3.3 The Case for Cautious Use: When Stopping May Be Worse

If the aggregate data suggests antidepressants are ineffective and risky, why do guidelines not ban them entirely? Because a clinically identifiable subgroup of bipolar patients responds profoundly to adjunctive antidepressants and deteriorates rapidly when they are removed.

The most compelling evidence comes from the discontinuation study by Altshuler and colleagues (Altshuler et al., 2003, American Journal of Psychiatry). This prospective, naturalistic (non-randomized) study tracked 84 patients with bipolar disorder who had achieved remission from depression by adding an antidepressant to a mood stabilizer. After remission, 43 patients discontinued the antidepressant within 6 months, while 41 continued it beyond 6 months. Because patients were not randomly assigned to continue or discontinue, factors such as illness severity, prescriber judgment, and patient preference may have influenced which group patients entered.

Over one year of follow-up, 70% of the discontinuation group suffered a major depressive relapse, compared to 36% of the continuation group. The overall rate of manic relapse across the entire cohort was only 18%, and manic switch was not significantly associated with continued antidepressant use (Altshuler et al., 2003, American Journal of Psychiatry). While the non-randomized design means these numbers cannot establish causation with the rigor of an RCT, the magnitude of the difference (70% vs. 36%) and the low rate of manic relapse (18%) represent the strongest available evidence that antidepressant continuation matters for a subset of patients.

3.4 Bipolar I vs. Bipolar II: Stratifying the Risk

The risk-benefit calculation changes substantially between bipolar subtypes. Data consistently show that the frequency and severity of antidepressant-associated mood elevation is significantly greater in bipolar I disorder than in bipolar II disorder. A systematic review found the relative risk of switch in bipolar I to be substantially higher than in bipolar II (RR = 1.78; 95% CI 1.24--2.58; p = .002) (Tondo et al., 2010, Acta Psychiatrica Scandinavica).

In bipolar II disorder, the switch risk appears substantially lower and, when it occurs, manifests as hypomania rather than full mania. Studies focusing exclusively on bipolar II patients treated with SSRIs found switch rates that were statistically indistinguishable from background rates in lithium monotherapy control groups (approximately 19--20%), suggesting the observed mood elevations reflected the natural illness course rather than iatrogenic drug effects (Altshuler et al., 2017, American Journal of Psychiatry).

This has practical implications. Modern consensus guidelines permit a wider therapeutic role for antidepressants in bipolar II depression, where the risk of full psychotic mania is functionally absent, the depressive burden is severe and chronic, and patients often tolerate and benefit from adjunctive antidepressant therapy.

3.5 Not All Antidepressants Are Equal: The Risk Hierarchy

When a clinician determines that an antidepressant trial is warranted, agent selection matters enormously. The ISBD Task Force and decades of meta-analytic data agree that different antidepressant classes carry vastly different switch risks (Pacchiarotti et al., 2013, American Journal of Psychiatry).

High-risk agents (avoid): Tricyclic antidepressants (TCAs) carry switch rates of 10--42%, with some agents (amitriptyline, imipramine) approaching 40%. SNRIs, particularly venlafaxine, carry switch rates of 13--29% and are grouped with TCAs as high-risk agents by all major guidelines (Vieta & Valentí, 2013, CNS Drugs).

Lower-risk agents (preferred): SSRIs (sertraline, paroxetine, escitalopram) show switch rates of 3--7%, often statistically indistinguishable from placebo when used with mood stabilizer coverage. Bupropion has a switch rate of approximately 4%, and its lack of serotonergic activity may confer additional stability (Sachs et al., 2007, New England Journal of Medicine; Leverich et al., 2006, American Journal of Psychiatry).

3.6 The ISBD Consensus Framework: 12 Statements for Clinical Practice

The International Society for Bipolar Disorders convened a global expert task force that established 12 consensus statements using the Delphi method, achieving at least 80% agreement among leading bipolar experts (Pacchiarotti et al., 2013, American Journal of Psychiatry). The key principles:

Absolute prohibitions:

Antidepressant monotherapy is contraindicated in bipolar I disorder. Antidepressants must be avoided in the presence of current mixed features (two or more concurrent manic symptoms), in patients with a history of rapid cycling (four or more episodes per year), and in patients with prior antidepressant-induced mood elevation.

Conditions for cautious use:

Adjunctive antidepressants may be considered when the patient has a documented prior positive response without mania induction, when first- and second-line agents (atypical antipsychotics, mood stabilizers) have been adequately trialed, and always with mood stabilizer coverage. If used, prefer SSRIs or bupropion over TCAs or SNRIs. Maintenance antidepressant treatment may be considered only when the patient has documented relapse upon discontinuation.

The CANMAT/ISBD positioning:

Traditional antidepressants are explicitly excluded from first-line treatment. In the CANMAT hierarchy, SSRIs and bupropion appear only as third-line adjunctive options, permitted after first- and second-line treatments have failed and only under antimanic coverage (Keramatian et al., 2023, Focus).

3.7 The Clinical Algorithm: A Step-by-Step Approach

Important Note: The Most Common Prescribing Error in Bipolar Depression

The single most common cause of iatrogenic mood destabilization in bipolar disorder is treating a bipolar depressive episode with the algorithm designed for unipolar major depressive disorder. When a patient with unrecognized or confirmed bipolar disorder presents in a depressive episode, the first pharmacological intervention should be a medication specifically approved or guideline-endorsed for bipolar depression (quetiapine, lurasidone, cariprazine, lamotrigine, lithium, lumateperone), not an SSRI. Antidepressants are a third-line option, used only after first- and second-line agents have been adequately tried. For clinicians considering an antidepressant trial after first-line failures, the following framework synthesizes CANMAT, NICE, and ISBD guidance: Step 1: Verify the foundation. Confirm the bipolar diagnosis is correct. Ensure the patient has adequate antimanic coverage (mood stabilizer or antipsychotic at therapeutic dose with confirmed adherence). Never prescribe an antidepressant as monotherapy in bipolar I. Step 2: Try first-line bipolar depression agents. The standard of care is to initiate a medication specifically indicated for bipolar depression before considering an antidepressant (see Section 2). Step 3: Screen with the ISBD exclusion checklist. If first-line agents have failed after adequate trials, assess for absolute contraindications to antidepressant use: current mixed features, rapid cycling history, or prior antidepressant-induced mood elevation. If any are present, do not prescribe an antidepressant. Step 4: Select, initiate, and monitor. Choose an SSRI or bupropion (never a TCA or SNRI as first-choice). Common clinical practice is to initiate at a lower dose than typically used for unipolar depression (e.g., sertraline 25--50 mg, escitalopram 5--10 mg, bupropion XL 150 mg), though specific bipolar-adjusted dosing thresholds are based on clinical convention rather than formal guideline recommendations. Schedule follow-up weekly or biweekly. Educate the patient and family to report early warning signs of switch: sudden decrease in sleep need, intense irritability, accelerated speech, uncharacteristic impulsivity. Step 5: Have an exit strategy. If a full manic switch occurs (clear mania with decreased sleep need, grandiosity, psychomotor acceleration), discontinue the antidepressant promptly and optimize the antimanic agent. For mild hypomanic switches, clinical judgment is required: abrupt discontinuation of short-half-life SSRIs (particularly paroxetine) can produce discontinuation syndrome that complicates the clinical picture, so a rapid but not abrupt taper over several days may be more appropriate than immediate cessation. Fluoxetine's long half-life makes discontinuation syndrome less of a concern. In either case, consider short-term benzodiazepines (lorazepam, clonazepam) for acute agitation and sleep restoration.

**In plain language for families: **You may have heard conflicting things about antidepressants and bipolar disorder. Here is the bottom line: standard antidepressants (the medications used for regular depression) do not work reliably for bipolar depression, and in some patients they can make things worse by triggering a manic episode or an unstable, irritable mood state. This does not mean they are never used. A small group of patients with bipolar disorder clearly benefits from adding an antidepressant to their mood stabilizer, especially people with bipolar II (the type with less severe highs). But antidepressants are not the first medication doctors try for bipolar depression. They are used only after other, safer options have been tried, always with a mood stabilizer on board, and with close monitoring for any signs of mood destabilization. If your doctor suggests adding an antidepressant, ask what the plan is if your mood starts going up instead of stabilizing.

PsychHQ

Antidepressant Use: Clinical Algorithms & Consensus Framework