Development and Validation of a Prognostic Model to Predict the Risk of In-hospital Death in Patients With Acute Kidney Injury Undergoing Continuous Renal Replacement Therapy After Acute Type a Aortic Dissection

BACKGROUND: This study aimed to construct a model to predict the risk of in-hospital death in patients with acute renal injury (AKI) receiving continuous renal replacement therapy (CRRT) after acute type A aortic dissection (ATAAD) surgery. METHODS: We reviewed the data of patients with AKI undergoing CRRT after ATAAD surgery. The patients were divided into survival and nonsurvival groups based on their vital status at hospital discharge. The data were analyzed using univariate and multivariate logistic regression analyses. Establish a risk prediction model using a nomogram and its discriminative ability was validated using C statistic and the receiver operating characteristic (ROC) curve. Its calibration ability was tested using a calibration curve, 10-fold cross-validation and Hosmer–Lemeshow test. RESULTS: Among 175 patients, in-hospital death occurred in 61 (34.9%) patients. The following variables were incorporated in predicting in-hospital death: age > 65 years, lactic acid 12 h after CRRT, liver dysfunction, and permanent neurological dysfunction. The risk model revealed good discrimination (C statistic = 0.868, 95% CI: 0.806–0.930; a bootstrap-corrected C statistic of 0.859, the area under the ROC = 0.868). The calibration curve showed good consistency between predicted and actual probabilities (via 1,000 bootstrap samples, mean absolute error = 2.2%; Hosmer–Lemeshow test, P = 0.846). The 10-fold cross validation of the nomogram showed that the average misdiagnosis rate was 16.64%. CONCLUSION: The proposed model could be used to predict the probability of in-hospital death in patients undergoing CRRT for AKI after ATAAD surgery. It had the potential to assist doctors to identify the gravity of the situation and make the targeted therapeutic measures.