Text version
Should you use antidepressants in a patient with bipolar depression? This is one of psychiatry’s most controversial questions. Debate continues on about just how often antidepressants precipitate hypomania or mania, induce rapid cycling, or cause mixed states. But an even more fundamental question is, do they work?
Hi! Jim Phelps here for the Psychopharmacology Institute. Several meta-analyses of antidepressant efficacy in bipolar depression are available, and they don’t agree. It looks like which studies you pick to enter into your meta-analysis determines what the results say. A recent one in Lancet Psychiatry 2016 concluded that antidepressants did not work significantly better than placebo in the short run and can worsen the course in the long run. Now comes one more randomized trial with the same conclusions, which should tip the meta-analytic balance scale toward greater doubt about the role of antidepressants in bipolar depression. This is crucial since community psychiatrists are still using antidepressants in nearly half their patients according to even relatively recent studies, as shown by a referenced study by Grande et al. last 2013.
So, let’s look at this new randomized trial by Nassir Ghaemi and colleagues which compare short- and long-term outcomes of citalopram and placebo in patients with bipolar depression. For the primary short-term outcome, which was a reduction in Montgomery-Asberg Depression Rating Scale (MADRS) scores over 6 weeks, citalopram was associated with a 14-point drop and placebo with a 12-point drop. That difference is not statistically significant nor does it reach usual standards of clinical significance.
Then there was a maintenance phase. Maybe the benefits of citalopram relative to placebo would show up over a year, but the study was already a little small, with a sample size of 46 completers in each group after that first 6-week phase. Over the year, they then lost two-thirds of those subjects, for a final sample size of 14 and 12 in the citalopram and placebo groups, respectively. Before quibbling about that, note that pharmaceutical clinical trials designed to get an FDA indication are only 8 weeks long or, lately, 6 weeks after some treatments lost separation from placebo in Weeks 7 and 8. So, no maintenance phase for them at all. That means that any data 1 year out are hard won and to be appreciated. For the remaining one-third of this study’s sample, then, the results were the same as in the acute phase—about a 2-point difference in the MADRS scores favoring citalopram but not statistically or clinically significant.
At this point, we’re down to a small enough group such that secondary analyses are pretty dubious, but it is of note that over a year in a rapid cycling subgroup, there were statistically significantly more mood episodes in the antidepressant continuation group, consistent with the concern that antidepressants can induce mood instability even with mood stabilizers on board. In other words, Ghaemi’s et al. randomized trial of citalopram in bipolar depression found no benefit relative to placebo.
What does this mean for clinicians? Well, to put this in a slightly broader context, compare the recommendations of 18 bipolar specialists regarding antidepressants in bipolar disorder compiled in 2019 by my coauthor Chris Aiken. Of the 18, 1 found antidepressants helpful. Ten would use them but only with a mood stabilizer. Six avoid them, using them only as a last resort, and 1 avoids them entirely. That’s a wide spread, although 17 out of 18 are expressing caution. Wait a minute. Caution? That’s a focus primarily on risk, not benefit. This new study from Ghaemi and colleagues questions benefit. If benefit is really no greater than placebo in the short or long term, then there is no role for antidepressants in the treatment of bipolar depression with or without a mood stabilizer. But remember, this is just 1 study. It might tip the meta-analytic balance scale toward greater doubt about any role for antidepressants in patients with bipolar I or II disorder. This study included both.
In summary, here’s my opinion, after having watched this debate for 20 years or more: A psychiatrist’s practice is not really very strongly determined by data. You’re the exception because you’re listening to this Quick Take. The few cases in which antidepressants are prescribed and the patient improves rapidly are likely to be enough for many clinicians to arrive at a conclusion that is not easily swayed by a new study or even a new meta-analysis. However, Dr. Ghaemi points out that the natural history of bipolar depression includes episode lengths of less than 6 months, often much shorter; so, a spontaneous recovery is very common. Witness the placebo group improvement in this study, for example. All that reminds us that, for those of us watching the literature at least, we should be cautious about reasoning from your own experience. It’s not useless, and it’s not always misleading, but it’s difficult to cross check.
For more on all this, see the 2016 meta-analysis by McGirr et al., which is included in the references.
Abstract
Citalopram for Acute and Preventive Efficacy in Bipolar Depression (CAPE-BD): A Randomized, Double-Blind, Placebo-Controlled Trial
S Nassir Ghaemi, Elizabeth A Whitham, Paul A Vohringer, Sergio A Barroilhet, Andrea Amerio, Oleksandr Sverdlov, Ashwin A Patkar
Objective: To assess the efficacy and safety of citalopram in the acute and maintenance phases of bipolar depression in a randomized, double-blind, placebo-controlled trial.
Methods: Between 2007 and 2014, 119 subjects with acute major depressive episodes diagnosed with DSM-IV bipolar disorder, type I or type II, were randomized blindly to citalopram or placebo, added to standard mood stabilizers. They were followed for 6 weeks for acute efficacy (primary outcome) and up to 1 year for maintenance efficacy (secondary outcome) using scores on the Montgomery-Asberg Depression Rating Scale (MADRS) and the Mania Rating Scale of the Schedule for Affective Disorders and Schizophrenia (MRS-SADS). The study was powered for a clinically meaningful effect size.
Results: Mean ± SD MADRS scores changed from a baseline value of 27.4 ± 9.1 to 13.1 ± 8.4 at the end of the acute phase for citalopram versus a change from 27.4 ± 7.3 to 15.2 ± 9.9 for placebo, a clinically and statistically nonsignificant difference. Maintenance efficacy also was not better with citalopram than with placebo. Acute manic/hypomanic episodes were similar in both groups, and subjects with type II illness did not have better outcomes than subjects with type I illness. In maintenance treatment, MRS-SADS scores were greater overall, especially in subjects with a rapid-cycling illness course, with citalopram versus placebo.
Conclusions: Citalopram, added to standard mood stabilizers, did not have clinically meaningful benefit versus placebo for either acute or maintenance treatment of bipolar depression. Acute mania did not worsen with citalopram, but maintenance treatment led to worsened manic symptoms, especially in subjects with a rapid-cycling course.
Download PDF and other files
Reference
Ghaemi, S. N., Whitham, E. A., Vohringer, P. A., Barroilhet, S. A., Amerio, A., Sverdlov, O., & Patkar, A. A. (2021). Citalopram for acute and preventive efficacy in bipolar depression (CAPE-BD). The Journal of Clinical Psychiatry, 82(1).
Related References
- McGirr, A., Vöhringer, P. A., Ghaemi, S. N., Lam, R. W., & Yatham, L. N. (2016). Safety and efficacy of adjunctive second-generation antidepressant therapy with a mood stabiliser or an atypical antipsychotic in acute bipolar depression: A systematic review and meta-analysis of randomised placebo-controlled trials. The Lancet Psychiatry, 3(12), 1138-1146.
- Aiken, C. (2019, May 13). Antidepressants in Bipolar II Disorder. Psychiatric Times, 35(5).
