Development and validation of a prediction model for respiratory failure in patients with sepsis associated acute kidney injury within 48 hours of admission
10.3760/cma.j.cn115455-20250508-00403
- VernacularTitle:脓毒症伴急性肾损伤患者入院48小时内呼吸衰竭风险预测模型的建立与验证
- Author:
Bin WANG
1
;
Fengxiang ZHANG
Author Information
1. 锦州医科大学附属第一医院重症医学科，锦州 121000
- Publication Type:Journal Article
- Keywords:
Sepsis;
Acute kidney injury;
Respiratory insufficiency;
Forecasting;
Risk factors
- From:
Chinese Journal of Postgraduates of Medicine
2025;48(10):894-900
- CountryChina
- Language:Chinese
-
Abstract:
Objective:To explore the risk factors of respiratory failure within 48 hours of admission in patients with sepsis associated acute kidney injury(SA-AKI)and establish a predictive model and verify it.Methods:A retrospective selection was made of 702 patients with SA-AKI admitted to Dongyang People's Hospital from June 2012 to October 2024, and they were randomly divided into the training set (492 cases) and the validation set (210 cases) in a ratio of 7∶3. The risk factors of respiratory failure within 48 h of admission in patients with SA-AKI were analyzed in the training set to establish a nomogram. The discriminative ability of the model was evaluated by using the receiver operating characteristic (ROC) curve, and the clinical effectiveness of the predictive model was evaluated by using decision curve analysis (DCA). Meanwhile, validation was conducted in the validation set. The Sequential Organ Failure Assessment (SOFA) and National Early Warning Score (NEWS) models were established, and the Delong test was applied to compare them with this prediction model.Results:The results of Logistic regression analysis showed that lactic acid, D-dimer, pro-B-type natriuretic peptide precursor, albumin, globulin, percutaneous blood oxygen saturation and pulmonary infection were independent risk factors for respiratory failure within 48 h of admission in patients with SA-AKI ( P<0.05). The results of ROC curve analysis indicated that the area under the curve (AUC) of this model for predicting respiratory failure within 48 h of admission in SA-AKI patients in the training set was 0.818 (95% CI 0.777 - 0.860), and that in the validation set was 0.795 (95% CI 0.723 - 0.860). The calibration curves showed that the P values were 0.973 and 0.864 respectively. The DCA curve was applied to evaluate the clinical effectiveness. The model curves were above the two extreme curves in both the training set and the validation set suggested that the model had good significance in discrimination, calibration and clinical effectiveness. The AUC of the SOFA model was 0.583 in the training set and 0.628 in the validation set. The AUC of the NEWS model was 0.601 in the training set and 0.618 in the validation set. The Delong test suggests that in both the training set and the validation set, compared with the SOFA and NEWS models, this prediction model had advantages in discrimination ability ( P<0.01). Conclusions:The nomogram model based on lactic acid, D-dimer, B-type brain natriuretic peptide precursor, albumin, globulin, percutaneous blood oxygen saturation and pulmonary infection can effectively predict the risk of respiratory failure within 48 h after admission in patients with SA-AKI.
