Your browser does not fully support modern features. Please upgrade for a smoother experience.
Please note this is a comparison between Version 1 by Paulina Gonzalez-Latapi and Version 2 by Vivi Li.
Cognitive impairment is a common non-motor symptom in Parkinson’s Disease (PD) and an important source of patient disability and caregiver burden. The timing, profile and rate of cognitive decline varies widely among individuals with PD and can range from normal cognition to mild cognitive impairment (PD-MCI) and dementia (PDD). Beta-amyloid and tau brain accumulation, oxidative stress and neuroinflammation are reported risk factors for cognitive impairment.
- Parkinson’s disease
- dementia
- mild cognitive impairment
- risk factors
- clinical profile
1. Introduction
The classic view of Parkinson’s disease (PD) has been dominated by the key misconception that it is a neurodegenerative disorder that predominately affects dopamine-producing neurons of the substantia nigra, resulting in motor symptoms. Now, we understand that PD is a heterogeneous multisystem disorder comprising a wide range of motor and non-motor symptoms. These non-motor symptoms are frequent and become increasingly prevalent with advancing disease.
Cognitive decline is among the most common non-motor symptoms. As one of the most severe and disabling non-motor symptoms of PD, cognitive impairment, particularly dementia, has been statistically associated with mortality [1]. In addition, cognitive impairment impairs social function, intensifies caregiver burden and the costs of disease-related medical care [2]. The pathophysiology of cognitive impairment in PD is complex and likely involves a disruption of multiple distinct neural networks occurring over time [3]. The timing, profile and rate of cognitive decline varies widely among individuals with PD. As such, it is important to better comprehend the clinical profile of cognitive impairment in PD as well as the risk factors for developing these cognitive changes. This understanding is crucial clinically for communicating prognosis and managing patients. A better understanding of these factors also allows for optimization of research efforts, particularly clinical trial design.
2. Clinical profile
2.1. Cognitive Profile in Parkinson Disease
The cognitive profile in PD can range from normal cognition to MCI and eventually dementia. Therefore, patients can have normal scores on neuropsychological testing with and without subjective complaints (normal), impairments on neuropsychological testing without (MCI) or with a substantial impact on daily instrumental life activities (dementia). Clinical diagnosis of PD is primarily based on the presence of motor symptoms [4]. Nonetheless, the neurodegenerative process underlying PD starts long before the onset of the motor symptoms [5], and cognitive changes can occur during these earlier phases of neurodegeneration. Accordingly, several reports have shown increased PD risk in people with cognitive changes, and cognitive deficits are included in the prodromal PD criteria [6][7][8][9][6,7,8,9]. In longitudinal studies including individuals without PD at baseline who later developed PD during the follow-up period, subtle cognitive changes occurred within 7 to 9 years before the diagnosis of PD [10][11][10,11], and worse cognitive performance has been associated with a higher probability of prodromal PD [7][12][7,12]. Although the presence of dementia before or close to the onset of parkinsonism is more suggestive of a clinical diagnosis of dementia with Lewy bodies (DLB) [13], these reports show that cognitive deficits at the level of subjective complaints and/or MCI can occur before the onset of PD and have predictive value for PD development.
After the diagnosis of PD, patients can present with subjective complaints which may or may not be accompanied by objective changes [14][15][14,15]. Even if those patients do not have impairment on an objective neuropsychological measure, they should be monitored for cognitive decline as subjective complaints can predict the development of MCI over a two-year follow-up in newly diagnosed PD patients [14]. In terms of the cognitive profile in early PD, the majority of patients have a non-amnestic single-domain cognitive decline with impairments in visuospatial functioning, attention, or executive functioning [16][17][18][16,17,18]. Additionally, impairments on tests assessing language and visuospatial functioning have a higher sensitivity for predicting dementia [19][20][19,20]. However, patients do not always have single-domain cognitive decline, and decline can be present across multiple and/or all those cognitive domains [17]. To explain this heterogeneity for the affected cognitive domains in PD, the “dual syndrome hypothesis” has been proposed [21]. This hypothesis suggests that in PD patients with more fronto-striatal network dysfunction, modulated by dopamine, attention/working memory and executive functions are predominantly affected, whereas in patients with more posterior cortical degeneration, memory, language, and visuospatial functioning are predominantly affected related to greater cholinergic loss. Therefore, comprehensive evaluation of individual cognitive domains is crucial for estimating the underlying pathophysiology and subsequently lead to development of effective treatments. Admittedly, the cognitive profile can be hard to discern if the patient’s cognitive decline is too advanced with severe impairment across all domains at the time of testing. Sex-specific patterns for PD-associated cognitive alterations have also been described, with verbal fluency and lack of facial emotions being more prevalent in males and a reduction in visuospatial cognition appearing more frequently in females [22].
As the neurodegenerative progression continues in PD, cognitive changes also advance; the frequency of patients with MCI increases [23], and those with MCI are at a higher risk for the development of dementia [24]. Although dementia may seem almost inevitable after 20 years following the diagnosis of PD, with dementia prevalence reaching 80% [25], the rate of progression to dementia is highly variable. The additional high variability for the affected cognitive domains and the severity of the cognitive decline in PD makes it challenging to make predictions about the progression and underlying pathologies. Although there are currently no pharmacological treatments to delay or prevent cognitive decline in PD, from a research perspective it is important to be able to detect the patients at risk before the onset of severe cognitive decline and to monitor the progression to determine the rate of decline. Detecting such patients and progression rates will enable the development of effective strategies to prevent or delay progression to dementia in PD.
2.2. Diagnostic Criteria
To promote uniformity amongst those who work on cognition in PD, clinical criteria for stages of cognitive decline (MCI and dementia) have been previously defined by expert panels for use in clinical and research settings [26][27][26,27]. These stages are based on a combination of patient and caregiver reports, clinician observations, and neuropsychological assessments of the patients. The diagnosis of PD-MCI is based on (1) the individual having a diagnosis of PD, (2) a gradual cognitive decline reported by the patient, informant, or the clinician, (3) cognitive decline on comprehensive neuropsychological testing or a scale of global cognitive abilities validated in PD and (4) cognitive decline not sufficient to interfere significantly with functional independence [27]. PD-MCI represents the prodromal stage of PD dementia (PDD) and can provide a window to prevent or delay the progression to PDD. The PD-MCI diagnostic criteria have two levels with Level I implying brief assessment and Level II implying a more comprehensive assessment with at least two tests for each of the five cognitive domains (attention, executive function, visuospatial functioning, language, memory) [27]. If Level II testing can be performed, the patient needs to show impairments on two tests in one cognitive domain or one impaired test in two different cognitive domains. Impairment is defined by performance 1-2 standard deviations (SDs) below appropriate norms, decline from prior neuropsychological testing, and decline from estimated premorbid levels [27]. Both Level I and II criteria are valid predictors of PDD [24][28][24,28], however, detailed neuropsychological evaluations can be more sensitive to detect cognitive decline [29]. Given the heterogeneous nature of the cognitive decline in PD, detailed neuropsychological evaluations assessing various cognitive domains can provide a more comprehensive view of the cognitive state.
The diagnosis of PDD is based on (1) the individual having a diagnosis of PD, (2) a slowly progressive cognitive decline that developed after an established PD diagnosis, and (3) impairment in more than one cognitive domain, which represents a decline from premorbid level and is severe enough to impair instrumental daily life activities [26]. Behavioral symptoms including hallucinations, delusions, apathy, depression, anxiety, personality changes, and excessive daytime sleepiness frequently co-occur with PDD but are not required for the diagnosis of PDD [26]. Nevertheless, these behavioral symptoms impact the quality of life of both the patient and the caregiver [30] and need to be addressed by the clinician.
For both PD-MCI and PDD, the presence of other conditions which may explain the cognitive impairment (e.g., delirium, stroke, trauma, metabolic abnormalities) and other PD-related comorbidities (e.g., motor impairment, severe depression, anxiety, excessive daytime sleepiness, psychosis) that may significantly influence neuropsychological testing need to be taken into account [26][27][26,27]. As global cognitive screening tools, The Montreal Cognitive Assessment, the Mattis Dementia Rating Scale Second Edition, and the Parkinson’s Disease-Cognitive Rating Scale were recommended for use in PD by the MDS Rating Scales Review Committee [31]. These three scales adequately represent relevant cognitive domains and were found to be reliable, valid and sensitive to change in PD. In terms of detailed neuropsychological testing, there is no optimal battery to detect cognitive changes in PD for the time being. Different neuropsychological tests and norms are used across centers and for different research studies and for clinical assessment. However, preliminary work has been performed to guide future efforts to establish an efficient neuropsychological battery. Out of 19 MDS-recommended scales representing five cognitive domains, the two best performing tests for each cognitive domain were determined (Table 1) with 2 SDs below norms on these tests suggested to be highly sensitive and specific for PD-MCI diagnosis [32]. However, a multi-site study by the MDS Study Group for Validation of MCI in PD, which included 2908 PD patients and 1247 healthy controls, did not provide strong support for this threshold [33]. Different sites were using different normative scores for neuropsychological tests, which led to a high variability in cognitive performance across study sites. This between-site variability was reduced when comparisons involved healthy controls that were matched with the patients for education, language, test version, test procedures, and source population within the studies. The study group outlined the need to unify existing guidelines for test procedures, including matched healthy controls, and developing more extensive normative data for tests to account for age, education, sex, ethnicity and race to allow comparability of findings in research. In a clinical setting, repeated neuropsychological testing to determine the change over time and consideration of various factors while interpreting the results including the patient’s age, sex, education, motor and non-motor symptom severity, as well as functioning in instrumental daily life activities will help reach a more reliable conclusion about their cognitive status.
| Attention/ Working Memory | Trail Making Test-Part A |
|---|---|
| Symbol Digit Modalities Test | |
| Executive function | Trail Making Test-Part B |
| Clock Drawing Test | |
| Visuospatial functioning | Judgment of Line Orientation |
| Intersecting pentagons | |
| Language | Boston Naming Test |
| Animal naming | |
| Memory | Free and Cued Selective Reminding Test |
| Figural Memory |
2.3. Prevalence and Incidence of Cognitive Impairment in Sporadic and Genetic Forms of PD
The reported prevalence of PD-MCI varies widely, which has been attributed to the use of differing criteria in many studies. Studies using the Movement Disorder Society Task Force Level II criteria, enrolling early stage (H&Y Stage 2) and de novo PD (within 2 years of symptom onset), report prevalence varying from 20% to 41% [23][34][23,34]. A population-based cohort of early-stage PD found that in those with normal cognition at baseline, the cumulative incidence of PD-MCI was 9.9% after 1 year and 28.9% after five years of follow-up [34]. A recent meta-analysis reported pooled PD-MCI prevalence to be 40% on a total sample of 7053 PD patients. As for type of PD-MCI, prevalence of multiple-domain was 31% [35]. Interestingly, only disease progression (assessed by H&Y scale) significantly moderated the prevalence rate of PD-MCI. Other variables including older age, longer disease duration, depression and apathy did not moderate prevalence estimates but were significantly different between PD with and without MCI. Interestingly, lower education levels also characterized PD-MCI [35].
About 10% of PD patients develop dementia every year, which is four to six times higher than that of non-PD patients [36], and the life-long prevalence of PDD is almost 80% [25][28][37][38][25,28,37,38]. Patients with PD-MCI are also more likely to eventually develop dementia [25] and to develop it earlier [39][40][39,40] than PD patients without cognitive impairment.
Subjective cognitive complaints are also associated with progression to dementia. A recent study showed that in patients with PD who expressed subjective cognitive complaints at baseline, a third had progressed to PD-MCI and PDD within 7.5 years [41][42][41,42].
Most cases of PD are sporadic, although a significant minority have strong genetic determinants (termed ‘monogenic PD’) and a proportion of individuals carry one or more genetic variants known to increase PD risk. Clinically, there are differences between the sporadic and monogenic forms of PD, and this includes the risk of cognitive decline. Some of the most studied genes include alpha-synuclein (SNCA), leucine-rich repeat kinase 2 (LRRK2), as well as the glucocerebrosidase (GBA) gene. Importantly, genetic risk variants may also have an influence on the incidence and rate of progression of cognitive impairment in sporadic PD and this topic is discussed later in our review.
SNCA mutations cause autosomal-dominant PD and are associated with severe effects on cognition. A recent review using data from the MDSGene database (www.mdsgene.org) concluded that 70% of SNCA carriers were affected by cognitive decline. SNCA mutation carriers also show a shorter time from emergence of motor symptoms to the onset of dementia and a considerably younger age at onset of dementia than sporadic PD. Even more, there appears to be a dosage effect, with 88% of SNCA triplication carriers showing cognitive changes, compared to 68% of duplication carriers [43]. Aside from disease-causing mutations associated with autosomal dominant inheritance patterns, a variant in intron 4 of SNCA has also been associated with a higher risk of PDD [44]. This raises the possibility that SNCA-related mechanisms are associated with cognition in sporadic PD, warranting additional studies in larger and also more ethnically diverse cohorts.
Mutations in the GBA gene are reported to be associated with faster rates of motor and cognitive progression compared with sporadic PD. Zhang et al (2015) showed that GBA mutations were associated with a 3.2-fold increased risk of dementia or cognitive impairment [45]. More recently, Creese et al (2017) confirmed these findings, showing a 2.4-fold increase in risk of cognitive impairment in PD subjects with GBA mutations [46]. Furthermore, GBA gene mutations are also associated with more rapid cognitive decline [47][48][49][47,48,49]. This has been described for both “severe” (L444P) and “mild” (N370S) GBA mutations [47][49][50][47,49,50]. Nonetheless, data from the Parkinson’s Progression Markers Initiative (PPMI) shows no difference in MoCA score or detailed neurocognitive battery between GBA PD and sporadic PD patients (n = 80 and 361, respectively) [51]; this may be due to shorter disease duration (3 years) compared to other cohorts. These data have interesting implications for the timeline of the dissociation of this phenotype from sporadic PD and suggest that the cognitive differences emerge predominantly at later disease stages. Since the PPMI analysis was cross-sectional, longitudinal data will be useful for defining the slope of progression and comparing it with sporadic PD.
Mutations in LRRK2 are the most common cause of late-onset autosomal-dominant PD, with LRRK2 Gly2019Ser as the predominant variant. LRRK2 G2019S mutations carriers are reported to have less non-motor disabilities and a slower rate of PD progression, compared to sporadic PD [52]. Recently, Simuni et al (2020) described no meaningful difference in cognitive performance between LRRK2 and sporadic PD subjects [51]. This is consistent with previous reports from other studies [52][53][52,53].
The effect of less frequent variants is more difficult to ascertain due to the small number of cases available for study, as well as inconsistent use of diagnostic criteria for PD-MCI and PDD. A systematic review summarizing cognitive and psychiatric manifestations of genetically determined PD, found that patients with PINK1 (n = 24) mutations had the highest incidence of cognitive decline, followed by SNCA (n = 151) and DJ1 (n = 8) mutation carriers. VPS35 (n = 24) and LRRK2 (n = 1625) had the lowest rates of cognitive impairment [54]. An earlier review that used MDSGene data and included a total of 1127 patients with causative Parkin (n = 958), PINK1 (n = 139) or DJ1 (n = 30) mutations suggests that cognitive impairment was present in only a small percentage of subjects with PINK1 associated PD [55]. Cognitive changes also appear to be rare in Parkin mutation carriers [56].
Further work in longitudinal cohorts of genetic PD in both pre-symptomatic and symptomatic individuals is needed to better elucidate the nature of the observed associations. Additionally, there are no available studies that specifically examine the association between neural network dysfunction, neuropathological findings and domain impairment in genetic PD.
3. Protective Factors
3.1. Exercise
Exercise is another important lifestyle consideration. The general health benefits of exercise are well-known and there is reliable evidence for significant benefits on cognition in healthy older adults, especially for executive control processes [57][161]. Since executive dysfunction is a key finding in PD-MCI and PDD, exercise may be a relevant factor preventing or modifying cognitive impairment in PD. In the Parkinson’s Environment and Gene study, higher lifetime average physical activity was associated with less decline on the MMSE [58][138]. A small study showed that patients with PD who participated in a multimodal exercise program had statistically significant improvement in executive function, measured by the Wisconsin Card Sorting Test, compared to a group of patients who kept to their usual daily routine [59][162]. A randomized single-blind, pilot trial of patients with mild to moderate PD also showed that treadmill training was associated with significantly improved global executive function, compared to a control group [60][163]. Resistance exercise may also benefit cognition in PD. After 24 months of twice-weekly progressive resistance exercise, adults with PD improved on working memory, inhibition, and attention [61][164]. Another group used a combined aerobic and strength-training regimen, with cognitive improvement following exercise training [62][165]. The above suggest that exercise could be a feasible intervention to improve executive function in PD. Interestingly, some studies suggest that high intensity exercise may lead to an increase in reactive oxygen species, which theoretically could lead to increased brain oxidative stress. This effect has not been observed with moderate intensity exercise [63][64][65][166,167,168]. Most of these studies have been conducted in murine models and their generalizability to humans remains to be seen. Nonetheless, this underscores the importance of determining the frequency, intensity, type, and timing of exercise that may be most effective in PD. Additionally, the effect of exercise on preventing cognitive impairment has not been fully elucidated and has not, to our knowledge, been studied in randomized trials. Future randomized controls trials in larger cohorts, as well as in prodromal PD groups are needed.
3.2. Diet
Diet is a complex lifestyle factor to study since it represents a combination of many individual food components as well as vitamin and mineral supplementation. Foods and other dietary components have received significant attention as potentially important environmental factors related to the pathogenesis of and protection against PD. There is also interest in describing which nutrients and food may impact the development of cognitive changes in PD, but the body of literature in this area is small. The Mediterranean diet has been of interest given its reported association with cognitive health in the general population and in other neurodegenerative disorders, such as AD [66][67][169,170]. In one study, patients with PD were randomly assigned to either an individually designed Mediterranean diet or usual dietary intake. After 10 weeks, those in the intervention group had significantly improved MoCA scores in the executive function, language, attention, concentration and active memory tasks, while they were unchanged in the control group. There was no significant change in visual-spatial ability, memory learning or navigation tasks [68][171]. The potential mechanisms regarding these effects are unknown, although a reduction in vascular risk factors, particularly hypertension and dyslipidemia [69][172], beneficial effect on oxidative stress and an anti-inflammatory effect may be playing a role.
In some studies, 25-hydroxyvitamin D 25(OH)D has been found to be lower in PD patients compared to controls [70][173] and low vitamin D levels might contribute to the development of dementia in the general population [71][174]. However, the relationship between 25(OH)D and clinical features of PD is inconsistent. In one study, PD patients had significantly lower 25(OH)D and MMSE scores when compared to healthy controls. Among PD patients, those with vitamin D deficiency had a greater impairment of global cognitive function [72][175]. Another study following neuropsychiatric function in 286 PD patients found that higher 25(OH)D levels were associated with better semantic fluency and memory [73][176]. Nonetheless, a different prospective observational study compared newly diagnosed PD with age-matched controls; while 25(OH)D levels were lower in PD patients both at baseline and at 36 months, there was no association with cognitive measures [74][177]. More studies are needed to ascertain the role of 25(OH)D in cognitive changes in PD, as well as the potential role of 25-(OH)D supplementation for possible benefits in cognitive function.
Several longitudinal studies support an inverse association between caffeine consumption and decreased memory impairment associated with aging, as well as reduced risk of dementia and AD [75][76][178,179]. Coffee has been inversely associated with PD risk [77][180]. A cross-sectional study involving treatment-naïve PD patients showed that coffee drinking was significantly associated with a reduced severity in the cognitive domain of the non-motor symptom assessment scale [78][181]. In the Parkinson’s Environment and Gene study PD patients also reported on their average caffeinated beverage consumption and ever coffee consumption was associated with less decline in MMSE [58][138]. In a cohort of newly diagnosed, treatment-naïve PD patients, there was an association between the number of cups of coffee per day and MMSE and MoCA scores, although statistical significance was lost after adjusting for confounders [79][182]. The Café-PD study was a multicenter parallel-group controlled trial where patients were randomized to caffeine or placebo. Interestingly, MoCA scores at 18 months were worse in the caffeine group; nonetheless, this was mostly accounted for by improvement in the placebo group, and this was an exploratory finding, which, due to the multiple comparisons in the study, may have been due to chance [80][183].