Abstract:
Background: Parkinson´s Disease (PD) is, after Alzheimers Disease, the second most common neurodegenerative disorder. Loss of cognitive functions occurs frequently in the disease process. Affection of cognition seems to be heterogeneous among PD patients and first alterations may already be present in the beginning of PD . However, patients in the early disease stages only have mild cognitive deficits (PD-MCI) and are less severely affected than patients with Parkinson’s disease dementia (PD-D). It has been shown in many studies that the occurrence of PD-MCI increases the risk of developing PD-D. Since therapeutic early intervention in PD-D patients is already partly available, the identification of an at-risk population and an early diagnosis of PD-D is most important.
To date, varying cut-off values are used to define the significance of the cognitive impairment to make the diagnosis of PD-MCI. Furthermore, subtypes of PD-MCI are classified according to quality and quantity of cognitive impairment. Little is known about the profile of cognition among the group of PD patients and how different neuropsychological tasks constitute separate cognitive entities. Most authors define different cognitive domains by using a predominately theoretical approach which are not validated by clinical data in many cases.
The aim of this study was twofold: first to compare the frequency of PD-MCI and subtypes of PD-MCI by use of varying cut-off values, and second, to analyse how this variation of cut-off values might affect the interpretation of the clinical profile investigated in the PD-MCI group. Moreover, a data-driven approach was used to identify distinct cognitive domains in a large cohort of demented and non-demented PD patients.
Methods: A cohort of 100 PD patients was investigated with a comprehensive neuropsychological test battery. The frequency of PD-MCI subjects and PD-MCI subtypes (i.e. amnestic/non-amnestic) were determined by use of varying healthy population-based cut-off values (1,0 / 1,5 / 2,0 standard deviations below the population mean). Demographic and clinical parameters such as quality of life, daily living function and neuropsychiatric symptoms were assessed to compare the clinical profile of the study groups. Exploratory factor analysis was used to characterize different cognitive domains.
Results: A seven-factor solution accounting for 69,5% of total variance fitted best to the data and revealed high internal consistencies (Cronbach’s alpha coefficients > 0,6).
Varying cut-off values for the definition of PD-MCI were found to affect frequency of PD-MCI subjects (3%-84%) and, maybe more important, lead to a “shift” of proportion of detected PD-MCI subtypes:
75% of the patients showed an amnestic subtype when the 1,0 SD criterion was used, but only 47% when the 2,0 SD criterion was applied. Using the 2,0 SD criterion 53% of patients were classified with non-amnestic PD-MCI but only 25% if the 1,0 SD criterion was applied.
However, variation of cut-off values had no relevant impact on the characterization of the clinical PD-MCI phenotype. PD-MCI and PD-D group differed independently of the cut-off applied in the following parameters: motor deficits, daily living function, neuropsychiatric symptoms,quality of life, Mini-Mental State Examination score and the score of the Parkinson neuropsychometric dementia assessment. Age and duration of the disease differed only when the 1,0 SD and the 1,5 SD criterion was used.
Discussion:
In our sample, frequency of PD-MCI and PD-MCI subtypes were relevantly dependent on the cut-off value used for definition of cognitive impairment. Therefore our results argues for the need of standardized cut-off values for the assessment of cognitive impairment in PD-MCI as well as a standardization of the neuropsychological test battery.