A significant proportion of patients with unipolar depression clinically develop manic or hypomanic switch during acute antidepressant treatment. Elucidation of its prevalence and predicting factors is of clinical relevance during acute antidepressant treatment of such patients. We retrospectively studied patients with unipolar depression who were admitted to our department during the 6-year period from 1997 to 2002 and who had fewer than 3 previous episodes before admission. The clinical background of the consecutive patients with manic/hypomanic switch (n = 37) who met Diagnostic and Statistical Manual of Mental Disorders, Fourth Edition criteria was compared with that of patients without manic/hypomanic switch (n = 245). The prevalence rate of manic/hypomanic switch was 13.1%. The switch group was composed of 23 men and 14 women, whose average age was 48.8 +/- 12.3 years (range, 26-78 years). Manic/hypomanic switch was most frequently observed between 2 and 3 weeks after the antidepressant was increased to the ongoing dose. Antidepressants were decreased in 13 patients and discontinued in 23. Manic/hypomanic episodes lasted from 1 to 8 weeks. The patients in the switch group included a greater proportion of male subjects and had a higher frequency of family history of bipolar disorders than those in the nonswitch group. The mean doses of antidepressants were not significantly different between these groups. The higher frequency of manic/hypomanic switch occurring around the period when antidepressants begin to show clinical effects and the higher frequency of family history of bipolar disorders might suggest a biological susceptibility to antidepressants in patients of the switch group.