摘要
目的基于分子生物学指标、肺计算机断层扫描(CT)构建慢性阻塞性肺疾病急性加重(AECOPD)伴Ⅱ型呼吸衰竭(Ⅱ-RF)患者预后转归的Nomogram预测模型并进行验证。方法选取南阳市第二人民医院2022年3月至2023年8月收治的200例AECOPD伴Ⅱ-RF患者开展前瞻性研究。出院后随访1年,失访5例,最终纳入195例患者。根据预后分为预后良好组(n=141)和预后不良组(n=54)。比较两组患者的一般资料及分子生物学指标、肺CT指标,应用最小绝对收缩与选择算子(Lasso)-逻辑回归(Logistic)分析AECOPD伴Ⅱ-RF患者预后不良的影响因素,根据影响因素构建AECOPD伴Ⅱ-RF患者预后不良的Nomogram预测模型,采用受试者工作特征(ROC)曲线、校准曲线、决策曲线评估预测模型的预测效能及价值,并利用Bootstrap法(重复抽样1000次)进行内部验证。另外选取2022年3月至2023年8月南阳市第二人民医院符合纳入和排除标准的AECOPD伴Ⅱ-RF患者100例(除外建模患者)作为验证样本进行外部验证。结果单因素分析结果显示,两组患者的年龄、吸烟史、急性生理及慢性健康状况评分Ⅱ(APACHEⅡ)、第1秒用力呼气容积(FEV_(1))/用力肺活量(FVC)、FEV_(1)占预计值百分比(FEV_(1)%pred)、机械通气治疗时间、血清同型半胱氨酸(Hcy)、β_(2)-微球蛋白(β_(2)-MG)和肽素(Copeptin)、可溶性髓系细胞触发受体-1(sTREM-1)、Cvin、Cvex比较差异均有统计学意义(P<0.05);Lasso-Logistic回归分析结果显示,APACHEⅡ评分、血清Hcy、β_(2)-MG、Copeptin、sTREM-1、Cvin、Cvex均为AECOPD伴Ⅱ-RF患者预后不良的危险因素,FEV_(1)%pred、FEV_(1)/FVC为保护因素(P<0.05);ROC曲线分析结果显示,该模型的C-index为0.907,AUC为0.907(95%CI:0.858~0.957);校准曲线显示,该模型预测值与实际值的良好一致性及拟合度;决策曲线显示,模型预测预后不良具有良好的净获益及临床效用;外部验证结果显示,该模型的预测敏感度为89.66%(26/29),特异度为78.87%(56/71),准确度为82.00%(82/100)。结论APACHEⅡ评分、血清Hcy、β_(2)-MG、Copeptin、sTREM-1、Cvin、Cvex、FEV_(1)%pred、FEV_(1)/FVC均为AECOPD伴Ⅱ-RF患者预后不良的影响因素,基于此所建立的Nomogram预测模型具有良好的预测效能、一致性及临床效用,可为早期预后转归预测提供参考依据。
Objective To construct and validate a Nomogram prediction model for the prognosis of patients with acute exacerbation of chronic obstructive pulmonary disease(AECOPD)accompanied by typeⅡrespiratory failure(Ⅱ-RF)based on molecular biology indicators and lung computed tomography(CT)scans.Methods A prospective study enrolled 200 patients with AECOPD accompanied byⅡ-RF admitted to Nanyang Second General Hospital from March 2022 to August 2023.After discharge,patients were followed up for 1 year,with 5 cases lost to follow-up,result-ing in the final inclusion of 195 patients.Patients were divided into a good prognosis group(n=141)and a poor progno-sis group(n=54)based on prognosis.The general information,molecular biology indicators,and lung CT indicators of the two groups were compared.Least absolute shrinkage and selection operator(Lasso)-logistic regression analysis was used to investigate the influencing factors of poor prognosis in patients with AECOPD accompanied byⅡ-RF,and a no-mogram was constructed accordingly.The predictive performance and value of the prediction model were evaluated us-ing receiver operating characteristic(ROC)curves,calibration curves,and decision curves;internal validation was per-formed using the Bootstrap method(with 1000 repeated samplings).An independent cohort of 100 eligible patients with AECOPD accompanied byⅡ-RF who met the inclusion and exclusion criteria at Nanyang Second General Hospital during the same period served for external validation.Results Univariate analysis showed significant group differences in age,smoking history,Acute Physiology and Chronic Health EvaluationⅡ(APACHEⅡ)score,forced expiratory vol-ume in one second(FEV_(1))/forced vital capacity(FVC),FEV_(1) percentage of predicted value(FEV_(1)%pred),duration of me-chanical ventilation treatment,serum homocysteine(Hcy),β_(2)-microglobulin(β_(2)-MG),Copeptin,soluble triggering recep-tor expressed on myeloid cells-1(sTREM-1),Cvin,and Cvex(all P<0.05).Lasso-Logistic regression analysis identified APACHEⅡscore,serum Hcy,β_(2)-MG,Copeptin,sTREM-1,Cvin,and Cvex as risk factors for poor prognosis,while FEV_(1)%pred and FEV_(1)/FVC were protective factors(all P<0.05).The model yielded a C-index of 0.907 and an area under the ROC curve of 0.907(95%CI:0.858-0.957).The calibration curve demonstrated good consistency and fit between the predicted and actual values of the model.The decision curve indicated that the model had good net benefit and clinical util-ity in predicting poor prognosis.The external validation results showed that the model had a sensitivity of 89.66%(26/29),specificity of 78.87%(56/71),and accuracy of 82.00%(82/100).Conclusion APACHEⅡscore,serum Hcy,β_(2)-MG,Copeptin,sTREM-1,Cvin,Cvex,FEV_(1)%pred,and FEV_(1)/FVC are all independent influencing factors for poor progno-sis in AECOPD patients accompanied byⅡ-RF.The nomogram prediction model developed from these factors exhibits good predictive performance,consistency,and clinical utility,providing a valuable tool for early prognosis prediction.
作者
胡虎
郝琳
康娟
HU Hu;HAO Lin;KANG Juan(The Second Ward of Respiratory and Critical Illness,Nanyang Second General Hospital,Nanyang 473000,Henan,CHINA)
出处
《海南医学》
2025年第13期1844-1849,共6页
Hainan Medical Journal
基金
河南省医学科技攻关计划(联合共建)项目(编号:LHGJ20191465)。
