Parkinson’s Disease: Cognitive Impairment

Cognitive impairments arise from a variety of brain diseases, disorders, and injuries; cognitive disorder syndromes are defined via clinical assessment of performance across a number of cognitive domains along with an appraisal of a patient’s level of everyday functioning. Diagnosis of dementia and subtle forms of cognitive impairment is often considered the sphere of specialist clinicians (e.g., neuropsychiatrists or specialty neurologists), but an appreciation of cognitive syndromes and their potential etiologies and a grasp of appropriate initial cognitive assessments are important aspects of general psychiatric practice. Evaluating cognition and differentiating possible causes of cognitive dysfunction require a working knowledge of definitions and criteria as they relate specifically to PD (5).

Dementia syndromes are characterized by a global decline in neurocognitive function with impairment in at least two cognitive domains (differentiating them from focal cognitive syndromes, e.g., aphasia). In addition, dementia syndromes involve a significant decline from baseline in day-to-day functioning and a normal level of consciousness (differentiating them from delirium), and their typical onset is in adulthood (differentiating them from lifelong mental retardation). The term mild cognitive impairment (MCI), thought likely to be a prodrome to dementia, refers to an intermediate stage of cognitive impairment between the expected cognitive decline of normal aging and dementia; it is detectable through objective testing and is unaccompanied by a significant impact on everyday functioning. After the MCI classification scheme was developed and validated for research on AD (6), similar classification schemes were devised to diagnose dementia and milder forms of cognitive impairment in other diseases with cognitive dysfunction as a common feature, including PD (7). The DSM-5 (8) adopted the terms major and minor neurocognitive disorder to capture the range of cognitive impairment due to various etiologies, with specifiers allowing the diagnosis to be ascribed to a specific etiology, such as “due to Parkinson’s disease.”

In PD, cognitive impairment is heterogeneous, and dysfunction occurs across the range of cognitive domains. The two main categories of cognitive dysfunction in PD are mild cognitive impairment (PD-MCI) and dementia (PDD) (9). Prior to evidence for PD-MCI, including early in the course of PD, patients often demonstrate selective cognitive impairments in executive, memory, attention, and visuospatial functions.

The typical presentation of cognitive dysfunction in PD is regarded as a subcortical profile of deficits, or frontal-subcortical dysexecutive syndrome. The term subcortical dementia, coined in 1974 (10), originally described a syndrome of slowed thinking, impaired ability to manipulate information, and forgetfulness (retrieval deficits) accompanied by apathy or depression in patients with progressive supranuclear palsy (PSP), a neurodegenerative disease thought to lack significant cortical involvement. The dementias of PD and Huntington’s disease (among others) were noted to present with similar deficits; thus, the subcortical dementia concept was invoked to contrast these dementias with the prominent memory and language deficits (e.g., amnesia, aphasia, and agnosia) present in cortical dementias such as AD. This division between cortical and subcortical dementias is controversial given that cognitive impairments are infrequently “clean” in a single patient and that subcortical and cortical deficits may overlap, which may be related to mixed pathologies (11). Nonetheless, placing PD-related cognitive impairment into the conceptual framework of a subcortical syndrome is useful to understand the deficits in attention, retrieval memory, visuospatial and executive functions, and accompanying neuropsychiatric phenomena typical in PD (2).

Heterogeneity in definitions for PD-MCI and PDD and limited applicability of AD-based criteria, which focus on memory deficits, prompted the Movement Disorder Society (MDS) to commission task forces to develop standardized definitions and diagnostic criteria for PD-MCI (9) and PDD (12, 13). These MDS criteria, based on systematic literature reviews and expert consensus, were designed to provide diagnostic guidelines for clinical practice or research that differentiate individuals with PD unaffected by cognitive impairments from individuals with PD-MCI and PDD.

Prototypical cognitive impairments in early PD involve fronto-striatal dysexecutive phenomena, which manifest in everyday life as disorganization, trouble establishing priorities, forgetfulness, and distractibility. These deficits can impact employment or independent living or cause demoralization but may not be problematic for some individuals. Isolated deficits in information processing (e.g., slowed thinking) and verbal fluency may also be evident in PD. The memory deficit in PD mainly affects explicit recall, with relative sparing of recognition memory and the absence of “rapid forgetting” seen in AD. In contrast to patients with AD, most PD patients have insight into their cognitive deficits and attempt to compensate. In this way, the impact of cognitive changes and associated disability is often subtle and context dependent (4).

The MDS diagnostic criteria for PD-MCI require a confirmed diagnosis of idiopathic PD; exclusion of other primary explanations for evidence of cognitive impairment; gradual cognitive decline from previous levels, as reported subjectively by either the patient or an informant or as observed by the clinician; absence of a significant impact on functional independence (though minor difficulties on complex functional tasks may be present); and demonstration of objective impairment either on a scale of global cognitive abilities (referred to as level I MDS Task Force criteria) or through comprehensive neuropsychological testing (i.e., level II MDS Task Force criteria) (9). In a cross-sectional sample of PD patients, 23.3% met the criteria for PD-MCI on the basis of level I assessments, and 60.5% met the criteria using level II assessments (14).

The level I and level II PD-MCI diagnostic categories provide different levels of diagnostic certainty with respect to the extent of the objective assessment (9). Level I criteria require an abbreviated neuropsychological assessment that uses either a scale of global cognitive abilities that is validated for use in PD or at least two tests from a limited neuropsychological battery (defined as one with fewer than two tests across each of the five cognitive domains—attention and working memory, executive function, language, memory, and visuospatial function—or one that assesses fewer than five domains). Level I assessments are intended for use by clinicians “at the bedside” and do not require special expertise in neuropsychological testing. With level I tests, impaired performance is defined by scores 1 to 2 standard deviations below the age-corrected mean score on two measures in one cognitive domain or on one measure in two cognitive domains and without impairments in activities of daily living.

Level II PD-MCI assessments are aimed at determining the severity of deficits, versus their presence, and, in general, are used in research settings. However, use of level II criteria increases rates of diagnosis (and thus can facilitate earlier interventions). Diagnoses based on level II criteria require a comprehensive neuropsychological assessment (defined as two tests within each of the five cognitive domains and impairment on at least two tests within one cognitive domain or one impaired test in two cognitive domains). To facilitate use of PD-MCI criteria across different populations and educational backgrounds, level II criteria define impairment as performance that is 1 to 2 standard deviations below relevant norms or performance that represents a significant decline on serial testing or from estimated premorbid abilities.

Level II assessments also enable PD-MCI subtype classification. The PD-MCI single-domain subtype involves abnormalities on two tests within a single cognitive domain (e.g., executive function) with unimpaired performance in other domains. In the PD-MCI multiple-domain subtype, abnormalities are present on at least one test in two or more cognitive domains (e.g., memory and visuospatial function).

Dementia syndromes in PD typically develop later in the disease course and involve global impairments that preclude independent living. In general, dementia syndromes in PD comprise three subgroups (13). The first involves intensification of selective deficits, especially in memory and information processing. In the second, there is greater involvement of cortical functions, including aphasia, apraxia, and memory, but the course and presentation are distinct from those of AD. A third presentation has features of both PD and AD, with pronounced language deficits.

As a dementia syndrome, PDD is diagnosed only when it develops in the context of an established diagnosis of idiopathic PD. Other core MDS criteria for PDD are exclusion of other possible diagnoses; development of a dementia syndrome represented by impairment in at least two of four cognitive domains, including attention, executive function, visuospatial function, and memory; decline from premorbid level; and deficits that are severe enough to impair daily functioning independent of impairments related to PD-related motor or autonomic phenomena (12). Behavioral symptoms, such as apathy, depressed or anxious mood, hallucinations, delusions, or daytime sleepiness, are supportive of the diagnosis of PDD.

Confidence in a PDD diagnosis also requires consideration of the presence or absence of “uncertainty factors.” These include the coexistence of other conditions that cause dementia syndromes, such as cerebrovascular disease, and whether there is information on the temporal sequence and time interval between development of motor and cognitive symptoms. The time course of motor versus cognitive symptoms helps to distinguish PDD from dementia with Lewy bodies (DLB), a neurodegenerative dementia syndrome frequently accompanied by parkinsonism, psychosis, and sleep and autonomic disturbances. “Probable PDD” has no uncertainty factors, whereas the presence of at least one uncertainty factor renders the diagnosis “possible PDD.” Similar to MDS criteria for PD-MCI, the extent of the objective cognitive assessment is the basis for the MDS criteria for PDD level I and level II assessment approaches.

DLB, the second most common neurodegenerative dementia, accounts for up to 20% of dementia cases late in life (15). PDD and DLB likely represent two ends of a disorder on a continuum of clinicopathology, although their relationship is debated (16). Cross-sectionally, DLB and PDD have overlapping features, namely, dementia and parkinsonism. Both synucleinopathies, PDD and DLB are categorized under the umbrella term “Lewy body dementias” and are considered to occur on a spectrum with respect to presence of dementia. However, some DLB patients do not meet clinical criteria for PD or never demonstrate parkinsonism. According to current diagnostic criteria (17), PDD requires an established diagnosis of PD prior to dementia, and DLB is characterized by a progressive cognitive decline appearing before or within one year of the onset of parkinsonian motor signs. However, this time interval between onset of motor symptoms and dementia is arbitrary and not founded on a clear clinical or pathological basis (18).

PDD and DLB share a mix of clinical symptoms, including prominent executive and visuospatial deficits, parkinsonian motor features, rapid eye movement (REM) sleep behavior disorder, visual hallucinations, and autonomic instability. The cognitive profile in DLB, as opposed to that in PDD, involves a prominent attentional disturbance, including fluctuating attentional states, executive and visuospatial deficits, and, later, memory dysfunction. DLB is also characterized by florid hallucinations, usually visual. Parkinsonism in DLB may respond to levodopa therapy, but levodopa can also aggravate psychosis. Features such as earlier visual hallucinations and psychosis, greater sensitivity to neuroleptics, greater fluctuations in cognition, less responsiveness to levodopa, and more symmetrical but fewer (or no) parkinsonian symptoms may be more characteristic of DLB. However, no single clinical characteristic clearly differentiates one Lewy body dementia from the other (17, 19).

The cornerstones of assessing PD-related cognitive impairment are a thorough clinical interview, objective cognitive screening measures, and additional diagnostic studies as indicated. To guide the exam and differential diagnosis, the first step is to confirm the diagnosis of PD. Deviations from the usual PD motor syndrome should trigger considerations of other explanations for parkinsonism, cognitive changes, and psychiatric phenomena. Ask patients to describe their initial motor symptoms, how they evolved, progressed, and responded to medications, and their impact on ADLs and other daily activities. Typically, in idiopathic PD, the characteristic motor symptoms of resting tremor, slowness, and rigidity begin unilaterally and spread and progress bilaterally. A robust motor response to dopaminergic replacement therapy supports a diagnosis of idiopathic PD. With ongoing use, complications of dopaminergic medications include on-off motor fluctuations and hyperactive dyskinetic choreoathetoid movements that can resemble the movements of Huntington’s disease or tardive dyskinesia. Since PD is a progressive disease, alternative diagnoses are suggested when motor symptoms do not worsen over time or when the patient never requires higher doses of antiparkinsonian medications to control motor symptoms. Inability to adhere to the usual PD medication regimen of dosing several times daily may be indicative of cognitive impairment rather than a nonresponse to medications. While early-onset imbalance and falls are inconsistent with the diagnosis of PD, their later emergence is associated with development of PD-MCI and PDD.

Screening should include the full range of NPS, as they themselves may cause cognitive dysfunction as well as aid in distinguishing PD from other possible diagnoses. Red flags of cognitive dysfunction include fatigue, depression, apathy, psychosis or hallucinosis, bradyphrenia, medication nonadherence, and gait changes (62, 63). Ideally, clinical interviews should be supplemented with information from one or more informants as the patient’s ability to provide an accurate history may be diminished due to cognitive impairment and because PD patients vary regarding their insight into NPS (64). Because PD patients or their families also may not be aware that PD has cognitive manifestations or may attribute deficits or significant decline and disability to normal age-related changes, cognitive deficits may not be appreciated or reported (22).

Assessment across the range of cognitive domains, including executive and visuospatial function, is essential in PD, as the typical “subcortical” pattern of deficits often does not include overt memory impairment as in AD. An important component of the clinical interview is to inquire about forms of cognitive dysfunction that occur in everyday life. First, determine when the patient or informant noted a change in the patient’s cognitive abilities compared with his or her adult baseline, the features of those initial changes (e.g., slowed thinking or trouble recalling names), how they progressed, and whether new difficulties emerged. In order to target interventions that are meaningful to the patient, it is often helpful to clarify the patient’s most bothersome deficits, the impact of those deficits on daily function, including employment and social activities, the effects of any prior interventions, including favorable or adverse effects of medications for PD or otherwise, and whether any compensatory strategies have been helpful.

When cognitive deficits are not mentioned spontaneously, the psychiatric interview of a PD patient should include inquiries about problems in everyday life that relate to cognitive dysfunction. These include gait and balance problems, which, as mentioned, are associated with PD-MCI and PDD. Executive functions, the mental processes related to the planning and execution of complex social and goal-directed behavior, are particularly affected in PD and influence compensatory strategies. As executive functions are context or task dependent, patients may not report troubles, especially if they have already ceased activities that tap executive functions, for example, set-shifting tasks such as paying the bills, taking part in group conversations, or driving. In PD, task and attentional demands during routine functions contribute to clinical symptoms and gait disturbances. For example, navigating through a crowded hallway while speaking with another person requires multitasking that may exceed the diminished cognitive capacity of a PD patient. Hence, patients use the expression that they “can no longer walk and talk at the same time.” The increased cognitive effort and the prospect of embarrassment because of difficulty or failure to complete tasks (or, simply, talk while walking) can cause increased stress and lead to significant anxiety, demoralization, and even social withdrawal because of anticipatory anxiety about encountering difficulties. Executive and attentional deficits further undermine the ability to ignore distractions and respond to the environment (i.e., selective attention and response inhibition), adjust actions and plans in response to new information (i.e., set-shifting and flexibility, such as moving aside or changing one’s path to avoid a collision or a fall), and develop and initiate a series of actions to achieve a desired goal (i.e., strategic planning) (62). This interaction of mood, motor, and cognitive deficits was nicely demonstrated in a study in which PD patients and healthy control subjects performed tasks of increasing motor and cognitive complexity (simple walking, walking and carrying a tray, walking and recalling a memory, and walking while carrying a tray and recalling a memory) (65). In that study, all participants, including controls, had reduced gait speed as the task complexity increased. However, greater reductions in gait speed were associated with cognitive deficits as well as with symptoms of depression, physical fatigue, and imbalance.

Deficits in other cognitive domains are frequently observable during the clinical interview. When patients are asked to describe their daily activities or medication regimens, slowness in the ability to process and respond to information presents as bradyphrenia, which affects speech latency, problem solving, and comprehension of information, especially of complex ideas. Language dysfunction may present as word-finding deficits or diminished spontaneity of speech. Even later in the disease, aphasia, agnosia, and apraxia are not as common or pronounced as they are in AD. Vague or discrepant answers raise the possibility of memory deficits. However, in PD, remote memory deficits present as relatively mild and subtle because memory encoding is relatively preserved until late in dementia, but retrieval of memories is impaired. Thus, in the context of a clinical interview with direct questioning, a patient’s relatively intact recognition memory (the ability to recall information when cued) may mask (or compensate for) impairments in explicit recall. Visuospatial dysfunction involves trouble perceiving, processing, discriminating, and acting on visual information in the environment. This affects navigation in the home, or from the waiting room to the office, or estimation of distances when reaching for objects, all of which can lead to falls, which are predictive of a transition to PDD.

As a minimum standard of care, American Academy of Neurology guidelines recommend formal evaluation of cognitive function initially and then at least annually along with documentation of memory and other cognitive complaints at every visit (66). In clinical settings, the purpose of screening is to minimize false negatives (i.e., to maximize sensitivity so that individuals with the condition are not overlooked). At the bedside, it is ideal for clinicians to use a single tool that is easy to administer, has high sensitivity as a screening tool, and can be used longitudinally to detect conversion to dementia and monitor its progression. A full neuropsychological test battery is often unnecessary to diagnose PD cognitive syndromes but may be used to formulate a treatment plan, clarify relative strengths and vulnerabilities, and serve as a baseline to monitor longitudinal change.

Several brief cognitive screening instruments, validated for use in PD, are recommended for level I assessments to determine whether MDS criteria are met for PD-MCI or PDD (67). For PD-MCI, the MDS Task Force originally recommended the Montreal Cognitive Assessment (MoCA), the Parkinson’s Disease-Cognitive Functional Rating Scale (PD-CRS), the Scales for Outcomes in Parkinson’s Disease-Cognition (SCOPA-COG), and the Mattis Dementia Rating Scale (MDRS) (9). While these tests can result in misclassifications depending on cut-off scores, all perform sufficiently to assess the spectrum of cognitive domains and are sensitive to early stages of cognitive impairment (68, 69). The SCOPA-COG and the PD-CRS (http://www.movementscales.com/pd-crs-en-sp) were developed to tap cognitive domains relevant to PD, in contrast to the focus on memory abilities in generic assessment tools, and are sensitive to diagnoses of PD-MCI as well as PDD (70).

Validation studies of the MDS-recommended criteria for PDD (13) show that the MDS eight-item diagnostic checklist for PDD has high specificity but low sensitivity (71) and may benefit from minor modifications to increase its sensitivity. The Mini-Mental State Examination (MMSE) is recommended as a screening tool for PDD but not for PD-MCI (72). The MMSE has limited ability to detect executive function, and it is not uncommon for PDD patients to actually score in the normal range (≥26). The Clinical Dementia Rating Scale (CDR), developed initially for AD, distinguishes normal cognition, PD-MCI, and PDD when PD-specific cut-offs are used (73). The PDD-Short Screen (PDD-SS), which combines components of the MMSE and the MoCA, has shown high sensitivity and specificity for PDD (74).

PD-specific tools to assess functional abilities were developed to distinguish the contributions of motor and cognitive deficits that are not supported by generic instruments. Two recently developed scales to assess instrumental ADLs are the Parkinson’s Disease–Cognitive Functional Rating Scale (PD-CRFS) (75) and the Brief Penn Daily Activities Questionnaire (PDAQ-15) (76). The Pill Questionnaire, which requires the patient to describe his or her daily medication regimen (which can be complex in PD) is sensitive to detection of PDD but not PD-MCI (67).

Since both PD-MCI and PDD are associated with increased rates of NPS, the full range of phenomena, including depression, anxiety, psychosis, impulse control disorders, apathy, parasomnias, and other sleep and wakefulness disorders should be queried for in the clinical interview or assessed with standardized assessment tools. The Neuropsychiatric Inventory (NPI) (77), administered by a clinician to an informant, assesses 12 behavioral domains commonly disturbed in neuropsychiatric conditions; it is frequently used to screen for the range of disturbances seen in PD. The full range of assessment tools recommended for research and clinical practice in PD, including psychiatric tools, is listed on the National Institute of Neurological Disorders and Stroke Common Data Elements website (https://www.commondataelements.ninds.nih.gov/PD.aspx#tab=Data_Standards).

A number of factors can aggravate cognitive dysfunction in PD and need to be reviewed as part of the standard psychiatric assessment. These include systemic medical and metabolic conditions. Medications commonly used to treat PD—anticholinergics, antidepressants, benzodiazepines, and dopamine agonists—also affect cognition. Furthermore, patients may experience nonmotor fluctuations in cognition and other NPS in relation to the timing of their dopaminergic dosing (78), and symptoms such as bradyphrenia (slowed thinking) in the “off” medication state may affect cognitive performance. Other treatments, such as deep brain stimulation, typically performed for severe motor complications of antiparkinsonian medications, may have both positive and negative effects on cognition. Further workup for delirium and other reversible causes of cognitive impairment should also incorporate laboratory studies including a metabolic panel, a complete blood count, Vitamin B12 and folate levels, and thyroid function tests, as well as testing for HIV, syphilis, and drugs of abuse in at-risk populations. Urinalysis and chest radiographs may be useful if infectious processes are suspected. Neuroimaging, including computed tomography (CT) and magnetic resonance imaging (MRI), is useful to elucidate brain structure and rule out strokes, tumors, hydrocephalus, or other lesions that may contribute to cognitive impairment. Single-photon emission computed tomography (SPECT), including its use for detecting dopamine transporters (DaT), may be useful in cases with diagnostic uncertainty.

AD and vascular dementia (VaD) are the first and second most common causes of dementia, respectively (79). Given the high prevalence of AD and VaD, they should be considered in the differential diagnosis of PDD, but with “common things being common” it is also possible for there to be overlap of PDD with AD or VaD in a single patient. Furthermore, there is convincing evidence from neuropathological studies that patients diagnosed with PDD, AD, or VaD commonly show mixed pathologies upon autopsy (80).

AD accounts for 60%−70% of clinically diagnosed dementia cases (79). AD presents with insidious onset of cortical cognitive deficits and with early short-term memory impairment accompanied by aphasia, agnosia, and apraxia as the disease progresses (in contrast to the subcortical deficits described in PDD). While extrapyramidal symptom risk increases as AD progresses (81), PDD, by definition, presents universally with parkinsonism. And while dementia onset typically begins later in the PD course, PD motor symptoms will have been present for the duration of the disease and may be relatively severe. NPS are common in both PDD and AD, but risk of psychosis is more common and more severe in PDD (82). However, the clinical presentations of PDD and AD become similar in the later or end-stages of either disease, when more global disease pathology leads to an overlap of cortical and subcortical deficits.

VaD, a heterogeneous disorder, is a controversial diagnosis, and agreement on a clear neuropathology is lacking. Subcortical white matter, basal ganglia, or brain stem infarctions can lead to development of a vascular parkinsonism (VP), with motor manifestations and often dementia at presentation. Clinically, patients with VP tend to be older, with shorter duration of parkinsonian symptoms, less evidence of tremor or levodopa responsiveness, and more frequent display of pyramidal signs, urinary incontinence, and pseudobulbar palsy. When a vascular contribution is suspected, structural neuroimaging can help identify a vascular lesion and DaT scanning will show greater asymmetry in tracer uptake in PD than in VP (83).

PSP is a rare (relative to PD) neurodegenerative illness characterized by symmetric parkinsonism with central rigidity, supranuclear palsy with downgaze paresis, and early postural instability and falls. Tremor is not a prominent feature, and psychosis is uncommon compared with PD. Bradyphrenia and apathy are prominent, and most patients advance to frank dementia with a subcortical picture. Typical age of onset is close to that of PD, at around 60 years old. Unlike PD, which has a male preponderance, both sexes are equally affected. PSP tends to be poorly responsive to levodopa, but a trial is warranted as motor deficits respond occasionally (84).

The hallmark symptoms of MSA include autonomic failure with parkinsonism (MSA-P) and/or cerebellar signs (MSA-C), as well as pyramidal symptoms (e.g., Babinski sign). The average age of onset, 55 years, is slightly younger than that for PD, and men and women are affected equally. About two thirds of MSA patients have MSA-P, characterized by parkinsonism with more severe autonomic symptoms (e.g., urinary incontinence, erectile dysfunction, blood pressure instability). Tremor is often present but does not have pill-rolling/rest qualities typical of PD, and levodopa response is limited. MSA-C is characterized by cerebellar signs including nystagmus, dysarthria, and ataxia. Like the Lewy body dementias, MSA is a synucleinopathy, and patients may exhibit REM sleep behavior disorder, which is thought to be associated with this particular pathology (85). Dementia in MSA is less common than in PD but may be equally severe when present (86).

A rare neurodegenerative condition (compared with those described above), CBD is characterized by unilateral parkinsonian rigidity, limb dyspraxia-apraxia (the so-called “alien hand syndrome”), myoclonus, and circumscribed cognitive deficits including executive dysfunction and parietal lobe dysfunction (i.e., cortical sensory and spatial integration deficits) (87). A jerky tremor may be present. Levodopa is rarely effective. Average age of onset is in the mid-60s, with males and females equally affected.

The differential diagnosis for cognitive dysfunction in PD should always include delirium, a cause of global cognitive impairment characterized by disturbances in attention and awareness as well as in other cognitive domains. The main risk factor for delirium is cognitive impairment at baseline, so PD patients with cognitive deficits are at greater risk. However, unlike dementia in PD, delirium has an acute onset and is always related to an underlying physiologic cause (e.g., infection, metabolic disturbances, brain trauma, alcohol withdrawal, and other delirium-inducing medications). The motor and nonmotor fluctuations related to antiparkinsonian medication dosing may cloud appreciation for the waxing and waning course of delirium. Fortunately, treating the underlying cause of delirium can lead to resolution of symptoms, although an episode of delirium can be associated with long-term worsening of cognitive function (88).

Several medications are available to treat cognitive dysfunction in PD, but there are no medications known to delay, stop, or reverse cognitive decline in PD. Recent reviews provide dosing recommendations for medications used to treat PDD (40) and details of clinical trials to date for PD-related cognitive impairment (93). Overall, the clinical trials data suggest modest benefits based on the outcome measures used. Interpretations of these results should note, however, that most medications used to treat PD-MCI and PDD were originally developed for treatment of AD and also used primary outcome measures developed for AD. Interpretations of clinical trial results should also consider the heterogeneity of the cognitive profiles for PD-MCI and PDD, the relative contributions of different etiologies to these profiles, and the specific properties of the drugs studied (94). For example, a PD-MCI patient with prominent executive dysfunction might be expected to benefit from drugs that target the noradrenergic system, versus those targeting the cholinergic system or another neurotransmitter system such as the glutamatergic, serotonergic, etc. Accordingly, treatment decisions in clinical practice should be based on an individual patient’s clinical profile, with targeted clinical trials of various agents that have been studied, whether originally developed for PD, AD, or some other cognitive condition, and on discussions of treatment preferences with the patient and family. From the standpoint of a caregiver, any improvements in cognitive processing and daily functioning that outweigh the negative aspects of an intervention can improve quality of life and reduce the burdens of the disease.

A pronounced cholinergic deficit in PD and PDD, in particular, supports the use of agents augmenting acetylcholine. Several clinical trials are underway or have studied cholinesterase inhibitors (donepezil, rivastigmine, and galantamine) for treatment of PDD and PD-MCI. Rivastigmine, a cholinesterase inhibitor, is the only medication approved by the U.S. Food and Drug Administration (FDA) to treat PDD (93, 95), but there is no evidence that one agent has advantages over another. In general, the clinical trials of cholinesterase inhibitors show small but significant improvements on clinicians’ global impression of change and on discrete cognitive measures with limited side effects (5, 96). NPS symptoms also improved. When prescribed using conventional approaches, the cholinesterase inhibitors are generally well-tolerated. Some patients experience nausea, vomiting, and increased tremor or bradycardia that may limit their use. There is the potential for increased tremor. As with all geriatric patients, eliminate or taper CNS-offending agents first, and then, according to the usual adage, “start low and go slow” in order to identify intolerance to the medication and increase the likelihood of adherence and an adequate trial duration to assess effects. Randomized controlled clinical trials of cholinesterase inhibitors for PD-MCI (97, 98) hint at potential benefits in terms of secondary outcome measures; the absence of effects on primary cognitive outcomes suggests a need for different study designs for PD-MCI trials (93, 94).

Several studies examined memantine, an N-methyl-d-aspartate receptor inhibitor that may act by decreasing glutamatergic transmission and limiting excitotoxicity. Results from placebo-controlled clinical trials including patients with PDD and DLB are varied; the impact of memantine on cognition is equivocal, but it is safe and well-tolerated, and some evidence indicates that it improves global impression of change and behavioral symptoms (91, 96). A recent meta-analysis (99) of six studies of memantine in PDD and DLB showed small overall effect sizes, in part attributed to clinical heterogeneity of samples, treatment response in the different patient groups (DLB vs. PDD), and limitations of outcome measures.

Degeneration of the locus coeruleus in PD and noradrenergic deficits in PD may underlie norepinephrine-mediated cognitive functions in PD, such as executive functions (93). Atomoxetine, a norepinephrine-reuptake inhibitor that is FDA approved to treat ADHD has been studied in PD in several clinical trials. The most recent placebo-controlled trial, which focused on PD-MCI (90), showed improvements in subjective measures of executive functions (attention and impulsivity) but no significant improvements on objective measures of cognitive function. This is consistent with the limited sensitivity of objective measures of executive function, in which patients tend to have better performance in structured settings.

More recent investigations are evaluating the cognitive effects of compounds developed specifically for study in PD patients (93). The SYNAPSE study, a proof of concept study of SYN210, a dual serotonin 5-HT6/5-HT2A antagonist, is underway for PDD patients taking a stable dose of a cholinesterase inhibitor. Theoretically, SYN210 facilitates cognition through its blockade of the 5-HT6 receptor and consequent increases in acetylcholine and glutamate and treats psychosis through 5-HT2a receptor antagonism, which is associated with an antipsychotic effect in PD. Rasagaline, an MAO-B inhibitor developed to treat motor aspects of PD, has been studied in nondemented PD patients, or those with PD-MCI, but has not shown compelling cognitive outcomes. N-acetylcysteine (NAC), an antidote to acetaminophen overdose, also has potential applications to neurodegenerative diseases. A precursor to the antioxidant glutathione, NAC modulates glutamatergic, neurotrophic, and inflammatory pathways by scavenging free radicals and maintaining cellular glutathione status.

A range of behavioral and neuromodulation interventions are being studied as treatments for cognitive impairment in PD. These have mostly addressed nondemented patients with PD and mild cognitive impairments to varying degrees, and treatment effects on instrumental ADLs, including driving and safety, have been evaluated in some studies, but the data are limited overall (93, 100). Studies conducted to date used a variety of methodologies, but there is increased emphasis on the need for scientifically rigorous randomized, controlled trials to determine the efficacy of these approaches. A systematic review of the effectiveness of nonpharmacological and noninvasive therapies in cognitively intact, cognitively impaired, and PDD samples identified only 18 publications from the prior decade that met bias and evidence-level criteria for inclusion in the review (100). These publications reported on neurostimulation, cognitive training, physical activity and exercise, or combinations of cognitive training and physical interventions; only six were randomized controlled trials, and there were no studies on PDD. “Brain training” or cognitive exercises often use computer-based programs or worksheets to target specific cognitive functions such as working memory or attention.

Brain stimulation studies include invasive and noninvasive techniques to modulate brain plasticity. Repetitive transcranial magnet stimulation (rTMS) and transcranial direct current stimulation (tDCS) are noninvasive interventions used in a variety of neurological conditions and mood disorders. Small studies reported variable effects of rTMS and tDCS on cognition when mood was the primary outcome measure. Studies on PD-related cognitive impairment need standardization of stimulation sites and parameters in order to interpret results (100). Currently, ongoing trials in PDD and DLB to investigate the effects of deep brain stimulation in the nucleus basalis of Meynert and the globus pallidus interna on cognitive function are targeting attention, learning, and memory processes subserved by these regions (93).