Antipsychotics for depression refers to using a second-generation antipsychotic, usually at low dose, as an add-on to an antidepressant when standard treatment alone has not worked. Four agents carry FDA approval for this purpose: aripiprazole, brexpiprazole, quetiapine extended-release, and olanzapine (paired with fluoxetine). Add-on aripiprazole lifts remission rates to roughly 26 to 28 percent versus 15 to 16 percent with placebo, a benefit of about 12 percentage points, per the STAR*D and VA augmentation literature. These drugs are not first-line. They enter the picture after one or two failed antidepressant trials.
How an antipsychotic relieves depression
Depression is not a serotonin deficit, and antipsychotics do not fix one. Their antidepressant effect comes from blocking two dopamine receptors and one serotonin receptor at the same time. At low doses, blocking the dopamine D2 receptor appears to normalize a sluggish reward circuit, the pathway that makes ordinary things feel worth doing.
The serotonin 5-HT2A block matters just as much. Blocking 5-HT2A releases the brake on dopamine and norepinephrine firing in the prefrontal cortex. That is why a drug designed to quiet psychosis can, at one-fifth of its usual dose, lift motivation and energy. In plain terms: these drugs turn the volume back up on a reward system that depression has turned down.
The strongest single mechanism signal comes from a 2022 network meta-analysis of 42 trials covering more than 8,000 patients. Aripiprazole, a partial dopamine agonist, produced the largest response, with an odds ratio around 1.9 against placebo. Drugs with heavier sedation and histamine blockade did less well. That pattern supports the dopamine-reward hypothesis over a simple sedative explanation.
The four approved drugs and their numbers
Each approved agent has its own dose range, evidence base, and side-effect profile. The table below compares them on the facts a clinician or patient actually needs.
| Drug | Typical add-on dose | Response vs placebo | Best-supported profile |
|---|---|---|---|
| Aripiprazole | 2 to 10 mg daily | About 26 to 28% vs 15 to 16% remission | Largest effect size; least sedation; higher akathisia |
| Brexpiprazole | 0.5 to 2 mg daily | About 24 to 26% vs 15 to 16% remission | Smoother dopamine partial agonist; weight gain |
| Quetiapine XR | 150 to 300 mg nightly | About 20 to 25% vs 15% remission | Strongest for insomnia and anxiety; sedation, metabolic risk |
| Olanzapine-fluoxetine | 5 to 12.5 mg olanzapine | About 25 to 30% vs 15% remission | Highest metabolic burden; reserved for refractory cases |
Doses are far below antipsychotic range. Aripiprazole for schizophrenia runs 10 to 30 mg; for depression, it starts at 2 to 5 mg. Quetiapine for schizophrenia can reach 800 mg; as an add-on it usually stops at 300 mg. The dose gap is the single most important practical fact: these are not antipsychotic doses, and the side effects are dose-proportional.
What the augmentation studies actually showed
The largest evidence set comes from the VA Augmentation and Switching Strategies trial, published in 2015, which enrolled 1,522 veterans with depression that had not remitted on an adequate antidepressant. Adding aripiprazole beat switching to another antidepressant by a margin of about 8 percentage points on remission. Quetiapine XR also outperformed switching but with more dropout from sedation.
An earlier sequence, STAR*D, enrolled 4,041 patients across four treatment steps. At each step, remission fell: roughly 37 percent at step one, then lower with each failure. That falling curve is why augmentation matters. Once a second antidepressant fails, the odds of spontaneous remission on a third single drug drop below 20 percent.
The olanzapine-fluoxetine combination earned approval in 2009 for treatment-resistant depression, based on trials showing separation from either drug alone. The effect is real but the metabolic cost is the highest in the class, with weight gain averaging 3 to 5 kg over 8 to 12 weeks in longer open-label follow-up.
Timeline: what happens after the first dose
Antipsychotic augmentation does not work overnight, and it does not fail overnight either. The time course below reflects what trials and clinical guidelines describe.
| Time point | What to expect | Action |
|---|---|---|
| Day 1 to 3 | Sedation, dizziness, or mild nausea | Start low; take sedating drugs at night |
| Week 1 to 2 | Sleep and appetite often change first | Recheck blood pressure for orthostasis |
| Week 2 to 4 | Early mood signal in responders | No dose increase yet |
| Week 4 to 6 | Main antidepressant effect emerges | Assess response objectively |
| Week 6 to 8 | Decision point: continue or stop | Discontinue non-responders by week 8 |
| Week 8 to 12 | Stable responders consolidate | Check weight, fasting glucose, lipids |
| Month 3 to 6 | Remission maintained on lowest dose | Consider tapering only after 6 months stable |
The 6-to-8-week assessment window is not arbitrary. A 2024 individual-participant analysis found most of the benefit appears by week 4 in eventual responders, and patients who show nothing by week 6 rarely convert later. In plain terms: if it has not started working by six weeks, it probably will not.
The side effects that decide whether people stay on it
The single most common reason patients stop is akathisia, an inner restlessness that is easy to mistake for worsening anxiety. Aripiprazole causes it in roughly 15 to 25 percent of patients at depression doses; quetiapine causes it in under 5 percent. The paradox is sharp: the drug with the best mood evidence also causes the most restlessness.
Metabolic effects arrive more slowly but matter more over years. Second-generation antipsychotics as a class raise weight by 2 to 6 kg within 3 months, with olanzapine and quetiapine at the top and aripiprazole and brexpiprazole lower. Fasting glucose can rise independent of weight.
Tardive dyskinesia, the involuntary movement disorder, is rare at these low doses and short durations but not zero. The FDA added a boxed warning for all antipsychotics in 2008. Metabolic monitoring is not optional. Professional guidelines call for weight, waist, blood pressure, and fasting glucose at baseline and again at 12 weeks.
What to actually do before starting:
- Get baseline weight, waist circumference, blood pressure, and fasting glucose or A1c.
- Ask specifically about prior akathisia, since a past episode predicts a repeat.
- Start at the lowest dose, usually half the target, and titrate once.
- Set the week-6 review date before the first pill.
- Agree in writing on what "working" means, using a score like the PHQ-9.
Common misconceptions, corrected
Misconception one: these are sedatives that just help sleep. Sleep improvement is real, especially with quetiapine, but the 2022 network analysis separated mood effect from sedation, and the least sedating drug had the biggest effect.
Misconception two: one failed antidepressant means you need an antipsychotic. It does not. Guidelines place dose optimization, switching, and psychotherapy ahead of augmentation. Antipsychotics usually rank third or fourth. In plain terms: they are a step up, not a first move.
Misconception three: low doses mean no metabolic risk. Weight gain still occurs, and guidelines do not relax monitoring for low-dose use. The 2015 VA trial reported weight gain even in the aripiprazole arm.
Misconception four: they must be taken forever. Most responders can taper after 6 to 12 months of stable remission, though relapse risk after stopping is real and merits a monitored taper rather than an abrupt stop.
Limits of the evidence and the honest caveats
Response rates near 26 percent mean roughly three in four patients still do not reach remission on any single augmentation. Combining an antipsychotic with the original antidepressant leaves 70-plus percent with residual symptoms.
Industry funding colors the literature. Most pivotal trials of the four approved agents were manufacturer-sponsored, and independent replication is thinner than the approvals suggest. The 2022 network meta-analysis relied heavily on these trials.
The evidence base skews adult and non-geriatric. In older adults with depression, antipsychotics raise mortality and stroke risk, and the FDA boxed warning for dementia-related psychosis applies. In bipolar depression, the risk-benefit calculus differs entirely, and this class is used differently. None of these drugs is approved for children with depression; their pediatric depression approvals do not exist, and use there is off-label and uncommon.
Finally, the comparison to psychotherapy augmentation is incomplete. Trials pitting an antipsychotic add-on against cognitive behavioral therapy add-on head-to-head are scarce, so claims that one beats the other rest on indirect comparison only.
Antipsychotics for depression work best as a precisely dosed, time-limited, monitored add-on in people who have failed at least one well-tried antidepressant. The mechanism is reward-circuit repair, not sedation. The evidence is strongest for aripiprazole, the window for judging response is 6 to 8 weeks, and the metabolic bill is real even at low doses.