By Shalom Feinberg, M.D.
Associate Clinical Professor of Psychiatry, Albert Einstein College of Medicine

A lot has happened in the four years since TCPR first looked at the role of second-generation antipsychotics (SGAs) in the treatment of depression (TCPR, Feb. 2003). The literature has mushroomed, and recently the FDA approved Seroquel (quetiapine) for the treatment of bipolar depression. In this article, we’ll take a hard look at the most recent data on the effectiveness of SGAs for depression in its various guises, including treatment-resistant depression (TRD) and bipolar depression (BD).

Seroquel for Bipolar Depression: The BOLDER Studies Seroquel was approved for bipolar depression on the evidence from two eight-week double-blind placebo-controlled studies called BOLDER I and II, which were funded by AstraZeneca. In its zeal to create a memorable acronym, AstraZeneca derived the study name in the following circuitous manner: BipOLar DEpRession (Calabrese, et al., Am J Psychiatry 2005;162:1351-1360; Thase et al, J Clin Psychopharm 2006;26:600-609).

Each study consisted of approximately 500 patients meeting criteria for bipolar I or II disorder and suffering from major depression at the time of enrollment. Patients were randomized to one of three conditions: Seroquel titrated to 300 mg/HS over four days, Seroquel titrated to 600 mg/HS over one week, or placebo. Exclusion criteria included substance abuse, acute suicidality, refractory depression, chronic depression lasting longer than 12 months, or other recently treated Axis I diagnoses. Unfortunately, a sizeable minority of our real-life patents would not meet the inclusion criteria, which limits the generalizability of the results.

At the eight-week endpoint, both studies showed a statistically significant benefit of Seroquel when bipolar I and bipolar II depression were combined: in BOLDER I there was a 58% response rate to Seroquel vs. a 36% response to placebo, while BOLDER II posted a 52% response rate to Seroquel vs. a 37% placebo response rate. Both studies found that dosing at 600 mg/day was no more effective than 300 mg/day. Treatment-emergent mania, a concern when prescribing antidepressants to bipolar patients, did not occur more frequently with Seroquel, at least over the short eight-week trial period.

While the Seroquel data appears to be robust, there are a few caveats to keep in mind. First, in breaking out the data it is clear that only one of the two studies – BOLDER II – found a significant benefit for Seroquel in bipolar II depression; BOLDER I did not. Second, Seroquel has not been directly compared with other active antidepressant agents or with mood stabilizers such as lamotrigine or lithium. Third (and crucially), there have been no non-industry sponsored replications of these studies. We emphasize this point because AstraZeneca fully planned the design and selected the statistical methodology used in BOLDER. In an apparent attempt to draw comparison to Lilly’s Symbyax study (discussed below) AstraZeneca used similar statistical methodology; it opted for the newer MMRM (mixed-effects model repeated measure) method rather than the standard LOCF (last observations carried forward) technique in its statistics. As TCPR discussed in a past article on Cymbalta (duloxetine; Jan. 2004), the MMRM approach may enhance reported effect sizes, making active treatments look better than they would otherwise appear if using LOCF.

The bottom line on Seroquel is that it may be a good option as monotherapy for bipolar depression, though more clearly for BP I than BP II patients. In otherwise medically healthy patients, we recommend a fairly rapid dosing schedule, increasing by 50 mg increments over four to seven days, up to 300 mg QHS. Warn your patients about initial sedation while reassuring them that most people accommodate to this intense side effect over time.

What Happened to Symbyax? Of course, Seroquel is not the first SGA found to be effective for depression: in 2003, the FDA approved Symbyax (a combination of Zyprexa and Prozac) for the treatment of bipolar depression. But we’ve rarely prescribed it, and chances are you haven’t prescribed it much either. Why not? First, it offers limited flexibility as a fixed-dose combination; second, Zyprexa is associated with the metabolic syndrome; and finally, there is concern that the Prozac component might precipitate antidepressant-induced mania in bipolar patients.

How does the Seroquel BOLDER data stack up against the Symbyax data? It’s pretty similar. The pivotal study reported a 56% response rate on Symbyax vs. a 30% response rate on placebo (Tohen, et al., Arch Gen Psychiatry 2003;60:1079-1088). However, the Symbyax study also compared Zyprexa monotherapy (that is, Symbyax minus the Prozac) with placebo, which is a fairer comparison with the Seroquel monotherapy data. Zyprexa alone did not do very well, posting a 39% response rate, only marginally (but statistically) better than placebo. The implication is that Seroquel monotherapy is a more effective antidepressant treatment than Zyprexa monotherapy, but only a head-to-head comparison of the two will tell.

How About Risperdal? Although Risperdal has been around longer than any other atypical (except for clozapine), there have been very few studies of this agent for bipolar depression. The only controlled study that we could find on this topic (Shelton and Stahl, J Clin Psychiatry 2004;65:1715-1719) was a double-blind trial in which patients continued on prior mood stabilizers and were randomized to either Paxil (paroxetine) or Risperdal monotherapy, or a combination of these two drugs. All three treatments were equally but modestly effective. However, the sample size of only 30 patients was meager, and the study lacked a placebo control, so it offers no definitive conclusions. In the recent multi-center NIMH-funded STEP-BD study of bipolar disorder, Risperdal fared rather poorly as add-on (or augmentation) treatment for refractory bipolar depression (i.e., a less than 5% response rate) in an open-label comparison to lamotrigine and inositol (Nierenberg et al, Am J Psychiatry 2006; 163:210-216).

Do SGAs work for Treatment-Resistant Depression? Until last year, the literature on SGAs for TRD primarily consisted of open studies showing a positive therapeutic role for every SGA as an augmentation agent to an antidepressant. There was only one published double-blind study; it endorsed Zyprexa augmentation of fluoxetine over continued monotherapy with either fluoxetine or Zyprexa (Shelton, et al., Am J Psychiatry 2001;158:131-134). But now Zyprexa has some competition. Controlled studies have been presented at major meetings describing successful augmentation of antidepressants with Risperdal and Seroquel, along with larger positive studies of Zyprexa (see review by Papokostas and Zarate, Primary Psychiatry 2007;14:59-65).

Do Geodon and Abilify Have Antidepressant Qualities? Clinicians who have used Abilify (aripiprazole) have noted that it tends to have a stimulating effect in many patients. For some patients, this is problematic, leading to agitation, akathisia, and, occasionally, worsening psychosis. On the positive side, for some patients this stimulation has an antidepressant effect, and some open data on Abilify augmentation has supported this theory. Bristol-Myers Squibb and Otsuka are currently sponsoring three double-blind studies of Abilify augmentation in TRD (www.clinicaltrials.gov). Like Abilify, Geodon suffers from Seroquel envy when it comes to data on the treatment of depression, though this hasn’t prevented many of us from using it off-label.

What’s the Bottom Line? For bipolar depression, Seroquel currently has the most convincing data among the SGAs and presents only a moderate liability for the metabolic syndrome. However, we suspect that, if put to an adequate test, all the SGAs would have some efficacy for this disorder. Similarly, Zyprexa augmentation for TRD has been the data leader among the SGAs, but others are right behind. Are all the SGAs equally effective for these illnesses? We don’t know yet, but stay tuned!


TCPR Verdict:

Second-generation antipsychotics as antidepressants? Probably.