Until recently there were no treatment studies for apathy in PD. Co-morbid psychiatric conditions (e.g. depression) were treated initially. Anecdotally, psychostimulants (e.g. methylphenidate) and stimulant-related compounds (e.g. modafinil) are used in clinical practice, but their effectiveness in PD-D patients is not known. In a small, randomized, placebo-controlled trial in PD patients with apathy but not dementia or depression, a rivastigmine patch significantly improved apathy compared with placebo . Based on the proposed pathophysiology of apathy, antidepressants and other medications that increase dopamine or norepinephrine activity (e.g. dual reuptake inhibitor antidepressants, bupropion, and atomoxetine) may be beneficial . In addition to pharmacological treatment, it is important to educate patients and families about the distinction between apathy and depression and to encourage steps that overcome patient inertia that may lead to improved functioning and quality of life .
Thobois et al.  found that symptoms of apathy occurring in PD patients after STN DBS surgery and withdrawal of dopaminergic therapy improved with piribedil (a D2/D3 receptor agonist) treatment in the context of a placebo-controlled study. Additionally, treatment with piribedil indicated a trend for improvement in depression, quality of life, and anhedonia. This clinical trial excluded PD-D patients, thus the results must be interpreted cautiously in this population.
Psychosis and agitation
Several studies have found a relationship between exposure to PD medication and the presence of some NPS, particularly psychosis and cognitive impairment [75, 80, 127-130]. If tolerated, a decrease in overall exposure to PD medication may lead to an improvement in mental status in PD-D patients. Based on expert opinion, medications are usually discontinued (if tolerated from a motor standpoint) in the following order: anticholinergics, selegiline, amantadine, dopamine agonists, catechol-O-methyltransferase (COMT) inhibitors, and, finally, a reduction in levodopa dosage [83, 131]. In a recent description of clinical outcomes in a small number of PD patients with psychosis, it was reported that a decrease in PD medications commonly led to improvement in psychosis, and the authors estimated that 30% of PD patients with psychosis may not require antipsychotic medication .
The effects of memantine, an N-methyl-D-aspartate (NMDA) receptor antagonist used in AD, has also been assessed in PD-D with mixed results; 20 mg of memantine daily was found to be safe and well-tolerated [133-135]. One double-blind, placebo-controlled, multicentre trial in a mixed population of DLB and PD-D patients found significant improvements in clinical global impressions of change (CGIC) scores; however, there were no improvements in NPI scores . In contrast, a larger randomized, double-blind, placebo-controlled trial, also in a population of PD-D and DLB patients, found that memantine did not improve cognition or NPI scores in PD-D patients, although NPI scores for psychosis, sleep, and appetite in DLB patients did improve .
Cholinesterase inhibitors (donepezil, rivastigmine, and galantamine) have also been studied for their effect on cognitive impairment in PD-D, and indirectly for their effects on NPS . In an open-label study examining the psychiatric benefits of donepezil in DLB and PD-D patients , 35 PD-D patients who completed at least 12 weeks of treatment had an average 12.0 point decrease in total NPI score (approximately a 50% decrease from the baseline NPI score) and reduced caregiver distress. In contrast, a small placebo-controlled crossover study of donepezil in PD-D did not report psychiatric benefits .
There is evidence that rivastigmine, approved for the treatment of PD-D based on the results of a large international study , may also have benefits for behavioural symptoms in this population. In that study, mean NPI scores decreased by 2.0 points in the rivastigmine-treated group over the course of treatment, compared with no change in the placebo-treated group. In addition, significantly more patients in the rivastigmine group had an improvement of at least 30% in NPI score (45.4 versus 34.6%). In a post hoc analysis of the data, it was found that those patients with visual hallucinations at baseline derived greater benefit from rivastigmine treatment relative to placebo treatment .
There is preliminary evidence that cholinesterase inhibitors may improve hallucinations in PD-D. A small open-label study of rivastigmine was conducted in PD patients with a range of NPS and significant cognitive impairment. Total NPI score, the hallucinations and sleep disturbance subscales specifically, improved significantly in those who completed the study (80% of the sample) over the course of 6 weeks of maximal treatment, and caregivers reported significantly less distress over time . Similarly, a small open-label study of galantamine in PD-D found benefit for NPS overall and psychosis specifically .
When psychosis does not improve with the aforementioned clinical interventions, it is sometimes necessary to introduce antipsychotic treatment. Recent research shows that approximately a third of older PD patients newly treated with dopaminergic agents will be prescribed an antipsychotic within a 7-year period . Four new-generation antipsychotics have been studied in the PD population: clozapine, risperidone, olanzapine, and quetiapine [80, 127, 128, 130]. Only clozapine has been shown to be efficacious for psychosis in PD, with few adverse motor effects [80, 127, 130]. Unfortunately, clozapine has a rare but serious side effect, agranulocytosis [80, 127, 130], which necessitates routine blood monitoring.
Risperidone and olanzapine are not recommended for use in PD, due to limited research and evidence that their use is associated with worsening parkinsonism [80, 127]. In one meta-analysis, worsening of motor symptoms was reported in 33% of PD patients treated with risperidone ; a similar meta-analysis found worsening of motor symptoms in 40% of patients treated with olanzapine [80, 145].
Quetiapine has become the first-line antipsychotic treatment for PD psychosis based on the results of open-label reporting of symptomatic improvement and good tolerability from a motor standpoint [80, 127, 130]. However, two randomized, placebo-controlled studies of quetiapine for psychosis in PD have been negative [146, 147]. In addition, a retrospective chart review of quetiapine for the treatment of hallucinations in PD found that PD-D patients were as likely as PD patients without dementia to experience a decrease in psychotic symptoms, but were more likely to report worsening of motor symptoms . Another retrospective analysis found that while approximately 80% of the PD patients experienced at least partial remission of psychotic symptoms with quetiapine treatment, the presence of dementia was independently associated with non-response . Finally, in an open-label study of quetiapine for psychosis in PD patients with and without dementia, those without dementia demonstrated a trend toward improvement on the Brief Psychiatric Rating Scale (BPRS) while PD-D patients showed no improvement. In addition, PD-D patients were more likely to experience motor worsening, required a longer titration period, and were eventually treated with a higher mean (SD) quetiapine dosage: 151 (90) versus 76 (59) mg/day .
In a recent study, a novel antipsychotic, pimavanserin [a serotonergic receptor (5-HT2A) inverse agonist] was found to be efficacious for the treatment of psychosis in PD in a randomized, placebo-controlled study . Pimavanserin was superior to placebo on the primary psychosis outcome measure [change in the Scale for Assessment of Positive Symptoms (SAPS)-PD score], on several secondary outcome measures (caregiver impression ratings, symptoms of sleep and wakefulness, and caregiver burden), and was well tolerated from a motor standpoint. This drug may be approved for the treatment of psychosis in PD in the near future.
The use of antipsychotics in PD-D is potentially of particular concern. In 2005 the US Food and Drug Administration (FDA) issued an advisory letter warning regarding increased morbidity and mortality in patients with dementia associated with the use of atypical antipsychotics . Specific causes of death reported were cardiovascular or infectious in nature, though prior studies did find significant linkages with cerebrovascular events [153, 154]. In 2008, the warning was extended to include conventional (i.e. typical) antipsychotics [155, 156]. One recent study suggests that the use of antipsychotics in PD may also be associated with increased mortality risk . Thus, given the frequent occurrence of dementia in PD, the associated morbidity and mortality with use of antipsychotics in PD-D is likely to be higher than previously thought .
The effects of risperidone and citalopram on a group of hospitalized individuals with DLB and NPS were recently examined in a randomized study. Discontinuation rates in the DLB group were 65-75%, with no difference between patients randomized to risperidone versus citalopram; NPI and CGIC scores worsened in DLB patients compared with a comparison group of AD patients .
For the reasons mentioned there has been interest in exploring other treatment options for psychosis and agitation in patients with dementia. Cholinesterase inhibitors, antidepressants, benzodiazepines, and mood stabilizers have all been studied to varying degrees for these indications, but there are currently no clear pharmacological alternatives to antipsychotics [160, 161].
Disorders of sleep and wakefulness
Treatment depends on the specific disorder and its aetiology. Sleep disturbances that are due to nocturnal worsening of parkinsonism may respond to adjustments in the PD medication regimen. RLS and PLMS are commonly treated with dopaminergic medications, and RBD is typically treated with clonazepam; melatonin may also help. Preliminary studies suggest that EDS can be treated successfully with modafinil [162, 163], and psychostimulants are also used in clinical practice. The role of other hypnotic or psychiatric medications in the treatment of sleep disturbances in PD has not been evaluated.
The combination of dextromethorphan hydrobromide and quinidine sulphate is FDA-approved for the treatment of PBA based on pivotal trials in patients with multiple sclerosis  and amyotrophic lateral sclerosis , but there have been no controlled studies in PD patients. Numerous small-scale studies have also found both TCAs and SSRIs to be efficacious in the treatment of PBA, although none included PD patients . Anecdotally, SSRIs and mood stabilizers (e.g. valproic acid) appear to be effective for this syndrome in PD, but no reports specifically in PD-D patients are available. In addition, it is important to educate patients and family members on the distinction between PBA and depression.