Ye Yong-Qiang, He Lan-Lan, Liu Gui-Ling, Zhang Jun, Long Lian-Sheng
Department of Radiology,South Taihu Hospital,Huzhou, Zhejiang 313000,China.
Neurology Center,South Taihu Hospital,Huzhou, Zhejiang 313000,China.
Zhongguo Yi Xue Ke Xue Yuan Xue Bao. 2022 Aug;44(4):636-642. doi: 10.3881/j.issn.1000-503X.14670.
Objective To investigate the pathogen distribution,imaging characteristics,and risk factors of pulmonary infection with multi-drug resistant organism (MDRO) in patients with severe craniocerebral injury,and establish and verify the risk prediction model. Methods A total of 230 patients with severe craniocerebral injury complicated with pulmonary infection were collected retrospectively.According to the 7∶3 ratio,they were randomly assigned into a modeling group (161 patients) and a validation group (69 patients).The risk factors of MDRO pulmonary infection were predicted with the data of the modeling group for the establishment of the risk prediction model.The data of the validation group was used to validate the performance of the model. Results Among the 230 patients,68 patients developed MDRO pulmonary infection.The isolated drug-resistant bacteria mainly included multi-drug resistant ,multi-drug resistant ,multi-drug resistant ,and methicillin-resistant ,which accounted for 45.21%,23.29%,16.44%,and 15.07%,respectively.The imaging characteristics included pleural effusion,lung consolidation,and ground-glass shadow,which accounted for 72.06%,63.24%,and 45.59%,respectively.Multivariate Logistic regression analysis showed that the independent risk factors for MDRO pulmonary infection included age ≥60 years (=0.003),history of diabetes (=0.021),history of chronic obstructive pulmonary disease (=0.038),mechanical ventilation ≥7 d (=0.001),transfer from other hospitals (=0.008),and coma (=0.002).A risk scoring model was established with the value (rounded to the nearest integer) corresponding to each index in the regression equation.Specifically,the values of age ≥60 years,history of diabetes,history of chronic obstructive pulmonary disease,mechanical ventilation ≥7 d,transfer from other hospitals,and coma were 1,1,1,2,2,and 1,respectively (value ≥4 indicated a high-risk population).The areas under the receiver operating characteristic curve of the modeling group and validation group were 0.845 and 0.809,respectively. Conclusions Multi-drug resistant is the most common pathogen of MDRO pulmonary infection in patients with severe craniocerebral injury.Pleural effusion,lung consolidation,and ground-glass shadow were the most common imaging characteristics.The established risk model has high discriminant validity in both the modeling group and the validation group.
目的 探讨重型颅脑损伤患者耐多药菌(MDRO)肺部感染的病原菌分布、影像学特征及危险因素,并建立和验证风险预测模型。方法 回顾性收集230例重型颅脑损伤合并肺部感染患者。按照7∶3的比例将其随机分为建模组(161例)和验证组(69例)。利用建模组数据预测MDRO肺部感染的危险因素以建立风险预测模型,用验证组数据验证模型性能。结果 230例患者中,68例发生MDRO肺部感染。分离出的耐药菌主要包括耐多药 、耐多药 、耐多药 及耐甲氧西林 ,分别占45.21%、23.29%、16.44%和15.07%。影像学特征包括胸腔积液、肺实变和磨玻璃影,分别占72.06%、63.24%和45.59%。多因素Logistic回归分析显示,MDRO肺部感染的独立危险因素包括年龄≥60岁( =0.003)、糖尿病史( =0.021)、慢性阻塞性肺疾病史( =0.038)、机械通气≥7 d( =0.001)、转自其他医院( =0.008)及昏迷( =0.002)。根据回归方程中各指标对应的 值(四舍五入取整)建立风险评分模型。具体而言,年龄≥60岁、糖尿病史、慢性阻塞性肺疾病史、机械通气≥7 d、转自其他医院及昏迷的 值分别为1、1、1、2、2和1( 值≥4表明为高危人群)。建模组和验证组的受试者工作特征曲线下面积分别为0.845和0.809。结论 耐多药 是重型颅脑损伤患者MDRO肺部感染最常见的病原菌。胸腔积液、肺实变和磨玻璃影是最常见的影像学特征。所建立的风险模型在建模组和验证组中均具有较高的判别效度。