Bipolar Antidepressant Agents: Shaken not stirred

Never say antidepressant again

The phase of bipolar disorder that confers the greatest burden is bipolar depression. It is therefore quite puzzling that there are no antidepressants specifically designed to treat depression in the context of bipolarity. A wide variety of medications are currently available for the management of bipolar disorder and while some are effective in treating acute bipolar depression, only a handful have reported efficacy in maintenance and prophylaxis. Notably, none of these agents have been developed primarily for the treatment of bipolar depression and it is therefore not surprising that clinically the term ‘bipolar antidepressant’ is seldom used. Instead, antidepressants are simply viewed as agents that treat depression, regardless of the context.

This is partly because bipolar depression is difficult to differentiate from unipolar/major depression, especially early on in the course of the illness, when there is often no hint of mania/hypomania. Thus, a pragmatic approach has been adopted and depressive symptoms are treated equally, irrespective of which type of depression they augur. This broad, non-specific approach would not be a problem, were it not for the fact that bipolar depression does not seem to respond to conventional antidepressants (e.g. SSRIs) in the same manner as major depression. In addition to these concerns regarding response, there is evidence to suggest that conventional antidepressants may worsen the course of bipolar disorder, and that on occasion they exacerbate mood instability and precipitate mania/hypomania.

If this is the case, it is a serious disadvantage. Bipolar disorder is already a difficult illness to treat successfully (Malhi et al., 2009). Effective management often requires careful longitudinal evaluation and judicious use of both psychological and pharmacological measures in conjunction with support and understanding from family, friends, and employers. Therefore, the administration of potentially harmful agents warrants careful consideration, and there is a pressing need for better informed practice, especially because currently the majority of patients with bipolar depression are routinely exposed to conventional antidepressants (Baldessarini et al., 2007). However, diagnostic uncertainty is not the only reason for this parlous state of affairs.

The ideal ‘double agent’

The ideal bipolar antidepressant agent is one that achieves rapid, robust remission of acute bipolar depression and then maintains long-term prophylaxis so that the likelihood of a recurrence is significantly diminished. The latter criterion is particularly important and is the key attribute of a mood stabiliser (Goodwin and Malhi, 2007). Alas, no single agent can genuinely claim such dual action not even special agents such as lithium, lamotrigine or quetiapine (Malhi et al., 2009; Malhi et al., 2012). Unfortunately, in the treatment of bipolar depression each of these medications has significant limitations in either efficacy or tolerability but even so, they remain preferable in comparison to conventional antidepressants.

In a highly cited meta-analysis which was limited by the small number of randomised controlled trials (RCTs) that it analysed and the inclusion of one particularly large study, the use of adjunctive antidepressants was shown to be effective for the treatment of acute bipolar depression (Gijsman et al., 2004), but a number of studies since have produced equivocal findings. However, many of the recent trials that have examined the effects of antidepressants in bipolar I and II disorder, either alone (Amsterdam and Shults, 2005; Amsterdam and Shults, 2008; Amsterdam and Shults, 2010; Agosti and Stewart, 2007) or in combination with mood stabilisers (Schaffer et al., 2006; Fonseca et al., 2006; Tamayo et al., 2009; Pilhatsch et al., 2010; Perlis et al., 2010), are methodologically compromised by small sample size and/or the lack of a placebo arm. Consequently, a recent meta-analysis (Sidor and MacQueen, 2011) that includes the findings of these studies reported no significant efficacy for antidepressants in the treatment of acute bipolar depression. It is important to note however, that antidepressant efficacy in this meta-analysis was very nearly significant with respect to both response and induction of remission. Furthermore, heterogeneity amongst the mood stabilisers to which the antidepressants were added (Malhi et al., 2011), and the fact that STEP–BD (Sachs et al., 2007) showed a negative treatment effect that favoured placebo make it difficult to be conclusive.

License to pill

It is quite remarkable that no new molecules that specifically target bipolar depression have been developed. Instead, on the basis of predominantly empirical evidence, agents that have been used for the treatment of major depression, such as conventional antidepressants, have been adopted into the pharmacotherapeutic armamentarium for bipolar depression. The few agents that have recently been formally trialled in the treatment of acute bipolar depression, such as the atypical neuroleptics, were initially manufactured for the treatment of schizophrenia (Malhi et al., 2011). Thence they gradually migrated to bipolar disorder, first to the treatment of acute mania and eventually to the treatment of bipolar depression. Even the anticonvulsants, which have been used in the management of manic-depressive illness for decades, are in effect ‘impostors’. Of these, lamotrigine has been the most successful in recent times and the fact that it has predominantly antidepressant properties has earned it ‘specificity’ for bipolar depression. In practice, its true efficacy is much contested, and its benefits in the long term remain unclear. Indeed, the only agent that can possibly claim specificity for bipolar disorder is lithium (Gershon et al., 2009), but even this element exerts only a modest effect when prescribed acutely, because of an inordinately slow onset of action (Malhi and Tanious, 2011). Arguably, its long-term efficacy surpasses all others and redeems its putative primacy (Malhi, 2010).

The doctors know

Interestingly, the actual efficacy of conventional antidepressants in the treatment of bipolar depression remains largely untested and therefore, as Professor Allan Young states, many patients with major depressive episodes who have an underlying bipolar diathesis are potentially in receipt of ineffective medications (Young, 2012). Thus, he humbly urges regulatory bodies to broaden their research focus to include all depressive episodes per se such that clinical trials include patients with both major and bipolar depression. This he argues, quite correctly, will allow a dimensional appraisal of bipolarity. However, this ‘modest’ proposition is likely to be considered by some as an ‘indecent proposal’, given that it brushes aside many decades of effort to meaningfully partition unipolar and bipolar mood disorders across psychosocial and neurobiological domains. But, given that 80 to 85% of bipolar depressed patients are treated with antidepressants, Professor Young’s yoking of the mood spectrum remains persuasive.

In contrast, Professor Michael Gitlin calls for greater focus on bipolar disorder pharmacotherapy, specifically the real-world use of antidepressants as add on agents to mood stabilisers (Gitlin, 2012). Further, he asks whether antidepressants could be used differentially across bipolar subtypes, and for how long they should be continued once remission has been achieved. Continuing to tantalise, he raises the prospect of clinically separable sub groups of bipolarity with differential antidepressant response and tolerability profiles. Professor Gitlin’s quest for answers concludes with what he concedes is a cliché - namely, a call for better, bigger, and brighter research.

This appeal for further research is echoed by Professor Nassir Ghaemi, who addresses clinical and cultural factors in addition to science, and in doing so ‘ups the ante’ on antidepressants emphasising the importance of carefully scrutinising the data that we have already garnered (Ghaemi, 2012). He explains clearly why greater value should be attached to randomised data and stresses that this should receive priority over and above clinical experience, which though undeniably important, should not obscure the findings from research. At the same time he concedes that randomised trials also have some limitations, because they derive from the study of ‘average’ patients and this limits the generalisability of findings. However, he also points out that the findings of RCTs do apply to the majority of patients, and therefore they are relevant and important.

Underwhelmed by the acute efficacy of antidepressants in bipolar depression either as monotherapy or in conjunction with a mood stabiliser, Professor Ghaemi warns that the use of antidepressants in bipolar disorder may in fact cause more harm than good.

Finally, Professor Joe Goldberg describes lucidly how misconceptions can arise and how clinician impressions can magically be transformed into accepted wisdom (Goldberg, 2012). He draws our attention to the fact that antidepressants perhaps do not stabilise all bipolar patients but only those that are perhaps more susceptible. He also questions quite rightly what the occurrence of an antidepressant-induced hypomania actually means and whether indeed it alludes to a bipolar propensity or is merely an adverse effect. With discerning argument, he notes that much of the literature fails to provide a robust operational definition of mania or hypomania and that most studies draw upon observational, non- randomised data. Further, he points out that there is no way of determining whether modest mood fluctuations are intrinsic to bipolarity or the consequence of treatment. He even questions whether antidepressants increase the frequency of cycling and encourage further occurrences of depression while at the same time noting that they do not have demonstrated efficacy over and above that of mood stabilisers alone.

Thus, given the many blind spots in our fields of view it would seem that consensus across the board regarding these issues is perhaps too ambitious a goal.

A quantum of science

Psychiatry as a whole and perhaps bipolarity in particular is in urgent need of psychiatric science and the explanations as to why specific bipolar antidepressants have not been developed should give each of us pause for thought.

First, for more than half a century the focus of bipolar disorder research has been skewed towards mania and hypomania, the aspects that essentially define bipolarity, and as a consequence an enormous amount of resources and intellectual energy have been devoted to the investigation and treatment of this phase of the illness. Only in the past decade or so has it become apparent that bipolar depression is the predominant state of the illness and that it is this phase of the disorder that confers the lion’s share of morbidity.

Second, it has become increasingly clear that in addition to mania and depression there are myriad bipolar admixtures ranging from subsyndromal to mixed states, not to mention the variations in course and progression that include rapid cycling and ultradian bipolarity. This has further thwarted research and obfuscated delineation of bipolar depression.

Third, and hand in hand with these issues of focus, our understanding of the neurobiology of bipolar disorder has been, and perhaps remains rather rudimentary. This has constrained our ability to develop new agents on the basis of known pathophysiology. Clearly, a logical concept driven approach is vital to educe treatments specific to bipolar depression.

Finally, right from the beginning an exaggerated fear that antidepressants may induce affective instability and precipitate mania/hypomania has discouraged their use in bipolar disorder and perhaps curtailed enthusiasm for testing and developing derivatives that may be better suited to bipolar depression. This apprehension is evident from the paucity of antidepressant studies that have been conducted in bipolar disorder.

All about Moneypenny?

The one reason that has not been discussed in detail even though it is perhaps the most critical is that of cost. It is probably the most apposite explanation as to why we have no shiny new bipolar antidepressants and it is also the most sombre. Developing a novel agent from scratch is immensely expensive, whereas adapting the use of an already existing agent albeit imperfect is much less of a risk and more favorable economically. Hence, the acquisition of new indications for existing medications is an obvious strategy but one that has clearly failed to bear fruit and is ultimately short term.

Q ‘action’

Bipolar disorder is a chronic illness that necessarily requires life long care. It is time that our treatments addressed this need.