重症监护室老年肺部感染病原学及风险预测模型的构建

Etiological features and construction of risk prediction model of pulmonary infection in elderly patients in intensive care unit

目的探讨重症监护室老年患者肺部感染病原学特点,建立风险预测模型。方法回顾性分析2019年4月-2020年4月南阳市第一人民医院重症监护室的老年患者805例,根据住院期间是否发生感染分为感染组未感染组,鉴定感染组患者病原菌种类。统计两组患者年龄、性别、吸烟史、饮酒史、合并疾病(糖尿病、高血压、冠心病、心力衰竭)、呼吸机使用情况、气管切开情况、血管内插管情况、抗菌药物使用情况、住院时间等临床资料。将上述临床资料纳入多因素Logistic回归分析,建立风险预测模型,利用Hosmer-Lemeshow检验评估模型拟合度,利用受试者工作特征曲线(ROC)对回归模型预测价值进行检测。结果805例ICU老年患者中,有108例出现肺部感染,感染率为13.42%;有93例(86.11%)患者的细菌培养为阳性,共分离出105株病原菌,革兰阴性菌占比最高,为62.86%,革兰阳性菌占30.48%,真菌占6.67%。多元Logsitic分析显示,年龄≥70岁(OR=2.119,P=0.003)、有糖尿病史(OR=2.187,P=0.001)、气管切开(OR=2.459,P<0.001)、联合使用抗菌药物≥2种(OR=1.588,P=0.049)、住院时间≥20 d(OR=2.719,P<0.001)均为影响老年ICU患者肺部感染的独立危险因素;建立风险预测模型的表达式为P=1/[1+e^(-(-0.434+0.751×(年龄)+0.783×(糖尿病史)+0.900×(气管切开)+0.462×(联合使用抗生素≥2种)+1.000×)(住院时间))],Hosmer-Lemeshowχ^(2)=11.366,P=0.182;ROC曲线结果显示,模型预测肺部感染的AUC为0.740,95%CI为0.688~0.792。结论ICU老年患者肺部感染率较高,其主要病原菌为革兰阴性菌,其次为革兰阳性菌,真菌较少;多因素Logsitic回归模型对ICU老年患者肺部感染具有良好的预测效果,临床可通过对ICU老年患者上述危险因素重点监测并加以干预以降低感染发生率。

OBJECTIVE To explore the etiological features of pulmonary infection in elderly patients in the intensive care unit(ICU),and to establish the risk prediction model.METHODS A total of 805 elderly patients in ICU of Nanyang First People’s Hospital from Apr 2019 to Apr 2020 were retrospectively analyzed.According to presence or absence of infection during hospitalization,they were divided into the infection group and non-infection group.The types of pathogens in the infection group were identified.The clinical data such as age,gender,histories of smoking and drinking,diseases(diabetes,hypertension,coronary heart disease,heart failure),ventilator usage,tracheotomy,endovascular intubation,antibiotics usage and hospitalization time in both groups were collected and analyzed by multivariate Logistic regression analysis to establish the risk prediction model.The fit of the model was evaluated by Hosmer-Lemeshow test.Predictive value of the regression model was tested by receiver operating characteristic(ROC)curves.RESULTS Among the 805 elderly patients in ICU,there were 108 cases(13.42%)with pulmonary infection and 93 cases(86.11%)with positive results of bacterial culture.And 105 strains of pathogens were isolated,of which the highest proportion was Gram-negative bacteria,accounting for 62.86%,followed by Gram-positive bacteria(30.48%)and fungi(6.67%).Multivariate Logsitic analysis showed that age not younger than 70 years old(OR=2.119,P=0.003),diabetes history(OR=2.187,P=0.001),tracheotomy(OR=2.459,P<0.001),types of antibiotics not fewer than 2(OR=1.588,P=0.049)and hospitalization time not shorter than 20 d(OR=2.719,P<0.001)were all independent risk factors affecting pulmonary infection in elderly ICU patients.The expression of the risk prediction model was as follows:P=1/[1+e^(-(-0.434+0.751×age)+0.783×(diabetes history)+0.900×(tracheostomy)+0.462×(types of antibiotics not fewer than 2+1.000×(hospitalization time)))],Hosmer-Lemeshowχ^(2)=11.366,P=0.182.The results of ROC curves showed that AUC and 95%CI of the model for predicting pulmonary infection were 0.740 and 0.688-0.792.CONCLUSION The incidence of pulmonary infection is higher in elderly patients in ICU.The main pathogens are Gram-negative bacteria,followed by Gram-positive bacteria,with less fungi.The multivariate Logsitic regression model is of good predictive effect on pulmonary infection in elderly ICU patients.Clinically,Clinically,the above risk factors can be monitored and intervened to reduce the incidence of infection in elderly patients in ICU.