The prevalence, incidence, natural history, and risk factors of upper dyspepsia (UD) were evaluated in a random sample of Danes. Of the 4581 eligible subjects, 79% responded. Five years later 85% of the survivors attended a similar follow-up study. Depending on the frequency of the symptoms, the prevalence of UD was 13-54% among men and 15-47% among women, and the incidence was 5-22% and 6-25%, respectively. Five years later 26-69% of subjects with UD were free from symptoms. As an independent factor psychic vulnerability was most strongly related to prevalence and incidence of UD, whereas experience of problems, smoking, and body mass index were only weakly associated with prevalence of UD. A comparison of the populations identified by various definitions of UD showed that these populations had rather few subjects in common. In conclusion, UD occurs frequently and fluctuates in the general population. Only psychic vulnerability was strongly related to UD incidence. The syndrome, however, still needs to be verified as a disease entity.