构建AECOPD患者风险模型的临床价值

Clinical Value of Constructing a Risk Model for Patients with AECOPD

目的对慢性阻塞性肺疾病(COPD)患者1年内发生急性加重的影响因素进行研究,建立和验证风险模型。方法收集2020年1月-2021年12月安徽医科大学附属巢湖医院呼吸内科612例AECOPD患者的一般资料,其中2020年1月-12月的324例AECOPD患者为建模组,2021年1月-12月的288例AECOPD患者为验证组,通过Logistic回归单因素分析选取用于建立AECOPD患者风险模型的因素,根据Logistic回归多因素分析危险预测因素,并据此建立最终的风险模型。通过受试者工作特征曲线(ROC)、校准曲线及临床决策曲线(DCA)评估此模型的预测价值。结果FEV1%预计值、既往加重次数、嗜酸性粒细胞百分比(EOS%)、血清白蛋白、三联用药史(ICS+LABA+LAMA)是患者1年内发生急性加重的影响因素,由此建立风险模型公式为:P=1/(1+e-x),x=2.796-0.018×FEV1%预计值+1.516×既往加重次数-0.098×血清白蛋白+0.206×EOS%-1.169×三联用药史,e为自然对数。ROC曲线下面积为0.839,以最大约登指数(0.579)选取最佳临界值为0.221,灵敏度为80.00%,特异度为77.90%。结论FEV1%预计值、既往加重次数、EOS%、血清白蛋白、三联用药史(ICS+LABA+LAMA)是患者1年内发生急性加重的影响因素,由此建立COPD患者急性加重的风险模型。此模型具有良好的区分度和校准度,且临床决策效益方面表现良好,可协助医护人员对加重风险较高的患者进行早期干预,防止病情进一步进展。

Objective To study the influencing factors of acute exacerbation in patients with chronic obstructive pulmonary disease(COPD)within 1 year,and to establish and verify the risk model.Methods The general data of 612 AECOPD patients in the Department of Respiratory Medicine of Chaohu Hospital of Anhui Medical University from January 2020 to December 2021 were collected.Among them,324 AECOPD patients from January to December 2020 were used as the modeling group,and 288 AECOPD patients from January to December 2021 were used as the validation group.The factors used to establish the risk model of AECOPD patients were selected by Logistic regression univariate analysis,and the risk predictors were analyzed by Logistic regression multivariate analysis,and the final risk model was established accordingly.The predictive value of this model was evaluated by receiver operating characteristic curve(ROC),calibration curve and clinical decision curve(DCA).Results FEV1%predicted value,previous exacerbations,EOS%,serum albumin,and the history of triple drug use(ICS+LABA+LAMA)were the factors influencing the occurrence of acute exacerbationsin the patients within one year.From this,the risk prediction model formula was established:P=1/(1+e-x),x=2.796-0.018×FEV1%predicted value+1.516×previous exacerbations-0.098×serum albumin+0.206×EOS%-1.169×history of triple medication,e was the natural logarithm.The area under the ROC curve was 0.839,and the optimal cut-off value was 0.221 based on the approximate index(0.579),with a sensitivity of 80.00%and a specificity of 77.90%.Conclusion FEV1%predicted value,previous exacerbations,EOS%,serum albumin,and the history of triple use(ICS+LABA+LAMA)are the factors influencing the occurrence of acute exacerbations in the patients within one year.Based on this,the risk prediction model of acute exacerbation in COPD patients is successfully constructed.The model has good discrimination,calibration,and clinical decision-making benefits,and can assist healthcare professionals in the early intervention of patients at high risk of exacerbation to prevent further progression.