Progression of Cognitive Decline in Parkinson’s Disease

Literature search

As a basis, a manual search of papers included in the meta-analyses of Muslimovic et al. (2007) [10] was conducted, as the authors had the same research question and included papers published until February 2007. Continuing our search, we used the identical search string used by Muslimovic and co-workers (2007) and reviewed papers from February 2007 to March 2017. The search was conducted in PubMed with the keywords Parkinsons disease or Parkinson’s disease in combination with cognition, or cognitive impairment, or memory, or executive function, or neuropsychological tests, and longitudinal studies, or prognosis, or progression. Titles and abstracts were screened according to the eligibility criteria (see below). Afterwards, the full-text articles of the studies meeting the inclusion criteria were further reviewed for inclusion in the systematic review.

Inclusion criteria

To be included in the review, studies had to meet the following criteria: The articles had to be published in English or German, and a diagnosis of idiopathic PD had to be made according to validated clinical criteria. In contrast with the criteria specified by Muslimovic and collaborators, who included all sizes of PD samples, we decided to include only samples with at least 50 PD patients examined prospectively on at least two occasions with a minimum of one standardized neuropsychological test as a dependent variable (longitudinal design). In cases of several time points of measurement, only the first and the last are reported in this review. Neuropsychological test scores had to be presented for the PD group twice at baseline and at follow-up, or other statistics had to be reported that could be converted to effect sizes. In cases where different papers reported data concerning the same group of patients, the study with the largest sample or with the longest follow-up period was included.

Exclusion criteria

Reports published only in abstract form were excluded. Furthermore, clinical trials were not considered (e.g. deep brain stimulation or other pharmacological or non-pharmacological interventions studies) to examine the course of cognitive decline in PD without systematic effects.

Outcome measures

Tests for overall cognitive functioning and tests assessing specific cognitive domains were reported in the review. In Table 2, tests were assigned into one of six cognitive domains (overall cognition, memory, executive function, attention and working memory, visual-spatial functions, and language) according to the definitions of Litvan and colleagues [11]. For tests that were not mentioned in the definitions of Litvan and co-workers, all authors conducted an expert rating to categorize the tests; in cases of heterogeneity, a consensus was made. In Table 3, assignment of tests to one specific cognitive domain referred to the descriptions of the authors in the reviewed studies if noted.

Calculation of effect sizes

From the data obtained in each study, the effect size Cohen’s d was calculated, indicating the mean difference between baseline and follow-up divided by the mean differences of the standard deviations at baseline and follow-up. A small effect is indicated when |d| ≤0.2, a medium effect is indicated when |d| > 0.2 and ≤ 0.5, and a large effect is indicated when |d| > 0.5 [12].

Inclusion and exclusion process of the studies

In total, 1368 studies were found through the database search and 25 studies by scanning the included studies in the systematic review and meta-analysis of Muslimovic and colleagues [10]. Figure 1 shows a PRISMA diagram illustrating the study selection process. After screening all records, 69 full-text articles were assessed for eligibility. Of these, 34 (49.3% ) were excluded because they did not report pre- and post-values, 16 (23.2% ) were excluded because of a low sample size, in four (5.8% ) studies no calculation of effect sizes was possible, two (2.9% ) studies were excluded because no full-text was available, and in one study (1.4% ), cognition was not the main outcome and therefore insufficient data was available for calculation. Finally, 12 (17.4% ) studies were included in the review. All studies were published in English.

OPEN IN VIEWER

Fig.1

Flow diagram of study selection process.

Demographic and clinical characteristics of included studies

An overview of study characteristics is outlined in Table 1. The baseline sample size varied greatly between studies from 60 [13] up to 1714 [14] PD patients. The mean age of the study sample was above 60 years in 10 out of 12 studies and ranged between 56.1 [15] and 70.0 years [16], with one study not reporting age at all [14].

Mean disease duration at baseline was heterogeneous with a range between 4 months [17] to 17.5 years [18]. Information on PD severity was given in all but three studies [14, 19]. One study reported data for PD-MCI patients at baseline [17].

Follow-up assessment

The mean maximum follow-up times of the studies ranged from one year [17] to seven years [15]. Sample sizes of the included studies at follow-up measurement ranged from 49 patients [19] to 316 patients [14]. Three studies reported effect sizes separately for PD and PDD patients at time of follow-up assessment [15–17].

Effect sizes of assessed cognitive domains

Table 2 displays the effect sizes reported in the different studies according to the conducted neuropsychological tests in the assigned domain. Surprisingly, only one study evaluated patients’ cognitive performance in all specified domains [20]. All other studies only evaluated specific aspects of cognitive worsening over time. Nine studies used screening tools to target overall cognition [13, 22]. Memory function was the most frequently targeted domain (6/12; [13, 23], followed by attention and working memory (5/12; [13, 22] and visual-spatial functions and language (both 5/12 in the same studies; [15, 23]. The domain executive function was only measured in 4 out of 12 studies [13, 23].

Use of neuropsychological tests

The choice of neuropsychological assessment was highly variable between the studies (see Table 3 for details). The most used screening tool for overall cognition was the MMSE in seven studies [13, 22]. The Wisconsin Card Sorting Test was assessed in four studies [13, 23]. Three studies assessed the Judgement of Line Orientation (JOLO) [18, 22] and the Wechsler Memory Scale [13, 23]. All other tests were only used in one or two studies. Furthermore, for nine neuropsychological tests, no information of assignment to a specific cognitive domain was given in the manuscripts.

Effect sizes of neuropsychological tests

Effect sizes of the neuropsychological tests show a large heterogeneity over studies according to duration of follow-up interval. Furthermore, as not every study assessed every domain, it is difficult to outline specific progression trends. However, as an important result, medium to strong effect sizes could only be observed in studies with follow-up intervals of four years or longer [15, 18]. Effect sizes of the MMSE as a measure of global cognitive state were reported in seven studies with ranges for the effect sizes between –0.03 [17] and –1.39 [16], in patients with PDD. Lower score values in the MMSE, which are indicative for greater cognitive decline, were more likely with an increase of the follow-up period, meaning that the longer the follow up period, the greater the decline in the MMSE scores. Only one study showed an exception to this trend [18]; however, this study reported a higher dropout rate than the others.