A review of the methodology and results of 9 controlled studies on the acute treatment of bipolar depression and the risk of switches into (hypo)mania is presented. There are indications but no proof for efficacy of mood stabilizers such as lithium, carbamazepine and valproate. Only lamotrigine has been shown to be effective, with a relative low risk of switching. Several antidepressants appear effective as well, but again there is no (placebo-controlled) proof of their efficacy. The only exception is tranylcypromine which has been found to be more effective than imipramine. The switch ratio into (hypo)mania by tricyclic antidepressants seems to be higher than by several other antidepressants, especially the selective serotonin reuptake inhibitors. In the acute treatment of bipolar depression, it is recommended to start with a mood stabilizer, and to add an antidepressant after 4-6 weeks in case of nonresponse. In severer cases, one might consider to start earlier with the combination of a mood stabilizer and an antidepressant, and in refractory patients, there is a place for tranylcypromine. In the maintenance treatment, there are indications that the combined treatment of a mood stabilizer (mostly lithium) and an antidepressant (TCA) is associated with an increased risk of switches into (hypo)mania, when compared to a mood stabilizer alone. Therefore, it is recommended to try whether a monotherapy with a mood stabilizer is effective, before combining it with an antidepressant.