The use of antidepressants has the remarkable effect of significantly reducing both the duration and symptoms of major depressive episodes. However, there were numerous reports and general consensus since the 60s indicating that the use of antidepressants can induce a manic/hypomanic episode, shortening the cycling length in bipolar disorders, or inducing a rapid-cycling bipolar disorder. We reviewed the data in the literature on this topic. But we were surprised to find that there were no clear conclusions, nor were there large scaled, prospective studies. Thus, in the current paper, we sorted out from the past literature the evidence supporting and against this point of view. We also pointed out the limitations of previous studies, and proposed several possible solutions for further research.