Treatment-induced manic switch in the course of unipolar depression can predict bipolarity: Cluster analysis based evidence

Background: Antidepressants are known to induce manic switch in patients with depression. Treatment-induced mania is not considered as bipolar disorder in DSM IV. The aim of this study was to assess whether clinical characteristics of patients with unipolar depression with a history of treatment-induced mania were similar to those of patients with bipolar disorder. Method: The study included 217 consecutive patients with DSM-IV mood disorders, diagnosed as: bipolar disorder type I (BP-I, n = 58) or type II (BP-II, n = 18) whose first episodes were depression, recurrent (unipolar) major depressive disorder with a history of antidepressant treatment-induced mania (switchers = stir); n = 61) and without such an event (rUD; n = 80). First, the groups were compared with regard to clinical features and course specifiers using variance and chi-square analysis. Variables that differed significantly between the four groups were included in two-step cluster analysis to explore naturally occurring subgroups in all diagnoses. Subsequently, the relationship between the naturally occurring clusters and pre-defined DSM-IV diagnoses were investigated. Results: Two-step cluster analysis revealed two different naturally occurring groups. Higher severity of depressive episodes, with higher rate of melancholic features, higher number of hospitalization and suicide attempts were represented in one cluster where switchers (77%), bipolar I (94.8%) and II (83.3%) patients clustered together. Conclusion: The findings of this study confirm that treatment-induced mania is a clinical phenomenon that belongs within-the bipolar spectrum rather than a coincidental treatment complication, and that it should be placed under "bipolar disorders" in future classification systems. Limitations: The study includes the limitations of any naturalistic retrospective study. (C) 2011 Elsevier B.V. All rights reserved.