Abstract: Objective: To observe the clinical features of patients with severe acute myocardial infarction (AMI) and explore the correlation factors associated with in-hospital death of the severe AMI and then construct and analyze the predicting model. Methods: one hundred and thirty-two patients presenting with acute myocardial infarction of Killip class Ⅲ and Ⅳ were enrolled in The Second Hospital of Tianjin Medical University between April 2006 and April 2010. The clinical data of all those patients were collected. Univariate and multivariate analysis were performed to evaluate the correlation factors associated with in-hospital death of the severe AMI and logistic regression model was constructed. Analyze the model with the ROC curves. Results: The total mortality of 132 patients was 51.5 %. Compared with the improved group, the death group was associated with a higher proportion of cardiogenic shock and arrhythmias. Glomerular filtrate rate (eGFR), systolic blood pressure (SBP) were significantly lower in the death group than in the control group, While blood urea nitrogen (BUN) , serum creatinine (CREA), white blood cell (WBC) count, the MB fraction of creatine kinase (CK-MB) and cardiac troponin I (CTnI) were higher in the death group than in the improved group. Multi-factorial logistic regression analysis showed that anterior and lateral myocardial infarction , arrhythmias and blood urea nitrogen might be independent risk factors of in-hospital death. The AUC values generated by the ROC curves for the logistic regression equation was 0.945. Conclusion: Severe Acute Myocardial Infarction had high mortality rate. and the predictive value of the model is high.

Key words: severe acute myocardial infarction, in-hospital death, predictors, ROC curves