Molecular genotyping studies often focus on clustered tuberculosis and recent transmission. Less attention has been paid to non-clustered tuberculosis. However, non-clustered cases also contribute significantly to the tuberculosis burden, especially in low-incidence countries. The objective of this study is to characterize non-clustered tuberculosis cases in Denmark and point out potential implications for tuberculosis control. The study is based on nationwide IS6110-RFLP genotyping of tuberculosis cases from 1992 through 2004, corresponding to 98% of culture verified cases. Of 3988 cases, 45% were non-clustered. Both Danes and immigrants had a peak incidence of non-clustered tuberculosis at older ages, 80-89 years (4.3 cases/10(5) population/year) and 60-69 years (28.8 cases/10(5) population/year), respectively. In addition, immigrants had a peak at 20-29 years (43.2 cases/10(5) inhabitants/year). In Danes, the incidence of non-clustered tuberculosis decreased during the study period and was predominantly found in elderly persons, presumably reactivating infection acquired during 1910-40, when tuberculosis incidence was high. In immigrants, the incidence was high at all ages, presumably reflecting reactivation of imported infections. In the future, the number of non-clustered tuberculosis cases will decrease, as older Danes die, and as time since primary infection increases for immigrants residing in Denmark. TB control should include focus on non-clustered cases.