From Wiki Journal Club
Jump to navigation Jump to search
Sachs GS, et al. "Effectiveness of adjunctive antidepressant treatment for bipolar depression". The New England Journal of Medicine. 2007. 356(17):1711-1722.
PubMedFull textPDF

Clinical Question

Among patients with bipolar disorder receiving mood stabilizing agents, does adjunctive antidepressant therapy reduce the symptoms of dipolar disorder without increasing mania?

Bottom Line

Among patients with bipolar disorder receiving mood stabilizing agents, the addition of an antidepressant did not yield improved symptom control compared to mood stabilizing therapy alone.

Major Points

Although it is common for patients of bipolar depression to receive antidepressants when they fail to respond adequately to mood stabilizers alone, the FDA has not judged any of the standard antidepressants as efficacious for use in this population. STEP-BD sought to prospectively evaluate the efficacy of antidepressant therapy among these patients in a randomized controlled setting.

A 2013 systematic review[1] found insufficient evidence for use of short- or long-term adjunctive antidepressants among patients with bipolar disorder.


As of June 2015, no guidelines have been published that reflect the results of this trial.


  • Multicenter, double-blind, randomized, parallel-group, placebo-controlled study
  • N=366
    • Mood stabilizer + antidepressant (n=179)
    • Mood stabilizer + placebo (n=187)
  • Setting: 22 centers in the United States
  • Enrollment: November 1999 to July 2005
  • Mean follow-up: 3.5 years
  • Analysis: Intention-to-treat
  • Primary outcome: Durable recovery, defined as euthymia for at least 8 consecutive weeks.


Inclusion Criteria

  • Patients who met the criteria for a major depressive episode associated with bipolar I or bipolar II disorder according to the Diagnostic and Statistical Manual of Mental Disorders, fourth edition (DSM-IV).
  • Age ≥18 years old

Exclusion Criteria

  • Patients with a known history of intolerance or non-response to both bupropion and paroxetine.
  • Patients who required current short-term treatment of a coexisting substance-abuse disorder.
  • Patients who required additional antipsychotic medication or dose adjustment of a long-term antipsychotic medication.

Baseline Characteristics

From the mood stabilizer+antidepressant group.

  • Demographics: Male sex 42%, age 40 years, white race 91%
    • Educational level:
      • Some or all of high school: 22%
      • Some post-high school: 32%
      • Associate, technical, college, or postgraduate degree: 46%
    • Income:
      • <$30K: 42%
      • $30K-$75K: 32%
      • ≥$75K: 21%
    • Marital status: Married 35%, never married 35%, divorced/widowed/separated 30%
  • Bipolar: Type I 69%, type II 31%
    • ≥10 prior manic episodes: 61%
    • ≥10 prior depressive episodes: 69%
    • Rapid cycling: 27%
    • Treatment-emergent affective switch: 39%
  • Anxiety disorder: Current 44%, lifetime 65%
  • Substance abuse: Current 17%, lifetime 58%
  • Rating scores: SUM-D 6.2, SUM-ME 1.1, MADRS 24.5, YMRS 5.8, GAF 55.95, CGI severity-of-illness subscale 3.9
  • Days in study before randomization: 197


Participants were randomized in a blinded fashion to either of the two groups listed below. In both groups, patients received a mood stabilizer (lithium, valproate, carbamazepine, or other FDA-approved medications).

  • Mood stabilizer+antidepressant -
    • The antidepressant used was either bupropion or paroxetine
    • Paroxetine was started at 10 mg daily and increased to a maximum of 40 mg daily
    • Bupropion (sustained-release) was started at 150 mg daily and increased to a maximum of 375 mg daily
    • Sarticipantsubjects were treated for 26 weeks
  • Mood stabilizer+placebo -

The randomization was equipoise-stratified, meaning that psychiatrists could preferentially choose from a strata of 1. placebo vs. bupropion, 2. placebo vs. paroxetine, or 3. placebo vs. bupropion or paroxetine


Presented as mood stabilizer+antidepressant vs. mood stabilizer+placebo.

Primary Outcomes

Durable recovery
Defined as 8 consecutive weeks of euthymia.
23.5% vs. 27.3% (P=0.40)

Secondary Outcomes

Treatment remission
17.9% vs. 21.4% (P=0.40)
Transient remission or durable recovery
41.3% vs. 48.7% (P=0.23)
Treatment-effectiveness response
32.4% vs. 38.0% (P=0.27)
Treatment-emergent affective switch
10.1% vs. 10.7% (P=0.84)
Discontinuation of study medication because of an adverse event
12.3% vs. 9.1% (P=0.32)

Subgroup Analysis

Adverse Events

Serious adverse events

Defined as events resulting in hospitalization, permanent disability, or death or requiring an intervention to prevent these outcomes.

4.5% vs. 5.3% (P=0.70)

Psychiatric hospitalization for suicidal ideation

2.2% vs. 2.7%

Increased frequency of suicidal ideation without psychiatric hospitalization

0% vs. 0.5%

Marked or grossly disabling adverse events

An adverse event was defined as marked when substantially impairing social or occupational functioning or requiring dose adjustment or discontinuation. A grossly disabling event is one which substantially impairs simple activities of daily living.

9.5% vs. 7.0% (P=0.37)


  • Low recruitment rate. In the study only 366 out of 2689 patients with at least 1 major depressive episode were enrolled. This is a potential source of bias.[2]
  • Short period of observation for primary outcome, which was defined as 8 consecutive weeks of euthymia.[3]
  • Many study subjects also received psycho social therapy during the study. The proportion of patients who received psycho-social intervention in the mood stabilizer + antidepressant and mood stabilizer+ placebo groups were 67.9% and 69.7%, respectively.[3]
  • Patients who had a manic episode shortly before the study may not have been enrolled due to concerns of switching from depression to mania or hypomania precipitated by antidepressant treatment.[2]
  • Some symptoms that are important in bipolar disorder, including mood instability and impulsivity, are not presented in the data.[4]
  • It is possible that only patients with good insight and motivation to receive treatment were involved in the study whereas patients with limited insight may have been understudied.[4]


  • The National Institute of Mental Health (NIMH)
  • Glaxo Wellcome and SmithKlineBeecham (now GlaxoSmithKline) donated the antidepressant medications.

Further Reading