Aripiprazole for Depression: Antipsychotic Add-On Evidence, Risks

Table of contents

• Evidence summary
• What antipsychotic augmentation is
• How it works
• What it is prescribed for
• What works and what does not
• Benefits by claim
• Risks and all side effects
• All interactions
• Who should avoid antipsychotic augmentation
• Dosage and how to take it (label doses only)
• Follow the money: who makes it and who funded the evidence
• Related research
• Frequently asked questions
• Sources and funding notes

Evidence summary

Independent evidence and credibility scorecard

Independence tiers: 1 = regulator or government body; 2 = academic or Cochrane team with no declared commercial funding for the work; 3 = independent authors analysing mostly manufacturer-run trials, or minor declared industry ties; 4 = manufacturer-funded or manufacturer-authored. Credibility runs from A (highest) to D.

Most of these sources are American. The two UK sources (NICE and Cochrane) and the Taiwanese meta-analysis reach broadly the same verdict: a real short-term effect, a small one, and a clear cost in side effects.

What antipsychotic augmentation is

"Augmentation" means adding a second medicine to an antidepressant that has not worked well enough, rather than replacing it. When that second medicine is an antipsychotic, the strategy is called antipsychotic augmentation. The drugs used are second-generation (atypical) antipsychotics, first developed for schizophrenia and bipolar disorder. NICE's further-line options list aripiprazole, olanzapine, quetiapine and risperidone as examples (NICE NG222).

In the United States, six products carry an official depression indication:

• Aripiprazole (Abilify): add-on to antidepressants for major depressive disorder in adults (label). Developed by Otsuka in Japan and co-promoted in the US with Bristol-Myers Squibb (DOJ).
• Quetiapine extended-release (Seroquel XR): add-on for major depressive disorder, made by AstraZeneca (label).
• Brexpiprazole (Rexulti): approved by the FDA in July 2015 as an add-on, discovered by Otsuka and co-developed with Lundbeck (Otsuka).
• Cariprazine (Vraylar): add-on approval in December 2022; developed by Gedeon Richter with AbbVie (AbbVie).
• Olanzapine-fluoxetine combination (Symbyax and generics): one capsule containing both drugs, for "treatment resistant depression", defined on the label as no response to two different antidepressants in the current episode (generic label).
• Lumateperone (Caplyta): approved by the FDA on 6 November 2025 as an add-on to antidepressants for major depressive disorder in adults, at 42 mg once daily; made by Intra-Cellular Therapies, now owned by Johnson & Johnson. The company reports MADRS differences of 4.9 and 4.5 points over placebo at 6 weeks in its two trials (J&J). These are manufacturer-reported results that the independent reviews in this article predate, so we do not grade them.

Legal status elsewhere differs. In the UK, NICE notes that in June 2022 antipsychotic use for depression was off-label for some of these drugs (NICE). In the EU, brexpiprazole (Rxulti) and cariprazine (Reagila) are authorised only for schizophrenia (EMA Rxulti; EMA Reagila). Off-label prescribing is legal, but it means the regulator has not reviewed the benefit-risk balance for that use.

This article covers adding these drugs for non-psychotic, unipolar depression. Depression with psychotic features and bipolar depression are different conditions with their own evidence; see our guides on psychotic depression and bipolar depression.

How it works

Nobody knows exactly why these drugs help depression. Every US label says the mechanism is unknown or unclear and then offers a hypothesis based on receptor binding (Abilify; Rexulti):

• Aripiprazole, brexpiprazole and cariprazine are partial agonists at dopamine D2 and serotonin 5-HT1A receptors and block 5-HT2A receptors. "Partial agonist" means they stimulate the receptor a little while blocking stronger stimulation (Vraylar label).
• Quetiapine blocks D2 and 5-HT2A receptors; its active metabolite, norquetiapine, acts more strongly at 5-HT2A (Seroquel XR label).
• Olanzapine-fluoxetine is thought to act through combined effects on serotonin, noradrenaline and dopamine systems (label).

What is proven in humans: that adding these drugs lowers depression rating-scale scores in short trials. What is not proven: which receptor action causes that effect. Their sedating and restlessness-causing effects can also shift rating-scale items such as sleep and agitation, which is one reason independent reviewers stress functioning and quality-of-life outcomes (Spielmans 2013).

What it is prescribed for

Licensed uses relevant to depression

In the US, aripiprazole, quetiapine XR, brexpiprazole, cariprazine and, since November 2025, lumateperone are licensed as add-ons for adults with major depressive disorder whose antidepressant has not worked well enough. Olanzapine-fluoxetine is licensed for treatment-resistant depression after two failed antidepressants (label). All carry other licences, mainly for schizophrenia and bipolar disorder.

Where guidelines place it

NICE NG222 is cautious. When depression has not responded to an antidepressant, it first asks clinicians to look for reasons: adherence problems, side effects, life circumstances, or another condition. It then lists options including group exercise, switching to psychological therapy, raising the dose or switching drug (recommendations 1.9.1 to 1.9.5). If a person prefers a combination of medicines, NICE says to explain the possible increase in side effects. Then consider referral to, or advice from, a specialist mental health setting. Adding a second-generation antipsychotic is one listed option alongside lithium, a second antidepressant, ECT, lamotrigine or liothyronine (1.9.8 and 1.9.9).

For our wider overview of next steps, see treatment-resistant depression.

What works and what does not

Benefits by claim

Does aripiprazole work for depression?

Yes, as an add-on, but by a modest margin. In the manufacturer's registration trial (362 people, sponsored by Bristol-Myers Squibb with Otsuka), MADRS scores fell by 8.8 points with aripiprazole vs 5.8 with placebo over 6 weeks, a 3-point difference on a 60-point scale (Berman 2007; registry). Independent pooling found remission OR 2.01 (95% CI 1.48 to 2.73) and response OR 2.07 (1.58 to 2.72); about 7 people needed treatment for one extra responder and 9 for one extra remission (Spielmans 2013).

The strongest real-world test is VAST-D, run at 35 Veterans Affairs medical centres. The 1,522 participants had not responded to at least one adequate antidepressant course; their mean age was 54 and 85% were men. After 12 weeks, remission was 28.9% with aripiprazole add-on, 26.9% with bupropion add-on and 22.3% after switching to bupropion. Aripiprazole beat switching (RR 1.30, 1.05 to 1.60), but not bupropion add-on. Response (a 50% or greater drop in symptoms) was higher with aripiprazole: 74.3% vs 65.6% and 62.4%. Relapse rates did not differ. The authors concluded the advantage was small and that the side effects meant cost-effectiveness needed study (Mohamed 2017).

Quetiapine XR for depression

The Cochrane review pooled three quetiapine add-on trials and found a small-to-moderate effect (SMD −0.32, 95% CI −0.46 to −0.18; high-quality evidence). But people were more likely to drop out on quetiapine (RR 1.57, 1.14 to 2.17), mainly because of side effects (Davies 2019). The review says nine of its ten trials were pharmaceutical-sponsored, but the sponsors are not named in the sections we read (Cochrane full text). Spielmans found remission OR 1.79 (1.33 to 2.42) and response NNT 10 (Spielmans 2013). In the licensing trials, 12.1% stopped quetiapine XR because of side effects vs 1.9% on placebo (label).

Brexpiprazole and cariprazine

These newer drugs were approved on manufacturer trials. In Otsuka's 6-week trial, MADRS fell 8.36 points with brexpiprazole 2 mg vs 5.15 with placebo. That result was reported in a population defined by amended randomisation criteria (Thase 2015; registry). In a Forest Laboratories trial (with Gedeon Richter) of cariprazine, 2–4.5 mg beat placebo by 2.2 MADRS points at 8 weeks, but 1–2 mg did not separate significantly (Durgam 2016; registry). Cochrane rated cariprazine's effect as high-quality evidence of a small benefit (MD −1.50 MADRS points, −2.74 to −0.25), with more dropouts (RR 1.68) (Cochrane). A 2026 bias-adjusted meta-analysis found the evidence for aripiprazole, quetiapine, brexpiprazole and cariprazine robust, but not for other antipsychotics (Hsu 2026).

Olanzapine-fluoxetine

Olanzapine-fluoxetine had the weakest effect in the independent analysis: remission OR 1.42 (1.01 to 2.0), NNT 19, and no significant effect on response. It also caused the most weight gain (Spielmans 2013). A network meta-analysis did find significant effects on both response and remission (Nuñez 2022).

Older adults

Two publicly funded trials studied people aged 60 and over. In IRL-GRey, 181 people who had not remitted on venlafaxine were randomised to aripiprazole or placebo for 12 weeks. Remission was 44% vs 29% (OR 2.0, 1.1 to 3.7; NNT 6.6, though the confidence interval runs from 3.5 to 81.8). Akathisia affected 26% vs 12%, and parkinsonism 17% vs 2% (Lenze 2015). In the open-label OPTIMUM trial (619 people), aripiprazole add-on improved well-being more than switching to bupropion (difference 2.79 points, 0.56 to 5.02). Remission was 28.9% with aripiprazole, 28.2% with bupropion add-on and 19.3% after switching. Falls were most frequent with bupropion add-on (Lenze 2023). See our guide to depression in older adults.

How does it compare with lithium?

Lithium is the older augmentation option. In the 2026 meta-analysis, lithium beat placebo in a conventional analysis (OR 2.06) but the result lost significance after publication-bias adjustment. The antipsychotic evidence was supported by trial sequential analysis (Hsu 2026). The lithium trials are older and small, so this reflects thin evidence rather than proof that lithium does not work. In OPTIMUM's second step, lithium add-on and switching to nortriptyline gave similar results (Lenze 2023). See our lithium review.

Risks and all side effects

Common side effects in the depression trials

These figures come from 6–8 week trials run by manufacturers. Independent reviewers warn that study design and reporting may have understated harms (Spielmans 2013).

Akathisia

Akathisia is an unpleasant inner restlessness and urge to keep moving. It is the signature side effect of aripiprazole, brexpiprazole and cariprazine, and rises with dose (Rexulti label). It can feel like worsening anxiety or agitation, so any new restlessness after starting one of these drugs should be reported promptly.

Weight, blood sugar and cholesterol

All labels warn of metabolic changes, including high blood sugar, abnormal lipids and weight gain. Severe high blood sugar has occasionally been linked to ketoacidosis, coma or death (Abilify label; olanzapine-fluoxetine label). Short trials understate this. In open-label follow-up of up to 52 weeks, 25–28% of people on aripiprazole 5 mg or more had medically significant weight gain, compared with 3.5% in a study using lower doses (Seshadri 2021). NICE advises checking weight, blood pressure, pulse, fasting glucose or HbA1c and lipids before starting. It then advises weekly weight checks for six weeks, then at 12 weeks, one year and yearly (NICE 1.4.33–1.4.34).

Tardive dyskinesia and other movement disorders

Tardive dyskinesia causes involuntary, often repetitive movements, usually of the face and mouth. The labels say risk rises with treatment length and total dose, that it can occur after short, low-dose use, and that it may be irreversible (Abilify; Seroquel XR). Rates below 1% at one year were reported in open-label aripiprazole studies (Seshadri 2021), but no controlled study has followed people long enough to measure it properly in depression. Parkinsonism (stiffness, slowness, tremor) affected 17% of older adults on aripiprazole in IRL-GRey (Lenze 2015).

Serious and rare risks

• Neuroleptic malignant syndrome: a rare, potentially fatal reaction with high fever, muscle rigidity and confusion (Rexulti label).
• Compulsive behaviours: the Abilify label warns of pathological gambling and other compulsive urges and advises dose reduction or stopping (label).
• Falls and low blood pressure on standing: sleepiness, dizziness and instability can cause falls and fractures (Vraylar label).
• Low white blood cell counts: leukopenia, neutropenia and agranulocytosis have been reported (Abilify label).
• Heart rhythm: QT prolongation has been reported after quetiapine overdose, and with fluoxetine (Seroquel XR label; olanzapine-fluoxetine label).
• Prolactin effects: NICE lists sexual and menstrual disturbance among side effects to monitor (NICE 1.4.34).

Stopping

NICE advises that antipsychotics for depression should only be stopped in specialist services or with their advice. The dose should be reduced gradually over at least four weeks, and longer after longer treatment (NICE 1.4.37). Do not stop either medicine suddenly without advice. Our antidepressant withdrawal guide covers the antidepressant side.

All interactions

Several of these drugs are broken down by liver enzymes that common antidepressants block, so the combination itself can be an interaction.

• Aripiprazole with fluoxetine or paroxetine: these strong CYP2D6 inhibitors raise aripiprazole levels. For other uses the label advises half the usual dose with strong CYP2D6 or CYP3A4 inhibitors, a quarter with both, and double the dose with strong inducers such as carbamazepine or rifampin. But when aripiprazole is given as an add-on for depression, the label says to use it without these dose adjustments, following the depression dosing of 2 to 15 mg a day (Abilify label). The registration trial capped the dose at 15 mg with fluoxetine or paroxetine, compared with 20 mg otherwise (Berman 2007).
• Brexpiprazole: no dose change is needed in depression with strong CYP2D6 inhibitors such as paroxetine or fluoxetine. Strong CYP3A4 inhibitors call for half the dose, and strong inducers for double (Rexulti label).
• Quetiapine: reduce to one-sixth of the dose with strong CYP3A4 inhibitors (for example ketoconazole, ritonavir). Inducers such as phenytoin, rifampin and St John's wort can mean up to five times the dose. Quetiapine also adds to the effects of alcohol (Seroquel XR label). See our St John's wort review.
• Cariprazine: reduce the dose with strong or moderate CYP3A4 inhibitors; use with CYP3A4 inducers is not recommended (Vraylar label).
• Olanzapine-fluoxetine with other serotonergic drugs: serotonin syndrome has been reported, especially when combined with other serotonergic agents (label).
• Other sedatives, alcohol and blood-pressure medicines: may add to drowsiness, dizziness and falls (Seroquel XR label).

NICE also tells prescribers to watch for interactions that raise antipsychotic levels and to adjust doses (NICE 1.4.34). A pharmacist can check a full medicine list.

Who should avoid antipsychotic augmentation

Not approved

• Older people with dementia-related psychosis: higher risk of death (boxed warning). The Abilify label also notes more strokes and mini-strokes in this group, including deaths (Abilify label).
• Children and teenagers with depression: the add-on depression licences are for adults only (Vraylar label; Rexulti label). See depression in children and adolescents.

Use with particular care

• Diabetes, prediabetes, obesity or high cholesterol: metabolic monitoring is essential (olanzapine-fluoxetine label).
• Older adults: higher risk of akathisia, parkinsonism and falls (Lenze 2015). The Vraylar label advises starting at the low end of the dose range (label).
• Heart or blood-vessel disease, or a tendency to low blood pressure: quetiapine needs particular caution (Seroquel XR label).
• Past movement disorder, low white cell count, or compulsive gambling, shopping or eating: see the risks above.
• Known CYP2D6 "poor metabolisers": aripiprazole and brexpiprazole doses need reducing (Abilify label).

Pregnancy and breastfeeding

Babies exposed to antipsychotics in the third trimester can have movement (extrapyramidal) or withdrawal symptoms after birth. For aripiprazole and quetiapine, published studies have not established a higher risk of major birth defects or miscarriage; for cariprazine, there are no human pregnancy data (Abilify; Vraylar). Decisions in pregnancy and breastfeeding need specialist advice; see perinatal depression.

Dosage and how to take it (label doses only)

These are the US label doses for the add-on depression indication, reported for information only. Doses are set by the prescriber.

Sources: Abilify; Seroquel XR; Rexulti; Vraylar; olanzapine-fluoxetine. Dose changes for interacting drugs are covered under interactions.

How long before it works, and for how long

The trials lasted 6 to 12 weeks. The Rexulti label asks prescribers to reassess periodically whether treatment is still needed (label). NICE asks for a review at every appointment of whether to continue, based on current physical and mental health risks (NICE 1.4.36). In open-label follow-up, lower maintenance doses of aripiprazole (2–5 mg) appeared to cause fewer side effects (Seshadri 2021).

Driving

Sleepiness and slowed reactions are common, especially with quetiapine. Do not drive or operate machinery until you know how the medicine affects you (olanzapine-fluoxetine label).

Follow the money: who makes it and who funded the evidence

Who makes and sells these drugs

• Otsuka (Japan): developed aripiprazole and co-promoted Abilify in the US with Bristol-Myers Squibb (DOJ 2008). Discovered brexpiprazole and co-markets Rexulti with Lundbeck of Denmark (Otsuka).
• AstraZeneca (UK/Sweden; US arm in Wilmington, Delaware): Seroquel and Seroquel XR (DOJ 2010).
• Gedeon Richter (Hungary) and AbbVie (US): cariprazine; Gedeon Richter led its research and AbbVie markets it in the US (AbbVie). The depression trial was run by Forest Laboratories with Gedeon Richter (registry).
• Eli Lilly (US): makes olanzapine as Zyprexa (DOJ 2009). Generic olanzapine-fluoxetine is sold by firms such as Teva (label).

Aripiprazole and quetiapine are now widely available as generics, while brexpiprazole and cariprazine are newer branded products.

Who funded the evidence

• Registration trials: sponsored by the companies that sell the drugs. Examples are Bristol-Myers Squibb and Otsuka for aripiprazole (NCT00095823), Otsuka for brexpiprazole (NCT01360645) and Forest for cariprazine (NCT01469377).
• Cochrane: 9 of its 10 included trials were industry-sponsored. The authors noted that such trials may be more likely to favour the drug tested (Davies 2019).
• Publicly funded trials: VAST-D was paid for by the US Department of Veterans Affairs. The VA took part in design and analysis but not the decision to publish, and Bristol-Myers Squibb supplied aripiprazole. One co-author reported a spouse employed by, and owning stock in, BMS, and others reported industry consulting (VAST-D full text). IRL-GRey was funded by NIMH and academic foundations (Lenze 2015). OPTIMUM was funded by PCORI (Lenze 2023).
• Independent reviews: The Spielmans funding statement and author disclosures were not verified in the full text (PLOS Med). Hsu was funded by Taiwanese public and hospital grants (BMJ Ment Health).

Documented integrity events

• AstraZeneca, Seroquel (April 2010): a $520 million civil settlement of allegations that from 2001 to 2006 the company promoted Seroquel for unapproved uses, with "depression" among them. The government also alleged ghostwritten articles and payments to doctors in breach of anti-kickback law. AstraZeneca entered a five-year Corporate Integrity Agreement (DOJ).
• Bristol-Myers Squibb, Abilify (September 2007): part of a settlement of over $515 million. The government alleged that from 2002 to 2005 BMS promoted Abilify for children and for dementia-related psychosis, both off-label (DOJ).
• Otsuka, Abilify (March 2008): paid over $4 million to settle similar allegations about children and dementia-related psychosis (DOJ).
• Eli Lilly, Zyprexa (January 2009): pleaded guilty to a misdemeanour and agreed to pay $1.415 billion, including a $515 million criminal fine. The plea concerned off-label promotion of Zyprexa for dementia in elderly patients (DOJ).

These cases concerned marketing, not the add-on depression trials themselves. We found no settlement specific to brexpiprazole or cariprazine.

What this means: The licences rest on short trials paid for by the companies that sell the drugs. The independent evidence confirms those trials found a real effect, but shows it is smaller and more costly in side effects than the original trial reports suggested. Publicly funded head-to-head trials suggest aripiprazole add-on is one reasonable option among several, not a clear winner.

Related research

For wider, evidence-graded context, see these Pure City Research guides:

• Treatment-resistant depression: assessment, treatment evidence and safety
• Major depressive disorder
• Bupropion: independent evidence
• Lithium: bipolar, suicide and kidney risk
• Esketamine (Spravato) for treatment-resistant depression
• Brain stimulation for depression
• Depression with psychotic features
• Bipolar depression
• Depression in older adults
• Depression relapse prevention
• Mirtazapine: depression, sleep and weight gain
• Antidepressant withdrawal and dependence

Frequently asked questions

Sources and funding notes

What we could not verify: the funding statements of the Nuñez 2022 network meta-analysis (only author disclosures were found), the Zhou 2015 network meta-analysis and the Seshadri 2021 long-term review. We also could not verify the funding of the individual open-label aripiprazole studies Seshadri pooled. We did not obtain the current owner of the Symbyax brand or citable sales figures for these products, so market size is omitted. We could not confirm the UK licence status of each individual drug for depression; NICE states only that the use was off-label for some. The Thase 2015 abstract does not explain why randomisation criteria were amended. The Cochrane review did not include aripiprazole or brexpiprazole trials, and we did not establish why. The DOJ settlements describe allegations; apart from Lilly's guilty plea, the companies did not admit liability in the materials we read. PubMed lists a 2025 correction notice for the Lenze 2015 IRL-GRey paper; we did not see what it corrects, and the figures we quote are from the PubMed abstract. The sponsors of the Cochrane trials, and the funding and author disclosures of Spielmans and Nuñez, were not confirmed in full text. We did not check lumateperone's full US label, its side-effect tables or who funded its trials beyond the company's own announcement.

1. NICE NG222. Depression in adults: treatment and management (29 June 2022) — UK guideline; NICE, sponsored by the Department of Health and Social Care (GOV.UK).
2. Abilify (aripiprazole) US prescribing information — FDA-approved label; Otsuka.
3. Seroquel XR (quetiapine) US prescribing information — FDA-approved label; AstraZeneca.
4. Rexulti (brexpiprazole) US prescribing information — FDA-approved label; Otsuka.
5. Vraylar (cariprazine) US prescribing information — FDA-approved label; Allergan/AbbVie.
6. Olanzapine and fluoxetine capsules US prescribing information — FDA-approved generic label; Teva.
7. EMA: Rxulti and EMA: Reagila — EU regulator; indications.
8. Mohamed S et al. VAST-D randomised clinical trial. JAMA 2017 (full text; NCT01421342) — VA Cooperative Studies Program; BMS supplied drug.
9. Lenze EJ et al. Aripiprazole augmentation for treatment-resistant late-life depression. Lancet 2015 (NCT00892047) — NIMH and foundations.
10. Lenze EJ et al. Antidepressant augmentation versus switch in treatment-resistant geriatric depression (OPTIMUM). NEJM 2023 (NCT02960763) — PCORI.
11. Davies P et al. Pharmacological interventions for treatment-resistant depression in adults. Cochrane 2019 (full text) — NIHR ARC West and NIHR Bristol BRC.
12. Spielmans GI et al. Adjunctive atypical antipsychotic treatment for MDD: meta-analysis. PLOS Med 2013 — funding and disclosures not verified in full text.
13. Hsu TW et al. Reappraising lithium, T3 and second-generation antipsychotic augmentation. BMJ Ment Health 2026 (full text) — Taiwan public and hospital grants.
14. Nuñez NA et al. Augmentation strategies for treatment resistant major depression: network meta-analysis. J Affect Disord 2022 — funding and author disclosures not verified in full text.
15. Zhou X et al. Atypical antipsychotic augmentation for TRD: network meta-analysis. Int J Neuropsychopharmacol 2015 — funding not verified.
16. Seshadri A et al. Long-term efficacy and tolerability of adjunctive aripiprazole. Prim Care Companion CNS Disord 2021 — funding not verified.
17. Berman RM et al. Aripiprazole as adjunctive therapy in MDD. J Clin Psychiatry 2007 (NCT00095823) — Bristol-Myers Squibb with Otsuka.
18. Thase ME et al. Adjunctive brexpiprazole 2 mg in MDD. J Clin Psychiatry 2015 (NCT01360645) — Otsuka.
19. Durgam S et al. Adjunctive cariprazine in inadequate responders to antidepressants. J Clin Psychiatry 2016 (NCT01469377) — Forest Laboratories with Gedeon Richter.
20. US DOJ: AstraZeneca to pay $520 million for off-label drug marketing (27 April 2010) — US government.
21. US DOJ: Bristol-Myers Squibb settlement (September 2007) — US government.
22. US DOJ: Otsuka to pay more than $4 million over Abilify (27 March 2008) — US government.
23. US DOJ: Eli Lilly agrees to pay $1.415 billion over Zyprexa (15 January 2009) — US government.
24. Otsuka: FDA approves Rexulti (July 2015) and AbbVie: FDA approves Vraylar for adjunctive MDD (December 2022) — company press releases; interested parties.
25. Johnson & Johnson: FDA approval of Caplyta (lumateperone) for adjunctive MDD (6 November 2025) — company press release; interested party.
26. Cochrane commercial sponsorship policy — Cochrane.

Educational research reviewed 6 October 2026. This article is not medical advice; talk to a qualified clinician or pharmacist before starting, stopping or combining treatments.