Development and validation of a predictive model for acute kidney injury in young, middle-aged and elderly patients with acute ischemic stroke
10.3760/cma.j.issn.1673-4165.2023.11.003
- VernacularTitle:年轻和中老年急性缺血性卒中患者急性肾损伤预测模型的构建和验证
- Author:
Beibei AN
1
;
Yan XU
;
Jiayu ZHANG
;
Ping LIU
;
Jie YU
;
Qingrong PENG
Author Information
1. 徐州医科大学附属连云港医院急诊科，连云港 222002
- Keywords:
Ischemic stroke;
Acute kidney injury;
Risk factors;
Predictive value of tests;
Young adult;
Middle aged adult;
Elderly
- From:
International Journal of Cerebrovascular Diseases
2023;31(11):815-823
- CountryChina
- Language:Chinese
-
Abstract:
Objective:To develop a risk prediction model for acute kidney injury (AKI) in young, middle-aged and elderly patients with acute ischemic stroke (AIS), and verify the predictive ability of the model.Methods:Patients admitted to the Department of Neurology, Lianyungang Hospital Affiliated to Xuzhou Medical University from January 2018 to August 2022 were retrospectively included as a modeling cohort, and patients with AIS from September 2022 to September 2023 were prospectively included as a validation cohort. Independent risk factors for AKI were determined by multivariate logistic regression analysis, and risk prediction models for AKI in young AIS patients group and middle-aged and elderly AIS patients group were developed. The predictive power of the model was tested using the receiver operating characteristic (ROC) curve. Results:The young group included 760 patients with AIS, of which 584 (76.84%) were in the modeling cohort, and 146 (25.00%) had AKI. Multivariate logistic regression analysis showed that anemia, systolic blood pressure, homocysteine, alcohol consumption, blood urea nitrogen, and National Institutes of Health Stroke Scale (NIHSS) score were independent risk factors for AKI (all P<0.05). ROC analysis showed that the area under the curve of the predictive model was 0.938 (95% confidence interval 0.912-0.963), the Youden index was 0.747, and the optimal cut-off value was 0.249. The sensitivity and specificity of predicting AKI were 84.8% and 89.9%, respectively. A total of 1 417 patients with AIS were included in the middle-aged and elderly group, of which 833 patients (58.79%) were in the modeling cohort and 230 (27.61%) had AKI. Multivariate logistic regression analysis showed that hypertension, atrial fibrillation, previous stroke history, smoking, infection, triglycerides, NIHSS score, use of antihypertensive drugs, use of loop diuretics, serum creatinine, and blood urea nitrogen were the independent risk factors for AKI ( P<0.05). ROC analysis showed that the area under the curve of the predictive model was 0.838 (95% confidence interval 0.808-0.868), the Youden index was 0.539, the optimal cut-off value was 0.242, and the predictive sensitivity and specificity were 78.3% and 75.6%, respectively. The Hosmer-Lemeshow goodness of fit test showed the predictive accuracy of the model was in good agreement with the actual occurrence of risk (the young group: χ2=8.968, P=0.345; the middle-aged and elderly group: χ2=11.250, P=0.188). The validation cohort analysis shows that the model had high prediction accuracy and credibility in two groups. Conclusion:The model can specifically predict the risk of AKI in young, middle-aged and elderly patients with AIS, with high sensitivity and specificity.
