We studied 433 patients with syncope to derive insights into the diagnostic evaluation and outcome of patients with this common problem. This study shows that the etiology of syncope was not found in approximately 41% of patients. When a cause of syncope was determined, it was most frequently established on the basis of initial history, physical examination and an electrocardiogram (EKG). Furthermore, many of the other entities (e.g., aortic stenosis, subclavian steal) were suggested by findings on the history and physical examinations that required directed diagnostic testing. Initial EKG was abnormal in 50% of patients but led to a cause of syncope infrequently (less than 7%). Prolonged electrocardiographic monitoring, which has assumed a central role in the evaluation of syncope, led to a specific cause in only 22% of patients. Other tests were less often helpful in assigning a cause of syncope. At 5 years, the mortality of 50.5% in patients with a cardiac cause of syncope was significantly higher than the 30% mortality in patients with a noncardiac cause or 24.1% in patients with an unknown cause. At 5 years, a mortality of 50.5% in patients with a cardiac cause of syncope was noted. There were 54 actual deaths in this group as compared to 10.7 expected deaths based on 1980-86 mortality data from Allegheny County, PA (standardized mortality ratio = 5.02). At 5 years, a 33.1% incidence of sudden death was noted in patients with cardiac cause of syncope, as compared with 4.9% in patients with a noncardiac cause and 8.5% in patients with an unknown cause. Mortality and sudden death remained significant for the first 3 years after which the survival curves were parallel. A cardiac cause of syncope was an independent predictor of sudden death and mortality. Recurrences were common but were not associated with an increased risk of mortality or sudden death. Major vascular events were also more frequent in patients with cardiac causes of syncope. The results of this study will be helpful in designing future studies to evaluate the usefulness of newer diagnostic techniques. Furthermore, short- and long-term outcome data will be useful in planning intervention strategies in these patients.