脑卒中相关性肺炎患者多药耐药菌感染影响因素分析及预测模型构建  

Influencing factors and predictive model construction of multidrug-resistant bacteria infection in patients with stroke-associated pneumonia

作　　者：高新梅 邢雷 王清涛[2] 母海蕾 朱东林[3] GAO Xinmei;XING Lei;WANG Qingtao;MU Hailai;ZHU Donglin(Department of Neurosurgery,First Hospital of Qinhuangdao City,Qinhuangdao,Hebei 066000,China;Department of Hand and Foot Surgery,First Hospital of Qinhuangdao City,Qinhuangdao,Hebei 066000,China;Department of Infection Management,First Hospital of Qinhuangdao City,Qinhuangdao,Hebei 066000,China)

机构地区：[1]河北省秦皇岛市第一医院神经外科,河北秦皇岛066000 [2]河北省秦皇岛市第一医院手足外科,河北秦皇岛066000 [3]河北省秦皇岛市第一医院感染管理处,河北秦皇岛066000

出　　处：《检验医学与临床》2025年第4期475-479,484,共6页Laboratory Medicine and Clinic

基　　金：河北省秦皇岛市重点研发科技支撑项目(202101157)。

摘　　要：目的观察脑卒中相关性肺炎(SAP)患者多药耐药菌感染情况,分析其影响因素并构建预测模型。方法选择2021年1月至2023年1月该院收治的202例SAP患者作为研究对象,所有患者均接受病原菌检测及药物敏感试验,根据多药耐药菌感染情况,将患者分为感染组与非感染组,对比2组患者临床资料,分析SAP患者多药耐药菌感染的影响因素,构建预测模型,绘制受试者工作特征(ROC)曲线,分析预测模型对SAP患者多药耐药菌感染的预测价值。结果202例SAP患者中113例患者发生多药耐药菌感染,发生率为55.94%;感染组美国国立卫生研究院卒中量表(NIHSS)评分高于非感染组(P<0.05),合并糖尿病、昏迷、有侵入性操作、预防性使用抗菌药物占比高于非感染组(P<0.05);Logistic回归分析结果显示,高NIHSS评分(OR=1.586,95%CI:1.393~1.806,P<0.001)、合并糖尿病(OR=3.259,95%CI:1.505~7.058,P=0.003)、昏迷(OR=2.823,95%CI:1.416~5.629,P=0.003)、有侵入性操作(OR=2.373,95%CI:1.299~4.336,P=0.005)、预防性使用抗菌药物(OR=2.941,95%CI:1.582~5.468,P=0.001)是SAP患者发生多药耐药菌感染的危险因素。根据Logistic回归模型得到回归方程Logit(P)=-2.034+0.461X_(NIHSS评分)+1.181X_(糖尿病)+1.038X_(昏迷)+0.864 X_(有侵入性操作)+1.079X_(预防性使用抗菌药物),将计算得到预测概率作为检验变量,绘制ROC曲线。该模型预测SAP患者多药耐药菌感染的曲线下面积为0.985,灵敏度、特异度分别为0.956、0.888。结论SAP患者多药耐药菌感染与诸多因素相关,临床应结合上述因素,早期预测SAP患者多药耐药菌感染,并采取相关干预措施,以改善患者预后。Objective To observe multidrug-resistant bacterial infections in patients with stroke-associated pneumonia(SAP),analyze their influencing factors and construct a prediction model.Methods A total of 202 patients with SAP treated at the hospital from January 2021 to January 2023 were selected as the study subjects.All patients underwent pathogen testing and drug susceptibility testing.Based on the presence of multidrug-resistant bacterial infections,the patients were divided into the infection group and the non-infection group.Clinical data between the two groups were compared to analyze the influencing factors of multidrug-resistant bacterial infections in SAP patients.Predictive model was constructed.Receiver operating characteristic(ROC)curve was plotted to assess the predictive value of the model for multidrug-resistant bacterial infections in SAP patients.Results Among the 202 patients with SAP,multidrug-resistant bacterial infections were observed in 113 cases(55.94%).The National Institutes of Health Stroke Scale(NIHSS)score in the infection group was higher than that in the non-infection group,and the proportions of patients with comorbid diabetes,coma,invasive procedures and prophylactic antibiotic use were significantly higher in the infection group than those in the non-infection group(P<0.05).Logistic regression analysis indicated that a higher NIHSS score(OR=1.586,95%CI:1.393-1.806,P<0.001),comorbid diabetes(OR=3.259,95%CI:1.505-7.058;P=0.003),coma(OR=2.823,95%CI:1.416-5.629,P=0.003),invasive procedures(OR=2.373,95%CI:1.299-4.336,P=0.005),and prophylactic antibiotic use(OR=2.941,95%CI:1.582-5.468,P=0.001)were identified as risk factors for multidrug-resistant bacterial infections in SAP patients.The regression equation Logit(P)=-2.034+0.461 X_(NIHSS score)+1.181X_(diabetes)+1.038X_(coma)+0.864X_(invasive procedures)+1.079X_(prophylactic use of antimicrobials) was derived from the Logistic regression model.The calculated predictive probabilities were used as test variables to plot the ROC curve.The area

关 键 词：脑卒中相关性肺炎 多药耐药菌 影响因素 预测模型 药物敏感试验 

分 类 号：R743.3[医药卫生—神经病学与精神病学] R563.1[医药卫生—临床医学]