Backgrounds: Use of implantable cardioverter-defibrillator(ICD) for prevention of sudden cardiac death(SCD) in heart failure is a well-established therapeutic modality. Depressed left ventricular ejection fraction (LVEF���30%) is the only criteria of ICD therapy for primary prevention and numbers needed to treat are still high. We sought to find adjunctive noninvasive parameters beyond LVEF for selection of ICD candidates in this study. Material and Methods: Thirty-six consecutive patients with ischemic or idiopathic cardiomyopathy who had ICD implantation were included. Clinical variables and parameters derived from pre-implantation 24hr ambulatory ECG were analyzed; heart rate variability, hear rate turbulence, presence of non-sustained ventricular tachycardia(NSVT), premature ventricular complex (PVC) counts(���10/hr) and T-wave alternans(TWA) using modified moving average method. The prognostic impact of all parameters on occurrence of appropriate ICD shock due to malignant ventricular arrhythmic event was tested. Results: During follow-up (mean, 23짹26 months), 10 VT, 2 VF and 2 SCD were observed in 13 patients (39.4%). The patients with NSVT on 24hr ambulatory ECG had more lethal arrhythmic event in unadjusted analysis (p=0.043). All parameters except NSVT were not associated with lethal arrhythmic event or SCD. However, multivariate analysis revealed that maximal TWA and incidence of NSVT were significant prognostic factors [Odd ratio = 3.65 (95% confidence interval 1.204 to 11.06), p=0.022; Odd ratio = 3.568 (95% confidence interval 1.046 to 12.169), p=0.042, respectively]. Conclusion: Both the presence of NSVT and maximal TWA were associated with occurrence of ICD shock due to lethal arrhythmic event. These noninvasive parameters would be adjunctive tool to select more appropriate ICD candidate. Key Words : Lethal arrhythmic events, heart failure, implantable cardioverter-defibrillator