August 2006, Vol. XXIII, No. 9
Some of the most challenging facets of bipolar disorder involve the management of its depressive symptoms. This article will focus on the pharmacologic management of this phase of bipolar disorder, describing the evidence base for controlled studies, emerging treatments, and expert recommendations.
An unmet need exists for acute and maintenance treatments that target depressive symptoms in bipolar disorder. Depression predominates over the course of both bipolar I (BP I) and bipolar II (BP II) disorders, occurring more than 3 times as frequently as manic and hypomanic symptoms combined in BP I,1 and nearly 40 times as frequently as hypomania in BP II.2 Patients repeatedly attribute greater impairment in their work, family life, and psychosocial functioning to depressive symptoms than to mania.3 Even more problematic, the majority of attempted suicides occur in states of depression, in which the rate is 20 times higher than in the general population.4
A significant challenge facing clinicians is the misdiagnosis or underdiagnosis of bipolar depression. Perhaps this is because it presents itself less dramatically and is more difficult to recognize. A large French multicentered study of patients experiencing a major depressive episode, as classified by DSM-IV, found that when systematic methods for identifying hypomania were employed, the rate of BP II disorder nearly doubled, increasing from 22% to 40% of such patients.5
Similarly, when the Mood Disorder Questionnaire was used in a primary care clinic, more than one fifth of patients receiving antidepressants for the treatment of depression screened positive for bipolar disorder.6 Before this assessment, bipolar disorder had not been diagnosed in the majority of these patients. It is clear that early discrimination between bipolar disorder and other psychiatric illnesses, particularly major depressive disorder, is essential to minimize misdiagnosis and provide safe and effective care.
Acute and maintenance effects of lamotrigine
A rigorous and methodical evaluation of a pharmacologic treatment for bipolar depression emerged in 1999, when Calabrese and associates7 published the first randomized, parallel group, placebo-controlled trial of lamotrigine in the treatment of patients with BP I depression. Conducted over 7 weeks, the study randomized patients to lamotrigine 50 mg/d (n = 66), lamotrigine 200 mg/d (n = 63), or placebo (n = 66). Significant improvements were seen in patients receiving both dosages of lamotrigine over patients in the placebo group in observed change on the Hamilton Depression Rating Scale (HAM-D) total scores.
On the Montgomery-sberg Depression Rating Scale (MADRS) total score, significant improvement over placebo was evident for only the lamotrigine 200 mg/d group (P < .05). However, a responder analysis—where response is defined as 50% or greater improvement in MADRS tota score—showed that 54% of the patients receiving lamotrigine 200 mg/d and 48% of the patients receiving lamotrigine 50 mg/d met criteria for response. Response rates for patients treated with both dosages of lamotrigine were significantly greater than the 29% response rate observed for placebo-treated patients (P < .05).
Not all controlled studies of lamotrigine have demonstrated superior efficacy over placebo.8,9 Some trial results suggested improvement with lamotrigine in the core symptoms of depression but failed to detach from placebo on the primary outcome measure. Collectively, results demonstrated that lamotrigine was effective in treating bipolar depression, although dosages of 200 mg/d may be needed for maximum benefit.
In terms of safety, lamotrigine has generally been shown to have a sideeffect profile similar to that of placebo. The primary concern is the rare development of a serious rash, such as Stevens-Johnson syndrome and toxic epidermal necrolysis.10
Perhaps even more challenging than the acute relief of depressive symptoms in bipolar disorder is the prevention of depression recurrence. To address this concern, lamotrigine was studied over an 18-month period in recently depressed patients with BP I.11 After completing an open-label phase with lamotrigine for 8 to 16 weeks, 463 patients were randomized to lamotrigine, 50, 200, or 400 mg/d; lithium, 0.8 to 1.1 mEq/L/d; or placebo.
Both lamotrigine and lithium were more effective than placebo at delaying the time to intervention for any mood episode, and lamotrigine (but not lithium) was more effective than placebo at delaying the time to intervention for a depressive episode. In a complementary fashion, lithium (but not lamotrigine) was superior to placebo at delaying the time to intervention for a manic episode. Similar results were observed in another 18-month maintenance study that evaluated recently manic or hypomanic patients.12 Taken together, these studies support the use of lamotrigine for the prophylaxis of depressive episodes in bipolar disorder.
Atypical antipsychotics for acute bipolar depression
In addition to lamotrigine, the atypical antipsychotics olanzapine13 and quetiapine14 have been evaluated in randomized, placebo-controlled trials for the treatment of bipolar depression. The antidepressant efficacy of olanzapine was studied over the course of 8 weeks in patients with BP I disorder who were experiencing a major depressive episode.13
Patients were randomized to olanzapine (n = 370), an olanzapine-fluoxetine combination (n = 86), or placebo (n = 377). The monotherapy group received a mean modal olanzapine dosage of 9.7 mg/d. The combination group received a mean modal olanzapine dosage of 7.4 mg/d and a fluoxetine dosage of 39.3 mg/d.
Beginning at week 1 and continuing throughout the duration of the study, patients receiving olanzapine and the olanzapine-fluoxetine combination demonstrated significantly greater mean improvement over placebo in MADRS total scores. Olanzapine monotherapy was superior to placebo on measures of response (39.0% vs 30.4%; P = .02) and remission (32.8% vs 24.5%; P = .02).
However, when fluoxetine was added to olanzapine, the antidepressant efficacy became even more vigorous. The combination of olanzapine and fluoxetine was significantly better on all efficacy measures of depression than either olanzapine monotherapy or placebo, including higher rates of response (56.1%; P= .006 vs olanzapine, P< .001 vs placebo), remission (MADRS < 12; 48.8%; P = .07 vs olanzapine, P< .001 vs placebo), study completion, and times to response and remission.
Treatment with olanzapine was associated with somnolence, weight gain, increased appetite, dry mouth, and asthenia. The side-effect profile of the olanzapine-fluoxetine combination was similar to that seen with olanzapine monotherapy, except for higher rates of nausea and diarrhea. Weight gain, cholesterol levels, and nonfasting glucose levels were higher in patients taking olanzapine or an olanzapine- fluoxetine combination, compared with placebo. No differences were observed among the 3 groups in the rate of treatment-emergent glucose elevations of 200 mg/dL or greater.
To date, quetiapine is the only other atypical antipsychotic studied in the treatment of bipolar depression.14 In an 8-week trial termed the BOLDER (BipOLar DEpRession) I study, patients with BP I and BP II disorder were randomly assigned to receive quetiapine, 600 mg/d (n = 180); quetiapine, 300 mg/d (n = 181); or placebo (n = 181). Patients with BP II disorder were not excluded as in the olanzapine-fluoxetine trial and composed 34% of the study population.
As early as week 1, both dosages of quetiapine resulted in significantly greater mean improvement in MADRS total scores compared with placebo and at all time points in the intent-to-treat group of patients (P < .001). The mean change in MADRS total score from baseline to last assessment was –16.73 in the 600 mg/d group, –16.39 in the 300 mg/d group, and –10.26 in the placebo group (P < .001 for both quetiapine dosages vs placebo). The response rate was approximately 58% for patients treated with either dosages of quetiapine, compared with 36.1% for placebo (P < .001). Remission rates were also significantly higher in patients treated with quetiapine, namely 52.9% for quetiapine at both 600 and 300 mg/d, compared with 28.4% for placebo (P< .001). Adverse effects that occurred at significantly higher rates with the drug than with placebo included dry mouth, sedation, somnolence, dizziness, constipation, headache, and weight gain.
Cross-study comparisons of olanzapine and quetiapine should be interpreted with caution because of methodologic and sample differences. However, some distinct differences have emerged among atypical antipsychotics in their efficacy for treating symptoms of bipolar depression. The magnitude of clinical effect in comparison to placebo can be calculated through measurement of Cohen’s deffect size.15 The effect size for 600 mg/d of quetiapine was 1.09 in patients with BP I depression and 0.91 for a dosage of 300 mg/d. Since both values were 0.8 or greater, they were considered to represent a large treatment effect. Olanzapine, on the other hand, achieved an effect size of 0.32, but this increased to 0.68 when it was combined with fluoxetine.
Other notable differences in medications relate to findings on the MADRS individual item analysis. Patients receiving quetiapine or an olanzapine-fluoxetine combination improved on the core mood items of depression: representing apparent sadness, reporting sadness, inability to feel, and pessimistic thoughts. Olanzapine monotherapy was not more effective than placebo at improving the core mood items of depression. In addition, both dosages of quetiapine were found to reduce suicidal thoughts compared with placebo, whereas both olanzapine groups did not differ from placebo in reducing suicidal thinking.
A replication study of quetiapine in the treatment of bipolar depression (BOLDER II) has recently been completed. 16 The results are yet to be formally published, but a preliminary analysis showed for a second time that quetiapine monotherapy at dosages of both 300 and 600 mg/d was superior to placebo at improving depressive symptoms. In contrast to the BOLDER I study, treatment effect sizes in the replication study were found to decrease in the BP I subgroup of patients. However, significantly greater mean improvement compared to placebo on the MADRS was observed not only in the BP I subgroup but also in patients with BP II disorder.