病毒性肺炎所致急性呼吸窘迫综合征早期风险模型的构建

Establishment of the early risk model for acute respiratory distress syndrome caused by viral pneumonia

目的探讨病毒性肺炎患者所致急性呼吸窘迫综合征(ARDS)的危险因素,并构建其早期风险预测模型。方法回顾性分析,选取2019年4月-2023年4月在我院ICU收治的病毒性肺炎患者400例,依据7∶3比例分为训练集(n=280)和验证集(n=120),根据训练集中患者有无ARDS的发生进行分组,分为ARDS组(n=97)和非ARDS组(n=183)。比较2组患者的临床资料,并采用单因素分析和多因素二元logistic回归分析相关危险因素,利用logistic回归分析建立早期风险预测模型,采用ROC曲线、校准曲线检验模型预测效果,采用验证集的数据绘制2种曲线对模型的预测效能进行外部验证。结果多因素分析结果显示:2组患者年龄(OR=1.057,95%CI=1.021~1.094,P=0.002)、合并糖尿病(OR=3.712,95%CI=1.693~8.141,P=0.001)、肺损伤预测评分(OR=1.375,95%CI=1.121~1.686,P=0.002)、英国国家早期预警评分(OR=1.134,95%CI=1.045~1.230,P=0.003)和肺炎严重指数(OR=1.029,95%CI=1.015~1.043,P<0.001)为病毒性肺炎所致ARDS的独立危险因素(P<0.05),在此基础上建立风险预测模型。模型公式Z=0.055×年龄+1.312×合并糖尿病+0.318×肺损伤预测评分+0.126×英国国家早期预警评分+0.028×肺炎严重指数-16.430。其ROC曲线下面积为0.878(95%CI=0.838~0.919,P<0.001),灵敏度为0.856,特异度为0.803,Youden指数为0.659。结论本模型可有效预测病毒性肺炎患者ARDS的发生,为医护人员制定早期预防病毒性肺炎患者ARDS发生的护理方案提供参考依据,具有良好的早期预警价值。

Objective To explore the risk factors of acute respiratory distress syndrome(ARDS)in patients with viral pneumonia,and to establish an early risk prediction model.Methods A total of 400 patients with viral pneumonia admitted to the ICU of our hospital from April 2019 to April 2023 were selected for retrospective analysis,and were divided into the training set(n=280)and the verification set(n=120)according to the ratio of 7∶3.The patients in the training set were grouped according to the presence or absence of ARDS,and divided into the ARDS group(n=97)and the non-ARDS group(n=183).The clinical data of 2 groups were compared,univariate binary logistic regression analysis multivarinte analyzsis were used to analyze the related risk factors,and the prediction effect of the model was tested by ROC curve and calibration curve,and the data of the validation set were used to draw 2 kinds of curves to externally verify the predictive efficacy of the model.Results The results of the multivariate analysis showed that patient age between 2 groups(OR=1.057,95%CI was from 1.021 to 1.094,P<0.002),combined with diabetes mellitus(OR=3.712,95%CI was from 1.693 to 8.141,P=0.001),lung injury prediction score(OR=1.375,95%CI was from 1.121 to 1.686,P=0.002),UK National Early Warning Score(OR=1.134,95%CI was from 1.045 to 1.230,P=0.003),and pneumonia severity index(OR=1.029,95%CI was from 1.015 to 1.043,P<0.001)were the independent risk factors for ARDS due to viral pneumonia(P<0.05),Based on this basis,the risk prediction model was established.Model formula Z=0.055 Age+1.312 combined diabetes+0.318 lung injury prediction score+0.126 UK National Early Warning Score+0.028 pneumonia severity index-16.430.The area under the ROC curve was 0.878(95%CI was from 0.838 to 0.919,P<0.001),with a sensitivity of 0.856,specificity of 0.803,and Youden index of 0.659.Conclusion This model can effectively predict the occurrence of ARDS in patients with viral pneumonia,and provide a reference for medical staff to formulate nursing plans for early prevention of ARDS in patients with viral pneumonia.It has good early warning value.