Antidepressants in bipolar depression: when less is more

A 54-year-old woman with bipolar disorder was referred for a consultation for refractory depression. She has a family history of bipolar disorder. The initial course of her illness is described as clearly episodic, with full recovery between minor depressive and hypomanic episodes that she had been experiencing since her teens. At the age of 26, she developed postpartum depression and was treated with a tricyclic antidepressant. Later, after a manic episode, she was diagnosed with bipolar disorder and treated with a combination of lithium (plasma levels 0.7–0.9 mmol/L) and various antidepressants (i.e., all available selective serotonin reuptake inhibitors, bupropion, doxepine, amitriptyline and venlafaxine). In addition, she had trials of carbamazepine, sodium divalproex, lamotrigine and several neuroleptics. Her course of illness gradually became more chronic. For the last 10 years, she has been mostly depressed with intermittent improvements. At the time of consultation, she was severely depressed, crying, with poor concentration and hypnagogic hallucinations; her medications were lithium carbonate, citalopram, lorazepam, L-THYROXINE AND ESTRADIOL.

This case suggests a good likelihood of prophylactic response to lithium treatment (Grof et al, Br J Psychiatry 1993;163(Suppl 21):16-9). This is supported by the history of recurrent illness with fully remitting course before treatment and by the family history of bipolar disorder. During antidepressant treatment, the clinical picture became less typical with a chronic fluctuating course of illness. The consultant recommended the antidepressant be discontinued, and this resulted in full recovery within 10 days. This made it possible to gradually decrease lorazepam. Six months later, she remains well. Should she experience symptoms of depression in the future, the recommended therapy would be electroconvulsive therapy or a careful trial of another mood stabilizer such as lamotrigine (without an antidepressant) or a monoamine oxidase inhibitor.

Antidepressants are often used in bipolar depression, and in some patients they may, indeed, be necessary. At the same time, their use is often considered problematic, mainly because of the risk of precipitation of mania or induction of rapid cycling. In the absence of these adverse reactions, a clinician may not suspect that an antidepressant may still be responsible for lack of improvement. One such effect is persistent depression not responding to treatment. There is relatively little information about such outcomes in the literature. However, Ghadirian (Biol Psychiatry 1986;21:1298-1300) described a sudden improvement of depression upon antidepressant discontinuation, and Sharma (J Affect Disord 2001;64:99-106) presented several cases of patients with bipolar or unipolar depression who lost response to antidepressants but showed sustained improvement after the withdrawal of these drugs and treatment with mood stabilizers. Interestingly, almost 50% of the patients with unipolar depression had a family history of bipolar disorder. Others (see Goldstein, J Clin Psychiatry 1999;60:563-7) have reported the development of mania or hypomania after antidepressant withdrawal, despite continued treatment with mood stabilizers.

In a patient such as the one described here, a clinician may speculate about the interaction between the antidepressant and lithium. Is it possible that in the absence of an effective mood stabilizer, the patient would, indeed, develop either mania or rapid cycling?

In patients with bipolar disorder, clinicians should use antidepressants sparingly, minimize long-term use and always consider the possibility of atypical or idiosyncratic response to an antidepressant in patients not responding to treatment.