We prospectively evaluated and followed 204 patients with syncope to determine how often a cause of syncope could be established and to define the prognosis of such patients. A cardiovascular cause was established in 53 patients and a noncardiovascular cause in 54. The cause remained unknown in 97 patients. At 12 months, the overall mortality was 14 +/- 2.5 per cent. The mortality rate (30 +/- 6.7 per cent) in patients with a cardiovascular cause of syncope was significantly higher than the rate (12 +/- 4.4 per cent) in patients with a noncardiovascular cause (P = 0.02) and the rate (6.4 +/- 2.8 per cent) in patients with syncope of unknown origin (P less than 0.0001). The incidence of sudden death was 24 +/- 6.6 per cent in patients with a cardiovascular cause, as compared with 4 +/- 2.7 per cent in patients with a noncardiovascular cause (P = 0.005) and 3 +/- 1.8 per cent in patients with syncope of unknown origin (P = 0.0002). Patients with syncope can be separated into diagnostic categories that have prognostic importance. Patients with a cardiovascular cause have a strikingly higher incidence of sudden death than patients with a noncardiovascular or unknown cause.