Treatment of Bipolar Depression
The algorithm presented for treatment-resistant bipolar depression (

Figure 45-3) attempts to balance the need for antidepressant treatment and the risk of treatment-induced mania or rapid cycling. These treatments are started at the time of diagnosis (i.e., the acute phase).

Each individual acute treatment trial is carried out to one of three endpoints:

Discontinuation because the patient is unable to tolerate the adverse effects of treatment

Discontinuation because the patient has not responded to a maximal trial of the treatment (including, if warranted, augmentation strategies)

Treatment success (i.e., the patient has improved during this treatment)

In the case of treatment success, therapy should be continued for a period of time to prevent relapse (i.e., continuation phase). At the conclusion of the continuation phase, acute treatments are gradually tapered and the patient is maintained on a treatment regimen intended to prevent recurrence. Relapse during the taper period should be managed by resumption of acute-phase strategies. Antidepressant doses should be raised to effective levels and another taper attempted following another continuation phase.

For most patients, maintenance treatment with antidepressants should be avoided. However, because a substantial number of individuals with bipolar illness are subject to frequent severe depressive relapses (Altshuler 1999), a prophylactic regimen that includes an antidepressant as well as a mood stabilizer may be optimal for patients with a history of multiple relapses during attempts to discontinue their antidepressant treatment. For such cases in which antidepressants are made a part of the prophylactic regimen, it is prudent to seek the lowest effective dose.

Use of Lithium
Although 8 of 10 early studies showed that lithium had greater antidepressant efficacy than placebo, this may not establish lithium as equivalent in efficacy to other antidepressants. The traditional antidepressants may be more effective than lithium in treating bipolar depression. For instance, the only direct comparison of lithium and a standard antidepressant with a sample consisting of more than four patients demonstrated that the effect of imipramine was greater, but the statistical significance of this finding is unclear (Fieve et al. 1968). However, among unselected patients in open practice, use of lithium monotherapy appears to be sufficient therapy for about one-third to two-thirds of bipolar depressed patients (Sachs et al. 1994b). Furthermore, patients whose illness is effectively treated by lithium are spared the risk of antidepressant-induced mania.

For patients with acute depression that does not respond to lithium, the combination of lithium and a standard antidepressant appears to carry less risk of affective switch than treatment with antidepressant alone (Quitkin et al. 1981). Because lithium does not induce mania and appears to reduce the risk of affective switch during treatment with standard antidepressant agents, initiating treatment with lithium is a reasonable first step in the management of episodes of depressive disorder in patients with bipolar disorder. In cases in which depression occurs despite ongoing adequate lithium maintenance, some experts (Jann et al. 1982; Price et al. 1984) suggest that raising lithium levels will optimize antidepressive efficacy. The efficacy of increased lithium has not been studied, however, and many cases remain refractory to high therapeutic lithium levels. L. T. Young et al. (2000) reported that breakthrough depression responded equally well to the addition of a second mood stabilizer (lithium or valproate) as to the addition of paroxetine. The rate of affective switch was also equivalent for the two treatment groups in this small double-blind study.

Treatment of Anxiety During Depressive Episodes
Anxiety is a very common feature of acute depressive episodes that may intensify dysphoria and—particularly when associated with insomnia—may interfere with treatment response or lead to affective switch. Therefore, use of an anxiolytic agent is often warranted both to ameliorate the anxiety itself and to facilitate response to antidepressant treatment. Use of high-potency benzodiazepines such as lorazepam, alprazolam, or clonazepam appears to produce rapid benefit by significantly relieving dysphoria in some bipolar depressed patients. It remains unclear whether this represents a true antidepressant effect in these patients or if it is simply the result of diminished anxiety.

For anxious patients with either failed response or contraindication to benzodiazepines, clonidine and propranolol may be useful adjuncts. These adrenergic blockers must be used with caution because they may be conducive to depression or interfere with antidepressants.

For patients who experience successful resolution of the acute phase and stability through the continuation phase, tapering of anxiolytic drugs is usually indicated. Some patients do, however, appear to benefit from maintenance anxiolytic therapy. Patients may be maintained on lithium and a benzodiazepine with prophylactic efficacy at least as great as that achieved by lithium and an antipsychotic or lithium and an antidepressant (Thibault et al. 1993). In one report, however, five of five extremely ill patients whose bipolar disorders were refractory to ECT or the combination of lithium and an antipsychotic also did not benefit from the combination of clonazepam and lithium (T. A. Aronson et al. 1989).

Standard Antidepressant Medications
Antidepressants effective for unipolar depression are also likely to be effective for bipolar depression. Unfortunately, because bipolar disorder is routinely excluded from most antidepressant treatment trials, only eight clinical reports with at least eight or more patients with bipolar disorder were found (

Table 45-5). These few studies and additional reports with very small samples appear to confirm the clinical impression that standard antidepressant agents also effectively treat the depressed phase of bipolar illness (Zornberg and Pope 1993). An overview of the antidepressants commonly used in the treatment of bipolar depression is provided in

Table 45-6.

Quitkin et al. (1984a) reported that up to two-thirds of such patients responded to a TCA. Among the TCA nonresponders, two-thirds responded to an MAOI. Consistent with this, Himmelhoch et al. (1991) found a higher response rate for tranylcypromine than for imipramine in the depressed phase of bipolar disorder. In contrast, Baumhackl et al. (1989) found that imipramine was as effective as the reversible MAOI moclobemide in manic-depressive illness. Fluoxetine (and perhaps the other SSRIs) appears to be a very potent treatment for bipolar depression. Cohn et al. (1989) found fluoxetine superior to imipramine in a study of depressive bipolar disorder. Sachs et al. (1994a) found desipramine and bupropion to possess equivalent antidepressant properties. Although they are sparse, the available data suggest little difference in antidepressant efficacy among standard agents.

As reviewed by Goodwin and Jamison (1990), rates of treatment-emergent mania in patients treated with TCAs alone range from 31% to 70%. All effective standard antidepressant agents have been associated with treatment-emergent mania. It is possible, however, that various antidepressants differ in their propensity to induce mania. Among the few studies comparing antidepressants for bipolar disorder, even fewer report rates of treatment-emergent hypomania/mania. The incidence of hypomania/mania in bipolar depression over 3-6 weeks of treatment with fluoxetine, imipramine, or placebo was 0%, 9.5%, and 7.7%, respectively (Cohn et al. 1989). These rates can be misleading, however, because lithium use was not consistent across groups (33.3%, 16.6%, and 20.7%, respectively). Cohn et al. also reported a 16% incidence of affective switch among the 25 subjects who crossed over to open treatment with fluoxetine after placebo or imipramine failed to produce a response.

Over the course of a year of double-blind treatment in 19 subjects, there was a significantly increased rate of hypomania/mania with desipramine compared with bupropion (50% vs. 11%, respectively) (Sachs et al. 1994a). In an uncontrolled open case series, however, 6 (55%) of 11 patients experienced hypomania or mania during treatment with bupropion (Fogelson et al. 1992). Design differences may account for the disparity between the results from these latter two studies. In the study by Fogelson et al. (1992), the sample likely included patients at higher risk for affective switch; 10 of the 11 subjects had experienced mania during prior antidepressant treatment, and the bupropion dosage was rapidly escalated (the maximum dosage of 450 mg/day was usually reached by day 6). In contrast, the study by Sachs et al. (1994a) excluded patients with a history of treatment-emergent mania and used a more gradual titration of dosage, with all subjects having had 2 or more weeks of treatment at 300 mg/day before further dosage increase.

Replication of the findings described above is necessary to ensure their validity. Nonetheless, the available data suggest that bupropion is less likely to induce hypomania/mania than reuptake inhibitor-type antidepressants.

Electroconvulsive Therapy
ECT is a potent and beneficial treatment for bipolar illness. Nine studies (3 of them controlled and double-blind) have addressed the efficacy of ECT for bipolar depression (reviewed in Zornberg and Pope 1993). As with antidepressant medications, ECT can cause an affective switch. Given the high rate of mixed episodes in treatment-refractory bipolar depression, such episodes are frequently misdiagnosed as pure depression. A good response to bilateral ECT but not to unilateral nondominant ECT, which worsened the course of mania, has been reported for these patients (Small et al. 1985). Perhaps patients referred for ECT with significant features of mood elevation should be offered initial treatment with bilateral ECT.

Nonstandard Somatic Treatments
Anticonvulsants  In psychiatry, anticonvulsants have come into common use when treating patients with treatment resistance in diverse diagnoses. There are eight reports with results for eight or more patients given a trial of an anticonvulsant for acute treatment of depression, but only four of these trials were double-blind (

Table 45-7). The results are quite modest if judged by the very low frequency of marked improvement in the double-blind trials. Given that the samples in these studies included many treatment-resistant patients, the moderate improvement rate justifies including the anticonvulsants as options for the acute-treatment phase, particularly for patients with a history of antidepressant-induced mania or hypomania. An overview of anticonvulsants used in the treatment of bipolar depression is provided in

Table 45-8.

Treatment of psychosis during depressive episodes  The presence of psychotic features appears to increase the likelihood of treatment resistance. Treatment of psychotic depression in bipolar disorder often includes three or more medications (lithium, an antidepressant, and an antipsychotic). In the absence of data specific to treatment of bipolar depression with psychotic features, the approach to treatment may be guided by experience with unipolar delusional depression.

Lithium may have antipsychotic activity apart from its impact on the specific mood disorder component of the illness and might lessen the need for antipsychotic medication (Biederman et al. 1979). Use of anticonvulsants and benzodiazepines also appears to lessen the need for antipsychotic medication. Nonetheless, acute treatment of bipolar depression with psychotic features often requires the use of antipsychotic medication. Patients with affective illness appear to be at greater risk for such adverse effects of antipsychotics as acute dystonia and tardive dyskinesia (Nasrallah et al. 1988; Wegner et al. 1985). Therefore, exposure to antipsychotics should be minimized to the dosage necessary to safely control the acute psychotic symptoms within the available resources for patient management. Some reports suggest that atypical antipsychotics (risperidone and olanzapine) produce improved treatment outcome for patients with refractory bipolar depression with and without psychotic features.

Maximization of mood-stabilizing medication  After controlling the symptoms in the acute phase, the next treatment priority is prevention of recurrence. An iterative approach is necessary for meeting this objective. Clinicians can use systematic follow-up and mood charting to evaluate the impact of each treatment on cycle length. It is difficult to judge the benefit of prophylactic treatments without a sufficient follow-up period. In most cases, reasonable assessment can be made on the basis of outcome over a time equivalent to three times the patient’s cycle length and not less than 6 months.