The incidence and the diagnostic and prognostic implications of recurrences in 433 patients enrolled in a prospective syncope study are reported. Over a mean follow-up of 30 months, 146 patients had recurrent syncope. Patients with an initial diagnosis of a cardiovascular cause of syncope had a recurrence rate of 31 percent, patients with a noncardiovascular cause had a recurrence rate of 36 percent, and patients with syncope of unknown origin had a recurrence rate of 43 percent at three years (these differences were not significant; the minimum for any two-way comparison was p greater than or equal to 0.11). In eight of the 191 patients in whom a cause of syncope could not be found on initial evaluation, a diagnosis was assigned in follow-up after recurrent syncope. Recurrences led to major morbidity in eight of 146 patients (5 percent) and minor trauma in 10 patients (7 percent). Using recurrence as a time-dependent variable in the Cox models, it was found that this variable was not a significant predictor of overall mortality or sudden death. It is concluded that recurrences are common in patients with syncope, but new diagnosis are rarely established on the basis of evaluation of recurrences. Recurrences are not predictors of mortality or sudden death.