基于合并症的住院慢性阻塞性肺疾病患者死亡风险预测模型及验证

Comorbidity-based prediction model and validation of mortality risk in hospitalized chronic obstructive pulmonary disease patients

目的采用列线图构建住院慢性阻塞性肺疾病(chronic obstructive pulmonary disease,COPD)患者的死亡风险预测模型并加以验证。方法回顾性分析2017年1月1日至2022年8月1日期间于北京大学第三医院呼吸与危重症医学科住院的1110例COPD患者的临床结局。采用随机抽样方法将纳入患者按照7︰3的比例分为训练集(777例)及验证集(333例)。在训练集中通过单、多因素Logistic回归分析COPD患者住院期间死亡的危险因素并建立列线图预测模型。通过受试者操作特征曲线(receiver operating characteristic curve,ROC曲线)及曲线下面积(area under the curve,AUC)评估所建立模型。结果急诊入院、心率大于100次/min、呼吸系统感染、心律失常、低蛋白血症、慢性肾脏病与COPD患者住院期间死亡相关,据此建立的列线图在训练集及验证集中AUC分别为0.899和0.871,判别能力较好。结论本研究建立的列线图预测模型可以较好地预测COPD患者住院期间死亡的风险。

Objective To develop and validate a nomogram for predicting in-hospital mortality risk in patients with chronic obstructive pulmonary disease(COPD).Methods The retrospective analysis was conducted on 1110 patients with COPD who were hospitalized in the Department of Respiratory and Critical Care Medicine of Peking University Third Hospital from January 1st,2017 to August 1st,2022.The patients were randomly divided into a training set(777 patients)and a validation set(333 patients).Risk factors for in-hospital mortality in patients with COPD were analyzed using univariate and multivariate logistic analyses in the training set.Nomogram prediction models were then developed and evaluated by nomogram prediction model,using the receiver operating characteristic curve(ROC)and area under curve(AUC).Results Emergency admission,tachycardia,respiratory infection,arrhythmia,hypoproteinemia and chronic kidney disease were associated with in-hospital mortality in patients with COPD.The nomogram demonstrated good discriminatory power with an AUC of 0.899 in the training set and 0.871,respectively in the validation set,indicating good discriminative ability.Conclusions The nomogram can effectively predict in-hospital mortality in patients with COPD.