AECOPD合并呼吸衰竭预后的预测因素及防范策略  

Prognostic factors and preventive strategies for AECOPD with respiratory failure

作　　者：黄雅璇 梅培培[1] HUANG Yaxuan;MEI Peipei(Department of Geriatric Respiratory,Jiangsu Provincid People's Hospital,Nanjing 210000,China)

机构地区：[1]江苏省人民医院老年呼吸科,江苏南京210000

出　　处：《东南大学学报(医学版)》2025年第2期334-339,共6页Journal of Southeast University(Medical Science Edition)

摘　　要：目的:探讨慢性阻塞性肺疾病急性加重期(AECOPD)合并呼吸衰竭患者的预后影响因素,并构建预测模型,为防范对策制定提供参考。方法:便利抽样法选择2022年3月至2024年5月在本院确诊并住院治疗的324例AECOPD合并呼吸衰竭患者,依据入院28 d预后划分为死亡组、生存组。通过二元Logistic回归模型确定AECOPD合并呼吸衰竭患者入院28 d死亡的影响因素,建立列线图预测模型。通过受试者工作特征(ROC)曲线评价模型区分度,校准曲线评价模型一致性,H-L检验评价模型拟合优度,决策曲线评价模型临床价值。结果:死亡组年龄、全身免疫炎症指数(SII)、超敏C反应蛋白/白蛋白值(HCAR)、动脉二氧化碳分压(PaCO_(2))高于生存组,酸碱度(pH)值低于生存组(P<0.05)。二元Logistic回归分析显示,年龄、SII、HCAR、PaCO_(2)均为AECOPD合并呼吸衰竭患者入院28 d死亡的危险因素(P<0.05)。ROC曲线显示,列线图预测模型预测AECOPD合并呼吸衰竭患者入院28 d死亡的曲线下面积(AUC)为0.969,95%CI为0.946~0.993,预测敏感度为87.80%,特异度为95.05%,准确度为94.14%,区分度良好。校准曲线、H-L检验与决策曲线显示,列线图预测模型具有良好的一致性、拟合优度及较高临床获益。结论:AECOPD合并呼吸衰竭入院28 d死亡的预测因素包括年龄、SII、HCAR、PaCO 2,依据上述因素构建的预测模型具有一定预测价值。Objective:To investigate the prognostic factors for patients with acute exacerbation of chronic obstructive pulmonary disease(AECOPD)complicated by respiratory failure,and to construct a prediction model to provide reference for preventive strategies.Methods:A convenience sample of 324 patients diagnosed with AECOPD complicated by respiratory failure and hospitalized from March 2022 to May 2024 was selected.Patients were divided into death and survival groups based on 28-day prognosis after admission.A binary Logistic regression model was used to determine the influencing factors for 28-day mortality in AECOPD patients with respiratory failure.A nomogram prediction model was established.The discrimination was assessed using the receiver operating characteristic(ROC)curve,consistency was evaluated using a calibration curve,model s goodness of fit was evaluated using the Hosmer-Lemeshow test,and clinical utility was assessed using a decision curve.Results:The death group showed significantly higher age,systemic immune-inflammation index(SII),high-sensitivity C-reactive protein/albumin ratio(HCAR),and pressure of arterialcarbon dioxide(PaCO_(2)),and lower acidity of alkalinity(pH)compared to the survival group(P<0.05).Binary Logistic regression analysis revealed that age,SII,HCAR,and PaCO_(2) were risk factors for 28-day mortality in AECOPD patients with respiratory failure(P<0.05).The ROC curve analysis demonstrated that the model had good discrimination in predicting 28-day mortality,with the area under the curve(AUC)of 0.969(95%CI 0.946-0.993),sensitivity of 87.80%,specificity of 95.05%,and accuracy of 94.14%.Calibration curves,Hosmer-Lemeshow test and decision curves indicated good consistency,goodness of fit and high clinical benefit of the nomogram prediction model.Conclusion:Age,SII,HCAR,and PaCO_(2) are predictive factors for 28-day mortality in AECOPD patients with respiratory failure.The prediction model constructed based on these factors has certain predictive value.

关 键 词：慢性阻塞性肺疾病急性加重期 呼吸衰竭 预后 影响因素 预测模型 

分 类 号：R563.9[医药卫生—呼吸系统]