Hallucinations in Parkinson’s Disease: Approach and Management

Hallucination is defined as “false sensory perception not associated with real external stimuli.” Hallucinations in patients with PD are reported as visual hallucinations in 30%,¹ auditory hallucinations in 10%,² and, rarely, tactile hallucinations.³ Visual hallucinations experienced by patients with PD are complex. They occur when the patients are awake and alert,^2,4 as opposed to hypnogogic hallucinations. Hallucinations can occur at any time of the day or night, but are more common at nighttime. Initially, patients may be aware of the nonreality of these experiences but gradually become paranoid and delusional.⁵ The most common are realistic images of animals, familiar or strange people, or children. These hallucinations are different from more common hallucinatory syndromes seen in patients with schizophrenia or in those who take illicit drugs. They are not accompanied by flashing lights, elaborate shifting patterns, distortion of time and space, or abstract repeating units.^6,7 Most hallucinations are pleasant and not threatening. Occasionally, visual images are incomplete and resemble hallucinations in the visually impaired patient (Charles Bonnet’s syndrome).⁸ Fenelon et al⁹ classified the hallucinations in patients with PD as minor hallucinations and formed visual hallucination. Minor hallucinations are subclassified into three types: presence hallucinations (64%), which are described as a perception such as “someone is behind me”; passage hallucinations (33%) such as brief vision of an object passing sideways; and illusions such as transformation of an object into an animal.

Risk Factors

Risk factors that have been reported include dementia, duration of therapy with dopaminergic medications, duration of disease, age of patients,^10,11 anticholinergic medications, and sleep disorders.^12-15 The mnemonic “SADDAD” can be used to recall these risk factors for hallucinations in PD:

• Sleep disorders
• Age of patients
• Dementia
• Duration of therapy
• Anticholinergic medications
• Duration of disease

Severe cognitive impairment or dementia is a major and independent predictive factor for visual hallucinations.^4,9,16 In one study, the prevalence of visual hallucinations was reported as 70% in patients with PD who have dementia.⁹ It is not clear whether cognitive impairment and hallucinations occur independently as the natural progression of disease or if they have causative links. Studies using positron emission tomography have shown marked occipital hypometabolism in patients with PD who have dementia.¹⁷ This suggests that degenerative process in the visual cortex can be a cause for visual hallucinations.

Several medications have been implicated in causing hallucinations. Common drugs that are implicated include anticholinergics such as benztropine, dopaminergics such as ropinirole,¹⁸ and amantadine. Because both anticholinergics (antimuscarinic) and dopaminergics may induce hallucinations, disturbances of dopaminergic-cholinergic balance have been hypothesized as the pathophysiology.⁵ No dose-response relation between dopaminergic drugs and hallucinations has been reported in recent prospective studies. Goetz et al¹⁹ showed that there is no simple relationship between visual hallucinations and high plasma levels of levodopa or changes in plasma level.

Increased age has been associated with the presence of hallucinations, although this factor can be confounded by duration of illness, which is described as a major risk factor by Fenelon et al.⁹ It was proposed hypothetically that most hallucinations occur in the “off” state, but Sanchez-Ramos et al⁴ showed that the majority of their patients (60%) were “on” (able to move) while experiencing their hallucinations. Graham et al²⁰ identified two subgroups susceptible to hallucinations in idiopathic PD: patients with a disease duration of 5 years or less who are at risk of developing hallucination without cognitive impairment; and patients with disease duration greater than 5 years who are at risk of developing hallucinations with cognitive impairment, postural instability, and a lack of depressive affect. Authors of this controversial report hypothesized that the mechanism involved in these two groups was different, so the response to the treatment of hallucinations may be different. Other studies reported that hallucinations occur predominantly during the later stages of idiopathic PD.^21,22

Moskovitz et al¹³ suggested that visual hallucinations start as vivid dreams and advance into hallucinations, progress to delusions, and eventually to confusion. This progression of symptoms, which has been termed “levodopa psychosis,” is due to a “kindling” mechanism secondary to chronic dopaminergic pharmacotherapy. Sleep disruption is an early feature of levodopa psychosis.¹⁶ Although a high prevalence of sleep disorders is reported in patients with hallucinations, we cannot conclude that both symptoms have the same mechanism.²³ There has been a report of visual hallucinations induced by deep brain stimulation (DBS), a surgical therapy for medically refractory advanced PD. The case patient had symptoms only when DBS was turned on and he was not taking medication; symptoms resolved when the stimulator was turned off, and they responded promptly to clozapine.²⁴

Auditory and Tactile Hallucinations

Inzelberg et al² reported auditory hallucinations in 8% of patients with PD in their series. All of these patients had accompanying visual hallucinations. Patients reported hearing human voices talking outside their heads. The content was often incomprehensible and nonimperative. Fenelon et al⁹ reported that auditory hallucinations occur with visual hallucinations simultaneously or at different times. They have been verbal, musical, or various other noises (such as the sound of steps). Tactile hallucinations are rarely reported in patients with PD. When they occur, they are usually accompanied by visual hallucinations.³

Treatment

Treatment of hallucinations in idiopathic PD remains a challenge because dopaminergic medications remain the key treatment option for motor disability for these patients. A stepwise algorithm is useful for optimal management of hallucinations in patients with PD (Table I). The initial step should focus on assessment of systemic disorders that can aggravate or cause hallucinations based on history and physical examination such as urinary tract infection or electrolyte abnormalities. Common medications with hallucinogenic side effects such as H2 blockers should be eliminated if possible. If motor symptoms are under good control and hallucinations are nondisabling, observation and patient/caregiver education may be appropriate.

Reduction of medications at bedtime may alleviate nighttime hallucination.²⁵ The next step will be reducing or discontinuing low-potency antiparkinsonian medications that have the potential of inducing hallucinations. Examples of such medications are amantadine, selegiline, and anticholinergics. Selegiline can be discontinued abruptly because of its long duration of action in the brain, but amantadine and anticholinergics should be tapered. Levodopa and dopamine agonists can be optimized to minimum tolerated doses so that hallucinogenic potential is limited. If these strategies fail, addition of antipsychotic medications can be considered. This algorithmic approach is appropriate for “classic” hallucinations seen in patients with PD.

If patients have nonformed hallucinations such as flickering light or scotoma, the possibility of other causes such as seizure, migraine, or anatomical lesions should be considered. These types of hallucinations need appropriate imaging or neurophysiologic studies. Typical antipsychotic medications can relieve hallucinations, but these medications antagonize dopamine in nigrostriatal and mesocorticolimbic systems. The use of typical antipsychotics is not recommended because of extrapyramidal side effects. The newer atypical antipsychotics have been used successfully with fewer extrapyramidal side effects (Table II). Clozapine is virtually free of extrapyramidal side effects but has a potential side effect of agranulocytosis, which requires frequent blood count. A study by the Parkinson Study Group suggested clozapine at daily doses of 50 mg or less is safe and significantly improves drug-induced psychosis without worsening parkinsonism.²⁶ Clozapine has also been reported to ameliorate motor impairment in PD, specifically tremor and dystonia.²⁷ For the newer atypical antipsychotic olanzapine, open-label studies in patients with PD who have hallucinations reported excellent antipsychotic efficacy. It has an atypical feature, binding to dopamine D4 and serotonin 5HT2 receptors in preference to D2 receptors and causes minimal extrapyramidal side effects.²⁸ Graham et al²⁹ reported suppression of hallucinations in PD with olanzapine, but they believed that starting doses as high as 5 mg may result in an unacceptable exacerbation of motor disability.

Another study that compared clozapine and olanzapine reported similar decline in motor function in patients with PD who were treated with olanzapine, even at the lower starting dose of 2.5 mg.³⁰ Authors reported improvement in parkinsonism symptoms with clozapine as reported by others.²⁷ It is hypothesized that low D2 receptor affinity of olanzapine is still sufficient to block nigrostriatal receptors.³⁰ In another study comparing olanzapine with clozapine, olanzapine seemed to aggravate motor symptoms of PD. Clozapine was superior to olanzapine in controlling hallucinations.³¹

Quetiapine is another atypical antipsychotic that has been used in the management of drug-induced psychosis in PD.³² Quetiapine interacts with D1 and D2 and occupies fewer serotonin receptors relative to clozapine and olanzapine. However, quetiapine possesses a higher 5HT2A and 5HT2C affinity relative to D2 occupancy, which may account for its low extrapyramidal side effects.³³ In a study of 87 patients with PD, 80% had partial to complete resolution of psychosis using quetiapine, with worsening of motor symptoms found in 32%.³⁴ Based on current literature, quetiapine appears to be a good first option as additive therapy to alleviate hallucinations in PD; however, there is no published result of any double-blind, randomized, controlled trial. If the hallucinations are already well controlled with olanzapine and clozapine, caution should be used to replace them with other antipsychotic medications. Because of transient liver-enzyme increases and orthostatic hypotension in the early days of treatment, a slow-dose titration schedule is recommended. It may cause weight gain and hyperglycemia, but to a lesser degree compared to clozapine and olanzapine.³⁵

Risperidone has both D2 and 5HT2A receptor-blocking properties, and it is reported to have a higher risk of extrapyramidal side effects compared to clozapine. In a study of 39 patients with PD, a complete or near complete resolution of hallucinations and delusions have been reported in 23 of those patients, and worsening of symptoms in six of them.³⁶ Recent reports of a higher incidence of cerebrovascular adverse events in elderly patients with dementia make this drug less favorable. Donepezil was assessed for efficacy and tolerability in the treatment of hallucinations in PD in an open-label study because cholinergic mechanisms may be partially responsible for hallucinations. The investigators reported significant improvement in all eight patients, but deterioration in motor disability occurred in two of them.³⁷

Rivastigmine is reported to improve hallucinations, sleep disturbance, and caregiver distress in addition to enhancing cognitive performance in advanced PD, according to a small study in the United Kingdom.38 It is hypothesized that cholinergic deficits may play a role in the development of psychosis in advanced PD. Researchers rated neuropsychiatric symptoms using the Neuropsychiatric Inventory (NPI), which is a validated informant-rated measure. The open-label nature of their study raises the possibility of placebo effects on the results of study.

Summary

Treatment of hallucinations in PD is relatively easy. The suggested algorithm can be effectively implemented in outpatient and inpatient settings by geriatricians and primary care providers in cooperation with neurologists. In our experience, we prefer to start with quetiapine as the first option because it has the fewest extrapyramidal and anticholinergic side effects. Initial dosage can be as low as 25 mg at bedtime with increments of 25-50 mg 2 times a day on the third day, as tolerated, to a target dose 200-300 mg divided 2 times a day in the elderly population. If the hallucinations fail to respond to quetiapine, consider clozapine as the next step. Currently, there are not enough data available about the efficacy of the newer medications in this group, such as ziprasidone or aripiprazole.

Acknowledgment
The authors acknowledge the help of Dr. Robert Palmer, Chief of Geriatric Section, Cleveland Clinic Foundation, in the preparation of this paper.