Brugada syndrome is characterized by distinguishing electrocardiogram (ECG) patterns (coved and saddle-back types with day-to-day variation) and occurrence of lethal tachy-arrhythmias. The appearance of coved type ECG (type 1) is required for the diagnosis of Brugada syndrome, whereas the significance of saddle-back type ECG (type 2), which is inadequate for the diagnosis, has not been fully established. We enrolled 34 consecutive patients with type 2 ECG on outpatient-clinic. Among them, 7 patients were ventricular fibrillation (VF) survivors who were diagnosed as Brugada syndrome with transient appearance of type 1 ECG, and showed type 2 ECG on their first outpatient-clinic visit after the VF event (VF group). The remaining 27 were asymptomatic and never showed type 1 ECG on repeated ECG examinations (control group). The VF group showed significantly longer RJ intervals in leads V1 and V2 and QTc intervals in lead V2 compared with the control group (P < 0.030, P < 0.017, and P < 0.030, respectively). Late potentials, detected on the signal-averaged ECG (SA-ECG), reflect conduction abnormalities and are known as one of the risk markers of arrhythmic events. Among the 34 patients, late potentials were negative in 12 patients belonging to the control group. In conclusion, the SA-ECG could be helpful to identify high-risk patients for its high negative predictive value as the first step, and ECG parameters, including RJ intervals in leads V1 and V2 and QTc interval in lead V2, could be useful for further risk stratification in patients with type 2 Brugada ECG.