Holter monitoring, electrocardiographic (ECG) signal-averaging, body surface potential mapping (BSPM) for PQRST isoarea maps, and electrophysiologic study (EPS) were performed in 100 patients with syncope. Coronary artery disease (CAD) was found in 46 patients and other heart disease was found in 19. EPS was diagnostic in 44 patients, while Holter monitoring suggested a diagnosis in only 21 patients. Abnormal BSPM was frequently seen (56%), especially in CAD (70%), or with inducible ventricular tachycardia (VT) (87%). Late potentials were recorded in 13 patients with CAD; five had inducible VT. In seven other patients with VT, they were either absent or bundle branch block (BBB) was found. Thirteen deaths (three sudden) occurred in our series. EPS-guided therapy resulted in a low rate of total cardiac death. In conclusion, EPS had a higher diagnostic yield than Holter monitoring regardless of cardiac pathology. ECG signal-averaging was useful in predicting VT only in patients with CAD without BBB. BSPM was abnormal in most patients with cardiac disease, but poorly predicted VT.