Zhongguo linchuang yanjiu (May 2024)

Preliminary construction of a prediction model for HBV-related acute-on-chronic liver failure combined with acute kidney injury

  • LIU Chunhui, DAI Liwei, YU Haitao

DOI
https://doi.org/10.13429/j.cnki.cjcr.2024.05.020
Journal volume & issue
Vol. 37, no. 5
pp. 751 – 755

Abstract

Read online

Objective To establish a model that can predict the occurrence of acute kidney injury (AKI) in elderly patients with HBV-related acute-on-chronic liver failure (HBV-ACLF), and to conduct a preliminary evaluation of the performance of the model. Methods A total of 276 patients with HBV-ACLF hospitalised in The First Affiliated Hospital of Harbin Medical University from January 2020 to January 2023 were retrospectively included and divided into 72 (26.09%) in the AKI group and 204 (73.91%) in the non AKI group, according to whether AKI occurred during hospitalisation. The clinical data of all patients were extracted and screened for independent risk factors for AKI during hospitalisation using multivariate logistic regression. A prediction model was constructed accordingly, and then the efficacy of the prediction model was evaluated using ROC curves. Results Multivariate logistic regression analysis showed that age ＞70 years (OR=1.404, 95%CI: 1.134～1.737), procalcitonin＞1 ng/L (OR=1.473, 95%CI: 1.074～2.019), model for end stage liver disease (MELD) score ＞34 (OR=1.702, 95%CI: 1.254～2.311), and combined upper gastrointestinal bleeding (OR=1.516, 95%CI: 1.123～2.047) were independent risk factors for the occurrence of AKI in HBV-ACLF patients during hospitalisation (P＜0.05). A prediction model for the occurrence of AKI in HBV-ACLF patients during hospitalisation was established based on the above four parameters, and the ROC curve showed that the AUC predicted by the model was 0.882 (95%CI: 0.838～0.917), with a sensitivity of 66.67%, a specificity of 93.14% and an accuracy of 86.23%. Conclusion Elderly, high MELD score and procalcitonin level, and concomitant upper gastrointestinal bleeding are independent risk factors for AKI in patients with HBV-ACLF during hospitalisation, and the model constructed accordingly can predict the risk of AKI, thus assisting clinical disease management.

Keywords