The depression is the part that is actually happening to you, month after month, and an antidepressant is the obvious thing to try. It is also the most-prescribed and least-supported intervention in the whole of bipolar treatment, and the argument about it has been running for twenty years without a clean answer.
The short answer: less than everyone assumes. In the largest randomised trial, 366 people with bipolar depression already on a mood stabiliser were given either an added antidepressant or a placebo for up to 26 weeks; 23.5 percent on the antidepressant reached durable recovery against 27.3 percent on placebo — a non-significant difference favouring, if anything, the placebo. A meta-analysis of six trials and 1,383 patients found a small improvement on clinician-rated symptom scores but no significant difference in response or remission rates, and no excess switching in the short term but a significant increase over 52 weeks. An international task force concluded that "there is striking incongruity between the wide use of and the weak evidence base for the efficacy and safety of antidepressant drugs in bipolar disorder," and that lithium, lamotrigine, olanzapine, quetiapine and lurasidone should be considered before antidepressants. None of this is a reason to stop anything you are taking without your prescriber.¹ ² ³
The trial
STEP-BD randomised people with bipolar depression, all on a mood stabiliser, to an added antidepressant — bupropion or paroxetine — or a matching placebo, for up to 26 weeks, under conditions designed to look like ordinary care. The primary outcome was durable recovery: eight consecutive weeks of euthymia.¹
| Outcome | Mood stabiliser + antidepressant | Mood stabiliser + placebo |
|---|---|---|
| Durable recovery | 42 of 179 (23.5%) | 51 of 187 (27.3%) |
| Statistical difference | P = 0.40 | |
| Treatment-emergent mood elevation | 10.1% | 10.7% |
The authors: "Modest nonsignificant trends favoring the group receiving a mood stabilizer plus placebo were observed across the secondary outcomes. Rates of treatment-emergent affective switch were similar in the two groups."¹ ⁴
Two conclusions follow, and they point in opposite directions from the usual arguments. The antidepressants did not clearly help. They also did not clearly cause switching, at least over six months in people already on a mood stabiliser.
What the pooled evidence says
A meta-analysis of six randomised trials and 1,383 patients found:²
- A small symptom-scale improvement: standardised mean difference 0.165 (95% CI 0.051–0.278, p = 0.004) on clinician-rated depressive symptoms.
- No difference in what people care about: clinical response (1.158, 0.840–1.597, p = 0.371) and remission (1.220, 0.874–1.703, p = 0.243) were not significantly different from placebo.
- On switching, a time-dependent answer: "Acute treatment was not associated with an increased risk of treatment-emergent mania or hypomania (0·926 [0·576-1·491], p = 0·753), but 52 week extension periods were associated with an increase in risk (1·774 [1·018-3·091], p = 0·043)."²
The authors' own recommendation: "these medications should be used only in the short term because prolonged use is associated with an increased risk of treatment-emergent mania or hypomania."²
That is a specific and actionable statement, and it is close to the opposite of what usually happens, which is an antidepressant started in a bad winter and still running four years later.
What about staying on one that seems to be working?
There is a randomised study of exactly that. Seventy patients who had responded to an antidepressant plus a mood stabiliser and been well for two months were randomly assigned to continue or discontinue the antidepressant for one to three years. Continuation "trended toward less severe depressive symptoms… and mildly delayed depressive episode relapse (HR = 2.13 [1.00-4.56]), without increased manic symptoms," but produced "No benefits in prevalence or severity of new depressive or manic episodes, or overall time in remission."⁵
And one finding worth carrying into a conversation with a prescriber: a rapid-cycling course predicted three times more depressive episodes with antidepressant continuation — 1.29 against 0.42 episodes a year (P = .04).⁵ If your course has been rapid-cycling, that is a specific thing to raise.
The switching question, honestly
The evidence genuinely conflicts, and the honest summary is about circumstances rather than a blanket rule.
Randomised short-term trials do not show an excess of switching. STEP-BD: 10.1 percent against 10.7 percent.⁴ The meta-analysis above: no acute excess.²
Naturalistic and meta-analytic data do. A meta-analysis across 109 trials and 114,521 patients found "The overall risk of mania with/without ADs averaged 12.5%/7.5%," with tricyclics riskier than serotonin-reuptake inhibitors — and, uncomfortably, that "Mood stabilizers had minor effects probably confounded by their preferential use in mania-prone patients."⁶ An older comparison put treatment-emergent switch at 11.2 percent with tricyclics and tetracyclics against 3.7 percent with SSRIs and 4.2 percent with placebo.⁷ In a prospective clinic cohort, switch occurred in 24.4 percent, most strongly predicted by a history of previous switches.⁸
The defensible synthesis: switch risk is real but concentrated — tricyclics and venlafaxine, antidepressant monotherapy without a mood stabiliser, bipolar I, mixed features, rapid cycling, a history of previous switching, and longer exposure. It is not a uniform property of all antidepressants in all people. And mood stabilisers are not reliably protective, which is the part most often assumed.⁶
What the guidance says
The international task force's report is the most complete statement, and its conclusions are unusually candid:³
"There is striking incongruity between the wide use of and the weak evidence base for the efficacy and safety of antidepressant drugs in bipolar disorder."
"Because of limited data, the task force could not make broad statements endorsing antidepressant use but acknowledged that individual bipolar patients may benefit from antidepressants."
"In bipolar I patients antidepressants should be prescribed only as an adjunct to mood-stabilizing medications."
"There was consensus in this ISBD Task Force that non-antidepressant treatments, including lithium, lamotrigine, olanzapine, quetiapine, and lurasidone, should be considered as monotherapy before using antidepressants in bipolar depression."
"the use of antidepressants to treat depressive phases or components of bipolar disorder can neither be condemned nor endorsed without carefully evaluating individual clinical cases and circumstances."
The UK's national guideline goes further by omission: it contains no recommendation endorsing antidepressants for bipolar depression at all. The word appears in its recommendations only twice, both about stopping them — "If a person develops mania or hypomania and is taking an antidepressant… as monotherapy: consider stopping the antidepressant," and the same instruction where the antidepressant is combined with a mood stabiliser.⁹ What it recommends instead for moderate or severe bipolar depression is fluoxetine combined with olanzapine, or quetiapine on its own, with olanzapine alone or lamotrigine alone as alternatives.⁹
What is actually approved in the United States
Worth knowing, because this is where the labels and the prescription pads diverge. Verified against current FDA prescribing information:¹⁰
| Drug | Approved for bipolar depression? |
|---|---|
| Quetiapine (immediate release and XR) | Yes — and its trials enrolled bipolar I or II |
| Lurasidone | Yes — bipolar I, monotherapy and adjunctive to lithium or valproate |
| Cariprazine | Yes — bipolar I, adults |
| Olanzapine–fluoxetine combination | Yes — acute depressive episodes in bipolar I |
| Lamotrigine | No. Approved for maintenance of bipolar I "to delay the time to occurrence of mood episodes," with an explicit limitation: "Effectiveness of LAMICTAL in the acute treatment of mood episodes has not been established" |
| Olanzapine alone | No — mania and maintenance only |
| Lithium | No — acute mania and maintenance only, despite being the first-line long-term drug in guidelines |
| Any standard antidepressant | No |
The lamotrigine row is the one most often stated wrongly on consumer sites. It is a maintenance drug for bipolar I, not an approved treatment for an acute depressive episode.
What to ask your prescriber
Bring these, and bring them as questions rather than conclusions.
- "Which of my medications is treating the depression, and which is preventing episodes?" Different jobs, different evidence.
- "Have the non-antidepressant options been tried?" The task force names lithium, lamotrigine, olanzapine, quetiapine and lurasidone as things to consider first.³
- "If I stay on an antidepressant, for how long, and when will we review it?" The meta-analysis found the switch risk emerging over 52 weeks, not in the first months.²
- "Does my course include rapid cycling or mixed features?" Both change the calculation, and rapid cycling predicted more depressive episodes on continued antidepressants.⁵ ⁶
- "Am I on a mood stabiliser alongside it?" For bipolar I, the task force says an antidepressant should be adjunctive only.³
- "What is the plan that is not a prescription?" Structured psychological treatment added to medication reduces recurrence, and depression is the outcome it moves most. What the psychotherapies do →
Do not stop or change anything on your own. Several of these drugs have withdrawal effects, and an unplanned discontinuation in the middle of a depressive episode is its own risk. Requesting your records, if you need the full medication history →
Q&A
Q: Do antidepressants work for bipolar depression? A: The largest randomised trial found 23.5 percent durable recovery with an added antidepressant against 27.3 percent with placebo, a non-significant difference.¹ A meta-analysis of six trials found a small symptom-scale improvement with no significant difference in response or remission.² They are not established as effective for this.
Q: Will an antidepressant trigger mania? A: In the short term, in people already on a mood stabiliser, the randomised evidence says no — 10.1 percent against 10.7 percent in STEP-BD.⁴ Over a year, the pooled evidence says the risk rises.² The risk concentrates in tricyclics and venlafaxine, monotherapy, bipolar I, mixed features, rapid cycling and a history of previous switching.⁶
Q: My antidepressant is helping. Should I stop? A: Not on the basis of a web page, and not on your own. What the one randomised discontinuation study found was a trend toward less severe depressive symptoms on continuation, and no benefit on the prevalence or severity of new episodes or on overall time well.⁵ That is a genuine conversation to have, with your own history in front of you.
Q: Is lamotrigine an antidepressant for bipolar? A: It is widely used that way and it is not approved for it. Its US indication is maintenance of bipolar I to delay recurrence, with an explicit statement that effectiveness in acute mood episodes has not been established.¹⁰
Q: What if I only have bipolar II? A: The task force notes that antidepressant-associated mood elevations appear less frequent and less severe in bipolar II than bipolar I, and that the adjunct-only rule is stated for bipolar I.³ That is a difference of degree, not a green light, and the efficacy evidence is no stronger.
Q: What treats the depression, then? A: For medication, the options with US approval for bipolar depression are quetiapine, lurasidone, cariprazine and the olanzapine–fluoxetine combination, and the guideline route in the UK is olanzapine with fluoxetine or quetiapine.⁹ ¹⁰ Alongside that: structured psychological treatment, which the trials show moves depression more than mania, and rhythm and sleep work. The full bipolar map →
Ready to find a therapist who works alongside your prescriber? Filter by approach, schedule and payment route → · What the psychotherapies actually do →
Sources
- Sachs GS, Nierenberg AA, Calabrese JR, et al., "Effectiveness of adjunctive antidepressant treatment for bipolar depression," New England Journal of Medicine 356(17), 2007, 1711–1722 — "Forty-two of the 179 subjects (23.5%) receiving a mood stabilizer plus adjunctive antidepressant therapy had a durable recovery, as did 51 of the 187 subjects (27.3%) receiving a mood stabilizer plus a matching placebo (P=0.40)" — doi.org.
- McGirr A, Vöhringer PA, Ghaemi SN, Lam RW, Yatham LN, "Safety and efficacy of adjunctive second-generation antidepressant therapy with a mood stabiliser or an atypical antipsychotic in acute bipolar depression: a systematic review and meta-analysis of randomised placebo-controlled trials," The Lancet Psychiatry 3(12), 2016, 1138–1146 — six trials, 1,383 patients; SMD 0.165 (95% CI 0.051–0.278, p = 0.004) on clinician-rated symptoms with no significant difference in response or remission; "Acute treatment was not associated with an increased risk of treatment-emergent mania or hypomania (0·926 [0·576-1·491], p = 0·753), but 52 week extension periods were associated with an increase in risk (1·774 [1·018-3·091], p = 0·043)" — doi.org.
- Pacchiarotti I, Bond DJ, Baldessarini RJ, et al., "The International Society for Bipolar Disorders (ISBD) task force report on antidepressant use in bipolar disorders," American Journal of Psychiatry 170(11), 2013, 1249–1262 — "There is striking incongruity between the wide use of and the weak evidence base for the efficacy and safety of antidepressant drugs in bipolar disorder"; "in bipolar I patients antidepressants should be prescribed only as an adjunct to mood-stabilizing medications"; "non-antidepressant treatments, including lithium, lamotrigine, olanzapine, quetiapine, and lurasidone, should be considered as monotherapy before using antidepressants in bipolar depression"; "can neither be condemned nor endorsed without carefully evaluating individual clinical cases and circumstances" — doi.org.
- Sachs GS, Nierenberg AA, Calabrese JR, et al., 2007, as above — "rates of mood elevations were indistinguishable (10.1% compared with 10.7%)" between the adjunctive antidepressant and placebo groups.
- Ghaemi SN, Ostacher MM, El-Mallakh RS, et al., "Antidepressant discontinuation in bipolar depression: a Systematic Treatment Enhancement Program for Bipolar Disorder (STEP-BD) randomized clinical trial of long-term effectiveness and safety," Journal of Clinical Psychiatry 71(4), 2010, 372–380 — 70 patients randomised to continue or discontinue; "No benefits in prevalence or severity of new depressive or manic episodes, or overall time in remission, occurred"; "rapid-cycling course predicted 3 times more depressive episodes with antidepressant continuation (rapid cycling = 1.29 vs non-rapid cycling = 0.42 episodes/year, P = .04)" — doi.org.
- Tondo L, Vázquez G, Baldessarini RJ, "Mania associated with antidepressant treatment: comprehensive meta-analytic review," Acta Psychiatrica Scandinavica 121(6), 2010, 404–414 — 109 trials, 114,521 adult patients; "The overall risk of mania with/without ADs averaged 12.5%/7.5%… Tricyclic antidepressants were riskier than serotonin-reuptake inhibitors (SRIs)… Mood stabilizers had minor effects probably confounded by their preferential use in mania-prone patients" — doi.org.
- Peet M, "Induction of mania with selective serotonin re-uptake inhibitors and tricyclic antidepressants," British Journal of Psychiatry 164(4), 1994, 549–550 — treatment-emergent switch 11.2% with tricyclics and tetracyclics against 3.7% with SSRIs and 4.2% with placebo — doi.org.
- Valentí M, Pacchiarotti I, Bonnín CM, et al., "Risk factors for antidepressant-related switch to mania," Journal of Clinical Psychiatry 73(2), 2012, e271–e276 — "Treatment-emergent affective switch was detected in 54 patients (24.4%)," most strongly associated with a higher rate of previous switches — doi.org.
- National Institute for Health and Care Excellence, "Bipolar disorder: assessment and management," clinical guideline CG185 — recommendations 1.5.2 and 1.5.7 on stopping antidepressants when mania or hypomania develops; recommendation 1.6.3, "offer fluoxetine combined with olanzapine, or quetiapine on its own," with olanzapine alone or lamotrigine alone as alternatives — nice.org.uk.
- US Food and Drug Administration, current prescribing information retrieved via the openFDA drug label API — SEROQUEL and SEROQUEL XR (quetiapine), LATUDA (lurasidone), VRAYLAR (cariprazine), olanzapine and fluoxetine capsules; LAMICTAL (lamotrigine): "Maintenance treatment of bipolar I disorder to delay the time to occurrence of mood episodes… Limitations of Use: Treatment of acute manic or mixed episodes is not recommended. Effectiveness of LAMICTAL in the acute treatment of mood episodes has not been established"; ZYPREXA (olanzapine) and lithium carbonate labelling for their bipolar indications — open.fda.gov.
Paid for by participating therapists. Inclusion is computed from availability data — never purchased. No ads, no data sold.