慢性阻塞性肺疾病并发呼吸衰竭风险预测模型的Meta分析  

Meta⁃analysis of risk prediction models for patients with chronic obstructive pulmonary disease complicated with respiratory failure

作　　者：刘欣 刘雪萍 焦玉丁 任茂玲 何永琴 Liu Xin;Liu Xueping;Jiao Yuding;Ren Maoling;He Yongqin(Department of Preventive Health Care,2 Department of Respiratory Medicine,3 Department of Orthopedics,Army Medical University(Third Military Medical University),Chongqing 400037,China)

机构地区：[1]陆军(第三)军医大学第二附属医院预防保健科,重庆400037 [2]陆军(第三)军医大学第二附属医院呼吸内科,重庆400037 [3]陆军(第三)军医大学第二附属医院骨科,重庆400037

出　　处：《中华肺部疾病杂志(电子版)》2025年第1期110-114,共5页Chinese Journal of Lung Diseases(Electronic Edition)

摘　　要：目的研究慢性阻塞性肺疾病(chronic obstructive pulmonary disease,COPD)并发呼吸衰竭风险预测模型。方法检索从建库至2024年9月中国知网数据库(China national knowledge infrastructure,CNKI)、万方数据知识服务平台、维普数据库、中国生物医学数据库、美国国立医学图书馆数据库(PubMed)、Cochrane循证医学数据库(Cochrane Library)、医学文摘数据库(EMbase)及Web of Science数据库收录的COPD并发呼吸衰竭风险预测模型文献。采用Revman5.3软件对纳入模型中共性预测因子进行Meta分析。结果共纳入5篇文献,文献的样本总量为177~25638例,结果事件为44~3844例,潜在预测变量为14~42个,模型间具有共性的前六位预测因子分别为血清白蛋白水平、第1秒用力呼气容积(force expiratory volume in 1 second,FEV1)、每年慢性阻塞性肺疾病急性加重期(acute exacerbation of chronic obstructive pulmonary disease,AECOPD)发生次数、白细胞计数、C反应蛋白、COPD病程。COPD患者29316例,并发呼吸衰竭4084例(13.93%)。包含5个风险预测模型,模型建模时受试者工作特征曲线下面积(area under curve,AUC)为0.645~0.950,4个模型的AUC≥0.8。预测模型研究的偏倚风险评估工具(prediction model risk of bias assessment tool,PROBAST)结果显示,纳入的5篇文献为高偏倚风险,主要原因包括未报告缺失数据的处理、模型效果评价不完整。Meta分析结果显示,白细胞计数(OR=1.97,95%CI:1.33~2.92)是COPD并发呼吸衰竭的预测因子。结论现有的COPD并发呼吸衰竭风险预测模型偏倚风险高,应参照PROBAST规范完善研究设计,开发、更新和验证模型,验证临床实践中的适用性和安全性。Objective To study the risk prediction model for respiratory failure in patients with chronic obstructive pulmonary disease(COPD).Methods Relevant literature on risk prediction models for respiratory failure in COPD patients published from the establishment of the databases to September 2024 was retrieved from CNKI,Wanfang Data Knowledge Service Platform,VIP Database,China Biomedical Database,PubMed,Cochrane Library,EMbase,and Web of Science.Meta⁃analysis was conducted on the predictive value of common predictors in the included models using Revman 5.3 software.Results A total of 5 articles were included,with a sample size ranging from 177 to 25,638 cases,and the number of outcome events ranging from 44 to 3,844 cases.The number of potential predictors ranged from 14 to 42.The top six common predictors among the models were serum albumin level,force expiratory volume in 1 second(FEV1),the number of annual acute exacerbation of chronic obstructive pulmonary disease(AECOPD)episodes,white blood cell count,C⁃reactive protein,and the duration of COPD.A total of 29,316 COPD patients were included,among whom 4,084(13.93%)developed respiratory failure.Five risk prediction models were included,and the area under the curve(AUC)of all models ranged from 0.645 to 0.950,with four models having an AUC≥0.8.The prediction model risk of bias assessment tool(PROBAST)results showed that all five included articles had a high risk of bias,mainly due to the lack of reporting on the handling of missing data and incomplete model performance evaluation.Meta⁃analysis results indicated that white blood cell count(OR=1.97,95%CI:1.33⁃2.92)was a predictor of respiratory failure in COPD patients.Conclusion The existing risk prediction models for respiratory failure in COPD patients have a high risk of bias.Future studies should follow the PROBAST guidelines to improve research design,develop,update,and validate such models,and further verify their applicability and safety in clinical practice.

关 键 词：急性加重期 肺疾病 慢性阻塞性 呼吸衰竭 风险预测 

分 类 号：R563[医药卫生—呼吸系统]