The addition of an antidepressant to a mood stabilizer did not add any benefit for patients with bipolar depression, according to an NIMH-sponsored research project reported in April in the New England Journal of Medicine.1 In the multicenter study, adding an antidepressant to the regimen also did not increase affective switching to mania.
The study, from the STEP-BD (Systematic Treatment Enhancement Program for Bipolar Disorder) research collaborative, examined whether adjunctive use of either bupropion (Wellbutrin) or paroxetine(Drug information on paroxetine) (Paxil) with a mood stabilizer could increase the rate of durable recovery—defined as 8 consecutive weeks of euthymia—from bipolar depression.
Gary S. Sachs, MD, of Massachusetts General Hospital in Boston, the lead author of the report, and colleagues noted that antidepressants are commonly used as adjuncts for this purpose, "despite limited evidence of the short-term and long-term efficacies and the putative risk of treatment-emergent mania or hypomania."1
This disparity may reflect what was referred to as a "division of opinion between the real world of treatment and the more refined world of evidence," in an earlier report from Harm Gijsman, PhD, Warneford Hospital, Oxford, United Kingdom, and colleagues that reviewed antidepressant trials and concluded that the agents are effective for bipolar depression.2 Judging the benefit and risk of an antidepressant for bipolar depression remains "one of the major current challenges in psychiatric management," the authors declared.
Practice guidelines in the United States place mood stabilizers as first-line treatment for bipolar depression before an antidepressant adjunct is considered, and then they recommend maintaining treatment with a mood stabilizer to mitigate possible antidepressant-induced mania or increased cycling.3 Studies have differentiated antidepressants by their liability to cause affective switching, including an NIMH and Stanley Medical Research Institute-sponsored study that associated higher switch rates with venlafaxine (Effexor) than with either bupropion or sertraline(Drug information on sertraline) (Zoloft).4
Antidepressants have been found by others to be beneficial for bipolar depression, however, and without substantial risk for inducing emergence of mania.2,5 They are commonly prescribed for bipolar depression, particularly outside North America. In addition, there are data relating risk for antidepressant-induced mania to patient characteristics rather than medication properties: less treatment-emergent mania has been found in bipolar type II patients than in type I patients6; and the risk may be influenced by whether the "index episode" before antidepressant treatment is manic or depressive.7
Assessing antidepressant adjunct in practiceWith the efficacy of antidepressants for bipolar depression in question and the assumption that they induce mania and/or hasten affective cycling unproved by placebo-controlled studies, the STEP-BD collaborative offers the largest randomized, placebo-controlled study to date to examine the effect of adding an antidepressant to a mood stabilizer to treat bipolar depression.
The study was designed to reflect many of the conditions encountered in clinical practice, so that the results could be better generalized than those from more narrowly structured clinical trials. Among the design elements was "equipoise-stratified randomization," which allowed patients to opt out of a treatment that they preferred not to receive. Study criteria allowed inclusion of patients with either bipolar type I or type II disorders, as well as coexisting anxiety disorders, substance abuse disorders, and psychotic symptoms, and patients could continue prestudy pharmacotherapy and psychotherapy. The mood stabilizer used could be lithium(Drug information on lithium), valproate(Drug information on valproate), or the combination; or carbamazepine(Drug information on carbamazepine) and, in the last year of the approximately 5-year recruitment period, any FDA-approved antimanic agent.
The study enrolled 366 patients from the STEP-BD study population of more than 4000 patients at 22 sites, with more than 2600 having at least 1 depressive episode while in the study. In an editorial accompanying the study report, Robert H. Belmaker, MD, of Ben Gurion University and Beersheva Mental Health Center in Israel, suggested that the relatively small proportion of STEP-BD participants recruited to this study could be a source of bias; those who had a history of severe mania or mania following use of an antidepressant were directed by their usual care providers, with whom they continued throughout the STEP-BD study, to not participate.8
Sachs and colleagues acknowledged this limitation: "Our results are likely to be applicable only to those patients with bipolar depression who are considered appropriate candidates for treatment with standard antidepressants."
Patients receiving maintainance mood stabilizer therapy were randomized among 3 strata: placebo versus bupropion, placebo versus paroxetine, and placebo versus either antidepressant. This allowed for inclusion of patients who had a clear preference for a given antidepressant. Paroxetine was started at 10 mg/d and increased to a maximum of 40 mg/d. Sustained- release bupropion was initiated at 150 mg/d, and increased to a maximum of 375 mg/d. In addition to the primary outcome measure of attaining and maintaining euthymia, treatment effectiveness was gauged by a mood rating score improvement of 50% or more with absence of mania or hypomania.
The investigators found no significant differences among the groups in treatment effectiveness, but noted that the mood stabilizer with placebo appeared to trend better than with an antidepressant adjunct. Durable recovery was attained by 23.5% of those receiving an antidepressant adjunct and by 27.3% of those receiving a mood stabilizer alone. Treatment response criteria were met by 32.4% of those receiving an antidepressant and by 38% of those not receiving an antidepressant.
Although not effective, the antidepressants were well tolerated. Treatment-emergent affective switch occurred in 10% of both the antidepressant/mood stabilizer and placebo/mood stabilizer groups. The rates and types of adverse reactions were also similar in the antidepressant and placebo groups; and the 12.3% of the antidepressant/ mood stabilizer group discontinuing the study because of adverse effects was not statistically significantly different from the 9.1% discontinuing after receiving placebo/mood stabilizer.
Belmaker commented, "These findings widen the gap between North America and Europe regarding the efficacy of antidepressants . . . [for] bipolar disorder but narrow the gap regarding the danger of mania."8
Sachs and colleagues speculate that the greater antidepressant benefit found in the meta-analysis by Gijsman and colleagues may reflect, in part, narrower inclusion criteria in the reviewed studies. The absence of antidepressant-induced mania may also be a function of study design, Sachs and colleagues caution, noting that higher switch rates could occur with antidepressants other than bupropion and sertraline. The study also did not assess antidepressant effect in the absence of a mood stabilizer.
In Belmaker's opinion, the results from this STEP-BD report will not resolve the opposing views on the benefits and risks of antidepressants for bipolar depression. While anticipating that the study report will reduce the use of antidepressants for this condition, he expects use of the agents to continue as treatment is appropriately individualized. "A patient's clinical history and past response to treatment will probably remain important factors in treatment decisions," Belmaker offered.
