Although most studies have reported rates of mania or other short-term effects in bipolar patients, Dr. Ghaemi said that the more important outcome measure should be long-term risks of antidepressant therapy on the course of bipolar disorder.

Dr. Ghaemi told the audience that bipolar patients generally come to the attention of doctors when they are depressed. They often are treated with antidepressants because their initial diagnosis is unipolar depression. The diagnosis of prior episodes of hypomania or mania in a patient presenting with depression is made by history; history taking should be thorough and include interviews with family members as well as patients. Clinicians should make a diagnosis of unipolar depression only after bipolar depression has been ruled out.

Dr. Ghaemi noted that accurate diagnosis does not ensure appropriate treatment. Pharmacy marketing data indicate that antidepressants are used twice as frequently as mood stabilizers for patients with a confirmed diagnosis of bipolar disorder. The question then becomes, Are doctors justified in using antidepressants this much? He believes not.

In a study conducted by the National Institute of Mental Health, treatment-resistance in bipolar disorder was attributable to antidepressant use in 25 percent of patients. Approximately 50 percent of fifty patients in another National Institutes of Mental Health study appeared to have rapidly cycling bipolar disorder associated with antidepressant treatment. Dr. Ghaemi believes that antidepressants act as mood destabilizers, cancelling the effects of mood stabilizing agents.

Dr. Ghaemi outlined an algorithm for guiding treatment decisions in depressed bipolar patients. The algorithm begins with mood stabilizers, moves to atypical antipsychotics, and then to atypical anticonvulsants. Antidepressants are not excluded, however. Paroxetine and bupropion appear in the algorithm because double-blind, controlled data in bipolar I disorder suggest lower rates of mood-switching with those drugs.

Dr. Ghaemi concluded that current literature suggests possible negative effects of even newer antidepressants in patients with bipolar illness. He believes that antidepressant treatment poses risks for patients with bipolar I disorder that should be examined more closely.