Psychosis in Parkinson’s disease: how should we treat it?

Psychosis (delusions and/or hallucinations) is a common and problematic non-motor feature of Parkinson’s disease (PD). Hallucinations are sensory perceptions that occur without external stimulation of the relevant sensory organ. In PD, a typical visual hallucination is a complex visual image experienced when the patient is alert and has their eyes open. The image usually appears somewhat blurred. There is a predominance of human or animal forms, some with emotional significance. The hallucinations are sometimes ‘friendly’ but may also consist of frightening images. Hallucinations are generally more common in the evening [1]. Most patients ‘know’ that they are hallucinating (benign hallucinosis), but approximately 5% experience delusions and hallucinations with lack of insight [2]. Auditory and tactile hallucinations may also occur, but are less frequent and usually co-exist with visual hallucinations [1]. A sense of presence is another common hallucinatory symptom.

Delusions are false beliefs based on incorrect inferences. They are commonly anteceded by hallucinations, but can occur in their absence. In PD, common themes of the delusions are spousal infidelity or persecution.

Estimates of prevalence vary. In one US study of 102 consecutive clinic attenders, a prevalence of visual hallucinations of 26% was found [3]. In a systematic review of clinic populations, the prevalence of hallucinations alone was between 21 and 46% [4]. A community survey in Norway yielded 9% prevalence for hallucinations and 6% for combined hallucinations and delusions [2]. In patients attending a specialist clinic, the figure for the combined symptoms was between 10 and 15% [5,6]. The combination of hallucinations and delusions together is regarded as a more serious problem as it usually results in subjective distress and behavioral disturbance.

Psychosis is an indicator of poor prognosis in PD [7] and is often the factor that leads to institutional care [8]. In addition, carer burden and mortality are high [7–10]. There is no consistent relationship with dose, drug type or on–off phenomena [11]. There is some evidence that the incidence of psychosis is higher in patients treated with dopamine agonists when compared with those receiving levodopa monotherapy [12–15].

Treating psychosis is one of the most difficult challenges for the parkinsonologist. Even in early monotherapy in new patients, drug- induced psychosis can be a major dose-limiting side effect. There is an increase in frequency with disease duration. Other risk factors are age, cognitive impairment and sleep disturbance [4,16]. There are some general principles for managing PD psychosis. The most important first step is to exclude delirium in all cases of first-presentation psychosis. Drug effects, concurrent physical illness, electrolyte imbalance and systemic infection should all be excluded. For fully effective management, a multidisciplinary approach is desirable. Ideally, this should be comprised of the patient and carer, PD physician, PD nurse, psychiatrist and social worker. Education and ‘insight-building’ work with the patient and carer will allow them to better understand and cope with the condition. Several simple measures may improve the situation; for example, improved lighting and avoidance of patterned furniture/carpets. Some patients find that staring directly at the hallucinations causes them to disappear. Occasional visual hallucinations with preserved insight often do not require anything other than reassurance.

If these simple approaches are unsuccessful, it is necessary to consider reduction of dose or elimination of one or more antiparkinsonian drugs. This is usually carried out in the following order: anticholinergics, amantadine, selegiline, dopamine agonists, catechol-O-methyltransferase (COMT) inhibitors and l-dopa. There is often a tricky balancing act between psychosis and motor function (the motion/emotion dilemma) [6]. It may be necessary to accept some deterioration in motor function.

Drug treatment is necessary if all the above measures fail. Attempts to control psychosis using the older ‘typical’ antipsychotic or neuroleptic drugs led to worsening of motor symptoms. The introduction of second-generation or ‘atypical’ antipsychotic drugs and cholinesterase inhibitors opened up a new range of therapeutic options. The atypical antipsychotics are loosely defined as a group of neuroleptic drugs that fail to cause catalepsy in rats (rigid maintenance of a body position over an extended period of time). Their characteristics in humans include a low incidence of extrapyramidal side effects and a lack of induction of significant prolactin elevation. They have a higher serotonin (5HT)-2 versus D2 receptor affinity than classic neuroleptic drugs. Of the atypical antipsychotics, risperidone, olanzapine, clozapine and quetiapine have all been used to treat psychosis in PD with varying success [17,18]. Risperidone and olanzapine are associated with worsening of motor function [19,20]. Quetiapine is the treatment favored by many psychiatrists and neurologists in clinical practice. However, the evidence for its efficacy is mixed. In two recent randomized trials where it was compared with clozapine, it was found to be at least as effective [21,22]. In addition, in two recent open-label trials, it showed benefit in the control of psychosis [23,24]. However, two recent placebo-controlled trials had negative results [25,26]. Further randomized, controlled trials are required to settle the debate. Clozapine is the most effective agent and does not worsen parkinsonism. However, the need for regular white blood cell monitoring owing to the risk of fatal agranulocytosis tends to restrict its use to patients who have not responded to other agents. In addition, patients usually need to be started on it in an inpatient setting, which is obviously costly. Aripiprazole and ziprasidone have also been used. However, the evidence for these is weak, as there have been no randomized, controlled trials. One case series using aripirazole was not particularly encouraging, causing a significant worsening of PD symptoms [27].

Recently, there has been concern regarding the safety of the atypical antipsychotics in certain patient populations. A recent meta-analysis examining 15 trials of atypicals in patients with dementia showed that death occurred more often among patients randomized to drug [28]. The odds ratio was 1.54 (95% confidence interval: 1.06–2.23; p = 0.01). The US FDA report a similar increased risk between 1.6 and 1.7. Deaths were most frequently due to cardiac-related events or infections such as pneumonia. Consequently, a black box warning has been added to the label of all atypical antipsychotics in the USA 101.

Cholinesterase inhibitors are another recently introduced treatment option for PD psychosis, particularly where there is associated dementia. The brains of PD patients show marked degeneration of cholinergic neurons [29]. This has also been shown in vivo using functional neuroimaging [30]. Relative dopaminergic over-activity in the mesocortical system combined with subcortical cholinergic neuronal loss could provide a more permissive environment for the development of psychosis. A review published in 2003 tentatively stated that cholinesterase inhibitors were a potentially useful treatment in PD psychosis [31]. This conclusion drew on the evidence from one double-blind, placebo controlled trial of rivastigmine in dementia with Lewy bodies patients [32] and from four small open-label studies in PD dementia [33–36].

Since 2003, four double-blind, placebo-controlled randomized, controlled trials have been reported examining the use of cholinesterase inhibitors in Parkinson’s disease with dementia (PDD) where neuropsychiatric symptoms were reported as secondary end points. A large randomized, controlled trial using rivastigmine [37] reported a statistically significant reduction in Neuropsychiatric Inventory (NPI) [38] scores compared with placebo. Nausea and tremor were common adverse events. It is important to note, however, that the initial NPI scores were rather low (the mean score in the rivastigmine group was 12.7 out of a possible maximum of 120). The mean change in this score at week 24 was of only 2 points and so it is difficult to be certain of the clinical significance of this. In addition, NPI scores represent a composite of various neuropsychiatric symptoms. It is therefore difficult to assign clinical meaning to a single change in this score. Three randomized, controlled trials with donepezil showed no significant improvement in NPI scores versus placebo [39–41]. An open-label study using donepezil in PDD reports significant improvements in hallucinations and delusions [42]. Aarsland and colleagues report a further open-label trial using galantamine in patients with PD dementia [43]. This trial again showed improvements in hallucinations.

With regard to tolerability, there is evidence from the above trials that cholinesterase inhibitors worsen motor function, especially tremor, although this is not always picked up by United Parkinson’s Disease Rating Scale scores [44]. In addition, nausea appears to be a significant problem, affecting 29% of patients receiving rivastigmine in the largest trial.

Further randomized, controlled trials are required to specifically assess the efficacy of cholinesterase inhibitors in PD psychosis.

In a recent paper discussing the etiology of visual hallucinations, Collerton and colleagues postulate a primary role for cholinergic dysfunction and a secondary role for dopamine dysfunction [45]. Therefore, cholinergic hypoactivity alone, or dopamine hyperactivity only if cholinergic hypoactivity is present, are the neurochemical mechanisms for hallucinations. This suggestion is supported by evidence that combined cholinergic and neuroleptic treatment is effective in reducing hallucinations in Alzheimer’s disease [46]. One future direction would be to carry out a similar study in PD patients. As the disease progresses, deterioration in cholinergic function combined with dopaminergic hyperactivity is one postulated set of factors that leads to the development of hallucinations.