Strakowski et al indicate that bipolar depression may need to be treated differently from unipolar depression because in bipolar depression antidepressants may be ineffective and may precipitate the risk of a switch to (hypo)mania (and induction of rapid cycling). While this could be a major argument to discriminate bipolar depression form unipolar depression, a critical appraisal of the available literature is indicated. First, are antidepressants ineffective in bipolar depression? Compared to around 1,500 randomized controlled trials (RCTs) with antidepressants in unipolar depression, there are around 15 RCTs in bipolar depression and only four of them exceeded 50 patients per treatment arm. Therefore, the main conclusion is that, compared to unipolar depression, bipolar depression is heavily understudied 1. Nevertheless, in a meta-analysis, antidepressants as a group appeared to be effective 1. Of the four larger studies, two reported positive results. The first study found that the combination of fluoxetine and olanzapine was more effective than placebo plus olanzapine 2. The second study reported response rates of 50-60% when sertraline, bupropion or venlafaxine was added to an ongoing treatment with mood stabilizers 3. The limitation of this study was the lack of a placebo arm. On the other hand, in a third study, paroxetine was not more effective than placebo, while two doses of quetiapine (300 and 600 mg/day) did separate from placebo 4. Finally, in the STEP-BD study, sertraline or bupropion as add-on to ongoing treatment with (among other medications) lithium or valproate was not found more effective than the addition of placebo 5. Although this is the largest study which investigated the efficacy of antidepressants in bipolar depression, it has major methodological limitations (e.g., patients were allowed to continue with another antidepressant during the first two weeks of the study; they could also use other drugs with an antidepressant effect, such as quetiapine, olanzapine and lamotrigine; 69% of the patients participated at the same time in a study comparing three different forms of psychotherapy with treatment as usual). It is clear that the conclusion that antidepressants are not effective cannot be drawn on the basis of this study. My overall conclusion is that, due to a lack of well designed RCTs, we can only state that it has not (yet) been proven that antidepressants are effective in bipolar depression. Second, do antidepressants cause a switch into (hypo)mania and rapid cycling? In our meta-analysis 1, we did not find that in the acute treatment of bipolar depression antidepressants were more often associated with a switch into (hypo)mania than placebo. However, in all studies comparing different antidepressants, the treatment arms with a tricyclic (TCA) were associated with a greater risk of switch into (hypo)mania than the treatment arms with other antidepressants, suggesting that TCAs have a greater risk of switch. Concerning the risk associated with long-term antidepressant treatment, we argued that there is a scarcity of randomized studies, and that the available studies all suffer from various forms of bias 6. Nevertheless, we concluded that antidepressants, when combined with a mood stabilizer, seemingly do not induce a switch into hypomania or mania. Ghaemi et al 7 presented a meta-analysis of seven RCTs in which antidepressants were used for at least 6 months. Three of these RCTs (total n=50) compared the effects of antidepressants monotherapy with placebo; five (total n=246) compared antidepressants plus a mood stabilizer versus a mood stabilizer alone (or in combination with placebo); and three (total n=108) compared antidepressants alone with a mood stabilizer alone. In most of the studies the antidepressant was a TCA. When combining all RCTs, the antidepressants yielded a significant 27% lower risk of a depressive recurrence versus control treatment without an antidepressant, but also a significant 72% greater risk for a manic recurrence. However, in RCTs with an antidepressant alone versus placebo, the only significant result was fewer depressive recurrences with the antidepressant, while in RCTs with an antidepressant alone versus a mood stabilizer alone (lithium), the only significant result was fewer manic recurrences with lithium. Therefore, my conclusion is that antidepressants do protect against depressive recurrences, while lithium does protect against manic recurrences. Whether antidepressants may accelerate episode frequency and/or cause other forms of destabilization in patients with bipolar disorder remains to be properly studied. In conclusion, there is not enough evidence to conclude that bipolar depression needs to be treated differently from unipolar depression. Especially in bipolar II depression, antidepressants still have a role even in monotherapy, as also suggested by recent guidelines 8,9. Therefore, bipolar I depression and unipolar depression should be seen as the ends of a continuum, with arbitrary demarcations and bipolar II depression in between.
Read full abstract