Lewy body degenerations as neuropsychiatric disorders

Introduction

In his 1817 monograph An Essay on the Shaking Palsy, James Parkinson rendered the phenomenology of the movement disorder that now bears his name. Much has since been learned about the pathogenesis and natural history of Parkinson’s disease (PD), including the profound clinical heterogeneity of the disorder, genetic and environmental risk factors, and cellular-metabolic dysfunctions. The discovery of Lewy bodies (LB)—neuronal inclusions featuring the protein α-synuclein (αSyn)—as a neuropathological hallmark of PD enabled its later co-classification with the less common disease Dementia with Lewy bodies (DLB), which often features parkinsonian motor signs.

Interestingly, Parkinson was nosologically specific when he famously stated that the “senses and intellect” are not affected by the disorder. Indeed, the fundamental view of PD as a movement disorder has remained largely stable. However, in recent decades, research has clarified that a plethora of non-motor and neuropsychiatric symptoms (NMNS) are core features of PD. In this review, we survey the state of this research with the intent of articulating the value of focusing on NMNS as clinically meaningful signs of disease. We also focus on the concept of prodromal PD as a model in which to evaluate the significance of NMNS in PD— e.g., as biomarkers of its variegated pathophysiological patterns or future progression. Finally, we compare PD to the course and manifestations of DLB insofar as it is helpful for relating the neuropathology of Lewy body degenerations to their neuropsychiatric complications.

PD as a neuropsychiatric disorder

What is to be our portrait of the prototypical PD patient? From the traditional neurological perspective, PD is clinically identified and defined by the demonstration of several classic motor signs of insidious onset. These four cardinal features include:

1. low-frequency, “pill-rolling” (supination-pronation) hand tremors at rest;

2. muscle rigidity during passive movement;

3. bradykinesia affecting the initiation and execution of internally generated movements;

4. loss of postural reflexes, manifesting as postural instability.¹

These objective, visible, and measurable motoric impairments are intrinsically linked to the nosology of PD and the clinical entity that the term evokes. The strength of these associations testifies to the influence of Jean-Martin Charcot, the “Father of Neurology,” in the characterization of PD. Charcot credited Parkinson for his initial sketches of the disease and popularized the use of “Parkinson’s disease” as a term. However, Charcot also took a special interest in the disease near the end of the 19^th century, leveraging his characteristically meticulous clinical precision to shape the construct of PD as a disease of movement.² In so doing, Charcot identified each cardinal feature of PD listed above and also cemented PD within the domain of neurologists and movement disorder specialists.

While the classic movement abnormalities of PD remain central to the clinical identification of the disorder, research in recent decades has shown that PD also causes a variety of neuropsychiatric problems for affected patients. For example, cognitive deficits are observed throughout the natural history of PD and dementia eventually affects the majority of patients with longstanding disease (15+ years).³ Given the progressive nature of this cognitive dysfunction, it is possible that its delayed recognition by the medical community is best explained by improved management and thus a more frequent engagement with longstanding disease. However, as will be discussed in this review, other frequent NMNS in PD occur prior to the onset of the movement signs and throughout the course of disease. Some frequently emphasized examples include psychosis, depression, anxiety, sleep disturbances, changes in vision, olfactory impairment, and gastrointestinal dysfunction. The relatively recent appreciation for these NMNS cannot be entirely explained by advances in medical care or increasing survival times for PD patients, suggesting that shifting medical ideologies and cultural attitudes among both patients and providers may be important factors. Increasing emphasis on quality-of-life measures is likely part of this trend, as NMNS are often rated as being more bothersome to patients than their motor impairment.^4,5 They are also prolific—one study found that 98.6% of PD patients endorsed at least one NMNS, with 8 being the average number of NMNS for each patient.⁶ Therefore, the historic emphasis on movement in PD probably reflects the understandable propensity to problematize the most visible, treatable, or quantifiable aspects of a patient’s presentation.

A practical impetus for focusing on motor impairment in PD research has also been that it is essentially the only disease manifestation that can be linked reliably to a specific neuropathological lesion. In the early 20^th century, converging discoveries from researchers across Europe identified that: (1) degeneration of the pigmented cells in the substantia nigra pars compacta (SNc) was involved in PD; (2) intracellular, eosinophilic, proteinaceous inclusions—called Lewy bodies (LB) after Fritz Heinrich Lewy, who described them in 1912—were present in the SNc of PD patients. Degeneration of dopamine (DA) producing cells of the SNc cause the motor signs of PD, especially rigidity and bradykinesia. These discoveries afforded a coherent disease model for motor impairment as a consequence of a hypodopaminergic striatum, which eventually led to the rational design of DA replacement therapies, such as levodopa. Motor impairment is attributable to loss of dopaminergic regulation in the putamen, the principal striatal component of the “motor circuit” connecting the cortex and basal ganglia; however, caudate regions that engage in cognitive loops with the orbitofrontal and dorsolateral prefrontal cortices exhibit comparable dopamine deficits.^7,8 Other nuclei exhibit degeneration early in PD, such as the noradrenergic locus coeruleus (LC), and the cholinergic neurons of the basal forebrain. Importantly, it is likely that neuromodulator dysfunction across complex and diffuse circuitry is more likely to engender NMNS in PD than specific neuronal lesions per se. Additionally, there is a challenge in that non-dopaminergic pharmacotherapies effective in the general population may be less effective in the context of PD, or vice versa. Parsing out these distinctions requires costly trials and/or obtaining approval for PD-specific indications, which has been a major challenge for the field.

Expanding the scope: Prodromal PD

Viewing PD as a neuropsychiatric disorder may lead one to pose the question: What is the utility of PD as a biomedical construct when divorced entirely from the motor signs of the disease? In other words, is it meaningful to talk about whether a person could have PD without ever exhibiting its characteristic movement abnormalities? In the current moment, this seems naïve given that these motor signs are the only reliable clinical indicator of PD, whereas many of the NMNS in PD are relatively non-specific. However, the distinction may lie in whether one limits the discussion to this clinically overt phase of the disease, or to expand our conception of the disease process to encompass the earliest manifestations of the pathological processes that ultimately cause PD as it is clinically recognized. Our understanding of PD pathophysiology has greatly increased alongside of the clinical appreciation of NMNS; as such, future improved diagnostic methods and/or biomarkers that enable earlier detection of the disease process may shift the window of diagnosis into a phase where motor signs are subtle or yet to appear. This idea was central to the recommendations of a 2015 International Parkinson and Movement Disorder Society (MDS) task force on the definition of PD.^9,10 At a broad level, the task force recommended dividing PD into three stages:

1. Preclinical PD is the stage during which latent neurodegeneration or other PD-specific processes have begun, but there are no recognizable signs or symptoms of the disease.

2. Prodromal PD is the stage during which pathological processes have progressed to the point where they are causing signs or symptoms, but these are not diagnostically sufficient or specific enough to define the disease.

3. Clinical PD is the stage of overt, clinically recognizable PD where cardinal movement abnormalities are present.

This recommendation perfectly encapsulates the thrilling atmosphere of clinical PD research today. Ideally, the expansion of PD-specific constructs into earlier events will help researchers identify biomarkers and neuroprotective therapies with relevance to pre-clinical and/or prodromal phases of the disease. Furthermore, in these phases, motor signs are absent by definition, meaning that NMNS may be key indicators of pathological subtypes or prognostic indicators of the rate of progression. For example, if biomarkers are found that enable detection of preclinical PD, then otherwise non-specific symptoms such as depression, apathy, or anxiety may carry more prognostic relevance and clearer pathophysiological correlations. More specifically, subtypes of these and other NMNS could theoretically be linked to specific underlying pathologies or neurotransmitter dysfunction through high-resolution MRI and PET studies.¹¹ As such, the neuropsychiatric elements of PD have never been more relevant. In the following sections, we outline several types and categories of NMNS that appear in clinical and/or prodromal PD.

Cognitive impairment and Dementia

Progressive cognitive impairment is a central element of the disease experience for PD patients. It can be accompanied by marked disruptions of social support, increased experience of stigma, and caregiver distress, making it a key aversive determinant of quality of life (QoL).¹² Early and pronounced cognitive deficits often occur in domains with functional proximity to or dependence on frontal-executive functions,¹³ or cognitive control in the Research Domain Criteria (RDoC) of the National Institute of Mental Health (NIMH).¹⁴ As such, PD is often viewed as a fronto-striatal (or subcortico-frontal) syndrome in which cognitive flexibility, working memory, and attention are affected by disruptions to dopaminergic neuromodulation.¹⁵ This manifests as poor performance on assessments that tax these domains, such as the Tower of London test, attentional set-shifting and task-switching tasks, and Stroop performance. The attractive fronto-striatal model predicts that these deficits might improve with dopamine replacement therapy, but studies investigating this hypothesis have provided mixed results.^15–17 As such, the mechanistic relation of dopaminergic deficits to frontal-executive function in PD—and how relevant deficits might be therapeutically addressed—remains unclear. Furthermore, PD is associated with abnormalities in neurotransmitter-specific circuits other than dopaminergic, such as cholinergic and catecholaminergic, which may explain, in part, the diversity of cognitive impairment.¹³ For example, PD autopsies reveal severe loss of noradrenergic neurons and cholinergic neurons (in the LC and nucleus basalis, respectively) that may ultimately be more severe than SNc degeneration.^18,19

Due to the heterogenous presentation of cognitive dysfunction in early- to mid-stage PD, a battery of neuropsychological assessments that assess specific cognitive domains are commonly employed in research studies and in formal diagnosis of cognitive impairment. However, until recently, the field has lacked a consensus understanding of how cognitive dysfunction manifests in PD prior to the onset of dementia. In the pursuit of such, formal diagnostic criteria for mild cognitive impairment in PD (PD-MCI) were put forth in 2012 by a task force commissioned by the Movement Disorder Society (MDS).²⁰ PD-MCI criteria focus on five cognitive domains: attention and working memory, executive function, language, memory, and visuospatial function. Impaired function on ≥ 2 neuropsychological tests within a single domain or on ≥ 1 test within two or more domains is required for PD-MCI diagnosis. As such, whether PD-MCI is a truly standardized approach to diagnosing and researching cognitive impairment in PD remains somewhat debatable. The concept of cognitive subtyping may provide a powerful auxiliary approach to understanding diversity within PD-MCI. For example, a cohort study (CamPaIGN) differentiated patients with frontostriatal executive deficits to those with a “posterior cortical” impairment profile distinguished by problems in semantic fluency, memory, and visuospatial skills.^21,22 The latter is more strongly associated with conversion to dementia and is speculated to be grounded in cholinergic dysfunction, possibly with concomitant Alzheimer disease (AD) neuropathology.²³

Regardless of their specific cognitive trajectory, a majority of patients with longstanding PD will eventually develop dementia (PDD).^3,24 The etiological basis of PDD remains profoundly elusive. For example, in many cases, it may be in fact caused by a co-morbid dementing illness such as AD. Neuropathological investigations have shown that a majority of PDD cases have at least some AD pathology;^25–27 furthermore, amyloidogenic “cross-seeding” interactions between the AD-linked protein tau and αSyn proteins have been demonstrated.^28,29 Alternatively, the fact that AD pathology is not universal in PDD may reasonably be construed as evidence that increasing cortical αSyn pathology may be sufficient to make PD per se a dementing illness.³⁰ The “elephant in the room” in this debate is that these traditional pathological markers—Lewy bodies in PD or DLB, neurofibrillary tangles (NFTs) and amyloid plaques in AD—are probably epiphenomena of compensatory and orderly cellular processes rather than the principal drivers of neurodegeneration.^31–33 The hypothesis that the accumulation of such proteins causes neuronal degeneration in these diseases has not been borne out in recent decades. Ultimately, deciphering the mechanistic basis of PDD will likely require alternative approaches that enable detection of early disease processes, such as high-resolution neuroimaging modalities.

Whereas its pathological basis is debatable, the impact of dementia on PD patients and their caregivers is clear. Most of the increased mortality risk that comes with a diagnosis of PD can likely be attributed to the eventuality of dementia, making it a critical turning point in the disease trajectory.³⁴ As might be expected, PDD is also a strong predictor of both lower patient QoL and caregiver burden.³⁵ Interestingly, the neuropsychiatric profile of PDD tends to be different from that of AD. For example, most AD patients fit a classical “cortical” cognitive profile with memory and language deficits, whereas PDD is more commonly a “subcortical” dementia (attention, executive, visuospatial dysfunction). Additionally, psychosis is a common feature of both PDD and AD, but hallucinations are twice as common (~45% prevalence) among patients with PDD than delusions (~25%).³⁶ For AD patients with psychosis, this pattern is reversed. Rivastigmine is the only FDA-approved drug that is considered efficacious by the MDS Evidence-Based Medicine Committee for improving or slowing the decline of cognitive function in PD, highlighting the functional relevance of diffuse degenerative processes in PD that affect multiple neurotransmitter systems.³⁷

Psychosis

PD psychosis (PDP) manifests principally as visual hallucinations (VH) and delusions. VH are the most frequent psychotic symptom, with 50% of PD patients reporting them at some point in the disease course.³⁸ Nonvisual hallucinations (auditory, olfactory, tactile, etc.) also occur, but are more common among patients who do not begin to hallucinate until late in disease.^39–41 This pattern can be contrasted with schizophrenia, where unimodal hallucinations are most likely to be auditory in nature. Interestingly, unimodal hallucinations are more common in late-onset schizophrenia, paralleling the convergence observed in late-onset PDP, but the significance of this is unknown.^42,43 In nondemented PD patients, delusions are rarer (~5%) than in schizophrenia (> 50%); they are mostly seen in severe advanced PDP.^44,45

In late-stage disease, psychosis is among the most disabling NMNS of PD. Because PDP is often accompanied by personality changes, confused states, and agitation, it has a dramatic impact on caregiver burden and represents a strong predictor of nursing home placement.^46,47 Some reports from early in the levodopa era suggested that levodopa may precipitate psychotic experiences,⁴⁸ and there is some meta-analytic evidence for increased risk of VH with non-ergoline DA agonists.⁴⁹ However, it has been noted that this distinction may be confounded by the lack of emphasis on VH and other NMNS in the era of first-generation ergoline DA agonists.⁴⁴ Regardless, large prospective and population-based studies have since failed to show appreciable correlations between psychosis and levodopa equivalent dose (LED).^50–53 High-dose levodopa infusions (with plasma level monitoring) were not found to precipitate hallucinations in PDP patients, even though they did initiate hyperkinetic motor signs (dyskinesias) that correlate well with high levodopa levels.⁵⁴ Finally, hallucinations are known to have occurred before the advent of levodopa.⁵⁵ On the other hand, multiple reports have found correlations between VH and day-to-day severity of “hyperdopaminergic” states and symptoms, such as dyskinesias, impulse control disorders (ICDs), and dopamine dysregulation syndrome (DDS).^56,57 This suggests that, although dopamine levels may not precipitate or temporally relate to PDP symptoms, other mechanisms may exist that make patients vulnerable to both PDP and the experience of hyperdopaminergic symptoms.⁵⁸ As such, it is routine clinical practice to consider LED-reduction—especially DA agonist dose—as a first step in the management of PDP.^59,60

Symptoms of PDP are not rare in PD; indeed, 50% of PD patients report experiencing VH at some point during the disease.³⁸ Whereas severe and debilitating psychosis is most often seen in the context of PDD, milder perceptual disturbances with insight often start prior to any severe cognitive deterioration. PDP is sometimes represented as having a spectrum of severity, with “benign” or “minor” VH (e.g. passage hallucinations) or illusions portending an eventual loss of insight and more serious symptoms, including delusions.^61,62 However, comprehensive longitudinal natural history studies to support this concept of PDP are lacking. There is some evidence to suggest that PDP is not necessarily a chronic condition and ever-worsening condition, but may exhibit remission and relapse for yet undetermined reasons.⁶³ Research criteria for PDP have been formulated by a joint NINDS/NIMH task force,⁶⁴ which may help standardize measurements across studies. Notably, the NINDS/NIMH criteria added two minor phenomena to the spectrum of PDP: illusions and false sense of presence. As might be expected, this increases prevalence estimates; one comparative study found an increase in prevalence from 43% to 60% for PDP using the NINDS/NIMH criteria.⁶⁵ It remains to be seen whether this updated construct will be of benefit to researchers and clinicians hoping to intervene early in the course of PDP.

Given its impact on patient QoL, improved strategies for managing PDP are greatly needed. Atypical antipsychotics are most commonly used in PD due to their reduced risk for worsening parkinsonian symptoms (i.e. reduced dopamine receptor blockade effects).^66–69 Among these drugs, the only option with high-quality RCT evidence for efficacy in PD is clozapine.³⁷ Quetiapine is also often used, but RCT-based evidence for its efficacy in PDP specifically is not conclusive. There is some evidence for increased mortality risk in PD with any antipsychotic use relative to non-use,^70,71 but the generalizability and significance of these findings to clinical practice is hotly contested.⁷² The antipsychotic utility of clozapine is thought to relate to its antagonism at serotonin (5HT) 2A receptors, rather than any activity on dopaminergic neurotransmission.^68,69,73,74 Recently, after being granted breakthrough therapy status by the FDA, pimavanserin became the first drug specifically approved for psychosis in the context of PD.^75–77 Pimavanserin is a highly selective 5HT_2A receptor inverse agonist—i.e., it reduces 5HT_2A receptor signal transduction to sub-basal levels.⁷⁷ Its efficacy has buttressed a “serotonergic hypothesis” of PDP and pimavanserin is also being investigated for utility in other contexts.^76,77 However, it is likely that PDP phenomenology ultimately entails disruption to multiple brain systems and neurotransmitter networks, including dopamine and acetylcholine.⁷⁸ For example, it has been suggested that serotonergic mechanisms are most relevant for affect-associated psychosis in PD (e.g., hallucinations occurring during depressive or manic episodes), whereas late-onset psychosis that occurs alongside MCI or dementia may be more grounded in cholinergic dysfunction.⁷⁹ Although no formal study of comparative efficacy between clozapine, pimavanserin, and quetiapine exists, the experience in our clinic is that pimavanserin has at least equal benefit for positive symptoms, is the easiest to use due to once daily dosing with no specialized monitoring, and is the best tolerated of the MDS recommended antipsychotic options.

Sleep-related disturbances

Sleep-related NMNS affect most PD patients;⁸⁰ examples include daytime somnolence, fatigue, insomnia, nocturia, vivid dreaming, sleep fragmentation, and REM behavior disorder (RBD).^81–83 Dopaminergic medications, especially DA agonists, may also interfere with sleep quality or increase a subjective feeling of sleepiness in patients.⁸⁴ However, rotigotine—a DA agonist delivered continuously via a transdermal patch—appears to be effective as a treatment for insomnia in PD.⁸⁵ Fatigue may be alleviated by rasagiline, but patients with symptoms related to obstructive sleep apnea may benefit from continuous positive airway pressure (CPAP) management with specialized monitoring.³⁷

More broadly, emerging research strongly implicates some connection between poor sleep quality and sleep-related behavioral disturbances in neurodegenerative conditions.⁸⁶ Often, this is explored in relation to the effects of disease on sleep centers or circuits. Sleep issues may also increase risk for neurodegenerative conditions, as sleep deprivation reportedly reduces the clearance of protein aggregates in both mice and humans.^87,88 However, more attention has been centered on how pre-motor latent pathologies might drive sleep disturbances, with a focus on articulating prognostic biomarkers. For example, the single most powerful predictor of incipient LB disease is polysomnography confirmed RBD.^10,89 Therefore, a model of how LB pathology could be inducing RBD or other sleep disturbances may enable recognition of PD far before the onset of overt motor dysfunction.

Among the sleep-related phenomena seen commonly in PD, RBD is the most dramatic example. RBD is a parasomnia distinguished by a failure to maintain muscle atonia.⁹⁰ Consequently, affected individuals are able to move and speak as they are inclined in response to dream scenarios. These are often in response to emotionally disturbing or violent dream experiences, leading to commensurately violent reactions in which patients may injure themselves or others; in an early case series of 93 patients, 32% reported injuring themselves and 64% had injured their spouses.^91,92 Whereas idiopathic RBD has an estimated point-prevalence of approximately 1% in the general population,⁹³ RBD is likely present in over 50% of PD patients, with similar conferred risk of injury.^94–96 Current REM theory holds that the critical atonia-generating circuitry involves the pontine sublaterodorsal nucleus (SLD alternatively locus subcoeruleus) and potentially groups of inhibitory neurons in the ventral medulla.^97–102 Pathology or changes in these and other brainstem nuclei have been observed in autopsy cases and imaging studies of PD patients with RBD.^103–105

Mood disorders and affective dysregulation

Affective symptoms and mood disorders are highly prevalent among PD patients. Given the clear impact on monoaminergic circuitry integrity and function,¹⁰⁶ it seems logical to infer that increased rates of mood dysregulation is likely to be rooted in the organic neuropathobiology of PD. However, there is certainly room to argue that PD per se does not universally increase risk for affective symptoms more than other chronic and disabling conditions. For example, well-controlled studies suggest that while the severity and prevalence of depression in PD patients is quantifiably higher than in healthy controls, they are not notably different from those with other chronically disabling diseases, such as rheumatoid arthritis.^107,108 Altogether, the relative contributions of pathobiology and the psychological impact of diagnosis are difficult to parse. Nonetheless, the field’s current consensus is that affective symptoms may be shaped into unique forms by the disease—and its pharmacologic management—that require a PD-specific therapeutic approach for maximal efficacy.³⁷ Furthermore, PD-associated depression may be strongest or most common in the prodromal phase, suggesting some association with early pre-motor disease processes.¹⁰⁹

Depression

Estimates for the prevalence of depression in PD varies widely based on the diagnostic approach. Early systematic reviews suggested a prevalence of about 30% on average, but cited figures as high as 76% in studies using depression scale cutoffs.^110,111 A meta-analysis of this effect found that studies using Diagnostic and Statistical Manual of Mental Disorders (DSM) criteria for major depressive disorder (MDD) reported a weighted mean prevalence of 7%.¹¹² The same study found that clinically significant depressive symptoms were present in 27% of patients when using DSM criteria, but 42% when using a depression scale cutoff.¹¹² However, the DSM may have limitations in this context. For example, MDD features overlap with the presentation of PD, including anhedonia, psychomotor retardation, fatigue, flattened affect, physical or somatic complaints, and impaired cognition. Therefore, it has been argued that a diagnostic approach that recognizes these symptoms as related to depression as well as PD may have higher validity and help identify a treatable condition.¹¹³

Prompt recognition and treatment of depression in PD can substantially improve patient QoL. In 2019, meta-analytic recommendations of an MDS task force listed pramipexole (a dopamine agonist) and venlafaxine as the only drugs known by high-quality (RCT) evidence to be efficacious and clinically useful for reducing depressive symptoms in PD.³⁷ Tricyclic antidepressants (TCAs) were deemed “likely efficacious”, whereas insufficient evidence is indicated for selective serotonin reuptake inhibitors (SSRI), monoamine oxidase (MAO) inhibitors, and non-pharmacologic neurologic interventions (e.g. electroconvulsive therapy) all lacking sufficient evidence of efficacy.³⁷ Recently, a quantitative RCT meta-analysis found that TCAs, MAO-inhibitors, and SSRIs all had significant effects on depressive symptoms, but this analysis was restricted to five studies that reported continuous outcome measures (i.e. depression scales).¹¹⁴ A similar review that included 20 studies was able to show efficacy only for SSRIs.¹¹⁵ The paucity of available data and inconsistency across meta-analyses highlights the need for more comprehensive and comparative efficacy studies, as well as a standardized understanding of PD depression. Notably, there is also evidence for the efficacy of psychosocial treatments, such as cognitive behavioral therapy (CBT), particularly for the emotional, cognitive, or behavioral elements of depression.^37,115–118

Anxiety

As is the case for depression, current thinking holds that PD may shape the expression of anxiety into disease-specific syndromes or subtypes,¹¹⁹ with potential implications for recognition and treatment. When applying DSM anxiety disorder criteria to PD patients, meta-analytic evidence¹²⁰ places the weighted prevalence of anxiety disorders at 31% in PD, but 31% of these patients have at least two comorbid anxiety disorders. Generalized anxiety disorder (GAD) and social phobia are the most common at 14% each, with agoraphobia and panic disorder being present in 9% and 7% respectively. Unlike for depression, this estimate does not vary much based on the application of DSM criteria or the use of anxiety scales. As such, research into PD-specific anxiety is focused on differentiating unique or diverging anxiety syndromes, rather than enhancing overall sensitivity to detection.^121–123 For example, panic disorder is more common in young-onset PD and associated with higher rates of therapy complications (fluctuating medication response, comorbid depression, and lower QoL scores).¹²⁴ This could be interpreted to mean that the pathophysiology of panic overlaps in part with that of latent PD neurodegeneration, or that one accelerates the other. Importantly, there is evidence that anxiety may itself be a risk factor for PD, but how distinct subtypes may modify this association is not clear.^125,126 Anxiety is also highly co-morbid with depression in PD, making it important to screen for anxiety disorders in depressed individuals.¹²⁷ In addition, certain types of anxiety might be uniquely associated with PD and therefore dissociable from depression, such as anxiety episodes that occur during fluctuations between the “on” and “off” dopamine medication states.^119,124

Remarkably, there have not yet been any quality RCT studies for anxiety treatments in PD.³⁷ In the general population, antidepressants and benzodiazepines are the standard of care. While several RCTs for the treatment of depression in PD have included anxiety as a secondary outcome, all failed to demonstrate improvement in anxiety despite using a variety of different antidepressants.^128,129 None of these studies were designed to detect changes in anxiety, so it is difficult to be confident about the lack of efficacy. Benzodiazepine use in PD is problematic for a number of reasons, chiefly an increased risk of falls and dementia.¹³⁰ Given the high prevalence and impact of anxiety on QoL in PD, RCTs with anxiety as the primary outcome are needed.

Apathy

Given the role of dopaminergic neurotransmission as a critical neuropsychological substrate of motivated behaviors,¹³¹ it is intuitive that PD patients might experience something like apathy. Apathy is a condition marked by reduced motivation—as indicated by a reduction in goal- or reward-directed activity and the expression of interest or emotions that often accompany them—that is not clearly the result of cognitive dysfunction, confusion, or distress.^132,133 As is the case for anxiety and depression, apathy is common in prodromal PD, with a prevalence of approximately 30% in de novo PD (i.e. before initiation of PD-specific medication).¹³⁴ Reduced executive function in PD may foster apathy by interfering with attention, memory, and planned behaviors, thus introducing roadblocks to the pursuit of otherwise motivating outcomes.¹³⁵ Although dopamine cells in the ventral tegmental area (VTA) are more often associated with affectivity and reward-oriented behavior than SNc neurons, selective ablation of SNc dopamine neurons was sufficient to cause motivational deficits in rats.¹³⁶

Interestingly, apathy is commonly noted in patients who have undergone surgery for subthalamic nucleus deep-brain stimulation (STN-DBS).¹³⁷ When occurring several months after surgery, this is typically related to post-operative reduction of dopaminergic medication dosage (e.g. levodopa), making it a form of “dopamine withdrawal.”¹³⁸ This is mechanistically supported by imaging data showing that post-operative apathy correlates with mesocorticolimbic dopaminergic denervation.¹³⁸ This withdrawal syndrome may be ablated by post-operative administration of dopamine receptor agonists with activity at D2 and D3 receptors.¹³⁹

Given the difficulty of dissociating apathy from related or overlapping disorders (depression, anxiety, cognitive impairment), casual clinical interactions are typically not sufficient to recognize apathy in the complicated context of PD.¹³² Similarly, no diagnostic criteria or treatment strategies have been uniformly adopted. However, rivastigmine is an effective treatment.¹⁴⁰

Impulse control disorders

In sharp contradistinction to the reduced activity observed in apathy, PD patients may develop impulse control disorders (ICDs), marked by excessive reward- or goal-directed activity. Whereas apathy is likely a function of diffuse dopaminergic denervation, ICDs are considered a hyperdopaminergic phenomenon, often iatrogenic in nature. ICDs have been reported in approximately 15% of Parkinson’s disease patients receiving dopamine receptor agonists, but are less common in patients on carbidopa/levodopa monotherapy.¹⁴¹

While ICDs have also been reported when dopamine agonists are used for non-PD conditions such as hyperprolactinemia and restless leg syndrome, they are still much less frequent than in PD. Further, PD patients newly diagnosed and never treated with dopaminergic medications compared to healthy controls do not appear to have an increased risk of ICDs.¹⁴² These findings suggest that both exposure to dopamine in the form of agonists (rather than levodopa) and the condition of PD are needed to maximally express ICD behaviors.¹⁴³ The mechanism for this outcome is unclear, however dopamine agonists have a high binding affinity for D3 receptors, which are especially dense in the limbic and forebrain regions compared to motor circuits. Compounding this issue of greater relative density of D3 receptors in circuits that mediate behavior, the pattern of dopamine loss in the SNc primarily affects the ventral neurons while sparing the dorsal neurons that modulate the limbic and forebrain systems, potentially leading to overstimulation.

Synucleinopathy and NMNS in LB diseases

Whereas NMNS have been a relatively recent focus of PD research, they represent the primary features of DLB, the neuropathological sister of PD. Lewy bodies are distinguished by the presence of the neuronal protein α-synuclein (αSyn), which forms amyoidogenic aggregates that spread throughout the brain in a stereotyped fashion in both disorders. According to the staging scheme devised by Heiko Braak,^144,145 these LB aggregates appear first in the autonomic and enteric nervous systems, medullary and pontine nuclei, spinal cord, and olfactory nucleus (stages I-II). Later αSyn pathology reaches the midbrain SNc, basal forebrain, and amygdala (stages III-IV), followed by extensive neocortical disease (stages V-VI). However, parkinsonian motor signs are thought to appear only after the disease has depleted well over half of the SNc dopamine neurons, which may require as much as a decade after LB pathology appears in the midbrain.¹⁴⁶ Variations in this process are thought to explicate part of the heterogeneity within and between LB diseases. For example, neuroinflammation, genetic factors, or environmental exposures might permit accelerated neocortical disease or neuronal vulnerability.¹⁴⁷ An extremely attractive hypothesis at the current moment is that different “strains” of amyloidogenic αSyn exhibit diverging propensity for cellular or regional spread.^148,149

Neuropsychiatric signs are among the core clinical features of DLB:

1. fluctuating cognition and alertness;

2. recurrent well-formed VH;

3. REM sleep behavior disorder (RBD);

4. one or more of the cardinal features of parkinsonism.¹⁵⁰

Additionally, there is extensive overlap between PD and DLB with respect to the repertoire of neuropsychiatric complications observed in patients. This naturally suggests that these α-synucleinopathies preferentially affect nuclei and cell types with relevance to these NMNS manifestations. However, the timecourse of NMNS and parkinsonian motor impairment in DLB and PD differ substantially. Nosologically, the essential “cortical” signs of LB disease—such as cognitive changes and dementia—often precede any parkinsonian motor impairment in DLB. A semi-arbitrary “one-year rule” is used to discriminate DLB from PD, where dementia onset ≤ 1 year after parkinsonian motor symptom onset supports a diagnosis of DLB over PD.¹⁵⁰ Furthermore, DLB with dementia as a presenting symptom is often misdiagnosed as AD, which exhibits co-morbidity with LB diseases as discussed earlier. Thus, LB disease is only one of the etiological spectra upon which PD and DLB rest.

Prospective studies have evinced that 75% of RBD cases will experience LB disease onset within about a decade, making it the strongest predictor of PD or DLB.⁸⁹ Theoretically, it is likely that all RBD patients have a prodromal LB disease that will manifest if given enough time. Therefore, the RBD patient population has become a important focus of therapeutic intervention to prevent or delay PD and DLB. These studies may also yield insights into the early or latent processes driving LB disease. For example, RBD patients exhibit increased glial activation in the brainstem and striatum, suggesting an early role for neuroinflammation in relevant nuclei.¹⁵¹ RBD may also portend more severe cholinergic dysfunction (i.e. cognitive impairment and autonomic dysfunction) as the dominant NMNS patients will face.¹⁰⁵ Studying RBD will hopefully continue to yield valuable prognostic information, such as factors associated with a motoric predominance to the LB disease (PD) or early dementia (DLB).

The New Parkinson’s Patient

The natural history of PD exhibits a significant degree of heterogeneity, but some key risk factors and the construct of prodromal PD enable us to suggest an updated portrait of the prototypical PD patient. This patient’s trajectory illustrates the impact of NMNS in determining patient quality-of-life and informing provider and caretaker decision-making.

The patient is a male who in his 30s experiences several bouts of major depression. He is treated with a selective serotonin reuptake inhibitor (SSRI) and continues an otherwise unremarkable adulthood. Around his 50^th birthday, he begins to note an uptick in general anxiety and difficulty sleeping at night, but attributes this to worries about his advancing age and the economics of retirement. In his mid-50’s, he begins to experience abdominal discomfort after large meals and is more frequently constipated than in the past, but his physician tells him this is normal and suggests that he take a non-prescription laxative for symptomatic relief. He continues to have intermittent difficulties staying asleep and at the age of 57 his partner notes that he has been “tossing and turning” more frequently. Concerned about an emerging sleep disorder, his physician orders a sleep study. The polysomnograph is consistent with a diagnosis of RBD and he is started on clonazepam, which helps with his sleep problems. At the age of 66, he starts to notice a slight tremor in his right hand while watching television. His physician orders a PET/SPECT scan which shows that dopamine transporter (DAT) ligand binding in the striatum is reduced to approximately 60% of what is typical for a person of his background. A diagnosis of clinical PD is made and he is referred to a neurologist to discuss management. One year later, he begins to exhibit right-sided rigidity in his arm and worsening tremor that makes it difficult to perform tasks such as eating, writing, and driving. His neurologist initiates treatment with a DA agonist, which ameliorates the symptoms for a while. Gradual worsening of his motor signs over the next couple of years leads to a need for levodopa-carbidopa treatment. About five years after the initial diagnosis, he begins to have some difficulty discerning his sleeping and awake states, with some dream-like images appearing during the daytime. These include recurrent hallucinations of snakes crawling on the couch, prompting his neurologist to prescribe clozapine, which is somewhat effective. Around the same time, he reports some difficulty remembering things and an increasing dependence on his partner to “jog his memory.” Over the next three years, his rigidity and postural instability worsen significantly. He also begins to exhibit rapid cognitive deterioration and to experience paranoid delusions consistent with psychosis. A diagnosis of PD dementia (PDD) is made and his family elects to relocate him to an assisted living facility soon after.

Summary

The movement disorder paradigm of PD that dominated in the 19^th and 20^th centuries is giving way to a holistic, neuropsychiatric conceptualization of the disease. This has been motivated by converging lines of research, particularly the neuropathological connection to DLB and the clinical significance of prodromal LB disease. Other means of understanding these diseases, such as large-scale investigations of genetic or proteomic data, may help us move to an even more granular understanding of heterogeneity within them. For now, we are optimistic that increasing recognition of NMNS and their impact on QoL for PD patients and caregivers will help ameliorate these disabling elements. We think it is also likely to justify further clinical trials to ascertain the most effective pharmacotherapeutic strategies.

Synopsis.

Parkinson’s disease (PD) has historically been conceptualized as a movement disorder. In recent decades, non-motor and neuropsychiatric symptoms (NMNS) have become increasingly recognized as being of paramount importance for PD patients. Neuropsychiatric phenomena dominate the course of the other major Lewy body (LB) disease, Dementia with Lewy bodies (DLB). In this review, we survey the clinical relevance of NMNS to the heterogeneous presentations of LB disease and their significance to ongoing research in this area. We also consider how the nature of LB neuropathology may help explicate the basis of NMNS in these two disorders

Key Points.

1. Non-motor and neuropsychiatric symptoms (NMNS) in Parkinson’s disease (PD) and Dementia with Lewy bodies (DLB) are often more burdensome than the classic motor features.

2. Key NMNS in PD include cognitive impairment, psychosis, sleep disorders, depression, anxiety, apathy, and impulse control disorders (ICDs).

3. Over 98% of PD patients endorse at least one NMNS and the average patient endorses 8 separate NMNS features.

4. NMNS may appear long before the onset of motor symptoms in PD, making them critical to the emerging constructs of preclinical and prodromal PD.

5. Different patterns of NMNS between PD and DLB may help understand the neuropathological underpinnings of these disorders in dopaminergic, cholinergic, noradrenergic, and serotonergic networks.