INTRODUCTION

Cognitive aspects of Parkinson’s disease (PD) have been largely ignored for many years. This is partly because of the original description that “the senses and intellect remain intact,” but also because of the short survival time of patients with PD in the past. Thanks to the modern treatment, patients with PD have substantially prolonged survival times, with life expectancies close to their healthy peers, and we now understand that cognitive changes, especially dementia, are largely age- and disease stage–dependent phenomena. In the last few decades, dementia associated with PD (PD-D) has been increasingly recognized and better understood. This chapter is devoted to description of epidemiologic features, clinical characteristics, underlying neurochemical and neuropathologic abnormalities, genetics, diagnosis, and treatment of PD-D.

EPIDEMIOLOGY

Both the prevalence and the incidence of dementia are substantially higher in patients with PD compared to age-matched controls. The prevalence figures for PD-D vary substantially across different studies, probably due to differences in study populations, assessment methods, diagnostic tools, and definition of dementia. Two large, cross-sectional, population-based studies have revealed prevalence figures ranging from 28% to 41% (1,2). A systematic review of 12 carefully selected studies revealed a point prevalence of 24% to 31%. The prevalence of PD-D in the general population aged 65 years and over was calculated to be 0.3% to 0.5%; 3% to 4% of patients with dementia in the general population were estimated to be due to PD-D (3).

Incidence figures usually provide a more reliable estimate of frequency for chronic conditions, as they are relatively free of survival bias. The risk of dementia in PD was reported to be 1.7 to 5.9 times higher compared to controls (4,5). The cumulative incidence of dementia in patients with PD was assessed in several prospective studies with a sizeable number of patients and lengthy follow-up times. In the majority of these studies, patients free of dementia at baseline were included and the occurrence of newly emerging dementia was assessed after a variable length of time. The incidence rate of dementia among patients with PD was reported to be six times higher than in controls in a study with 5-year follow-up (6). In another study during 5-year follow-up, dementia was diagnosed in 62% of patients with PD, who were free of this condition at baseline, compared to 17% of controls (7). In a similar study, the cumulative incidence of dementia was 53% after 14 years (8). In a prospective study conducted in Norway, 8-year cumulative incidence of dementia was 78.2%, with 26% of cases being demented already at baseline (9). The 12-year frequency of dementia in this cohort was 60% at the end of the follow-up period (10). In a community-based study in the United Kingdom, 36% of newly diagnosed patients were found to have some degree of cognitive impairment at the time of diagnosis, while 57% of this cohort developed cognitive deficits within 3.5 (+/−0.7) years. In the same cohort, 21 incident dementia cases were identified over 5.2 years of follow-up, corresponding to a dementia incidence estimate of 38.7 per 1,000 person-years of observation (11). In the prospectively followed-up Sydney cohort, 48% of surviving patients had developed dementia 15 years after the diagnosis (12) and the cumulative incidence had risen to 83% 20 years after the diagnosis (13). In the Rotterdam study, which employed a door-to-door survey, 15% of the PD patients developed dementia compared the 4.9% of the control group, during a mean follow-up time of 4.3 years for the incident and 6.9 years for the prevalent PD group (4).

A number of risk factors have been associated with PD-D, both in cross-sectional studies as associated features and in prospective studies as baseline characteristics predictive of incident dementia (14). Most established risk factors include old age at disease onset or at the time of evaluation; severe motor disability; presence of mild cognitive impairment at the time of diagnosis; and atypical neurologic features such as early autonomic symptoms, symmetrical disease at presentation, and poor response to dopaminergic treatment (Table 12.1).

Both cross-sectional and prospective studies have found advanced age as a prominent risk factor. In a population-based study, the prevalence was zero in patients below the age of 50, but 69% above the age of 80 (1). Similarly, in a group of patients entered in a prospective observational study, the prevalence after 5 years was 62% and 17% in patients whose disease had begun after or before the age of 70, respectively (7). Patients with combined severe motor impairment and old age had a 12-fold increased dementia risk compared to young patients with mild disease demonstrating the combined effect of these risk factors (15).

Low cognitive scores at baseline, early development of confusion, hallucinations or psychosis on dopaminergic medication, axial involvement, including speech impairment and postural imbalance, presence of depression, smoking, and excessive daytime sleepiness (EDS) may also be associated with increased risk of dementia in PD. Rapid eye movement (REM) sleep behavior disorder (RBD) is frequently seen in PD and may be more frequent in patients who eventually develop dementia. In one study, PD patients with RBD had a sixfold higher occurrence of dementia than those without (16). In a recent population-based cohort study, RBD was associated with a 2.2-fold increased risk of developing PD-MCI over 4 years (17). Neuropsychological features such as poor verbal fluency as well as poor performance on verbal memory, and the presence of subtle impairment in executive functions at baseline were significantly associated with incident dementia (18). It was suggested that not all patients with cognitive impairment would progress to dementia: impairment in cognitive tests relying on frontal executive functions were associated with a lower risk of dementia, whereas impairment in those tests tapping more posterior cortical functions was associated with a higher risk (11). In the 10-year follow-up of the CamPaIGN study, 46% of patients developed dementia, baseline predictors were age, motor impairment, “posterior–cortical” cognitive deficits, and MAPT genotype (19). In the Norwegian ParkWest study, in which 182 patients with incident PD were monitored for 3 years, significantly more patients with mild cognitive impairment than without at baseline (27.0% versus 0.7%) progressed to dementia during follow-up, suggesting that mild cognitive impairment at PD diagnosis predicts a highly increased risk for early dementia (20). In one study, reduced cerebrospinal fluid (CSF) β-amyloid levels, an established CSF biomarker in Alzheimer’s disease (AD), was found to be related with cognitive decline in PD patients (21); this finding, however, needs confirmation. In another study, white-matter hyperintensities were associated with cognitive decline in PD patients regardless of age, sex, education status, duration or severity of PD symptoms, and vascular risk factors (22).

In terms of motor phenotype, dementia is associated with postural instability and gait disorder phenotype (PIGD type); tremor-dominant patients have lower risk of developing dementia, unless they convert to PIGD form in their disease course. In addition to constituting risk factors for dementia, PIGD phenotype and longer disease duration were also found to be risk factors for mild cognitive impairment believed to be representing a predementia state in PD (23).

Although several environmental risk factors have been associated with PD, less is known about their role in PD-D. Smoking was associated with a fourfold higher risk for dementia in PD (24). Another study with a mean follow-up of 3.6 ± 2.2 years found a twofold increase in the risk of dementia in PD patients with a history of smoking. In the same study, there was no significant association between head injury, diabetes mellitus, and incident dementia (25). Estrogen replacement therapy was found to be protective in one study (26).

CLINICAL FEATURES

Clinical features of PD-D include cognitive, behavioral, autonomic symptoms, and disturbances of sleep–wake cycle. In typical cases, the profile of dementia can be best described as a dysexecutive syndrome with prominent impairment of attention, executive and visuospatial functions, moderately impaired memory, as well as neuropsychiatric symptoms, including apathy and psychosis (Table 12.2).

COGNITIVE FEATURES

Cognitive impairment without dementia is designated as mild cognitive impairment of PD, or PD-MCI, where the activities of daily living (ADL) are largely preserved. Diagnostic criteria for PD-MCI have recently been described by a Movement Disorder Society Task Force (27). Transition from MCI to dementia is gradual, both in terms of symptom severity as well as temporal course and not all patients with PD-MCI may convert to dementia (11). The profile of cognitive deficits in PD-MCI is variable, the most frequent subtype is single-domain nonamnestic MCI, whereas the most frequent single deficit is memory impairment (28).

Theoretically, patients with PD can be afflicted by all types of etiologies that can cause dementia in the population at large, including other degenerative dementias such as AD. The clinical profile of dementia in such cases will be compatible with the underlying etiology. PD is, however, associated with a highly significant increase in the prevalence and incidence of dementia. This close association strongly suggests that the disease process itself also underlies dementia accompanying PD, which has characteristic clinical features. These features can be best summarized as a dysexecutive syndrome with prominent impairment of attention, visuospatial functions, and accompanying behavioral symptoms (Table 12.3). Several recent studies compared cognitive profiles in patients with PD-D, dementia with Lewy bodies (DLB), and AD. The results confirmed the similarities between PD-D and DLB, and differences between these two and AD. These features are summarized in the following paragraphs.

Impairment of attentional functions and working memory is an early and prominent feature of patients with PD-D. Reaction time and vigilance are impaired, and fluctuating attention is similar to those seen in patients with DLB. Cognitive slowing (bradyphrenia) is more prominent in PD-D patients as compared to patients with AD, and patients with PD-D have longer response durations both in measures of simple and choice reaction time, suggesting that their central processing time is prolonged. Impaired attention is an important determinant of ADL in PD-D; the measure of vigilance and focused attention was the single strongest cognitive predictor of ADL status, matching the strength of the effect of motor function on ADL (29).

Impairment in executive functions (defined as the cognitive ability to plan, organize, and perform goal-directed behavior) is one of the core features of PD-D. Deficits involve tasks requiring concept formation, rule finding, planning, problem-solving, set elaboration, set shifting, and set maintenance. In the Mattis Dementia Rating Scale, PD-D patients had lower initiation, perseverance, and construction, but higher memory subscores compared to patients with AD (30). Executive dysfunction usually emerges early in the course of PD-D and is prominent throughout the disease course. Insight that mental functions are impaired is usually preserved in PD-D, in contrast to patients with AD, where deficits are usually denied. Patients have more difficulties with internally cued behavior, that is, when they have to develop their own internal strategies; their performance improves substantially when external cues are provided.

All types of memory are impaired in PD-D, including working memory, explicit memory (both verbal and visual), and implicit memory, such as procedural learning. The severity and the profile of impairment, however, differ from the amnesia seen in AD. Patients with PD-D have been reported to have deficits in learning new information; however, these deficits are less severe than those seen in patients with AD. PD-D patients have impaired free recall, but their recognition is significantly better, implying that new information was stored but not readily accessed. Accordingly, when structured cues or multiple choices are provided, retrieval is usually facilitated. In one study, memory scores in patients with PD-D were found to be correlated with executive function test scores; it was suggested that memory impairment may be due to difficulties in accessing memory traces, reflecting an impairment in the ability to generate encoding and retrieval strategies due to the executive dysfunction (31). This is in contrast to the “limbic-” or “hippocampal-type” memory disorder with impaired storage and consequent deficits of recall, as well as recognition as the key neuropsychological feature in AD. On the other hand, there are other studies in which impaired storage as well as recognition has been found resembling memory impairment seen in AD (32). In fact, the type (AD- vs. LB-type), severity, and topography of lesions may determine the severity and the profile of memory impairment seen in PD-D patients.

An early and prominent deficit in visual–spatial functions is another characteristic feature of PD-D. Tasks that require visuospatial analysis and orientation seem to be the most affected, suggesting that impairment in visual perception may be the core of the problem. Impairment, especially in visuoperceptual abilities, is more severe compared with AD patients with similar global dementia severity. Visuospatial abilities such as object assembly are more impaired in PD-D, whereas visuospatial memory tasks are worse in AD. Impairment becomes especially evident in more complex tasks that require planning and sequencing of response or self-generation of strategies so that deficits in visuomotor tasks may be due partly to problems in sequential organization of behavior and thus to deficits in executive functions.

Language, especially core language functions, is less impaired in patients with PD-D compared to those with AD. Impaired verbal fluency is the main feature, and it is usually more severe than that seen in patients with AD. Anomia (naming or word-finding difficulties) is frequent in the more advanced phases of PD-D, other deficits include decreased information content of spontaneous speech and impaired comprehension of complex sentences, these occur to a significantly lesser extent than that seen in patients with AD. The anomia typically progresses to a “transcortical-type” aphasia with disease progression. It was suggested that some of the language deficits, such as impaired verbal fluency and word-finding difficulties, may not reflect a true involvement of core language functions, but may rather be related to executive dysfunction, such as impairment of self-generated search strategies.

BEHAVIORAL FEATURES

PD-D is associated with prominent behavioral symptoms and changes in personality. The most common symptoms in PD-D are depression, apathy, anxiety, hallucinations, delusions, and insomnia; at least one neuropsychiatric symptom is present in more than 90% of the patients (33). Hallucinations and delusions commonly follow treatment with dopaminergic agents, and occur disproportionately frequently in patients with dementia. Visual hallucinations are common, and when minor forms such as feeling of presence were included, they were found in 70% of patients with PD as compared to 25% of those with AD (34). Delusional misidentification syndromes (such as home not being own home) are found in 17% of PD-D patients, and are associated with hallucinations, and more severe memory and language deficits (35). Likewise, depressive features were found to be more common in patients with PD than in those with AD. A comparison of patients with PD-D and AD revealed that 83% of those with PD-D, compared to 95% with AD, had at least one psychiatric symptom: Hallucinations were more severe in PD-D, whereas increased psychomotor activity such as aberrant motor behavior, agitation, disinhibition, and irritability were more common in AD. In PD-D, apathy was more common in mild stages, while delusions increased with more severe motor and cognitive dysfunction (36). In a longitudinal study, patients with PD-D were discriminated by the presence of cognitive fluctuations, visual and auditory hallucinations, depression, and sleep disturbance from patients with AD, whereas these features were identical to those observed in DLB patients (37).

MOTOR FEATURES

In PD-D patients, motor symptoms are frequently described as being more symmetrical with predominance of bradykinesia, rigidity, and postural instability. Such features are also correlated with more rapid cognitive decline, whereas tremor dominance has been associated with relative preservation of mental status. In a cross-sectional study, PIGD subtype was overrepresented with 88% in patients with PD-D in contrast to 38% in nondemented patients (38). It was also found that in nearly all dementia cases, dementia was preceded by PIGD-dominant PD, or by a transition from tremor-dominant to PIGD-type PD (39). In the CamPaIGN study, severity of akinetic-rigid type was associated with a higher risk for dementia independent of age (40). PD patients with falls are more likely to have lower Mini Mental State Examination (MMSE) scores than those without falls and also more likely to have dementia. Levodopa (L-dopa) responsiveness may diminish as cognitive impairment emerges, although this assumption is largely based on retrospective clinical data. Mechanisms underlying relative loss of L-dopa response may include development of α-synuclein pathology in striatum and loss of striatal dopamine D₂ and D₃ receptors. On the other hand, this may simply reflect the development or predominance of nondopaminergic axial features, such as postural instability.

AUTONOMIC FEATURES

Autonomic disturbances in PD, including constipation, urinary incontinence, orthostatic and postprandial hypotension resulting in syncope and falls, excessive sweating, reduced heart rate variability predisposing to ventricular arrhythmias, and sexual dysfunction, are frequent and may significantly contribute to disability in patients with PD-D. In a comparative study, cardiovascular autonomic dysfunction was more frequent in patients with PD-D as compared to those with DLB, vascular dementia and AD, PD-D patients demonstrating consistent impairment of both parasympathetic, and sympathetic function tests as compared to controls (41). In a prospective study of autonomic dysfunction, PD-D patients with persistent orthostatic hypotension had a significantly shorter survival compared to those with no or nonpersistent orthostatic hypotension; patients with constipation and/or urinary incontinence, in addition to persistent orthostatic hypotension, had a poorer prognosis compared to those with isolated persistent orthostatic hypotension or no orthostatic hypotension (42).

SLEEP DISTURBANCE

RBD is common in PD-D. In return, half of patients with RBD develop a neurodegenerative disease, mainly PD and DLB over a 10 year of follow-up period. Presence of RBD in patients with PD is also associated with cognitive deficits: in nondemented patients with PD, only those with concomitant RBD had impaired performance on neuropsychological tests, specifically on measures of episodic verbal memory, executive function, visuospatial, and visuoperceptual processing (43). In some patients, RBD can be an early indicator of incipient dementia and may antedate the onset of dementia for many decades. EDS and poor sleep quality are more common in patients with PD, PD-D, and DLB as compared to AD.

NEUROCHEMICAL DEFICITS

In PD-D, degeneration of the subcortical nuclei results in various neurochemical deficits, including cholinergic, dopaminergic, serotoninergic, and noradrenergic losses. The most profound and the most strongly associated one with cognitive impairment is the cholinergic deficit, although the others may also contribute to some aspects of mental dysfunction.

The predominant neurochemical impairment in PD, nigrostriatal dopaminergic deficit, was initially assumed also to underlie the cognitive impairment. However, many with PD, especially young patients, may not show any cognitive impairment despite considerable motor dysfunction. In addition, clinical experience demonstrates that dementia does not improve with L-dopa treatment; L-dopa may even worsen behavioral and cognitive functions, especially in demented patients. These observations suggest that dopaminergic deficit is unlikely to play a major role in PD-D. It was proposed that some cognitive deficits may be due to dopaminergic dysfunction, especially early in the disease process, whereas dopaminergic stimulation may be detrimental in later stages (44). Several experimental findings support this hypothesis. For example, reduced ¹⁸F-fluorodopa uptake in the caudate nucleus and frontal cortex correlated with impairment in neuropsychological tests measuring verbal fluency, working memory, and attention, indicating that ascending dopaminergic projections may be involved in mediating some of the cognitive dysfunction in PD (45). Similarly, using fluorodopa and positron emission tomography (PET), a bilateral impairment in the anterior cingulate area and ventral striatum, as well as in the right caudate nucleus was shown in PD patients with dementia, as compared to those without an impaired mesolimbic and caudate dopaminergic function was suggested to be associated with PD-D (46). In another study, dopamine levels in neocortical areas were found to be decreased to a greater level in demented than in nondemented PD patients (47), suggesting some role for the degeneration of the mesocortical dopaminergic system in cognitive impairment. Neuronal loss was observed in ventral tegmental area as well, which provides dopaminergic input to mesolimbic and prefrontal cortex and medial substantia nigra. In one study, degree of dementia was found to be associated with the neuronal loss in medial substantia nigra (48).

There is substantial evidence that cholinergic deficits due to degeneration of the ascending cholinergic pathways significantly contribute to cognitive impairment and dementia in patients with PD. There is loss of cholinergic neurons in the nucleus basalis of Meynert (nbM) in PD-D (49); the extent of neuronal loss was found to be greater in patients with PD than in those with AD (50); this loss was not associated with AD-type pathology in the cerebral cortex (51). In parallel to these morphologic findings, biochemical deficits in the nbM and in the cerebral cortex were also described: cholinacetyltransferase (ChAT) activity was found to be decreased in the frontal cortex and the nbM of patients with PD, and the decrease was greater in the frontal cortex of PD patients with dementia (52). Extensive reductions of ChAT and acetylcholinesterase (AChE) in all examined cortical areas were described, and ChAT reductions in the temporal neocortex were correlated with the degree of mental impairment, but not with the extent of plaque or tangle formation. In addition in PD, but not in AD, the decrease in neocortical ChAT correlated with the number of neurons in the nbM suggesting that primary degeneration of these cholinergic neurons may be related to declining cognitive function in PD (53). Reductions in ChAT activity were found to be more extensive in the neocortical (especially temporal) region than in the archicortical region. In a comparative study of various pathologic and chemical indices, only presynaptic cholinergic markers (including the number of neurons in the nbM) were related to dementia in PD (54). In a comparative study of patients with AD, DLB and PD mean mid-frontal ChAT activity was found to be markedly reduced in PD and DLB, compared to normal controls and AD: The activity was reduced to almost 20% of controls in DLB and PD, whereas in AD it was reduced to 50% of the activity in normals (55). Imaging studies of cortical cholinergic function using PET revealed similar findings: compared with controls, mean cortical AChE activity was lowest in patients with PD-D (−20%), followed by patients with PD without dementia (−13%) and patients with AD (−9%) (56). These and other studies indicate that the severity of cholinergic deficiency is greater in PD-D than in AD and that these deficits may occur earlier in the clinical course of PD-D. In addition, in contrast to AD, PD-D is also associated with neuronal loss also in the pedunculopontine cholinergic pathways that project to structures such as the thalamus (57). Another important finding is that nicotinic receptor binding was found to be reduced in striatum in patients with PD, suggesting a reduced risk of parkinsonism with cholinergic treatment through stimulation of striatal cholinergic receptors (58). Functional imaging findings also support the association between dementia and cholinergic deficits: Using vesicular acetylcholine transporter ([¹²³I] iodobenzovesamicol [IBVM]) as a marker of cholinergic integrity, single-photon emission computed tomography (SPECT) demonstrated reduction in IBVM binding in parietal and occipital cortices in nondemented PD patients, while demented PD cases have a more extensive decrease in cortical binding, similar to patients with early-onset AD (59). PET studies demonstrated that compared with controls mean cortical AChE activity was lowest in patients with PD-D, followed by patients with PD without dementia and AD patients with equal severity of dementia (60). A subsequent study revealed that the degree of cortical cholinergic deficits correlated particularly well with typical cognitive deficits found in PD-D, for example, impaired performance on tests of attention and executive functions (61).

Deficits in other ascending monoaminergic systems, namely noradrenergic and serotoninergic pathways, were also suggested to cognitive impairment. Locus coeruleus, main source of noradrenergic input to the forebrain and cortex, shows neuronal loss especially in demented and depressed PD patients (62); degeneration in serotonergic dorsal raphe nucleus was also reported and suggested to be related with symptoms of dementia (63). Thus, although the strongest evidence indicates cholinergic deficits as the main biochemical correlate of dementia, it was suggested that impairment in other neurotransmitter systems may also contribute to behavioral and cognitive symptoms. It was suggested that dopaminergic deficits may partly be responsible for dysexecutive syndrome, cholinergic deficits for impairments in memory, attention, and frontal dysfunction; noradrenergic deficits may contribute to impaired attention and serotoninergic deficits to depressive mood (64).

NEUROPATHOLOGY

The site and the type of pathology underlying PD-D have been somewhat controversial. Three types of pathologies have been suggested, including degeneration in subcortical structures, coincident Alzheimer-type pathology, and Lewy body (LB)-type pathology in limbic and cortical areas (14).

Dopaminergic cell loss was suggested to underlie dementia in PD by several groups (63,65), which found that cellular loss in the medial part of the substantia nigra (SN) correlated with dementia, and this correlation was still significant after accounting for amyloid burden (63). A comparison of neuronal loss in the SN of demented and nondemented patients, by others, however, showed no difference and no correlation with dementia (66). Another subcortical structure was proposed as a potential site of pathology. Components of thalamus assigned to the limbic loop were found to be severely affected by PD-related pathology (LBs and Lewy neurites), compared to a mild pathology in other thalamic nuclei; it was suggested that damage to the thalamic components of the limbic loop nuclei may contribute to cognitive, emotional, and autonomic symptoms in patients with PD (57).

Alzheimer-type pathology has been suggested as either the cause of or associated with dementia in PD in a number of studies from the early 1980s onward (67). In fact, senile plaques seem to be present in most cases with advanced dementia and demonstrate high specificity, but they are absent in many cases with cognitive impairment and have low sensitivity (68). An interesting finding is the significant correlation between neocortical LB counts and senile plaques, as well as neurofibrillary tangles, suggesting either common origins for these pathologies or that one may trigger the other (69).

The primary role of AD-type pathology in PD-D has been challenged by more recent studies in which both AD-type and LB-type pathology were evaluated concomitantly. As a result of these studies, cortical and limbic LB-type degeneration has been suggested to be the main pathology underlying dementia in PD, as first proposed by Kosaka et al. (70), although in a recent study, combination of LB-type and AD-type pathology was found to be a better predictive of dementia than the severity of a single pathology (71). The evidence supporting LB-type degeneration to be the predominant pathology has been provided by studies using α-synuclein antibodies to identify LBs, a more sensitive method than the conventional ubiquitin staining. In four studies using this method, and assessing both AD-type and LB-type pathology, a similar conclusion was reached in all: α-Synuclein–positive neocortical or limbic LBs were found to be associated with cognitive impairment, independent of AD-type pathology (68,69,72,73); the presence of LBs in the cortex or limbic areas showed the highest correlation with the occurrence of dementia. Another strong evidence that LB-type pathology alone can induce dementia is provided by genetic studies. Patients with familial PD with a triplication of the α-synuclein gene commonly develop dementia, whereas those with duplication do not, or do so rarely (74–77). This supports a primary and possibly a “dose-dependent” role for synuclein-based pathology in the development of PD-D. The presence of limbic or cortical LBs, however, may not always be associated with dementia in patients with PD (78). Therefore, not only the presence but also the topography and burden of LBs may be crucial for the development of dementia.

It was suggested by Braak et al. that pathologic changes in PD follow an ascending order, sequentially involving cerebral structures, as the disease advances, this temporospatial pattern may provide a plausible explanation for the late emergence of dementia in the course of PD (79,80). They described that the initial lesions in PD occur in certain, susceptible brain stem and anterior olfactory nuclei, less-vulnerable nuclear grays, and cortical areas gradually become affected thereafter. As the disease progresses, cellular loss in subcortical nuclei projecting to areas involved in cognition also increase. Some support for this “bottom-up” hypothesis is also provided by other studies (81): PD patients with relatively long disease duration prior to dementia onset had lower levels of cortical choline acetyltransferase than those with a short disease duration before dementia onset, implying greater loss of ascending cholinergic projections. In contrast, a more “top-down” pathologic process with greater burden of cortical pathology in PD patients with a more malignant disease course and short time before dementia onset has also been described (80,81). In fact, a clinicopathologic study arising from the Sydney cohort suggested that there may be three types of pathologic constellations associated with three different clinical phenotypes, in particular with regard to the temporal course of dementia. Younger patients who developed dementia late in the disease process seem to have a predominance of α-synuclein pathology with little amyloid pathology, whereas those patients with a late age of onset and rapid progression to dementia seem to have mixed α-synuclein and amyloid pathology: such patients which constituted 25% of the sample with dementia had severe neocortical LD disease, more consistent with a “DLB-like” phenotype, but also with a high amyloid burden (80). It is currently unknown which factors determine this clinical and pathologic variability, age of onset is probably one of them. Further complicating are findings from another clinicopathologic study: approximately 55% of subjects with widespread α-synuclein pathology (Braak PD stages five to six) lacked clinical signs of dementia or extrapyramidal signs antemortem (82). It is unclear why these subjects could “tolerate” high levels of synuclein deposition without developing symptoms.

In summary, although the majority of recent studies suggest LB-type degeneration to be the primary pathologic substrate of PD-D, AD-type pathology may also contribute. In fact, AD-type and LB-type pathology do not need to be mutually exclusive, as there are interactions between different protein aggregations. For example, α-synuclein can induce phosphorylation and fibrillization of tau, β-amyloid deposits in cerebral cortex enhance α-synuclein induced damage, presence of β-amyloid may promote the aggregation of α-synuclein, and exacerbating α-synuclein induced neuronal dysfunction (83). Hence protein aggregation may be synergistic, with one protein promoting the aggregation of the other, although the consequences of these protein aggregations in terms of cellular function are not clearly established. In a recent review of studies conducted at the Queen Square Brain Bank, cortical Lewy- and Alzheimer-type pathologies were associated with milestones of poorer prognosis and with nontremor predominance, which in turn, are linked to dementia. The combination of these pathologies was found to be the most robust neuropathologic substrate of PD-related dementia, with cortical Aβ burden determining a faster progression to dementia (84).

GENETIC ASPECTS

Genetics of PD-D include both susceptibility genes as well as occurrence of dementia in monogenic PD. In a community-based study, siblings of PD-D patients were found to have threefold increased risk of history of AD (85). The data on the association of ApoE4 with PD-D have been inconsistent; along with ApoE4, ApoE2 was also suggested to be associated with PD-D (86). In a recent study, autopsy subjects were classified into five groups: dementia with high-level AD without LB pathology (AD group), dementia with LB and high-level AD pathology (LBD-AD group), dementia with LB but no or low levels of AD pathology (pure DLB group), PD-D with no or low levels of AD pathology, and controls. ApoE4 allele frequency was significantly higher in the AD (38.1%), LBD-AD (40.6%), pDLB (31.9%), and PD-D (19.1%) groups compared with the control group, suggesting that ApoE4 allele is a risk factor across the LBD spectrum and occurs at an increased frequency in pDLB relative to PDD; it increases the likelihood of presenting with dementia in the context of a pure synucleinopathy (87). In another pathologic study in patients with a clinical diagnosis of PD, ApoE4 genotype was positively associated with dementia (88).

Variations in the tau (MAPT) gene seem to be a genetic risk factor for PD-D. MAPT H1/H1 haplotype has been associated with a greater rate of cognitive decline and dementia in PD patients (89,90), being a strong predictor of dementia with an odd ratio of 12.1 in the CamPaIGN cohort. In a Spanish case–control study consisting of PD, PD-D, DLB, AD patients and control subjects, H1 haplotype was found to be strongly associated with PD and has a strong influence on the risk of dementia in PD patients but not in other neurodegenerative diseases such as DLB and AD (91).

A significantly higher frequency of heterozygote mutations in the glucocerebrosidase (GBA) gene has been reported in PD and DLB compared with control subjects. GBA mutations may exert a large effect on susceptibility for LB disorders at the individual level, but they are associated with a modest (approximately 3%) population-attributable risk in individuals of European ancestry (92). Up to half of the PD patients, heterozygous for GBA mutations developed cognitive impairment later in their disease in one series (93). Recently, PD patients with GBA mutations were found to be at higher risk of dementia with an odds ratio of 5.8 (94).

There are rare forms of monogenic PD with dementia. Altered expression of or missense mutations in the α-synuclein gene have been linked to early-onset familial PD, sometimes associated with dementia. The most robust association is between triplication and less so with duplication of the α-synuclein gene (75,95). The additional copies of α-synuclein gene result in an excess of the wild-type protein, indicating that the magnitude of increase in the expression of α-synuclein, the major component of LBs, is important for the development of dementia.

Mutations in parkin, PINK1, and DJ-1 genes cause autosomal recessive PD. Dementia rates seem to be lower in patients with PINK1 and DJ-1 mutations, whereas the rate in parkin mutations may be more similar to idiopathic PD patients. There are case reports of cognitive impairment associated with the G2019S LRRK2 mutations where the inheritance pattern is autosomal-dominant (96). Frequency of dementia in monogenic forms of PD does not appear to be higher, and indeed may be lower than in sporadic PD. Relatively younger age of patients, especially in recessive forms may be one reason for this observation (97).

NEUROIMAGING

Structural and functional imaging studies suggest several features associated with PD-D. None of these features, however, are specific or sensitive enough to be relied on as a diagnostic tool in individual patients in routine clinical practice.

In structural imaging studies, the presence of dementia in patients with PD was not found to be associated with any specific pattern of MRI abnormalities (98). There is frontal, occipital, and parietal gray matter loss and an increased rate of whole brain atrophy in PD-D patients compared with control subjects. Atrophy of gray and white matter in PD-D is less prominent compared to DLB patients. A recent volumetric study revealed a relationship between decrease in caudate volume (but not in hippocampus) and cognitive decline (99). There are also studies showing an atrophy of medial temporal lobe structures such as hippocampus, enthorhinal cortex, and amygdala in PD-D; these are, however, not as prominent as those found in AD; there is more severe atrophy of the thalamus and occipital lobe in PD-D. Supratentorial white-matter hyperintensities were also found to be independently associated with cognitive decline in PD-D patients (24).

Reduced fractional anisotrophy (FA) was found in the substantia nigra of the nondemented PD patients with diffusion tensor imaging (DTI) technique. PD-D patients showed significant FA reduction in the bilateral posterior cingulate bundles compared with nondemented PD patients (100); both FA and mean diffusivity values in cingulate and corpus callosum showed significant correlations with cognitive parameters (100,101). In a recent resting state functional MRI study, corticostriatal connectivity was found to be selectively disrupted in PD-D patients (102).

Functional imaging studies have been performed using different methods. Perfusion deficits in SPECT have been consistently reported in patients with PD-D. Although the reports have been variable, the most consistent findings were temporoparietal perfusion deficits in patients with PD-D, as compared to those without dementia, in some studies also involving frontal and occipital areas. In a review of SPECT studies performed in PD patients, Bissessur and coworkers (103) reached a similar conclusion: in PD-D, rCBF assessments often demonstrate frontal hypoperfusion or bilateral temporoparietal deficits. Perfusion deficits in precuneus and inferior lateral parietal regions, areas associated with visual processing, were described in patients with PD-D, whereas patients with AD showed a perfusion deficit in the midline parietal region, in a more anterior and inferior location (104). More severe abnormalities in temporoparietal regions of demented PD patients, compared to those without dementia, were also observed with fluorodeoxyglucose PET (FDG-PET) studies, in some studies also including frontal association cortices, the posterior cingulate cortex, or the visual cortex (105).

In SPECT studies using FP-CIT, a marker of dopamine transporter (which is located on presynaptic dopaminergic terminals), patients with PD, PD-D, and DLB were found to have significant reductions in FP-CIT binding in the caudate and the anterior and posterior putamens compared to patients with AD and controls. Transporter loss in DLB was of similar magnitude to that seen in PD, and the greatest loss in all three areas was seen in patients with PD-D (106). This method can be useful to differentiate patients with PD-D from those with AD, when there is doubt on clinical grounds.

Iodine-123 meta-iodobenzylguanidine (¹²³I-MIBG) is an analog of noradrenaline; it can be used in SPECT imaging to quantify postganglionic sympathetic cardiac innervation. The heart-to-mediastinum ratio (H/M ratio) is lower in PD than in other akinetic-rigid syndromes and normal controls. In complex cases, cardiac ¹²³I-MIBG can be used to distinguish between DLB, PD-D, and AD; patients with PD or DLB have reduced H/M ratios, whereas in AD tracer uptake is normal (107).

PET reveals hypometabolism in parietal and temporal cortex similar to the pattern seen in AD; hypometabolism in visual areas and frontal lobe were also described. PET with N-[11C]-methyl-4-piperidyl acetate (MP4A) can be used to assess cholinergic innervation of the cortex. PD-D patients exhibited a severe cholinergic deficit in various cortical regions, including frontal and temporoparietal cortices (108). In PET studies with the Pittsburgh B compound (11C-PIB), which shows amyloid burden, mean cortical levels of amyloid were increased twofold in AD and by 60% in DLB (109). In PD-D, mean cortical amyloid load was comparable to controls and nondemented PD patients (110). In another study, 83% (10/12) of PDD patients had “normal” PIB uptake, whereas 85% (11/13) of DLB patients had significantly increased amyloid load in one or more cortical regions (111). In a further PIB study, an increased cortical amyloid burden was found in DLB similar to AD, not in PD-D; striatal PIB retention in the DLB and PD-D groups was associated with less-impaired motor function (112). These findings suggest that global cortical amyloid burden is low and infrequent in PD-D compared to DLB and AD.

DIAGNOSIS OF PD-D

The diagnostic process in patients with PD and suspected dementia can be subsumed into two main steps: first, ascertaining the presence of dementia, that is, differentiating it from conditions mimicking dementia; second, the differential diagnosis as to the etiology of dementia, that is, if the dementia is due to the neurodegenerative process associated with PD or if it involves other etiologies such as vascular disease (Table 12.4).

The diagnosis of dementia in patients with PD may be confounded by several factors. First, the apparent impairment in certain cognitive domains, such as language, may be difficult to differentiate from the consequences of motor dysfunction. Second, it may be difficult to decide if impairment in ADL, a necessary criterion for the diagnosis of dementia, is due to cognitive or motor dysfunction. Comorbid conditions such as depression, systemic disorders, or adverse effects of drugs may also mimic symptoms of dementia. Along with a detailed history, especially elucidating the onset, course, pattern, and chronology of the cognitive and behavioral symptoms, it is important to administer appropriate neuropsychological tests and behavioral questionnaires, which include tests sensitive to attentional, visuospatial, memory, and executive dysfunction, which can also differentiate the types of deficits in certain cognitive domains, for example, storage versus retrieval deficit in memory performance.