The pharmacological management of the neuropsychiatric symptoms that affect the majority of patients with advanced Parkinson's disease.
Cognitive impairment is present in up to 80% of patients with Parkinson's disease over the course of the illness and represents one of its most disabling non-motor features. The cholinergic deficit in PD dementia is even more severe than in Alzheimer's disease, making cholinesterase inhibitors a rationally grounded and evidence-supported treatment strategy in this population.
Parkinson's disease dementia (PDD) develops in the majority of patients who survive long enough, with a prevalence approaching 80% by 20 years of disease. The pathological substrate is cortical Lewy body deposition combined with cholinergic nucleus basalis degeneration that exceeds the cholinergic deficit seen in Alzheimer's disease. The cognitive profile is characterized by prominent executive dysfunction, visuospatial impairment, and attention fluctuations, with memory impairment typically less severe than in Alzheimer's disease at comparable stages. Mild cognitive impairment (MCI) in PD affects approximately 25% of newly diagnosed patients and confers a threefold increased risk of progression to dementia.1
Rivastigmine is the only cholinesterase inhibitor with an FDA indication specifically for Parkinson's disease dementia. The pivotal EXPRESS trial (Emre et al., 2004) demonstrated that rivastigmine capsules 3–12 mg/day produced significant improvement in the primary cognitive endpoint (ADAS-cog) and the clinician's global impression of change compared to placebo over 24 weeks. The transdermal patch formulation (9.5 mg/24 hr) is generally preferred over the capsule because it produces more stable plasma concentrations and substantially fewer gastrointestinal adverse effects; nausea and vomiting are the primary dose-limiting toxicities of oral rivastigmine. Donepezil is frequently used off-label and shows benefit in uncontrolled and some controlled data, but lacks a specific PDD approval. Galantamine carries the least evidence base in PDD specifically.12
Psychosis in PD presents most commonly as formed visual hallucinations, which are typically non-threatening (people, animals, insects) and occur in a clear sensorium in early stages. As the disease progresses, paranoid delusions, passage hallucinations, and illusions become more frequent. PD psychosis (PDP) is driven by a combination of dopaminergic excess from medications, cholinergic deficits, and intrinsic Lewy body pathology. The first management step when psychosis develops is a systematic review of the medication burden: anticholinergics, amantadine, and dopamine agonists should be tapered or discontinued before antipsychotic therapy is initiated, followed by MAO-B inhibitors and COMT inhibitors if psychosis persists, with levodopa dose reduction as a last resort among dopaminergic adjustments.3
Pimavanserin (Nuplazid) is the first and only FDA-approved medication specifically for hallucinations and delusions associated with PD psychosis. It is a selective inverse agonist at serotonin 5-HT2A and 5-HT2C receptors with no dopamine receptor binding, meaning it does not worsen motor function. The pivotal ACP-103-020 trial demonstrated a 37% reduction in the Scale for Assessment of Positive Symptoms adapted for PD (SAPS-PD) score compared to placebo after six weeks of treatment with 34 mg once daily. Pimavanserin carries a black box warning for increased mortality in elderly patients with dementia-related psychosis, identical to the class warning carried by all atypical antipsychotics. QTc prolongation is an additional concern and a baseline electrocardiogram is appropriate before initiation. Clozapine at low doses (12.5–50 mg/day) remains the best-evidenced conventional antipsychotic for PDP and does not worsen parkinsonism at these doses, but mandatory agranulocytosis monitoring limits its practical use.34
Step 1: Taper/discontinue anticholinergics and amantadine. Step 2: Reduce or stop dopamine agonists. Step 3: Reduce MAO-B and COMT inhibitors. Step 4: Reduce levodopa dose if steps 1–3 insufficient. Step 5: Add pimavanserin 34 mg once daily (FDA-approved for PDP) or low-dose clozapine (12.5–50 mg). Never use first-generation antipsychotics or risperidone/olanzapine in PD — they cause severe motor worsening.
All first-generation antipsychotics (haloperidol, chlorpromazine) and most second-generation agents (risperidone, olanzapine) are contraindicated in PD psychosis due to D2 blockade causing severe, potentially irreversible motor worsening. Quetiapine is sometimes used at very low doses (12.5–50 mg) despite limited efficacy data; it has less motor risk than risperidone but more than pimavanserin or clozapine.
Depression as the most prevalent non-motor symptom, anxiety as an underrecognized companion, and the autonomic failures that erode quality of life in advanced disease.
Depression affects approximately 35% of patients with PD at any given time and is often the non-motor symptom most predictive of quality of life impairment. Its pathophysiology in PD extends beyond reactive sadness and encompasses intrinsic serotonergic and noradrenergic degeneration as part of the ascending Braak staging of Lewy body pathology, making antidepressant pharmacotherapy physiologically rational even in the absence of reactive life stressors.
The evidence base for antidepressants in PD is smaller than the clinical need warrants, but several agents have controlled trial support. Selective serotonin reuptake inhibitors (SSRIs) and serotonin-norepinephrine reuptake inhibitors (SNRIs) are the most widely used first-line agents. The SIC trial demonstrated paroxetine and venlafaxine XR both significantly reduced depression scores compared to placebo, with venlafaxine showing a numerically larger effect size. Tricyclic antidepressants (TCAs) such as nortriptyline have demonstrated efficacy in controlled trials and may additionally benefit pain and sleep quality, but their anticholinergic adverse effect profile makes them poorly tolerated in older patients and those with cognitive impairment. The interaction between MAO-B inhibitors and serotonergic antidepressants requires clinical attention: while the risk of serotonin syndrome at standard doses of selective MAO-B inhibitors is substantially lower than with non-selective monoamine oxidase inhibitors (MAOIs), the combination of rasagiline or selegiline with SSRIs or SNRIs still warrants monitoring.5
Anxiety affects 30–40% of PD patients and frequently co-occurs with depression and motor fluctuations. Non-motor wearing-off, in which anxiety, sweating, inner restlessness, and dysphoria occur as levodopa concentrations fall, is a common and under-recognized form of anxiety in PD that may respond to optimizing the levodopa regimen rather than adding anxiolytic medications. When pharmacological anxiolytic treatment is needed, SSRIs and SNRIs are appropriate first-line choices. Benzodiazepines should be used with caution given the risks of falls, cognitive impairment, and excessive sedation in this population, but low-dose clonazepam may be appropriate for specific indications such as REM sleep behavior disorder (discussed in Section 3). Buspirone is an alternative with less fall risk, though evidence in PD anxiety specifically is limited.5
Orthostatic hypotension (OH) is one of the most clinically consequential autonomic manifestations of PD, occurring in 30–50% of patients and contributing significantly to fall risk, syncope, and reduced quality of life. The pathophysiology involves both disease-related degeneration of postganglionic sympathetic neurons and the vasodilatory and hypotensive effects of dopaminergic medications. Non-pharmacological measures should be instituted first: increased salt and fluid intake, compression stockings, elevation of the head of the bed, avoiding prolonged standing, and spacing large meals. When pharmacological treatment is required, midodrine (a peripheral alpha-1 agonist, 2.5–10 mg two to three times daily) and droxidopa (a synthetic norepinephrine precursor, 100–600 mg three times daily) are the two agents with FDA approval for neurogenic orthostatic hypotension. Fludrocortisone (0.1–0.2 mg/day) is widely used off-label as a mineralocorticoid to expand plasma volume but requires monitoring for supine hypertension, heart failure exacerbation, and hypokalemia.6
Sialorrhea (drooling) results from impaired swallowing automaticity rather than true hypersalivation and affects up to 75% of patients with advanced PD. Glycopyrrolate (a quaternary anticholinergic that does not cross the blood-brain barrier) 1–2 mg two to three times daily reduces salivary output without central anticholinergic effects, making it preferable to tertiary anticholinergics. Botulinum toxin injection into the parotid and submandibular glands provides effective sialorrhea control for three to six months and is particularly useful in patients with significant cognitive impairment where oral anticholinergics carry cognitive risk. Constipation, affecting up to 80% of PD patients, reflects both intrinsic enteric nervous system Lewy body pathology and dopaminergic medication effects; macrogol (polyethylene glycol) laxatives, increased dietary fiber, and adequate hydration are first-line measures, with prucalopride (a 5-HT4 agonist) an option for refractory cases. Urinary urgency and frequency are best treated with beta-3 adrenoreceptor agonists (mirabegron) rather than anticholinergic bladder agents, which worsen cognitive impairment.6
Non-pharmacological first: increased salt/fluid, compression stockings, head-of-bed elevation, small frequent meals. Pharmacological: midodrine 2.5–10 mg two to three times daily (avoid supine dosing within 4 hours of bedtime) or droxidopa 100–600 mg three times daily. Fludrocortisone 0.1–0.2 mg/day off-label; monitor supine hypertension, heart failure, hypokalemia.
REM sleep behavior disorder, excessive daytime sleepiness, restless legs syndrome, and insomnia: how sleep pathology in PD requires targeted rather than generic pharmacological approaches.
Sleep disorders affect more than 90% of patients with PD at some stage and encompass a mechanistically heterogeneous group of conditions, each requiring specific pharmacological consideration. Treating sleep in PD cannot rely on generic sedative hypnotics, which often worsen cognitive function and fall risk without addressing the underlying pathophysiology.
REM sleep behavior disorder (RBD) is characterized by loss of normal skeletal muscle atonia during REM sleep, resulting in dream enactment behavior ranging from vocalisation and limb movements to violent behavior with injury risk to the patient and bed partner. RBD is not merely a PD complication; it is now recognized as a prodromal feature of synucleinopathies, often preceding the motor symptoms of PD by a decade or more and carrying high specificity for subsequent Lewy body disease. Pharmacologically, clonazepam 0.25–1.0 mg at bedtime is the most widely used treatment, reducing the frequency and severity of enactment behavior in the majority of patients, though evidence is largely observational. Melatonin 3–12 mg at bedtime is an alternative with fewer cognitive and sedation risks and is preferred in patients with concurrent cognitive impairment or those at high fall risk.7
Excessive daytime sleepiness (EDS) and sudden onset sleep episodes represent a significant patient safety issue in PD, affecting up to 50% of patients and creating driving hazard. EDS in PD is multifactorial: it reflects both intrinsic hypocretin/orexin degeneration in the lateral hypothalamus (a feature of PD neuropathology), nocturnal sleep fragmentation, and the sedative effects of dopaminergic medications, particularly dopamine agonists. Where dopamine agonist sedation is the primary contributor, switching from a more sedating agonist (pramipexole) to a less sedating formulation or agent may help. Modafinil 100–400 mg in the morning is the most widely used pharmacological treatment for EDS in PD, with randomised trial evidence supporting improved subjective sleepiness, though objective sleep architecture measures have been less consistently improved. Sodium oxybate has been evaluated in PD and improves nocturnal sleep consolidation as well as daytime function, but its Schedule III classification and complex distribution system limit widespread use.8
Restless legs syndrome (RLS) and periodic limb movements of sleep (PLMS) occur with increased frequency in PD and share pathophysiological links through dopaminergic striatal circuits. When RLS is mild, low-dose dopaminergic therapy already prescribed for PD motor symptoms typically provides adequate symptomatic relief. When additional treatment is needed, low-dose dopamine agonists at RLS-appropriate doses (pramipexole 0.125–0.5 mg in the evening) are effective, though the risk of augmentation (worsening of RLS symptoms with dose escalation over time) must be considered. Alpha-2-delta calcium channel ligands such as gabapentin enacarbil or pregabalin are increasingly preferred for RLS given their freedom from augmentation risk, though they add sedation and fall risk to a population already burdened by both.7
Insomnia in PD is common and frequently multifactorial: nocturnal akinesia (inability to turn in bed due to loss of dopaminergic tone), nocturia, pain, depression, and early morning off-period dystonia each contribute and require individually targeted approaches. Controlled-release carbidopa/levodopa at bedtime addresses nocturnal akinesia directly and may reduce the frequency of nocturnal awakening due to motor symptoms. Melatonin 3–5 mg at bedtime is a safe, low-risk approach to sleep initiation difficulty. When pharmacological sedation is genuinely needed, low-dose quetiapine 12.5–25 mg at bedtime is frequently used in clinical practice, providing sedation without significant dopamine receptor blockade at these doses, though formal insomnia indication evidence in PD is limited. Z-drugs and benzodiazepines increase fall and cognitive impairment risk and should be used with caution and only for clearly defined short-term indications.8
RBD often predates motor PD by years and is a prodromal synucleinopathy marker. Immediate priority: environmental safety (bed rails, floor padding, removing sharp objects). Pharmacological: clonazepam 0.25–1.0 mg at bedtime (most effective; cognitive caution) or melatonin 3–12 mg at bedtime (preferred if cognitive impairment or fall risk). Advise patient not to drive until symptoms are controlled.
Pain as the most underrecognized symptom of Parkinson's disease, its classification into mechanistic subtypes, and the evidence basis for pharmacological management.
Pain affects 60–85% of patients with Parkinson's disease, yet it is underrecognized and undertreated in clinical practice. Pain in PD is not a single entity: it encompasses musculoskeletal pain related to rigidity and dystonia, neuropathic pain, central pain, and radicular pain, each requiring a different therapeutic approach. The first clinical step is always determining whether the pain is fluctuation-related and therefore dopaminergically responsive.
The King's PD Pain Scale and the PD Pain Classification System describe at least six mechanistically distinct pain subtypes in PD. Musculoskeletal pain from rigidity and akinesia is the most common subtype and may respond directly to optimization of dopaminergic therapy. Dystonic pain, occurring most commonly as early morning off-period foot dystonia before the first levodopa dose, is frequently responsive to bedtime controlled-release levodopa or botulinum toxin injection into the affected muscle groups. Radicular or peripheral neuropathic pain may or may not be PD-related and should be evaluated on its own merits; gabapentinoids or TCAs are appropriate pharmacological approaches. Central pain in PD is diffuse, difficult to characterize, often described as burning or aching, and may be partially responsive to dopaminergic therapy during on periods. Opioids for PD-related central pain carry the standard risks of constipation and sedation, which are amplified in this population; tramadol should be avoided in combination with MAO-B inhibitors.9
Fatigue is reported by up to 50% of PD patients as one of their most disabling symptoms, yet it is distinct from excessive daytime sleepiness and from depression, though all three frequently coexist. PD fatigue is poorly responsive to standard antidepressant or stimulant therapy. Optimization of motor function with dopaminergic therapy may reduce fatigue if it is related to the physical effort of moving against rigidity, but central fatigue does not reliably respond. Methylphenidate at low doses has some controlled trial evidence for PD fatigue, and modafinil, though primarily used for EDS, may have some benefit for fatigue in subgroups. Rivastigmine may improve fatigue in PDD patients as part of its broader effect on cholinergic function and arousal systems.9
Anosmia (loss of smell) is among the earliest non-motor features of PD, often predating motor symptoms by years, and has no effective pharmacological treatment. Olfactory impairment is now incorporated into prodromal PD research criteria but does not require specific treatment in established PD. Hyperhidrosis, most commonly occurring during off periods as part of the non-motor wearing-off complex, typically responds to levodopa dose optimization rather than to anticholinergic sweat suppressants; the latter carry cognitive risk in PD patients and should be avoided. Seborrheic dermatitis, related to dopaminergic dysregulation of sebaceous gland secretion, is treated with standard topical antifungal and anti-inflammatory approaches and does not require specific antiparkinson medication adjustment.10
Before prescribing analgesia for pain in PD, determine: is the pain fluctuation-related? Does it occur during off periods and resolve in on periods? If yes, the primary intervention is levodopa optimization (COMT inhibitor, MAO-B inhibitor, or dose-interval adjustment) rather than analgesic escalation. Dystonic off-period pain may also respond to botulinum toxin. Only pain that persists across motor states warrants non-dopaminergic analgesic therapy.
A framework for managing the complexity of polypharmacy, drug interactions, and competing therapeutic goals in the full PD patient.
Non-motor symptoms of PD are not simply additive complications to be treated in isolation. They interact with one another, with motor symptoms, and with the medications used to treat both. The clinician managing a patient with advanced PD must maintain an integrated picture of the entire pharmacological burden, the competing risks of each agent, and the hierarchy of clinical priorities.
The central tension in managing non-motor symptoms alongside motor symptoms is the frequent bidirectionality of pharmacological effects. Dopaminergic therapy that controls motor symptoms can worsen psychosis, impulse control, daytime sleepiness, and orthostatic hypotension. Antipsychotic therapy for psychosis can worsen motor function. Anticholinergic therapy for bladder symptoms can worsen cognitive function. Cholinesterase inhibitors for dementia can worsen tremor through increased cholinergic tone and may interact with the beta-agonist mechanism of mirabegron. Every added agent must be evaluated for its impact on the existing symptom and medication constellation, not just its target symptom in isolation.10
A practical hierarchy of management priorities in advanced PD places cognitive safety at the top. Cognitive impairment and dementia, once established, dramatically narrow the range of safe pharmacological options: anticholinergics become contraindicated, many sedatives become hazardous, and the margin between therapeutic and toxic dosing for dopaminergic agents narrows. The management of any new symptom in a cognitively impaired PD patient must begin with a review of the existing medication burden for drugs that could be contributing to cognitive worsening before new agents are added. This includes over-the-counter medications, bladder agents, antihistamines, and any drug with anticholinergic burden rating above zero on standard scales such as the Anticholinergic Cognitive Burden scale.10
The management of non-motor wearing-off deserves specific emphasis because it is frequently misdiagnosed. When a PD patient reports anxiety, sweating, inner restlessness, palpitations, or dysphoria that occur predictably before a scheduled levodopa dose, these symptoms represent non-motor wearing-off driven by sub-therapeutic levodopa levels rather than a new psychiatric or autonomic disorder. The appropriate response is dopaminergic optimization, not the addition of anxiolytics, antidepressants, or cardiovascular medications. Asking about the relationship of non-motor symptoms to dose timing at every clinic visit is a key clinical skill that is under-practiced and has major therapeutic implications. Validated instruments such as the Non-Motor Symptoms Scale and the Wearing-Off Questionnaire-19 can systematize this assessment.5
Multidisciplinary management is indispensable for non-motor PD but does not diminish the need for rigorous pharmacological oversight. Physiotherapy, occupational therapy, speech therapy, and neuropsychology contribute essential non-pharmacological interventions for cognition, mobility, swallowing, and communication. However, the pharmacological management of PD remains highly specialized and the integration of new medications by non-specialist providers without awareness of the full PD-specific interaction landscape is a common source of preventable harm. Particular vigilance is warranted when PD patients are admitted to hospital, where dopaminergic medications may be withheld or delayed (risking acute akinesia and dysphagia), and where antipsychotics, antiemetics, and anticholinergics are commonly prescribed without awareness of their contraindication in PD.10
Never withhold dopaminergic medications in hospitalised PD patients — acute akinesia can progress to aspiration and death within hours. Avoid metoclopramide (D2 blocker) as antiemetic; use domperidone or ondansetron instead. Avoid first-generation antipsychotics and high-potency second-generation agents for delirium. Ensure the admitting team is aware of the patient's complete antiparkinson regimen and its time-critical dosing schedule.
Ask at every visit: do you experience anxiety, sweating, dysphoria, inner restlessness, or palpitations that occur predictably before your next levodopa dose and resolve after taking it? If yes, this is non-motor wearing-off and the primary intervention is levodopa optimization, not new psychiatric or cardiovascular medications. Use the WOQ-19 or NMS Scale to systematize assessment.