AIMS: The present study aimed to assess whether there are differences in risk indicators for perpetuating ventricular tachyarrhythmias (pVT) and self-terminating ventricular tachyarrhythmias (stVT). METHODS AND RESULTS: Patients with acute myocardial infarction (AMI) and baseline left ventricular ejection fraction </= 40% (n = 292) received an implantable electrocardiogram loop recorder from 5 to 21 days after AMI and were followed up for 24 months to document arrhythmic events in the Cardiac Arrhythmias and Risk Stratification after Acute Myocardial Infarction (CARISMA) study. Several risk markers, such as the inducibility to sustained ventricular tachycardia during programmed electrical stimulation (PES), the signal-averaged ECG QRS duration (SAECG-QRS), heart rate variability (HRV) and turbulence (HRT), T-wave alternans, and non-sustained ventricular tachycardia on Holter were analysed at 6 weeks after the AMI. During the follow-up, 26 patients (9%) experienced an stVT (>/= 16 beats and < 30 s), and 21 patients (7%) a pVT. The occurrence of non-sustained ventricular tachycardia on Holter significantly predicted stVT [hazard ratio (HR) = 2.90, 1.26-6.67, 95% confidence interval (CI), P = 0.01], but not pVT during the follow-up. The inducibility during PES (HR = 5.02, 1.85-13.60, 95% CI, P = 0.001), SAECG-QRS >/= 130 ms (HR = 8.73, 3.38-22.56, 95% CI, P < 0.001), the short-term scaling exponent HRV parameter </= 0.77 (HR = 5.65, 2.12-15.10, 95% CI, P = 0.001), and HRT slope </= 1.75 ms/NN (HR = 4.57, 1.80-11.59, 95% CI, P = 0.001) were significant predictors of pVT, even after adjustments with relevant clinical parameters (P from < 0.01 to < 0.001), but did not significantly predict the occurrence of stVT (P from 0.35 to 0.75). CONCLUSION: Self-terminating ventricular tachyarrhythmia and pVT have differences in electrophysiological substrate and arrhythmia modifiers in post-AMI patients with moderate left ventricular dysfunction.