脑出血术后AKI发生的危险因素分析及预测模型建立  

Risk factors analysis and nomogram establishment for patients developing AKI after intracerebral hemorrhage surgery

作　　者：肖增丽[1] 杜安琪[1] 孙瑶 赵慧颖[1] 安友仲[1] Zengli Xiao;Anqi Du;Yao Sun;Huiying Zhao;Youzhong An(Department of Intensive Care Unit,Peking University People's Hospital,Beijing 100044,China)

机构地区：[1]北京大学人民医院重症医学科,北京100044

出　　处：《中华重症医学电子杂志》2024年第2期157-163,共7页Chinese Journal Of Critical Care & Intensive Care Medicine(Electronic Edition)

摘　　要：目的分析脑出血术后患者发生急性肾损伤(AKI)的危险因素并建立预测模型。方法回顾性收集2014年10月至2021年10月间北京大学人民医院ICU收治的因脑出血行外科手术的207例患者的临床资料,根据术后是否发生AKI将患者分为AKI组(35例)和非AKI组(172例)。分析2组患者的基线资料[性别、年龄、身高、体质量、基础疾病、入院途径等],术前状态评估[格拉斯哥昏迷量表(GCS)评分、术前实验室检查、美国麻醉医师协会(ASA)分级等]、术中(出血量、是否输血、低血压时间等)及入ICU后情况[心率(HR),血压,体温,急性生理学和慢性健康状况评价(APACHEⅡ)评分,实验室检查,治疗用药等]。通过Lasso-logistic回归分析与脑出血术后发生AKI相关的独立危险因素,构建脑出血术后发生AKI的风险预测模型,并通过受试者工作特征曲线(ROC)和Hosmer-Lemeshow(H-L)检验判断模型的预测能力。结果术前状态评估显示,AKI组GCS评分明显低于非AKI组(Z=4.225,P<0.001),ASA分级>3级的比例明显高于非AKI组(χ^(2)=17.988,P<0.001),差异均有统计学意义。术后情况比较,AKI组患者入ICU时的HR及APACHEⅡ评分明显高于非AKI组,差异均有统计学意义(Z=2.730,P=0.006;Z=4.743,P<0.001);术后治疗用药上,AKI组患者血管活性药物使用比例明显高于非AKI组,差异有统计学意义(χ^(2)=6.571,P=0.010)。207例患者中,共计27例患者院内死亡,总体病死率为13.04%,其中AKI组病死率明显高于非AKI组,差异有统计学意义(χ^(2)=36.249,P<0.001);AKI组患者的机械通气时间及ICU停留时间均明显高于非AKI组,差异均有统计学意义(Z=3.317,P=0.001;Z=3.271,P=0.001)。Lasso-logistic回归分析结果显示,术前GCS评分(OR=0.89,95%CI:0.80~0.99,P=0.031)、ASA分级>3级(OR=2.87,95%CI:1.12~7.39,P=0.029)、入ICU时HR(OR=1.02,95%CI:1.00~1.04,P=0.016)是脑出血术后发生AKI的独立危险因素,联合此3项指标构建Nomogram风险预测模型。通过验证认为Objective To identify risk factors of acute kidney injury(AKI)after intracerebral hemorrhage surgery and establish a predictive model for predicting postoperative AKI in these patients.Methods Clinical data of 207 patients admitted to the ICU between October 2014 and October 2021 were retrospectively collected,and patients were divided into AKI(35 cases)and non-AKI(172 cases)based on diagnosis of postoperative AKI.Baseline data of patients in 2 groups(gender,age,height,body quality,basic diseases),preoperative status assessment[Glasgow coma scale(GCS)score,preoperative laboratory results,American Association of Anesthesiologists(ASA)classification],intraoperative(blood volume loss)and during ICU[heart rate(HR),blood pressure,body temperature,acute physiology and chronic health status evaluation(APACHEⅡ)score,laboratory results,treatment and medication]were recorded.Independent risk factors were identified by Lasso-Logistic regression analysis and a nomogram was established.Calibration and predictive analysis were established to evaluate the nomogram.The predictive ability of the model was evaluated with receiver operating characteristic(ROC)curve and Hosmer-Lemeshow(H-L)test.Results Preoperative evaluation showed that GCS score in the AKI group was significantly lower than the non-AKI group(Z=4.225,P<0.001),and the proportion of patients with ASA grade over 3 was significantly higher than that in non-AKI group(χ^(2)=17.988,P<0.001),which were statistically significant.Postoperative conditions showed that,the HR and APACHEⅡscores in AKI patients were significantly higher than those in non-AKI patients,and the differences were statistically significant(Z=2.730,P=0.006;Z=4.743,P<0.001);the proportion postoperative vasoactive treatment in AKI patients was significantly higher(χ^(2)=6.571,P=0.010).Among 207 patients,a total of 27 patients died in hospital,and the overall fatality rate was 13.04%,which is significantly higher in AKI group(χ^(2)=36.249,P<0.001);length of mechanical ventilation and length of ICU s

关 键 词：脑出血 急性肾损伤 危险因素 预测模型 

分 类 号：R651.1[医药卫生—外科学]