A prediction model to predict in-hospital mortality in patients with acute type B aortic dissection

BACKGROUND: Acute type B aortic dissection (ABAD) is a life-threatening cardiovascular disease. A practicable and effective prediction model to predict and evaluate the risk of in-hospital death for ABAD is required. The present study aimed to construct a prediction model to predict the risk of in-hospital death in ABAD patients. METHODS: A total of 715 patients with ABAD were recruited in the first affiliated hospital of Xinjiang medical university from April 2012 to May 2021. The information on the demographic and clinical characteristics of all subjects was collected. The logistic regression analysis, receiver operating characteristic (ROC) curve analysis, and nomogram were applied to screen the appropriate predictors and to establish a prediction model for the risk of in-hospital mortality in ABAD. The receiver operator characteristic curve and calibration plot were applied to validate the performance of the prediction model. RESULTS: Of 53 (7.41%) subjects occurred in-hospital death in 715 ABAD patients. The variables including diastolic blood pressure (DBP), platelets, heart rate, neutrophil-lymphocyte ratio, D-dimer, C-reactive protein (CRP), white blood cell (WBC), hemoglobin, lactate dehydrogenase (LDH), procalcitonin, and left ventricular ejection fraction (LVEF) were shown a significant difference between the in-hospital death group and the in-hospital survival group (all P < 0.05). Furthermore, all these factors which existed differences, except CRP, were associated with in-hospital deaths in ABAD patients (all P < 0.05). Then, parameters containing LVEF, WBC, hemoglobin, LDH, and procalcitonin were identified as independent risk factors for in-hospital deaths in ABAD patients by adjusting compound variables (all P < 0.05). In addition, these independent factors were qualified as predictors to build a prediction model (AUC > 0.5, P < 0.05). The prediction model was shown a favorable discriminative ability (C index = 0.745) and demonstrated good consistency. CONCLUSIONS: The novel prediction model combined with WBC, hemoglobin, LDH, procalcitonin, and LVEF, was a practicable and valuable tool to predict in-hospital deaths in ABAD patients.