Background: Decades of research now link TWA with inducible and spontaneous clinical ventricular arrhythmias. This bench-to-bedside foundation makes TWA, NT-ProBNP a very plausible index of susceptibility to ventricular arrythmia, and motivates the need to define optimal combination of TWA and NT-ProBNP in predicting ventricular arrythmia in myocardial infarction patients. We research this study with 2 targets: 1. To evaluate the role of TWA in predicting sudden cardiac death in myocardial infarction patients. 2. To evaluate the role of NT-ProBNP in predicting sudden cardiac death in myocardial infarction patients 3. Evaluate the role of the combined NT-ProBNP and TWA in predicting sudden cardiac death in myocardial infarction patients. Methods: Prospective study with follow up the mortality in 2 years: 71 chronic myocardial infarction patients admitted to hospital from 5/2009 to 5/20011 and 50 healthy person was done treadmill test to caculate TWA; ECG, echocardiography, NT-ProBNP. Results: Cut-off point of NT-ProBNP in predicting sudden cardiac death is 3168 pg/ml; AUC = 0,86 (95% CI: 0,72 - 0,91); Cut-off point of TWA in predicting sudden cardiac death is 107 µV; AUC = 0,81 (95% CI: 0,69 - 0,87); NT-ProBNP can predict sudden cardiac death with OR= 7,26 (p<0,01); TWA can predict sudden cardiac death with OR= 8,45 (p<0,01). The combined NT-ProBNP and TWA in predicting ventricular arrythmia in heart failure patients: OR= 17,91 (p<0,001). Conclusions: The combined NT-ProBNP and TWA have the best predict value of sudden cardiac death in myocardial infarction patients, compare to NT-ProBNP or TWA alone