急性脑梗死后长期卧床的老年患者肺部感染预测模型的建立和验证

Establishment and validation of a prediction model for pulmonary infection in elderly patients lying in bed for a long time after ACI

目的建立急性脑梗死后长期卧床的老年患者肺部感染的预测模型,精准识别肺部感染的高危人群。方法回顾性选择2020年1月至2021年1月内蒙古民族大学附属医院呼吸与危重症医学科收治的急性脑梗死后长期卧床的老年患者292例,随访1年,根据是否发生肺部感染分为肺部感染组112例和对照组180例。将数据集又随机分为训练集204例,验证集88例。分析急性脑梗死后长期卧床的老年患者肺部感染的危险因素,基于相关危险因素,建立Nomogram模型,分析并验证模型对急性脑梗死后长期卧床的老年患者肺部感染的预测价值。结果与对照组比较,肺部感染组年龄、糖尿病、慢性阻塞性肺疾病、贫血、意识障碍、吞咽困难、卧床时间明显增加,白蛋白水平明显降低(P<0.05,P<0.01)。年龄≥80岁、卧床时间≥6个月、糖尿病、慢性阻塞性肺疾病、贫血、吞咽困难是急性脑梗死后长期卧床的老年患者肺部感染的独立危险因素,白蛋白≥30 g/L是急性脑梗死后长期卧床的老年患者肺部感染的独立保护因素(P<0.05,P<0.01)。训练集曲线下面积为0.831(95%CI:0.775~0.887),验证集曲线下面积为0.758(95%CI:0.656~0.860)。在验证集里对模型进行Hosmer-Lemeshow拟合优度检验(χ^(2)=6.664,P=0.573)。结论急性脑梗死后长期卧床的老年患者肺部感染发生率较高,通过Nomogram模型可以有效识别肺部感染的高危人群。

Objective To establish a prediction model of pulmonary infection in elderly patients lying in bed for a long time after acute cerebral infarction(ACI)in order to accurately identify high-risk population of the infection.Methods A total of 292 elderly patients who had been lying in bed for a long time after ACI and admitted in our hospital from January 2020 to January 2021 were enrolled in this study.According to the occurrence of pulmonary infection within 1 year of follow-up,they were divided into pulmonary infection group(n=112)and control group(n=180).We randomly divided the data into a training set of 204 cases and a validation set of 88 cases.The risk factors of pulmonary infection in these patients were analyzed,and a nomogram prediction model was established based on the relevant risk factors.The predictive value of the model for pulmonary infection were analyzed and verified in these elderly patients.Results Compared with the control group,older age,larger proportions of diabetes,chronic obstructive pulmonary disease,anemia,disturbance of consciousness and dysphagia,longer bedridden time and lower albumin level were observed in the pulmonary infection group(P<0.05,P<0.01).Age≥80 years,bedridden time≥6 months,comorbidities of diabetes,chronic obstructive pulmonary disease,anemia and dysphagia were independent risk factors for lung infection in these elderly patients,and albumin≥30 g/L was independent protective factors(P<0.05,P<0.01).The AUC value of the training set was 0.831(95%CI:0.775-0.887),and that of the validation set was 0.758(95%CI:0.656-0.860).In the validation set,Hosmer-Lemeshow Goodness of Fit test showed the model had good value and reliability(χ^(2)=6.664,P=0.573).Conclusion The incidence of pulmonary infection is high in the elderly patients lying in bed for a long time after ACI.Our established nomogram model can effectively identify the high-risk population of pulmonary infection.