Mapping Actuarial Criteria For Parkinson’s Disease-Mild Cognitive Impairment Onto Data-Driven Cognitive Phenotypes Part 1

Abstract: 

Prevalence rates for mild cognitive impairment in Parkinson's disease (PD-MCI) remain variable, obscuring the diagnosis' predictive utility of greater dementia risk. A primary factor of this variability is inconsistent operationalization of normative cutoffs for cognitive impairment. 

Parkinson's disease is a common neurodegenerative disease that mainly affects the body's muscles and motor control. Although it is generally believed that Parkinson's disease only affects people's motor ability, in fact, Parkinson's disease also has an impact on memory. However, we should face these effects positively instead of falling into negativity.

Studies have shown that Parkinson's patients have some memory difficulties, and they may experience memory loss or forgetting important information. This is very similar to the cognitive decline of the elderly. However, it is worth noting that these effects are usually only present in early Parkinson's patients, and these effects will gradually decrease over time.

Although Parkinson's patients face some challenges in memory, they still have positive coping strategies. For example, they can record and remind themselves of things they need to remember by using calendars or small notes. Helping and supporting each other is also a very effective method, especially in daily life.

In addition, Parkinson's patients should adopt an active and healthy lifestyle to keep their bodies and brains healthy. This includes maintaining an appropriate amount of exercise, a good nutritious diet, and adequate sleep. In these ways, they can improve their intelligence and cognitive abilities and reduce the negative impact of Parkinson's disease on their lives.

Therefore, although Parkinson's disease may affect memory, we should take a positive attitude to face it and take effective methods to help ourselves do better in this regard. At the same time, we should support and encourage Parkinson's patients and give them more care and support when they face difficulties and challenges in life. It can be seen that we need to improve memory, and Cistanche can significantly improve memory because Cistanche is a traditional Chinese medicine with many unique effects, one of which is to improve memory. The efficacy of Cistanche comes from its various active ingredients, including tannic acid, polysaccharides, flavonoid glycosides, etc. These ingredients can promote brain health in many ways.

Click know 10 ways to improve memory

We aimed to determine which cutoff was optimal for classifying individuals as PD-MCI by comparing classifications against data-driven PD cognitive phenotypes. 

Participants with idiopathic PD (n = 494; mean age 64.7 ± 9) completed comprehensive neuropsychological testing. Cluster analyses (K-means, Hierarchical) identified cognitive phenotypes using domain-specific composites. 

PD-MCI criteria were assessed using separate cutoffs (−1, −1.5, −2 SD) on ≥2 tests in a domain. Cutoffs were compared using PD-MCI prevalence rates, MCI subtype frequencies (single/multi-domain, executive function (EF)/non-EF impairment), and validity against the cluster-derived cognitive phenotypes (using chi-square tests/binary logistic regressions). 

Cluster analyses resulted in similar three-cluster solutions: Cognitively Average (n = 154), Low EF (n = 227), and Prominent EF/Memory Impairment (n = 113). The −1.5 SD cutoff produced the best model of cluster membership (PD-MCI classification accuracy = 87.9%). It resulted in the best alignment between PD-MCI classification and the empirical cognitive profile containing impairments associated with greater dementia risk. 

Similar to previous Alzheimer's work, these findings highlight the utility of comparing empirical and actuarial approaches to establish concurrent validity of cognitive impairment in PD.

Keywords: Parkinson's disease; mild cognitive impairment; movement disorders; cluster analysis; prevalence.

1. Introduction

The experience of Parkinson's disease (PD) encompasses not only the prototypical motor symptoms but also a plethora of non-motor symptoms including cognitive changes. Past research estimates that approximately 40% of people with PD have mild cognitive impairment (PD-MCI) at any given time, and up to 80% of individuals with PD will develop dementia after living with the disease for 15–20 years [1,2]. 

However, the trajectory of cognitive changes can differ among individuals with clear diagnoses of idiopathic PD- with some declining more rapidly than others [3]. 

Therefore, while the endpoint of the trajectory is known for many individuals with PD, the question remains who is most at risk for a more rapid transition to Parkinson's disease dementia (PDD)? 

Two lines of research aim to answer this question. Some studies take an empirical approach by statistically examining neuropsychological data to see what patterns of cognitive performance arise. 

This is often done via the use of cluster analytic techniques, resulting in distinct clusters or cognitive phenotypes. Others take an a priori classification approach, meaning that mild cognitive impairment (MCI) is designated by specific impairment criteria, which are then used to identify patterns of deficits across cognitive tests or domains. 

One way of doing this is via "actuarial classification criteria", defined as using objective, pre-established numerical definitions of impairment, rather than a consensus diagnosis or clinical judgment. Both empirical (cluster analytic) and actuarial/clinical theoretical classification approaches aim to characterize distinct cognitive profiles in PD with the hope of subsequently determining if certain cognitive profiles or subtypes connote a greater risk of developing PDD at faster rates. 

Recently, a comparison of the predictive utility of these two types of approaches (cluster analysis vs. a priori classification) has gained traction among researchers in the MCIAlzheimer's disease (AD) literature. Indeed, recent studies have found that cognitive phenotypes derived from cluster analyses are more strongly correlated with AD biomarkers and are more strongly linked to dementia progression than traditional a priori classification methods [4,5]. 

To date, few studies have compared these two approaches in individuals with Parkinson's disease or addressed some of the psychometric issues inherent when comparing these approaches to each other [6,7]. 

Historically, the cognitive sequelae of Parkinson's disease have been linked to deficits in executive function (e.g., planning, inhibition, problem-solving), processing speed, and working memory and attributed to dopaminergic depletion in frontostriatal networks [8,9]. 

Even so, various studies have found less prevalent, yet still pronounced, deficits in other cognitive domains such as memory [10,11], visuospatial skills [12,13], and semantic language function [14]. 

These varying cognitive sequelae of PD play out in both data-driven and classification approaches. The use of data-driven approaches (e.g., cluster analyses) has resulted in some variability in the patterns of PD cognitive phenotypes across studies. Some reveal phenotypes that primarily differ in the level and breadth of cognitive impairments [15–17]. 

Yet, other studies identify clusters that differ in the "types" of cognitive domains that are impaired [18–20]. For example, Crowley and colleagues [21], in a recent cluster analytic study, with prospectively recruited individuals with PD, identified three cognitive phenotypes-those showing low executive function, those with low episodic memory performance, and those with no deficits relative to age-matched controls. 

There is also variability in the rules of the road used by various a priori classification approaches for identifying "mild cognitive impairment" in individuals with PD [22]. 

Most MCI classification approaches differ in terms of stringency of psychometric criteria such as number of cognitive tests used, use of composite scores, and impairment cutoff criteria. 

In 2012, the Movement Disorders Society (MDS) published consensus criteria for PD-MCI [23]. The Level II "comprehensive" criteria, which requires more extensive neuropsychological testing beyond a cognitive screener, defined impairment as having two or more tests falling 1–2 standard deviations (SD) below the normative mean or demonstrating a relative decline from previous evaluation [23]. 

While a diagnosis of PD-MCI using the MDS criteria is associated with a greater risk of developing PDD [24], even with unified criteria, the prevalence rates of PD-MCI continue to range from 25–65% across studies [25,26]. Such disparate estimates of the portion of individuals with PD-MCI limit this diagnosis' effectiveness at predicting the clinical trajectory of dementia.

In part, the variability in the prevalence of PD-MCI across studies results from methodologic differences (i.e., community vs. clinical sample, sample sizes, which neuropsychological tests that are used). 

However, beyond that, the operationalization of impairment (e.g., use of −1, −1.5, or −2 SDs) is a critical issue. Moreover, variable use of cutoff criteria relates to the notion of "decline" from a previous level, but this hinges on the assumption that test "norms" are inadequate to capture change in certain demographic sectors. 

Currently, results remain mixed over which cutoff criteria best identifies who is at greatest risk for impending dementia [8,24,27]. Previous work comparing empirical approaches to a PD-MCI classification, based on a prespecified impairment cutoff (−1.5 SD), found greater portions of PD-MCI participants in the more broadly impaired or amnestic phenotypes [6,7].

The overall goal of the present study was to address the issue of "cutoff" criteria head-on by comparing clinical classification and data-driven approaches in a large clinical sample of idiopathic PD patients without dementia. 

We specifically wanted to learn which cutoff was optimal for classifying individuals as PD-MCI. This is important as it works towards establishing more consistent prevalence rates of PD-MCI. 

To achieve this goal, the current study first examined the influence of using different SD impairment criteria on PD-MCI prevalence rates and subtypes. Next, we identified data-driven cognitive phenotypes using cluster analysis in this same clinical sample. 

These two approaches enabled us to determine how well the PD-MCI classifications mapped onto the cluster-derived cognitive phenotypes using each of three common SD impairment cutoffs. 

Based on previous literature [28–32], we predicted that the following empirically based phenotypes would emerge from cluster analyses: normatively average cognition, isolated executive function impairment, and broader cognitive impairment across multiple domains, particularly executive function, memory, and visuospatial. 

We predicted that a greater proportion of the PD-MCI cases would be represented in the cluster with broad cognitive impairment due to the involvement of cortical systems underlying lower memory, visuospatial, and executive performance. Impairments in these domains have previously been shown to put individuals with PD at greater risk of developing PDD [33–35]. 

Finally, we planned to determine which impairment cutoff jointly maximized the model's sensitivity and specificity and produced the highest classification accuracy.

For more information:1950477648nn@gmail.com