We sought to determine the epidemiology and clinical significance of paroxysmal supraventricular tachycardia (PSVT) in the general population. Current knowledge of PSVT has been derived primarily from otherwise healthy patients referred to specialized centers. We used the resources of the Marshfield Epidemiologic Study Area, a region covering practically all medical care received by its 50,000 residents. A review of 1,763 records identified prevalent cases as of July 1, 1991 and all new cases of PSVT diagnosed from that day until June 30, 1993. A mean follow-up period of 2 years was completed in all incident patients. Patients without other cardiovascular disease were labeled as having "lone PSVT." The prevalence was 2.25/1,000 persons and the incidence was 35/100,000 person-years (95% confidence interval, 23 to 47/100,000). Other cardiovascular disease was present in 90% of males and 48% of females (p = 0.0495). Compared with patients with other cardiovascular disease, those with lone PSVT were younger (mean 37 vs. 69 years, p = 0.0002), had a faster PSVT heart rate (mean 186 vs. 155 beats/min, p = 0.0006) and were more likely to have their condition first documented in the emergency room (69% vs. 30%, p = 0.0377). The onset of symptoms occurred during the childbearing years in 58% of females with lone PSVT versus 9% of females with other cardiovascular disease (p = 0.0272). There are approximately 89,000 new cases/year and 570,000 persons with PSVT in the United States. In the general population, there are two distinct subsets of patients with PSVT: those with other cardiovascular disease and those with lone PSVT. Our data suggest etiologic heterogeneity in the pathogenesis of PSVT and the need for more population-based research on this common condition.