Prescribing Iloperidone

Why Iloperidone Matters

Iloperidone binds the alpha-1 adrenergic receptor more tightly than it binds anything else. Its Ki there is 0.36 nM, against 5.6 nM at 5-HT2A and 6.3 nM at D2. Several of its problems trace to that blockade: orthostatic hypotension with dizziness and a fast heart rate, priapism, and floppy iris during cataract surgery. The week-long titration exists to keep blood pressure from dropping. It also lengthens QTc, by about 9 msec at 12 mg twice daily and about 19 msec when both of its metabolic pathways are blocked, so screen the heart and the medication list before starting.

In exchange, it rarely causes movement side effects. Akathisia in the schizophrenia trials ran at or below placebo (1.7% to 2.3% vs 2.7%). It was approved for schizophrenia in 2009 and, in April 2024, for acute manic or mixed episodes of bipolar I disorder. A randomized withdrawal trial shows it prevents relapse in schizophrenia.

Its reputation has been lukewarm from the start. Novartis's early trials left it looking weaker than haloperidol and risperidone, and it has generally been treated as a second-line drug. Weight gain is moderate, prolactin rises, and it must be taken twice a day. Its major active metabolite was approved as a separate drug, milsaperidone (Bysanti), in February 2026.

Part 1: Indications

FDA-Approved Uses

• Schizophrenia (adults). The label's evidence includes a long-term randomized withdrawal trial.
• Acute manic or mixed episodes of bipolar I disorder (adults), added April 2024. Studied only as monotherapy; there are no data adding it to lithium or valproate

Not approved: bipolar depression, bipolar maintenance, adjunctive treatment of major depression, patients under 18, and dementia-related psychosis (class boxed warning). The current label (DailyMed version of May 2026) lists no other indications.

Off-Label / Emerging

• No off-label use has controlled-trial support. Adjunctive MDD is being studied with milsaperidone, not with iloperidone itself (below).

The Evidence by Indication

Schizophrenia, acute. Two short-term trials support it. In Study 1 (6 weeks, n=706), both 12 to 16 mg/day and 20 to 24 mg/day beat placebo on the BPRS, but risperidone appeared superior in the first 2 weeks. Risperidone's faster titration probably explains at least part of that gap, according to the label, and among patients who stayed in for at least 2 weeks the two looked comparable. In Study 2 (4 weeks, n=604), 24 mg/day beat placebo on the PANSS and looked similar to ziprasidone, another drug that needs time to reach its dose. PANSS improved by 12.0 points on iloperidone, 12.3 on ziprasidone and 7.1 on placebo.

The earlier Novartis trials were less flattering. In the first, 8 and 10 mg/day failed, and haloperidol improved PANSS by 13.9 points against 9.9 for iloperidone 12 mg/day; in the other two, risperidone beat placebo at every dose while iloperidone won at three of four. The Vanda-funded authors blamed the slow titration. That may be fair, though a drug needing two weeks to ramp up is, in practice, less desirable.

Schizophrenia, relapse prevention. In REPRIEVE (Weiden et al., CNS Drugs 2016), 303 stable outpatients on 8 to 24 mg/day were randomized to stay on iloperidone or switch to placebo for up to 26 weeks. Kaplan-Meier relapse rates were 20.4% on iloperidone and 63.4% on placebo (hazard ratio 4.7 for placebo), and the data monitoring committee stopped the trial early for efficacy.

Bipolar mania. One trial supports the indication (Stahl et al., J Clin Psychiatry 2024; n=414 randomized). Iloperidone 12 mg twice daily (6 mg twice daily in CYP2D6 poor metabolizers) beat placebo on the YMRS at week 4 by 4.0 points (95% CI 5.70 to 2.25). The YMRS separated from placebo by day 14, and CGI-S and CGI-C scores were significantly better at day 28. MADRS scores improved numerically but not significantly. The trial tested iloperidone alone, so there are no data on adding it to lithium or valproate. The labeled mania schedule (Pack B) reaches full dose by day 5, faster than the schizophrenia schedule. This is a single positive trial with no active comparator. Risperidone, olanzapine and haloperidol are generally the more reliable choices when raw antimanic power is the goal.

Who Is a Good Candidate?

• A patient with schizophrenia or bipolar I mania who developed akathisia or parkinsonism on risperidone, aripiprazole, cariprazine, lurasidone or a first-generation drug
• A stable patient with schizophrenia who is already doing well on it (relapse after withdrawal was common in REPRIEVE)
• A patient with no cardiac history, normal electrolytes and no QT-prolonging co-medications
• A patient who will reliably take a twice-daily drug and can follow a written titration

Who Is a Poor Candidate?

• An agitated patient who needs control this week: the titration delays full dose, and risperidone pulled ahead in the first 2 weeks of Study 1
• Known QT prolongation, arrhythmia, recent myocardial infarction or uncompensated heart failure (the label says to avoid it), or another QT-prolonging drug that cannot be stopped
• Orthostasis risk: older adults, dehydration, antihypertensives, and especially alpha-1 blockers (prazosin, doxazosin, terazosin, tamsulosin), which the label says to avoid
• Obesity, diabetes or dyslipidemia, where weight gain matters (35% gained 7% or more of body weight in the mania trial)
• Erratic adherence: more than 3 days off means titrating again from day 1
• Planned cataract or glaucoma surgery (the label does not recommend starting it) or breastfeeding (the label advises against nursing during treatment)

Part 2: Mechanism

The mechanism in schizophrenia and bipolar I disorder is unknown; the label suggests combined D2 and 5-HT2 antagonism. It is a full antagonist (not a partial agonist like aripiprazole or cariprazine). Measured affinities:

• Alpha-1 adrenergic antagonist: Ki 0.36 nM, the strongest binding in its profile. The label links it to orthostatic hypotension, priapism and intraoperative floppy iris syndrome.
• 5-HT2A, D2 and D3 antagonist: Ki 5.6, 6.3 and 7.1 nM.
• Moderate affinity: D4 (25 nM), 5-HT7 (22 nM), 5-HT6 (43 nM).
• Low affinity: 5-HT1A (168 nM), D1 (216 nM), H1 (437 nM). No meaningful muscarinic binding (Ki above 1000 nM). The low H1 binding may explain its relatively modest sedation and weight gain, though the label tables still show both.

Metabolites. The liver clears iloperidone by carbonyl reduction, CYP2D6 hydroxylation and CYP3A4 O-demethylation. P88 is active, with a receptor profile like the parent's. P95 binds only 5-HT2A and alpha-adrenergic receptors and is about 48% of circulating drug-related exposure in extensive metabolizers.

Pharmacokinetics. Absorption is good (relative bioavailability of the tablet 96% versus solution), with a peak at 2 to 4 hours. A high-fat meal does not change Cmax or AUC, so food is optional. Half-lives in CYP2D6 extensive metabolizers are 18 hours for iloperidone, 26 for P88 and 23 for P95; in poor metabolizers they stretch to 33, 37 and 31 hours. Steady state arrives in 3 to 4 days. Exposure rises more than proportionally with dose, so the later steps of a titration raise levels more than the milligrams suggest. It is not a CYP1A2 substrate, so smoking does not matter.

Part 3: Before You Start: Workup and Candidacy

Add a cardiac and blood pressure screen to the usual antipsychotic baseline.

Ask about fainting, palpitations, planned eye surgery, prazosin or prostate drugs, fluoxetine, paroxetine or bupropion, azole antifungals, clarithromycin and methadone. For men, mention the ejaculatory side effects and priapism before the first dose.

Part 4: How to Start and Dose

The label settles the schedules and the dose cuts but leaves several practical questions open: how to slow a titration, how to cross-taper, when to get an ECG, how to restore a dose after an inhibitor stops, and how to stop. The answers on those points in this guide are ordinary clinical practice, not label requirements.

The Titration Schedules

Both schedules start at 1 mg twice daily. The mania schedule takes bigger steps and gets to full dose two days sooner.

The labeled schizophrenia range is 6 to 12 mg twice daily, so a patient doing well at 6 mg twice daily can stay there. Study 1 found both 12 to 16 and 20 to 24 mg/day effective, and the label's dose-related adverse reactions (dizziness, tachycardia, hypotension, weight gain, abdominal discomfort, musculoskeletal stiffness) were at least twice as common at 20 to 24 mg/day.

Timing and Food

Give it twice daily, with or without food. When dizziness is the problem, a slower climb is an option: hold a step for a day or two before moving on. That fits the label's reasoning that speed drives orthostasis.

Restarting After a Gap

Repeat the full titration whenever a patient has been off iloperidone for more than 3 days (label). Tell patients this at the start, and keep a titration pack or written schedule on hand. Restarting at full dose after a week off invites the orthostasis and syncope the titration is meant to prevent.

Switching From Another Antipsychotic

No cross-taper schedule is labeled. Run the iloperidone titration while tapering the old drug over one to two weeks, and watch for additive orthostasis. Iloperidone will not reach target until day 4 to 7, so expect a few days of thinner coverage.

Dose Adjustments and Special Populations

• CYP2D6 poor metabolizers (about 7% of White and 2% of Asian and African American patients): halve the dose. Their exposure to iloperidone and P88 is 47% and 85% higher.
• Strong CYP2D6 or CYP3A4 inhibitors: see Part 8. Fluoxetine lingers for weeks after it stops, so restore the full iloperidone dose gradually.
• Hepatic: mild impairment needs no change. Moderate impairment raised free P88 exposure about twofold, so a reduction may be needed. Severe impairment was not studied, and the label does not recommend use.
• Renal: creatinine clearance under 30 mL/min had minimal effect on peak levels (AUC up 24% for iloperidone and 52% for P95). No adjustment is labeled.
• Geriatric: only 25 of 3,210 schizophrenia patients were 65 or older, and none were 75 or older. No separate schedule is labeled; climb more slowly.

Part 5: Monitoring

No fixed schedule is labeled. This one follows usual atypical-antipsychotic practice, with orthostatic vitals and heart rate up front during titration.

Hematocrit, hemoglobin, WBC, total protein and albumin fell together in the trials, which the label attributes to hemodilution, as with other alpha antagonists.

Part 6: Side Effects and How to Manage Them

Discontinuation for adverse reactions was 5% vs 5% in the schizophrenia trials and 8.7% vs 5.3% in mania (liver enzymes, nausea and vomiting, dizziness, hypotension). Rates below are drug vs placebo from the label; schizophrenia figures are for 10 to 16 and 20 to 24 mg/day.

Dizziness, Orthostasis and Tachycardia

Dizziness was 10% and 20% vs 7% in schizophrenia and 12% vs 1% in mania. Orthostatic hypotension was 3% and 5% vs 1%, and syncope 0.4% vs 0.2% even with slow titration. Tachycardia was 3% and 12% vs 1% in schizophrenia and 23% vs 5% in mania, where the combined term included postural orthostatic tachycardia. Dizziness and tachycardia were at least twice as common on 20 to 24 mg/day as on 10 to 16 mg/day.

• Management: titrate as labeled or slower, rise slowly, keep fluids up, review antihypertensives and alpha-1 blockers, and stay at the low end of the range when it works. Palpitations or fainting need an ECG and cardiac evaluation (label).

Weight and Metabolic

Weight gain is moderate and dose-related. In the schizophrenia trials, mean change was 2.0 kg and 2.7 kg vs -0.1 kg, and 12% and 18% gained 7% or more vs 4%. The mania trial was worse: 4.6 kg vs 1.6 kg in 4 weeks, with 35% vs 14% gaining 7% or more. Fasting glucose shifted from normal to 126 mg/dL or higher in 10.7% vs 2.5% in a 4-week schizophrenia trial; in longer pooled data, cholesterol and triglycerides fell on average. For weight, it sits in a middle tier with asenapine, above aripiprazole, lurasidone and ziprasidone and below olanzapine and clozapine.

• Management: early and frequent weight checks, dietary counseling, and a lower dose if it works. If gain continues, switch to a lower-weight agent.

Sedation

Somnolence was 12% vs 5.3% at 10 mg/day or more in schizophrenia and 8% vs 3% in mania. The usual driving caution applies.

Dry Mouth, Nasal Congestion, GI

Dry mouth was 8% and 10% vs 1% (9% vs 2% in mania). Nasal congestion was 5% and 8% vs 2%; warn patients so it does not get mistaken for a cold. Nausea ran close to placebo (7% to 10% vs 8%), and diarrhea was 5% to 7% vs 4%.

Sexual and Urinary Effects

Ejaculation failure was 2% vs under 1% in schizophrenia. In the mania trial, sexual dysfunction (ejaculation failure, erectile dysfunction, retrograde ejaculation and delayed ejaculation combined) was 4% vs 0.5%. Retrograde ejaculation is also a postmarketing report. Urinary urgency and frequency appeared in 3% vs 0% in mania, and urinary incontinence was listed as frequent across the schizophrenia database.

• Management: ask directly at follow-up. If it is intolerable, switch drugs; nothing in the label suggests a treatment for it.

Extrapyramidal Symptoms and Akathisia

Movement side effects are iloperidone's strength. All EPS-related events were 13.5% and 15.1% vs 11.6% on placebo in schizophrenia, and akathisia was 1.7% and 2.3% vs 2.7%. Parkinsonism was 0.2% to 0.3%. The mania trial was less clean, with all EPS events at 8.3% vs 0% and akathisia 4.4% vs 0%. Iloperidone is among the atypicals with the lowest akathisia rates, and it causes less akathisia than risperidone.

Prolactin

Iloperidone raises prolactin, but modestly. In a 4-week schizophrenia trial, prolactin rose 2.6 ng/mL vs a 6.3 ng/mL fall on placebo, and 26% vs 12% had levels above 1.15 times the upper limit. In mania the rise was larger (15.7 ng/mL, with 35% vs 1% elevated). Clinical events were rare: gynecomastia in 0.1% and galactorrhea in 0.2% (0.5% on placebo) across 3,210 patients.

Liver Enzymes

ALT rose to at least 3 times the upper limit in 9.2% vs 1.5% in the mania trial, asymptomatically, and AST rises were less common. Liver enzyme rises were one of the more common reasons for stopping in that trial.

Part 7: Toxicity, Overdose, and Boxed Warnings

QT Prolongation

QT prolongation sits in the warnings section; it has no boxed warning. In an open-label study of 160 patients, 12 mg twice daily prolonged QTc by 9 msec. Paroxetine (2D6 inhibition) or ketoconazole (3A4 inhibition) increased the effect, and with both, the mean QTcF increase was about 19 msec. No torsades or severe arrhythmia occurred before marketing. In REPRIEVE, mean QTcF change during stabilization was 6.4 msec. In the mania trial, QTcF rose 8.3 msec by day 28, and 3 patients on iloperidone vs none on placebo had a rise of 60 msec or more (Stahl 2024). In the head-to-head trial, QTc rose 7.2 msec on iloperidone vs 6.1 on ziprasidone, so its QT liability is at least equal to ziprasidone's.

• Avoid with other QT-prolonging drugs (Part 8), and in congenital long QT, arrhythmia history, recent MI and uncompensated heart failure.
• Risk rises with bradycardia, hypokalemia and hypomagnesemia. Stop the drug if QTc stays above 500 msec.

Contraindication and Other Serious Risks

• Hypersensitivity to iloperidone or any component is the only contraindication. Anaphylaxis, angioedema, throat tightness, facial and tongue swelling, urticaria and rash have been reported.
• Priapism: 4 cases in the premarketing program (3 in schizophrenia, 1 in mania), attributed to alpha-adrenergic blockade. Severe cases may need surgery, so tell male patients that a prolonged erection needs emergency care.
• Intraoperative floppy iris syndrome: seen with alpha-1 blockers. Patients should tell their ophthalmologist they take or have taken iloperidone. Stopping before surgery does not appear to help.
• Class warnings: neuroleptic malignant syndrome (stop the drug and treat intensively), tardive dyskinesia (highest risk in older women; use the lowest effective dose), leukopenia and agranulocytosis (stop for ANC under 1000/mm3), cerebrovascular events in dementia, falls, seizures, dysphagia and heat intolerance.

Overdose

Eight premarketing overdoses (48 to 576 mg) caused no deaths. One patient who took 438 mg over 4 days had EPS and a QTc of 507 msec without cardiac sequelae. Expect drowsiness, tachycardia and hypotension.

Part 8: Drug Interactions

Iloperidone's levels move with CYP2D6 and CYP3A4 inhibitors, and its QT and alpha-blocking effects add to other drugs. It changes other drugs' levels very little: dextromethorphan exposure rose 17% and midazolam exposure less than 50%, which the label judges unlikely to matter.

Part 9: Special Populations

Pregnancy

Human data are too limited to establish a risk of birth defects or miscarriage. It was not teratogenic in animals, but rats had prolonged labor, more stillbirths (even at 1.6 times the maximum human dose) and lower pup survival. Third-trimester exposure to any antipsychotic risks neonatal extrapyramidal and withdrawal symptoms (agitation, hypertonia, hypotonia, tremor, somnolence, respiratory distress, feeding problems).

• Framework: untreated schizophrenia and mania carry real risks to mother and pregnancy. Iloperidone has little human pregnancy data, so in someone planning pregnancy, weigh a switch to an antipsychotic with more data against the risk of destabilizing her.
• Enroll patients in the National Pregnancy Registry for Atypical Antipsychotics (1-866-961-2388).

Lactation

There are no human milk data. In rats, milk radioactivity 4 hours after a dose was nearly 10 times the plasma level. Because of the potential for serious adverse reactions in the infant, the label advises women not to breastfeed during treatment.

Fertility

Iloperidone reduced fertility in rats, and prolactin elevation can impair reproductive function.

Elderly

• Boxed warning applies. Not approved for dementia-related psychosis.
• Older adults are named for orthostatic monitoring. Trial experience is almost nil: 25 patients 65 or older in schizophrenia and 2 in mania.

Renal Impairment

No adjustment is labeled. See Part 4 for exposure numbers.

Hepatic Impairment

Mild (Child-Pugh A): no change. Moderate (B): may require dose reduction. Severe (C): not recommended.

Pediatric / Adolescent

Safety and effectiveness have not been established. Any use under 18 is off-label.

Part 10: Discontinuation and Taper

No withdrawal syndrome or taper schedule is described. With half-lives of about a day (longer in poor metabolizers), the drug clears within several days.

• Schizophrenia: REPRIEVE is the argument against stopping casually. Stable patients switched to placebo relapsed at 63.4% vs 20.4% over up to 26 weeks.
• Mania: there is no maintenance data. Decide whether to continue it, taper it under a mood stabilizer, or switch to a drug with maintenance evidence.
• How to stop: taper over one to two weeks; the label does not require it.
• Restarting: after more than 3 days off, repeat the titration from day 1. Patients should still tell an eye surgeon they once took it.

Part 11: Iloperidone vs the Alternatives

Milsaperidone (Bysanti)

The FDA approved milsaperidone, iloperidone's major active metabolite, in February 2026 for schizophrenia and acute manic or mixed episodes of bipolar I disorder in adults. It interconverts with iloperidone in the body, and the approval drew on its bioequivalence to iloperidone and iloperidone's clinical record. Vanda expected a launch in the second half of 2026 and has a Phase III trial of milsaperidone as once-daily adjunctive treatment for MDD, with results expected in the first half of 2027. Paliperidone, risperidone's metabolite, never won a bipolar indication, while milsaperidone did.

Long-Acting Injectable

No long-acting injectable iloperidone is approved. Vanda's March 2026 investor presentation listed a Fanapt long-acting injectable Phase III program as ongoing. If a patient needs an injectable now, choose among the approved ones (risperidone, paliperidone, aripiprazole, olanzapine, haloperidol, fluphenazine).

Cost

Fanapt is still brand-only. A Mylan generic iloperidone label sits on DailyMed (2025), but patent settlements hold generic entry to November 2027.

The Bedside Cheat Sheet