首页 | 本学科首页   官方微博 | 高级检索  
检索        

脑卒中肺部感染危险因素及血清免疫炎性因子的预测价值
引用本文:张爱莲,柯慧,陈夏,黄晨,周准,王建林,郑俊俊.脑卒中肺部感染危险因素及血清免疫炎性因子的预测价值[J].中华医院感染学杂志,2021(6):827-831.
作者姓名:张爱莲  柯慧  陈夏  黄晨  周准  王建林  郑俊俊
作者单位:江西省上饶市人民医院神经内科;江西省上饶市人民医院检验科
基金项目:江西省卫生计生委科技计划基金资助项目(20181977)。
摘    要:目的基于主成分分析(PCA)-决策树模型分析脑卒中肺部感染的危险因素及血清免疫炎性因子的预测价值。方法回顾性分析2014年12月-2019年12月江西省上饶市人民医院确诊的176例脑卒中患者,根据是否发生肺部感染,分为感染组40例和未感染组136例。记录患者临床特征,检测血白细胞(WBC)计数、血清C-反应蛋白(CRP)、降钙素原(PCT)和可溶性白细胞介素-2受体(SIL-2R)水平。采用Logistic回归和卡方自动交叉检验(CHAID)模型分析影响脑卒中后肺部感染的危险因素;受试者工作特征曲线(ROC)分析血WBC、CRP、PCT、SIL-2R预测肺部感染的价值。结果脑卒中肺部感染与患者年龄、卒中类型、是否合并糖尿病、吸烟史、吞咽困难、意识障碍、侵入性操作有关(P<0.05);感染组患者血WBC、CRP、PCT和SIL-2R水平高于未感染组(P<0.05);糖尿病、侵入性操作、CRP和PCT均为影响脑卒中患者肺部感染的独立危险因素(P<0.05);PCA-ROC分析显示,血WBC、CRP、PCT、SIL-2R和联合预测主成分1(PC1)的曲线下面积(AUC)分别为0.670、0.816、0.854、0.686、0.918;PCA-卡方自动交互检测(CHAID)模型分析显示,WBC、CRP、PCT、SIL-2R和PC1预测肺部感染的准确率为77.33%、84.72%、86.90%、79.02%和89.24%。结论血清PCT水平可用于预测脑卒中患者肺部感染的发生,临床可重点关注合并糖尿病、意识障碍和侵入性操作的患者。

关 键 词:决策树  主成分分析  脑卒中  肺部感染  危险因素

Risk factors for pulmonary infection in stroke patients and predictive value of serum immune inflammatory factors
ZHANG Ai-lian,KE Hui,CHEN Xia,HUANG Chen,ZHOU Zhun,WANG Jian-lin,ZHENG Jun-jun.Risk factors for pulmonary infection in stroke patients and predictive value of serum immune inflammatory factors[J].Chinese Journal of Nosocomiology,2021(6):827-831.
Authors:ZHANG Ai-lian  KE Hui  CHEN Xia  HUANG Chen  ZHOU Zhun  WANG Jian-lin  ZHENG Jun-jun
Institution:(Shangrao People's Hospital,Shangrao,Jiangxi 334000,China)
Abstract:OBJECTIVE To analyze the risk factors for pulmonary infection in the stroke patients and observe the predictive value of serum immune inflammatory factors based on principal component analysis(PCA)-decision tree model.METHODS A total of 176 patients who diagnosed with stroke and were treated in Shangrao People’s Hospital from Dec 2014 to Dec 2019 were retrospectively analyzed and were divided into the infection group with 40 cases and the non-infection group with 136 cases according to the status of pulmonary infection.The clinical characteristics of the patients were recorded,the levels of white blood cell(WBC),serum C-reactive protein(CRP),procalcitonin(PCT)and soluble interleukin-2 receptor(SIL-2 R)were detected.The risk factors for pulmonary infection in the stroke patients were analyzed by means of Logistic regression and chi-squared automatic interaction detector(CHAID)model,and the values of serum WBC,CRP,PCT and SIL-2 R in prediction of pulmonary infection were analyzed by receiver operative characteristic(ROC)curve.RESULTS The pulmonary infection in the stroke patients was associated with the age,types of stroke,complication with diabetes mellitus,smoking history,dysphagia,disturbance of consciousness and invasive procedure(P<0.05).The levels of serum WBC,CRP,PCT and SIL-2 R of the infection group were significantly higher than those of the non-infection group(P<0.05).Diabetes mellitus,invasive procedure,CRP and PCT were the independent risk factors for the pulmonary infection in the stroke patients(P<0.05).PCA-ROC analysis showed that the areas under curve(AUC)of serum WBC,CRP,PCT,SIL-2 R and the combination with principal component 1(PC1)were 0.670,0.816,0.854,0.686 and 0.918,respectively.PCA-CHAID model analysis indicated that the accuracies of WBC,CRP,PCT,SIL-2 R and PC1 were respectively 77.33%,84.72%,86.90%,79.02%and 89.24%in prediction of pulmonary infection.CONCLUSION The serum PCT level can be used for prediction of pulmonary infection in the stroke patients.It is necessary for the hospital to focus on the patients complicated with diabetes mellitus,disturbance of consciousness and invasive procedures.
Keywords:Decision tree  Principal component analysis  Stroke  Pulmonary infection  Risk factor
本文献已被 CNKI 维普 等数据库收录!
设为首页 | 免责声明 | 关于勤云 | 加入收藏

Copyright©北京勤云科技发展有限公司  京ICP备09084417号