Studies suggest empirically derived subtypes of nicotine dependence exist in young adult populations with short smoking careers. It is not known if classes of dependence exist in middle aged smokers with longer smoking careers and whether these classes reflect quantitative or qualitative differences. It is not known if psychiatric disorders are associated with classes of nicotine dependence. Nicotine dependence symptoms were obtained from a 1992 administration of the Diagnostic Interview Schedule. Latent Class Analyses (LCA) was computed using data from 5440 members of the Vietnam Era Twin Registry. LCA was used to derive significantly different classes of nicotine dependence, which were assessed for their association with smoking history, nicotine dependence, and other psychiatric disorders. The LCA model which best fit the data was a 4 class solution characterized by severity. Age onset of regular smoking decreased with more severe classes. Cigarette consumption, failed cessation and psychiatric disorders were associated with more severe classes. Empirically derived subtypes of nicotine dependence are mostly characterized by increasing severity. Suggestions for refinement of nicotine dependence diagnostic criteria are discussed.