模型进行评价。结果共纳入381例脓毒症患者,114例在ICU住院期间出现谵妄,谵妄发生率为29.9%。两组单因素分析结果显示:年龄≥65岁、输血、使用咪达唑仑、使用去甲肾上腺素、急性生理与慢性健康评分Ⅱ(APACHEⅡ)评分>15、序贯器官衰竭评估(SOFA)评分>4、代谢性酸中毒、尿素>7.1 mmol/L、凝血功能障碍、乳酸、血小板等因素在ICU脓毒症患者谵妄组与非谵妄组之间差异有统计学意义(P<0.05)。多因素分析结果显示:年龄≥65岁、使用咪达唑仑、APACHEⅡ评分>15、代谢性酸中毒、尿素>7.1 mmol/L、凝血功能障碍为ICU脓毒症患者谵妄发生的独立危险因素。对预测模型进行评价,ROC曲线下面积为0.813,HL拟合优度检验P值为0.957>0.05,Bries评分为0.149<0.25。临床决策曲线及临床影响曲线显示预测模型具有较好的临床适用度。结论ICU脓毒症患者谵妄的发生与年龄≥65岁、使用咪达唑仑、APACHEⅡ评分>15、代谢性酸中毒、尿素>7.1 mmol/L、凝血功能障碍6个因素密切相关。本脓毒症谵妄预测模型具有较好的临床预测能力及临床适用度。Objectives To identify the risk factors associated with delirium in intensive care unit(ICU)hospitalization of sepsis patients and construct a clinical prediction model to to provide a reference for the prevention and control of delirium in sepsis patients.Methods Data were collected of sepsis patients admitted in the Intensive Care Unit in the Second Affi liated Hospital of Soochow University from September 2020 to August 2022.The patients were divided into delirium group and non-delirium group according to whether delirium occurred or not.Comparing of the differences in general and clinical data between the two groups,the independent risk factors for delirium were screened by backward stepwise regression method,and the delirium risk prediction model was constructed and evaluated.An independent risk factor analysis for delirium was conducted using a backward stepwise regression approach to identify significant predictors.A delirium risk prediction model was constructed based on the identifi ed risk factors,followed by a comprehensive evaluation of the model's performance.Results A total of 381 sepsis patients were included in the study,114 patients(29.9%)developed delirium during the ICU hospitalization.Univariate analysis revealed statistically signifi cant differences(P<0.05)between the delirium and non-delirium groups for several factors including age≥65years,blood transfusion,use of midazolam,use of adrenaline,APACHEⅡscore>15,SOFA score>4,metabolic acidosis,urea>7.1 mmol/L,coagulation disorders,lactate levels,and platelet count.Multivariate analysis identifi ed age≥65 years,use of midazolam,APACHEⅡscore>15,metabolic acidosis,urea>7.1 mmol/L,and coagulation disorders as independent risk factors for delirium in sepsis patients during ICU hospitalization.The predictive model was evaluated with an area under the ROC curve of 0.813,a non-signifi cant HosmerLemeshow goodness-of-fi t test(P=0.957>0.05),and a Brier score of 0.149(<0.25),indicating good predictive performance and calibration.Clinical decision

关 键 词：重症监护室 脓毒症 谵妄 年龄 咪达唑仑 急性生理与慢性健康评分II 代谢性酸中毒 尿素 危险因素 预测模型 

分 类 号：R459.7[医药卫生—急诊医学]