Bipolar depression is one of the least studied depressive illnesses. The standard practice for many doctors is to use antidepressant medicines, but there are few studies on the long-term results of these medicines. The goal of this study is to look at how effective and safe these medicines are in treating bipolar depression when taken with a mood stabilizer medicine.

The drug being studied is citalopram, also known as Celexa. Celexa is FDA approved for the treatment of major depression, but is not FDA approved for the treatment of bipolar depression. It is, however, standard practice for many doctors is to use antidepressants, like Celexa, to treat their patients with bipolar disorder depression.

The drug will be studied in three ways. We will see if it helps treat depressive symptoms. We will see how the drug affects the brain using PET and fMRI scans. Finally, we will look at the possibility that there may be a gene that could predict if a person would get better taking the drug using genetics.

The problem of interest

Depression is a common serious illness, with great personal suffering and a 10% or more risk of suicide. There are two kinds of depression, bipolar (depression alternating with mood swings) and unipolar ("simple" depression, or major depressive disorder). In both kinds of depression, antidepressants are commonly used. However, unlike unipolar depression, where a good deal of research supports this use, there is very little research on antidepressant use in bipolar disorder. Some studies support benefit with antidepressants for bipolar depression; i.e., if one is depressed, antidepressants can help a patient get better in the short-term. However, long-term studies are limited; the few available studies with older antidepressants did not demonstrate long-term preventive benefit. Some clinicians think new generation antidepressants (like Prozac and other serotonin reuptake inhibitors) are safer and more effective than the older antidepressants. Yet there are no rigorous long-term studies of new generation antidepressants in bipolar disorder. Recent observational studies with new antidepressants suggest that they may be harmful to some patients with bipolar disorder.

Clinicians are thus left in a quandary. Antidepressants appear to be effective in the short term, but should the antidepressants (especially new generation antidepressants) be continued long-term? Some studies support both approaches. Importantly, some evidence exists that antidepressants can make many patients worse, with more and more depression or mania over time.

This is a major public health problem, since clinicians prescribe antidepressants for the long-term in up to 80% of patients with bipolar disorder. This represents standard treatment, despite the limitations of the available evidence and the suggestion that in some patients such antidepressant use may be harmful.

This study is also looking at possible biological predictors that may reflect an individual patients' likelihood that he or she will respond to a specific treatment.

How the problem will be studied

This project is one of the most rigorous studies of new generation antidepressants in long-term treatment of bipolar disorder. We will recruit patients with bipolar disorder who are currently in a depressive episode and are taking or eligible and interested in taking a mood stabilizer such as lithium. Subjects will then be randomly put into one of two groups. The first group will receive the generic antidepressant, citalopram while the other group will receive placebo, sugar pill. Patients will be closely followed and monitored by the research psychiatrist and through a series of safety labs. Subjects will have an MRI and PET scan for the biological predictors section of the study. The key question is: After the acute recovery, should they continue antidepressants or not?

As there is limited scientific data, and opposing kinds of clinical experience, there is no clear rationale to making this decision. Our study seeks to provide a scientific basis for making that decision. We plan to follow subjects for a goal of 1 year to obtain long-term outcome data on which approach is best.

How the research will advance scientific knowledge or human health

Since there are no rigorous long-term studies with new generation antidepressants in bipolar disorder, this study will be a major advance in that knowledge. Clinicians will have some evidence on which to base that decision, rather than simply their own opinions or patients' preferences. In the light of recent evidence that antidepressants are potentially harmful to some patients, this study will provide more information about how new antidepressants work specifically for bipolar patients rather than unipolar depression (also known as major depression). The biological portion of the study will shed light on possible biological factors that could show how an individual may respond to a specific treatment. This could potentially help doctors to decide on which treatments are more or less likely to work for their individual patients rather then using a "hit or miss" type method.

Inclusion Criteria:

• Current age ≥18 years
• DSM-IV diagnosis of BPD, type-I, or type-II
• Current major depressive episode using DSM-IV criteria, lasting 8 weeks or longer.
• Use of lithium, divalproex, carbamazepine, or lamotrigine at therapeutic serum levels or doses for ≥4 weeks prior to study entry, or willingness to accept one of these agents.
• Prior to initial evaluations, each subject must provide competent, written, informed consent.

Exclusion Criteria:

• Past non-response to a therapeutic trial of R,S-citalopram (≥100 mg/day for ≥8 weeks).
• Previous intolerance of R,S-citalopram;
• Diagnosis of unipolar depression
• Diagnosis of schizoaffective disorder
• Serious medical illness with acute instability (cardiac, respiratory, hepatic, renal), based on hospitalization in the past month
• Abnormal thyroid function tests
• Previous allergic reaction to or inability to tolerate lithium, divalproex, or carbamazepine at therapeutic serum levels.
• Current or past renal dysfunction if taking lithium
• Current or past hepatitis or other liver disease if taking divalproex
• Current or past hematologic disease if on carbamazepine
• Severe suicidal ideation, plan or intent, as documented by a score of ≥4 on the Montgomery Åsberg Depression Rating Scale suicidality item (Item 10).
• Presence of psychosis
• Cognitive impairment sufficient to impair ability to give informed consent.
• Current pregnancy, or inability to utilize contraception
• The presence of any metallic implants
• History of claustrophobia

These subjects met inclusion/exclusion criteria and were randomized to citalopram versus placebo.

Adverse events are reported for 1 year of follow-up during the randomized trial.

All patients were systematically interviewed for the adverse events reported.