Why Pimavanserin Matters
Pimavanserin treats psychosis without binding dopamine receptors. In vitro it is an inverse agonist and antagonist at 5-HT2A receptors (Ki 0.087 nM), with lower affinity for 5-HT2C, and it has no appreciable affinity for D2, muscarinic, histaminergic or adrenergic receptors. That matters in Parkinson's disease, where most antipsychotics worsen motor symptoms and the dopaminergic drugs that treat Parkinson's can worsen the psychosis. In the pivotal trial, motor scores on pimavanserin did not differ from placebo.
It has one FDA indication, unchanged since its 2016 approval: hallucinations and delusions associated with Parkinson's disease psychosis. Every attempt to widen it has failed. The FDA turned down dementia-related psychosis in April 2021 and Alzheimer's disease psychosis in August 2022. Phase 3 trials missed as an add-on for major depression (2020), as an add-on in schizophrenia with inadequate response (2019), and for negative symptoms of schizophrenia (March 2024), after which Acadia said it would run no further pimavanserin trials. The current label (revised April 2026) also describes a negative pediatric trial in irritability with autism.
The benefit in Parkinson's psychosis is real and modest: 3.06 points over placebo on a 45-point scale in the one positive 6-week trial. The safety issues to plan around are QT prolongation, peripheral edema, confusion, and the class boxed warning on mortality in dementia-related psychosis. It is brand-only, costs about $5,000 a month at cash prices, and patent rulings in 2025 may keep generics away until about 2038.
Use pimavanserin for Parkinson's psychosis that still needs treatment after you have trimmed the Parkinson's drugs and handled the nondrug causes, when a D2 blocker would cost the patient motor function and clozapine's blood monitoring is not workable. Clear the QT and CYP3A4 lists before you write it, expect a modest effect that builds over weeks, and stop it if the psychosis has not clearly improved.
Part 1: Indications
FDA-Approved Uses
- Hallucinations and delusions associated with Parkinson's disease psychosis (adults)
Read the boxed warning's wording closely. Pimavanserin is "not approved for the treatment of patients with dementia who experience psychosis unless their hallucinations and delusions are related to Parkinson's disease." Psychosis in Parkinson's disease dementia sits inside the label. The pivotal trial enrolled Parkinson's patients with or without dementia, but required an MMSE of 21 or higher and the ability to self-report symptoms.
Not approved: psychosis in Alzheimer's disease or any other dementia outside Parkinson's, schizophrenia (adjunctive or negative symptoms), major depression, and anyone under 18.
Off-Label / Emerging
- Dementia-related psychosis outside Parkinson's disease, including dementia with Lewy bodies: one positive relapse-prevention trial (HARMONY), rejected by the FDA as the basis for approval
- Negative symptoms of schizophrenia: one positive trial with a small effect (0.21), then a failed phase 3
The Evidence by Indication
| Trial | Population | Result |
|---|---|---|
| Study -020 (Cummings, Lancet 2014) | Parkinson's psychosis, 6 weeks, 199 randomized | SAPS-PD -5.79 vs -2.73 (difference 3.06, 95% CI 1.20 to 4.91). Basis of the 2016 approval |
| Study -012 (2009) | Parkinson's psychosis, 298 patients | Missed its primary endpoint |
| Ballard, Lancet Neurol 2018 | Alzheimer's psychosis, nursing homes | Beat placebo at week 6 (primary endpoint, NPI-NH psychosis -3.76 vs -1.93, p=0.045); no difference by week 12 |
| HARMONY (Tariot, NEJM 2021) | Dementia-related psychosis, relapse prevention | Relapse 13% vs 28% (HR 0.35); stopped early. FDA declined approval |
| ENHANCE (2019) | Schizophrenia, add-on after inadequate antipsychotic response, 396 patients | Missed on PANSS total (p=0.094) |
| ADVANCE (Bugarski-Kirola, Lancet Psychiatry 2022) | Negative symptoms, add-on, 403 patients, 26 weeks | NSA-16 -10.4 vs -8.5, effect size 0.21; no change in function |
| ADVANCE-2 (2024) | Negative symptoms, add-on, 454 patients, 34 mg | NSA-16 -11.8 vs -11.1, effect size 0.07; failed |
| CLARITY (phase 2, 2019; phase 3, 2020) | MDD, add-on to SSRI or SNRI | Phase 2 positive (n=207); combined phase 3 missed on HAMD-17 at week 5 (p=0.296) |
Parkinson's disease psychosis. The label rests on one 6-week outpatient trial. Patients were 40 or older, had Parkinson's for at least a year, and had psychosis that began after the diagnosis and was bad enough to warrant an antipsychotic. Their Parkinson's drugs were held stable for 30 days before and throughout. The outcome was the SAPS-PD, a 9-item, 0 to 45 scale adapted from the hallucinations and delusions domains of the SAPS. Baseline scores were about 15. Pimavanserin beat placebo by 3.06 points, with a separate effect on hallucinations (2.01) and a smaller one on delusions (0.94, with a confidence interval that just excluded zero). The gap kept widening through week 6. An earlier 298-patient phase 3 trial missed its primary endpoint in 2009 (PharmaTimes).
A meta-analysis of four randomized trials found only a 2.3-point reduction on the 100-point SAPS hallucinations and delusions scale, a change that resolving one mild delusion could produce. The trials also carry some risk of bias: incomplete outcome reporting in one arm, thin detail on randomization and blinding, an unvalidated rating scale, and exclusion of patients with an MMSE under 21. Many of these patients stay psychotic despite antipsychotic treatment, so reducing dopaminergic drugs and nondrug measures stay central.
Dementia-related psychosis (off-label). HARMONY enrolled 392 patients with psychosis in any of five dementia types (Alzheimer's, Lewy body, frontotemporal, Parkinson's disease dementia and vascular) on open-label pimavanserin 20 to 34 mg/day. The 217 responders were randomized to continue or switch to placebo, and an interim analysis stopped the trial early. Relapse occurred in 12 of 95 on pimavanserin (13%) and 28 of 99 on placebo (28%), a hazard ratio of 0.35 per Acadia. Only 44 and 35 patients completed the planned 26 weeks. Asymptomatic QTc prolongation was more frequent on the drug. Some "relapses" on placebo may have been withdrawal effects, and the mix of dementia types makes it hard to know who responds. The FDA issued a complete response letter in April 2021, citing a lack of statistical significance in some dementia subgroups and too few patients with the less common subtypes. Acadia refiled for Alzheimer's psychosis alone. An FDA advisory committee voted 9 to 3 against effectiveness in June 2022, and a second complete response letter followed in August 2022.
Schizophrenia (off-label). As an add-on for patients with an inadequate response to their antipsychotic, ENHANCE missed its primary endpoint. For negative symptoms, ADVANCE found a small effect on the NSA-16 (0.21, larger at 34 mg in post-hoc analysis) with no change in functioning. ADVANCE-2 then failed in March 2024 with an effect size of 0.07.
Major depression (off-label). A 207-patient phase 2 trial of pimavanserin added to an SSRI or SNRI showed a medium effect (0.5 to 0.6) that needed replication. The combined CLARITY-2 and CLARITY-3 phase 3 study did not separate from placebo, so pimavanserin appears ineffective in depression.
Before you prescribe, the Parkinson's regimen is the first lever: stop unnecessary drugs and reduce dopaminergic medications, tapering the less potent ones first. Alongside that, teach caregivers to redirect, keep lights on during the day to cut visual hallucinations, and treat hearing loss or tinnitus for auditory ones.
Who Is a Good Candidate?
- A patient with Parkinson's psychosis, with or without Parkinson's disease dementia, whose hallucinations or delusions still need treatment after the regimen has been simplified
- A patient whose motor function cannot absorb a D2 blocker
- A patient or family who cannot manage clozapine's blood count monitoring
- A normal QTc and no QT-prolonging drugs that have to stay
- Insurance coverage in place, usually through a prior authorization
Who Is a Poor Candidate?
- Known QT prolongation, a history of cardiac arrhythmia, symptomatic bradycardia, hypokalemia, hypomagnesemia or congenital long QT; the label says to avoid it in all of these
- A patient who needs a QT-prolonging drug, such as a class 1A or class 3 antiarrhythmic
- A patient on a strong or moderate CYP3A4 inducer, such as carbamazepine, phenytoin or rifampin; the label says to avoid the combination
- Psychosis in a dementia unrelated to Parkinson's disease; it is not approved there and the boxed warning applies
- A patient whose psychosis is dangerous now; benefit builds over about 2 weeks
- A patient for whom the copay is prohibitive; the manufacturer's copay card covers commercial insurance only
Part 2: Mechanism
How pimavanserin treats psychosis is unclear. The label suggests the effect could come from inverse agonist and antagonist activity at 5-HT2A receptors and, to a lesser extent, 5-HT2C receptors. An antagonist blocks the receptor; an inverse agonist binds it and pushes its activity below baseline, the opposite of an agonist. The rationale is that overactive 5-HT2A signaling contributes to psychosis in Parkinson's disease, and the clinical hint came from low-dose clozapine, which blocks 5-HT2A with minimal dopamine blockade. What is measured:
- 5-HT2A inverse agonist and antagonist: Ki 0.087 nM, the highest-affinity target.
- 5-HT2C: Ki 0.44 nM, lower affinity.
- Sigma-1: low binding (Ki 120 nM).
- No appreciable affinity (Ki above 300 nM) for 5-HT2B, dopamine receptors including D2, muscarinic, histaminergic or adrenergic receptors, or calcium channels. That fits the absence of motor worsening and the short list of side effects.
Pharmacokinetics. Median Tmax is 6 hours, and a high-fat meal has no significant effect on exposure. CYP3A4 and CYP3A5 do most of the metabolism, with smaller roles for CYP2J2, CYP2D6 and FMO enzymes, and CYP3A4 forms the major active metabolite, AC-279. The mean half-life is about 57 hours for pimavanserin and about 200 hours for AC-279. Less than 1% of a dose of either appears in urine. Age, sex, ethnicity and weight do not change exposure in a clinically relevant way.
Pimavanserin does not inhibit or induce CYP3A4 to a clinically significant degree, is not an irreversible inhibitor of the major CYPs, and did not change midazolam or carbidopa/levodopa levels. Transporters play no significant role in its disposition.
By simple half-life arithmetic (not label text), the parent drug reaches steady state in about 12 days and the active metabolite takes several weeks. The same arithmetic runs in reverse: QT effects and side effects do not vanish the day after you stop, and a single missed dose barely moves the level.
Part 3: Before You Start: Workup and Candidacy
No baseline labs are required. The warnings section has two entries, the boxed mortality warning and QT prolongation, so the workup is about the heart, the medication list and the diagnosis. Items marked practice are common practice, not label text.
| Assessment | Why |
|---|---|
| Psychosis history and timing | The pivotal trial required psychosis that began after the Parkinson's diagnosis. Rule out delirium and recent drug changes first (practice) |
| Full medication review | QT-prolonging drugs, CYP3A4 inhibitors and inducers, and dopaminergic or other drugs that could be reduced |
| ECG | Practice: get one if the QTc is unknown, there is cardiac history, or another QT-prolonging drug is in the picture. The label says to avoid the drug in known QT prolongation |
| Potassium and magnesium | Practice: check if the patient is at risk (diuretics, vomiting, poor intake). The label lists hypokalemia and hypomagnesemia as reasons to avoid it |
| Cognitive screen (MMSE or similar) | Baseline for judging later confusion. Trial patients had an MMSE of 21 or higher |
| Symptom baseline | Assess psychotic symptoms carefully before starting so you can judge change afterward. Record frequency and distress in the patient's and caregiver's words |
| Leg exam for edema | Practice: peripheral edema is the most common adverse reaction, so document the baseline |
| Gait and fall history | Gait disturbance and falls appear in the adverse reaction data |
| Renal function | Exposure rises in severe impairment (CrCl under 30 mL/min); the label advises caution there and in ESRD |
| Insurance check | Most Medicare and commercial plans cover it, often with a prior authorization |
Weight, glucose and lipids are not label requirements. In the 26-week ADVANCE trial there were no clinically relevant effects on vital signs, weight, glucose or lipids. Ordinary medical care covers them.
Part 4: How to Start and Dose
| Parameter | Value |
|---|---|
| Formulations | Capsules 34 mg; tablets 10 mg (once daily, with or without food) |
| Dose | 34 mg once daily. No titration |
| Strong CYP3A4 inhibitor | 10 mg tablet once daily |
| Strong or moderate CYP3A4 inducer | Avoid |
| Capsule administration | Swallow whole, or open and sprinkle the entire contents on a tablespoon (15 mL) of applesauce, yogurt, pudding or a liquid nutritional supplement; take at once without chewing, do not store |
| Hepatic impairment | No dose adjustment |
| Renal impairment | No dose adjustment, mild through ESRD; caution in severe impairment and ESRD |
| Elderly | No dose adjustment |
| Pediatric | Safety and effectiveness not established |
| Boxed warning | Increased mortality in elderly patients with dementia-related psychosis |
One Dose
Start and stay at 34 mg once daily; there is no titration. The 10 mg tablet is for patients on a strong CYP3A4 inhibitor: population PK modeling showed that 10 mg taken with ketoconazole gives about the same steady-state exposure as 34 mg alone. It is not used for titration. Trials outside Parkinson's psychosis used flexible doses (20 to 34 mg in HARMONY, 10 to 34 mg in ADVANCE), and none of that is labeled.
Older references may describe two 17 mg tablets. That was the 2016 formulation; the 34 mg capsule and 10 mg tablet came with the 2018 label.
Timing and Food
Any time of day, with or without food. The sprinkle option helps a patient who has trouble swallowing capsules. Somnolence is not in the label's trial table (it shows up in postmarketing reports), so the label gives no reason to insist on bedtime dosing.
How Long Before You Judge It
The benefit builds over about 2 weeks, and the label's trial curve kept separating through week 6. Reassess at about 6 weeks, matched to the trial length (practice), and consider stopping it if there is no clear improvement.
Switching From Quetiapine or Clozapine
There is no label cross-taper schedule. Because pimavanserin takes about 12 days to reach steady state and its benefit builds over weeks, an abrupt switch can leave a gap in coverage, but none of my sources gives a cross-taper schedule. Any overlap pairs two QT-prolonging drugs, which the label says to avoid: it names ziprasidone, chlorpromazine and thioridazine, and both quetiapine's and clozapine's own labels carry QT warnings. If you do overlap, keep it short, get an ECG, and check potassium and magnesium (practice, not label).
Dose Adjustments and Special Populations
- CYP3A4: 10 mg with a strong inhibitor. Avoid strong or moderate inducers.
- Hepatic: no adjustment in the current label, based on a hepatic impairment study. The 2016 label said not to use it in hepatic impairment; that has been replaced.
- Renal: no adjustment; exposure is higher in severe impairment, and dialysis removed less than 10% of a dose.
- Geriatric: no adjustment. This is the population the drug was studied in (Part 9).
Part 5: Monitoring
No monitoring schedule is set in the label. The table below is practice built around the label's warnings and adverse reactions.
| Parameter | Baseline | Follow-up |
|---|---|---|
| Hallucinations and delusions (patient and caregiver report) | ✓ | At about 2 weeks and at 6 weeks; stop if there is no clear improvement |
| Confusion and cognition | ✓ | Every visit; confusional state was 6% vs 3% |
| Peripheral edema | ✓ | Every visit; 7% vs 2% |
| Gait, falls, motor function | ✓ | Every visit |
| ECG / QTc | If at risk | When a CYP3A4 inhibitor is added, with new cardiac symptoms, or if a QT-prolonging drug cannot be avoided (the label says to avoid the combination) |
| Potassium, magnesium | If at risk | With vomiting, diuretic changes or poor intake |
| Medication list | ✓ | Every visit; the label asks patients to report any prescription or OTC change |
| Weight, glucose, lipids | Not required | Routine medical care |
Mean QTc rose about 5 to 8 ms on 34 mg in the 6-week trials. In the thorough QT study, the maximum mean increase was 13.5 ms (upper bound of the 90% CI 16.6 ms) at twice the therapeutic dose, and the prolongation was concentration-dependent within the therapeutic range. Sporadic QTcF values of 500 ms or more, and increases of 60 ms or more, occurred at rates generally similar to placebo, and there were no reports of torsade de pointes. Anything that raises the level (a strong CYP3A4 inhibitor at full dose) or adds its own QT effect raises the concern.
Part 6: Side Effects and How to Manage Them
The adverse reaction list is short. In the 6-week trials (34 mg, n=202; placebo, n=231; mean age about 71), 8% on pimavanserin and 4% on placebo stopped for adverse reactions. Those that led to stopping in more than one patient at twice the placebo rate were hallucination (2% vs under 1%), urinary tract infection (1% vs under 1%) and fatigue (1% vs 0%). Rates did not differ by age (75 or younger vs over 75), sex, or MMSE (under 25 vs 25 or higher).
Peripheral Edema
The most common adverse reaction: 7% vs 2%.
- Management: examine the legs at each visit and look for other causes, such as heart failure, venous disease and other drugs (practice). Persistent, bothersome edema is a reason to reconsider the drug.
Confusion and Hallucinations
Confusional state ran 6% vs 3%, and hallucination 5% vs 3%. A drug given for hallucinations can cause new ones. Postmarketing reports add agitation and aggression.
- Management: compare against the baseline you recorded. New or worse confusion or hallucinations in the first weeks, without another cause such as infection or a Parkinson's drug change, point toward stopping (practice).
Nausea and Constipation
Nausea was 7% vs 4% and constipation 4% vs 3%. In healthy-volunteer studies, nausea and vomiting limited how high the dose could go.
Gait and Falls
Gait disturbance was 2% vs under 1%, and falls appear in postmarketing reports. Motor function did not worsen: the UPDRS Parts II and III change was -1.40 on pimavanserin and -1.69 on placebo, on a 0 to 160 scale.
- Management: ask about falls at every visit and loop in physical therapy or the neurologist when gait changes (practice).
Somnolence and Other Postmarketing Reports
Somnolence, rash, urticaria, angioedema-type reactions and fecal incontinence have been reported after approval, with no reliable frequency. In ADVANCE, headache and somnolence were the most common adverse effects, mostly mild, and overall rates were similar to placebo (35% to 40%).
What Is Missing From the List
None of the usual antipsychotic class warnings appears in the label: no warning for neuroleptic malignant syndrome, tardive dyskinesia, metabolic changes, orthostatic hypotension, leukopenia or seizures. The 26-week ADVANCE data showed no clinically relevant effects on weight, glucose or lipids, in younger adults with schizophrenia rather than older Parkinson's patients.
Part 7: Toxicity, Overdose, and Boxed Warnings
Increased mortality in elderly patients with dementia-related psychosis. Across 17 placebo-controlled trials of antipsychotics in dementia-related psychosis (modal duration 10 weeks, mostly atypicals), deaths ran about 4.5% vs 2.6% on placebo, a risk 1.6 to 1.7 times higher, mostly cardiovascular (heart failure, sudden death) or infectious (pneumonia). Pimavanserin is not approved for dementia-related psychosis unless the hallucinations and delusions are related to Parkinson's disease.
Mortality Data Specific to Pimavanserin
After postmarketing reports of deaths, the FDA reviewed the data in 2018 and found no new or unexpected safety findings. It attributed the reports to a population with high baseline mortality (advanced age, severe disease, comorbidities) and to a specialty distribution network that made reporting more likely (Pharmacy Times, 2018). An FDA-authored Medicare cohort study (Mosholder et al., Am J Psychiatry 2022) compared Parkinson's patients starting pimavanserin with those starting other atypical antipsychotics. Mortality was about 35% lower on pimavanserin in the first 180 days (HR 0.65), with no difference after 180 days (HR 1.05) and no advantage in nursing home residents. That is observational data, and the boxed warning stands.
Contraindication and Other Serious Risks
- Hypersensitivity to pimavanserin or any component is the one contraindication. Rash, urticaria and angioedema-type reactions (tongue swelling, circumoral edema, throat tightness, dyspnea) have been reported. The 2016 label listed no contraindication.
- QT prolongation: avoid in known QT prolongation, arrhythmia history, symptomatic bradycardia, hypokalemia, hypomagnesemia and congenital long QT, and with other QT-prolonging drugs (Part 8).
- Animal findings: phospholipidosis in multiple organs, worst in lungs and kidneys, in mice, rats and monkeys, with lung inflammation in rats at 10 or more times human exposure. The label calls its relevance to humans unclear.
- Abuse: not a controlled substance, and not systematically studied for abuse or dependence; trials showed no increase in drug-seeking.
Overdose
There is no specific antidote and trial data do not describe overdose symptoms. Start cardiovascular monitoring at once, including continuous ECG for arrhythmias. If an antiarrhythmic is needed, do not use disopyramide, procainamide or quinidine, which can add to the QT effect. Plan for the long half-life (about 57 hours) and consider other drugs taken. Poison Control (1-800-222-1222) can advise.
Part 8: Drug Interactions
Two lists matter: drugs that change pimavanserin's level through CYP3A4, and drugs that add to its QT effect. Pimavanserin itself did not change midazolam or levodopa levels.
| Interaction | Effect | Risk | Action |
|---|---|---|---|
| QT-prolonging drugs: class 1A antiarrhythmics (quinidine, procainamide), class 3 (amiodarone, sotalol), certain antipsychotics (ziprasidone, chlorpromazine, thioridazine), certain antibiotics (gatifloxacin, moxifloxacin) | Additive QT prolongation and arrhythmia risk | HIGH | Label: avoid the combination |
| Strong or moderate CYP3A4 inducers: rifampin, efavirenz (label); carbamazepine, phenytoin, St. John's wort, phenobarbital, primidone (FDA list) | Rifampin cut AUC 91% and Cmax 71%; efavirenz was modeled to cut AUC about 70% | HIGH | Avoid; the drug may stop working |
| Strong CYP3A4 inhibitors: ketoconazole (label); itraconazole, clarithromycin, ritonavir, nefazodone, posaconazole, voriconazole (FDA list) | Ketoconazole raised AUC 3-fold and Cmax 1.5-fold | HIGH | Pimavanserin 10 mg tablet once daily. Clarithromycin, ketoconazole, voriconazole and posaconazole also prolong QT on their own labels; the lower dose does not address that |
| Moderate CYP3A4 inhibitors (diltiazem, verapamil, erythromycin, fluconazole) | Not addressed in the label; expect a smaller rise in level than with ketoconazole | Not rated | No label dose change. Diltiazem and verapamil can cause bradycardia, and erythromycin and fluconazole prolong QT, both reasons the label gives to avoid pimavanserin |
| Carbidopa/levodopa | No clinically important interaction in PK studies | NONE | No dose adjustment |
| CYP3A4 substrates (midazolam) | No effect on midazolam pharmacokinetics | NONE | No adjustment |
Watch antibiotic prescriptions. Moxifloxacin and gatifloxacin are on the label's QT list, and clarithromycin is a strong CYP3A4 inhibitor on the FDA list. Tell patients and caregivers to mention pimavanserin to whoever prescribes an antibiotic. Check antidepressants too: citalopram's own label carries a QT warning, which makes it one of the QT-prolonging drugs the pimavanserin label says to avoid. Weigh another antidepressant before combining them.
Part 9: Special Populations
Pregnancy
There are no human data to assess risk of birth defects or miscarriage. Rats and rabbits showed no developmental harm during organogenesis at up to 10 and 12 times the maximum human dose. In rats dosed through pregnancy and lactation, maternal toxicity with lower pup survival and weight appeared at 2 times the maximum human dose.
Lactation
There is no information on presence in human milk, effects on a breastfed infant, or milk production. Weigh the benefits of breastfeeding against the mother's need for the drug and possible infant effects.
Elderly
- This is the trial population. Mean age was 71, with 49% aged 65 to 75 and 31% over 75. No dose adjustment is needed.
- In the pooled 6-week trials, 27% had MMSE scores of 21 to 24, and safety and effectiveness did not differ meaningfully from those scoring 25 or higher. Patients below 21 were not studied.
- Boxed warning applies to dementia-related psychosis outside Parkinson's disease.
- Falls and gait disturbance deserve attention in this age group (Part 6).
Renal Impairment
No dose adjustment from mild impairment through ESRD. Exposure rose in severe impairment (CrCl under 30 mL/min), so the label advises caution in severe impairment and ESRD. Dialysis does not meaningfully remove it.
Hepatic Impairment
No dose adjustment, based on a hepatic impairment study. This reverses the 2016 label, which said not to use it in hepatic impairment. The 2016 label also differed in two other places: it called use in severe renal impairment "not recommended" (now caution), and for strong CYP3A4 inducers it said to monitor and possibly raise the dose (now avoid strong and moderate inducers). Older drug references may still carry all three.
Pediatric / Adolescent
Safety and effectiveness have not been established. In a 6-week trial of 237 patients aged 5 to 17 with irritability associated with autism, pimavanserin did not beat placebo. Juvenile rats showed reversible slowing of growth and delayed sexual maturation in males.
Part 10: Discontinuation and Taper
No discontinuation syndrome is described, and no taper is given. With a 57-hour half-life and a longer-lived active metabolite, levels fall over days to weeks after the last dose, which works like a built-in taper (pharmacology, not label text).
- No clear response by about 6 weeks: consider discontinuation. Keep working on the Parkinson's regimen and nondrug measures.
- Responders: the only randomized withdrawal data in my sources come from HARMONY, in mixed dementias. Relapse after a switch to placebo was 28% vs 13%, though some of those relapses may have been withdrawal effects. A patient with dementia and psychosis who responds over 12 weeks should probably continue.
- Stopping for QT or side effects: stop it, and remember that QT effects and edema may take days or longer to clear.
Part 11: Pimavanserin vs the Alternatives
No head-to-head trial compares pimavanserin with clozapine or quetiapine in Parkinson's psychosis in my sources. The 2019 Movement Disorder Society evidence-based review rated pimavanserin and clozapine efficacious, clozapine with specialized monitoring; it rated quetiapine as insufficient evidence on efficacy but "possibly useful" in practice, and olanzapine as not efficacious.
| Comparison | Pimavanserin Advantage | Alternative Advantage |
|---|---|---|
| vs Low-dose clozapine | No blood count monitoring; no dopamine binding | Generic; outperformed placebo at 12.5 to 50 mg in two randomized trials without worsening motor symptoms, but carries clozapine's medical risks |
| vs Quetiapine | A positive placebo-controlled trial; quetiapine's evidence in Parkinson's psychosis is minimal or mixed | Generic; widely used for its low EPS liability |
| vs Other D2 blockers (risperidone, olanzapine, haloperidol) | Motor scores no different from placebo | Most antipsychotics worsen Parkinson's motor symptoms, so they are rarely a real alternative here |
| vs Adjusting the Parkinson's regimen and nondrug measures | Treats psychosis without giving up motor control | No drug risk; the usual first step |
Cost
Cash prices checked in October 2026 ran about $5,000 to $5,500 for 30 capsules (Drugs.com, GoodRx), up from above $3,000 a month in 2020 and about $4,000 in 2023. Most Medicare and commercial plans cover it, often after a prior authorization; the copay card works only for commercial insurance, so patients on government plans may face a prohibitive copay. Acadia won patent rulings in 2025 that, if they hold, keep generics off the market until about 2038.
Choose pimavanserin when motor function rules out a D2 blocker and blood monitoring rules out clozapine. Otherwise low-dose clozapine, a generic with the same efficacy rating in the 2019 MDS review, costs far less.
The Bedside Cheat Sheet
What it is
- 5-HT2A inverse agonist / antagonist, less at 5-HT2C; no appreciable D2 binding
- Half-life ~57 h; active metabolite ~200 h
- Only indication: hallucinations and delusions of Parkinson's disease psychosis
- FDA declined dementia-related and Alzheimer's psychosis; failed in MDD and schizophrenia trials
Starting and dosing
- Capsules 34 mg; tablets 10 mg
- 34 mg once daily. No titration. With or without food; capsule can be sprinkled on soft food
- Strong 3A4 inhibitor: 10 mg. Strong or moderate inducer: avoid
- No hepatic or renal dose change; caution in severe renal impairment and ESRD
- Judge it at about 6 weeks; stop if no clear benefit
Side effects that matter
- Peripheral edema 7% vs 2%
- Confusion 6% vs 3%; hallucinations 5% vs 3%
- Nausea, constipation, gait disturbance; falls postmarketing
- No motor worsening on UPDRS; no metabolic signal at 26 weeks
Don't forget
- QT: ~5 to 8 ms at 34 mg; avoid with QT-prolonging drugs and QT risk factors
- Boxed: dementia-psychosis mortality; Parkinson's disease dementia is within the label
- Check every new antibiotic and antiarrhythmic
- Brand-only, about $5,000 a month cash; prior authorization is common
Write down how often the hallucinations happen and how much they upset the patient, in the caregiver's own words, on the day you start, and compare against that note at 6 weeks.