| 丁平,李权华.慢性阻塞性肺疾病急性加重期并 Ⅱ型呼吸衰竭病人无创通气治疗结局预测模型的构建及评价[J].安徽医药,2025,29(11):2271-2276. |
| 慢性阻塞性肺疾病急性加重期并 Ⅱ型呼吸衰竭病人无创通气治疗结局预测模型的构建及评价 |
| Construction and evaluation of a non-invasive ventilation outcome prediction model for patients with acute exacerbation of chronic obstructive pulmonary disease and type Ⅱ respiratory failure |
| |
| DOI：10.3969/j.issn.1009-6469.2025.11.033 |
| 中文关键词: 肺疾病，慢性阻塞性 Ⅱ型呼吸衰竭 高碳酸血症 超敏 C反应蛋白 白蛋白 无创通气 预测模型 |
| 英文关键词: Pulmonary disease, chronic obstructive Type Ⅱ respiratory failure Hypercapnia Hypersensitivity to C-reactive pro-tein Albumin Non-invasive ventilation Forecasting model |
| 基金项目: |
|
| 摘要点击次数: 520 |
| 全文下载次数: 173 |
| 中文摘要: |
| 目的探讨慢性阻塞性肺疾病急性加重期( AECOPD)并 Ⅱ型呼吸衰竭病人无创通气( NIV)治疗结局的危险因素并构建预测模型加以评价。方法回顾性选取 2020年 5月至 2023年 5月在资阳市精神病医院就诊的 AECOPD并Ⅱ型呼吸衰竭病人 142例，均接受 NIV治疗。根据治疗 24 h内 NIV治疗结局将病人分为失败组( n=53)和成功组( n=89)。比较两组超敏 C反应蛋白(hs-CRP)、白蛋白( Alb)和 hs-CRP/Alb比值及临床资料。采用多因素 logistic回归分析探索 NIV治疗失败的独立危险因素，独立危险因素引入 R软件构建风险模型并进行模型评价。结果失败组 hs-CRP和 hs-CRP/Alb值( 2.14±0.50比 0.53±0.12)高将于成功组， Alb水平则低于成功组(均 P<0.05)；与成功组相比，失败组病人的呼吸频率值、舒张压、急性生理学和慢性健康状况评价 Ⅱ(APACHEⅡ)评分［(22.48±4.17)分比( 18.72±3.91)分］、序贯器官衰竭评估( SOFA)评分、白细胞计数及治疗前和治疗 1 h后动脉血二氧化碳分压( PaCO2)水平［( 76.11±12.22)mmHg比( 58.83±9.92)mmHg］更高，失败组病人治疗前的动脉血氧分压(PaO2)水平、氧合指数及治疗前和治疗 1h后的 pH值( 7.29±0.18比 7.40±0.21)更低(均 P<0.05)。多因素 logistic回归模型显示， hs-CRP/Alb值高( OR=2.75)、 APACHEⅡ评分高( OR=1.61)、治疗 1 h pH值低( OR=1.68)和 PaCO2高( OR=2.45)均为影响 NIV治疗失败的独立危险因素( P<0.05)；根据独立危险因素建立 NIV治疗失败的预测模型方程: Logit(P)=.0.06+0.47×APACHEⅡ评分+0.52×治疗 1 h pH值+0.90×治疗 1 h PaCO2+1.01×hs-CRP/Alb值，采用 Bootstrap法内部验证显示，预测模型区分度良好；校准曲线显示模型拟合度好；构建的 logistic预测模型进行回顾性验证，结果显示其预测 NIV治疗失败的受试者操作特征曲线(ROC曲线)下面积( AUC)为 0.93，预测模型的预测效能优于 hs-CRP/Alb值单独预测效能。结论 hs-CRP/Alb值高、 APACHEⅡ评分高、治疗 1 h pH值低和 PaCO2高是 AECOPD并Ⅱ型呼吸衰竭病人 NIV治疗失败的独立危险因素，引入独立危险因素构建的风险预测模型区分度、拟合度好，对 NIV治疗结局具有较高的参考价值。 |
| 英文摘要: |
| Objective The risk factors of non-invasive ventilation (NIV) outcomes in patients with acute exacerbation of chronic ob-structive pulmonary disease (AECOPD) and type Ⅱ respiratory failure were evaluated with a predictive model.Methods A total of 142 patients with AECOPD and type Ⅱ respiratory failure who were treated at Ziyang Psychiatric Hospital from May 2020 to May 2023were retrospectively selected and all received NIV treatment. Patients were divided into failure (n=53) and success (n=89) according to NIV outcomes within 24 h of treatment. Hypersensitive C-reactive protein (hs-CRP), albumin (Alb), hs-CRP/Alb ratio and clinical dataof the two groups were compared. Multivariate logistic regression analysis was used to use the independent risk factors for NIV treat-ment failure, and the independent risk factors were introduced into R software to construct a risk model and evaluate the model.Re. sults The values of hs-CRP and hs-CRP/Alb in the failure group (2.14±0.50 vs. 0.53±0.12) were higher than those in the success group, while the Alb level was lower than that in the successful group (all P<0.05). Compared with the successful group, the respiratoryrate, diastolic blood pressure, acute physiology and chronic health evaluation Ⅱ (APACHEⅡ) score [(22.48±4.17) points vs. (18.72±3.91) points], sequential organ failure assessment (SOFA) score, white blood cell count and the partial pressure of carbon dioxide (PaCO2)level [(76.11±12.22) mmHg vs. (58.83±9.92) mmHg] before and 1 h after treatment were higher in the failure group, and the level of oxy-gen partial pressure (PaO2), oxygenation index before treatment, pH before and 1 h after treatment (7.29±0.18 vs. 7.40±0.21) were lower (all P<0.05). The multivariate logistic regression model shows that High hs-CRP/Alb value (OR=2.75), high APACHE Ⅱ score (OR= 1.61), low pH value (OR= 1.68) and high PaCO2 (OR=2.45) at 1 hour of treatment all affected independent risk factors for NIV treat-ment failure (P<0.05). The prediction model equation for NIV treatment failure was established based on independent risk factors: Logit (P)=.0.06+0.47×APACHEⅡ score+0.52×pH value at 1 h of treatment+0.90×PaCO2 at 1 h of treatment+1.01×hs-CRP/Alb value. Inter-nal validation using the Bootstrap method showed that the predictive model had good discrimination. The calibration curve showed thatthe model fits well. The constructed logistic prediction model was retrospectively validated. The results showed that the area under thereceiver operating characteristic (ROC) curve (AUC) for predicting the failure of NIV treatment was 0.93, and the predictive efficacy ofthe prediction model was superior to that of the hs-CRP/Alb value alone. Conclusions High hs-CRP/Alb value, high APACHE Ⅱ score, low pH value and high PaCO2 at 1 h after treatment are independent risk factors for NIV treatment failure in patients with AECOPDand type Ⅱ respiratory failure. The risk prediction model constructed by introducing independent risk factors has good differentiationand fit, which has high reference value for NIV treatment outcomes. |
|
查看全文
查看/发表评论 下载PDF阅读器 |
| 关闭 |
|
|
|
