颅脑创伤住院并发肺部感染患者多指标联合预警诊断模型的建立  

Establishment of multi-index joint early warning model for diagnosis of lung infection in hospitalized patients with craniocerebral trauma

作　　者：陈晓艳[1] 张燕萍[2] 苏星[1] 吴小波[1] 石慧[1] 王娅 仲悦萍[1] 陆凤英[3] CHEN Xiao-yan;ZHANG Yan-ping;SU Xing;WU Xiao-bo;SHI Hui;WANG Ya;ZHONG Yue-ping;LU Feng-ying(Affiliated Hospitalof Nantong University,Nantong,Jiangsu 226001,China)

机构地区：[1]南通大学附属医院神经外科监护病区,江苏南通226001 [2]南通大学附属医院感染管理办公室,南通226001 [3]南通大学附属医院外科,江苏南通226001

出　　处：《中华医院感染学杂志》2020年第18期2785-2789,共5页Chinese Journal of Nosocomiology

基　　金：江苏省南通市市级科技计划基金资助项目(JCZ19055)。

摘　　要：目的建立临床多指标联合预警诊断颅脑创伤住院患者并发肺部感染模型。方法选取2017年1月-2019年12月江苏省南通大学附属医院收治的206例中重型颅脑创伤患者为研究对象,其中肺部感染组107例和未感染组99例。比较两组患者一般资料、急性生理与慢性健康评分(APACHEⅡ)、格拉斯哥昏迷评分(GCS)、生命体征、血常规、凝血功能、血清超敏C-反应蛋白、血清降钙素原(PCT)等指标差异,多因素logistic回归分析,建立肺部感染的多指标联合预警诊断模型,受试者工作特征(ROC)曲线分析各指标诊断肺部感染的临床价值。结果多因素logistic回归模型显示,APACHEⅡ评分(OR=3.031)、体温(OR=3.983)、呼吸频率(OR=1.709)、hs-CRP(OR=2.328)、PCT(OR=2.809)是颅脑创伤住院患者并发肺部感染的独立危险因素。根据回归系数由大到小依次排序建立回归方程为Y=-2.405+1.382×体温+1.109×APACHEⅡ评分+1.033×PCT+0.845×hs-CRP+0.536×呼吸频率,该模型对颅脑创伤住院患者并发肺部感染的预测价值:当Y≥2.30,灵敏度为92.50%,特异度为88.90%。此模型预测本研究纳入的206例患者发生肺部感染的概率为53.40%(110/206),实际发生率为51.94%(107/206),一致性较高(Kappa系数=0.815)。各指标对肺部感染的预测价值由高到低依次为:体温、PCT、hs-CRP、APACHEⅡ评分、呼吸频率。结论由体温、APACHEⅡ评分、PCT、hs-CRP和呼吸频率建立的多指标联合预警诊断模型准确度高,对颅脑创伤住院患者并发肺部感染的早期预警具有较高的临床应用价值。OBJECTIVE To establish a multi-index joint early warning model for diagnosis of lung infection in hospitalized patients with craniocerebral trauma.METHODS A total of 206 patients with moderate-to-severe craniocerebral trauma who were treated in Jiangsu Nantong University Affiliated Hospital from Jan 2017 and Dec 2019 were recruited as the study objects and divided into the infection group with 107 cases and the non-infection group with 99 cases.The baseline data,acute physiology and chronic health evaluationⅡ(APACHEⅡ)scores,Glasgow Coma Scale(GCS)scores,vital signs,blood routine,blood coagulation function indexes,serum high-sensitivity C-reactive protein and serum procalcitonin(PCT)were compared between the two groups of patients.The multivariate logistic regression analysis was performed,the multi-index joint early warning model for diagnosis of pulmonary infection was established,and the clinical value of the indexes in diagnosis of pulmonary infection was analyzed by means of receiver operating characteristic(ROC)curve.RESULTS The multivariate logistic regression analysis showed that the APACHEⅡscore(OR=3.031),body temperature(OR=3.983),respiratory rate(OR=1.709),hs-CRP(OR=2.328)and PCT(OR=2.809)were the independent risk factors for the pulmonary infection in the hospitalized patients with craniocerebral trauma.A regression equation:Y=-2.405+1.382×body temperature+1.109×APACHEⅡscore+1.033×PCT+0.845×hs-CRP+0.536×respiratory frequency was established based on regression coefficients in a descending order;when Y≥2.30,the sensitivity and specificity of the mode were respectively 92.50%and 88.90%in prediction of the pulmonary infection in the hospitalized patients with craniocerebral trauma.The model predicted that of the 206 enrolled patients,the probability of pulmonary infection was 53.40%(110/206),and the actual incidence rate was 51.94%(107/206),with the consistency high(Kappa coefficient=0.815).The predicative values of the indexes,from high to low,were respectively body temperature,PCT,hs-CRP,

关 键 词：颅脑创伤 肺部感染 多指标 预警诊断模型 

分 类 号：R651.15[医药卫生—外科学]