Abstract
Keywords
INTRODUCTION
Since the 1960s dopaminergic drugs have been successfully used to treat motor symptoms in Parkinson’s disease (PD). Cognitive impairment in PD has also begun to attract interest, and the majority of PD patients show mild cognitive impairment, although such impairment goes undetected by common tests such as the Mini-Mental State Examination (MMSE). Together with others we have reported a correlation between motor and cognitive function in progression of PD from the de novo to the treatment stage using dopaminergic agent [1, 2]. In particular, we reported that gait and postural stability correlate with executive [1, 3] and visuospatial [1] functions. Following these findings, we hypothesized a correlation in dopaminergic drug effects between both motor and cognitive functions. Although dopamine replacement therapy is an established treatment for motor symptoms, its effects on cognitive function in PD patients are uncertain. Some reports show levodopa improved frontal lobe functions such as working memory [4], planning, and set shifting [5]. In previous studies using brain functional imaging with [18F]fluorodopa PET, dopaminergic deficit was observed in caudate nucleus that likely mediates impairment of frontal/executive function [6]. Dopaminergic medication is also known to affect frontal/executive function [7]. Cools et al. [8] show that both too little or too much dopamine impairs working memory, and an optimum level is required. Therefore, dopamine has a defined relationship with executive function in PD patients. Previous studies of the effects of such medication on cognition in PD patients concentrated mainly on frontal/executive function, and the effects on other cognitive domains were not well studied. Furthermore, changes in both motor and cognitive function in relation to dopaminergic medication have not been compared in previous studies. Thus, we examined its effects on some cognitive domains and compared these with motor symptom changes.
PATIENTS AND METHODS
Subjects
Twenty-seven, drug-naïve, de novo PD patients were recruited from outpatient and inpatient groups diagnosed at Showa University Hospital and Showa University East Hospital, Tokyo, Japan. Diagnosis of PD was made using the United Kingdom Parkinson’s Disease Society Brain Bank [9]. Patients were first diagnosed by clinical history and neurological findings before medication, and then confirmed as PD patients after demonstrating improved motor symptoms on administration of dopominergic drugs. Motor symptoms of all de novo patients improved, and the diagnosis was unchanged. All patients had a Mini-Mental State Examination (MMSE) score of 25 or more. No patient was taking an anti-dementia drug, for example an acetylcholinesterase inhibitor or N-methyl-D-aspartic acid (NMDA) receptor antagonist. No patient was administered an anti-cholinergic drug such as trihexyphenidyl. None had depression likely to affect cognitive assessment. None had a history of impulse control disorder. None had a disease except PD that affected motor and cognitive functions. The Ethics Committee of Showa University School of Medicine approved the study, and it was performed according to the Declaration of Helsinki.
Clinical assessment
Dopaminergic drugs (levodopa, dopamine agonist, selegiline) were given to 27 drug-naïve, PD patients and increased to a dose sufficient to obtain optimal improvement in motor symptoms. Drugs were chosen according to patient background and condition. Drugs were given either for single use or in combination. The titer was calculated using Levodopa Equivalent Dose (LED) as described by Tomlinson et al. [10]. Patients were tested prior to, and 4–7 months after drug initiation. Motor function was assessed using the Unified Parkinson’s Disease Rating Scale (UPDRS) [11] as a total score of parts II and III (UPDRS total score), and the subscores of tremor (sum of items 16, 20, and 21), rigidity (item 22), bradykinesia (sum of items 23–26 and 31), gait (sum of items 13–15 and 29), and postural instability (sum of items 27, 28, and 30). Cognitive function was assessed using both the Japanese version of the Montreal Cognitive Assessment (MoCA-J) and the Neurobehavioral Cognitive Status Examination (COGNISTAT-J). MoCA-J includes subdomains of visuospatial function, executive function, attention, language, delayed recall, and orientation. This categorization was presented as the ‘domains based subscore’ in the original MoCA proposal [12]. The tests are shown in Table 1. Together with others we have confirmed the sensitivity of MoCA-J to detect early cognitive impairment in PD patients without apparent dementia [13, 14]. The test-retest reliability of MoCA-J has been demonstrated [12]. Furthermore, it takes only 10 minutes to perform and is easily applied in a clinical situation. We used the raw MoCA-J score for the total score and for the subscores of each domain. On the other hand, COGNISTAT-J is a screening test involving 10 separate subtests of major domains of cognitive function, including orientation, attention, language (comprehension, repetition, and naming), construction, memory, calculation, similarity, and judgement [15]. Information on each subtest is shown in Table 2. The raw score of each subtest in COGNISTAT-J is converted to a standard score, in which average score for normal controls is set at 10, and standard deviations (SD) in healthy controls is set at 1 [16]. A standard score of 8 signifies the mean minus 2 SD and the score of 8 or below is defined as impairment. We previously confirmed the reliability of COGNISTAT-J to detect early cognitive impairment in PD patients without apparent dementia [13]. Lamarre et al. [17] have established the test-retest reliability of COGNISTAT. We used the standard score of each subtest for assessment using COGNISTAT-J. All assessments were done blind and those after initiation of dopaminergic drugs were performed in the on-medication state. The increased score over baseline, in which a positive number indicates improvement and a negative number indicates deterioration, was compared for each motor and cognitive assessment.
The Kolmogorov-Smirnov test for each score improvement showed that the parameters were non-normally distributed, therefore the Spearman’s correlation coefficients were used to determine correlation between motor and cognitive score improvements. The level of significance was set at p < 0.05 (two tailed probability).
RESULTS
Patient backgrounds are shown in Table 3.
Changes in mean motor and cognitive scores after drug treatment are shown in Table 4. All motor assessments, as defined by UPDRS total score and sub-scores for tremor, rigidity, bradykinesia, gait, and postural instability, significantly improved with medication. Cognitive assessments using MoCA-J total score and sub-score of delayed recall significantly improved. However, mean scores of visuospatial, executive, attention, language, and orientation showed no significant change with dopaminergic treatment. For cognitive assessments using COGNISTAT-J, no subtest showed significant improvement. Spearman’s correlation coefficients for each motor and cognitive score improvement are shown in Table 5. The gait score improvement showed a significant positive correlation with scores improvements in the language subdomain of MoCA-J (p < 0.05) and in the language-comprehension subtest (p < 0.05) and the similarity subtest (p < 0.01) of COGNISTAT-J. These scores are shown in Fig. 1, and demonstrate that as gait function improved, so did cognitive function. The postural instability score improvement showed a significant positive correlation with score improvement in the COGNISTAT-J attention subtest (p < 0.05) [Table 5]. Score changes in each UPDRS, MoCA-J, and COGNISTAT-J assessment showed no correlation with LED of the medications used.
Furthermore, we performed a multiple regression analysis to examine whether gait and language improvements were directly related. The analysis used improvement from baseline in MoCA-J language score as a dependent variable, while the independent variables were both patient background (current age, duration of PD, age at PD onset, education and LED) and score improvement from baseline in all motor assessments. We found that the improvement in MoCA-J language subdomain correlated significantly with LED (R2 = 0.62, p < 0.05). In another multiple regression analysis, the improvement from baseline in gait score was used as dependent variable and both patient background (see above) and improvement from baseline in all COGNISTAT-J subtests score were used as independent variables. We found that the improvement in gait score showed significant correlation with improvement in language-comprehension subtests of COGNISTAT-J (R2 = 0.74, p < 0.05).
DISCUSSION
In our present study dopaminergic treatment improved all motor assessment scores and these findings agree with established dopamine replacement therapy for motor symptoms in PD. We demonstrated that certain cognitive assessments (mean score of MoCA-J total score, subscore of delayed recall) also demonstrated improvement with dopaminergic drugs. These results indicate that dopamine deficiency causes both motor and cognitive dysfunction. Although the mean language subdomain score in MoCA-J and the language-comprehension subtest in COGNISTAT-J did not show significant improvement in relation to dopaminergic medication, improvements in these language-related functions correlated significantly with improvement in gait score. This suggests that these language-related functions are also influenced by dopamine. Furthermore, multiple regression analysis showed that score improvement in the language domain of MoCA-J correlated with LED. Thus, some aspects of language function may involve a dopaminergic mechanism. This hypothesis is supported by previous studies that show dopamine modulates language functions such as verbal fluency and semantic activation [18, 19].
With regard to localization, language deficits (e.g. comprehension, verbal fluency, naming, and sentence processing) are linked to fronto-striatal deficits [20, 21]. The basal ganglia, which are the location of the main lesion for motor symptoms in PD, play a role in executing sentence repetition [22]. Thus we believe that motor and such language functions share common problems. The two cognitive tests showing improvement that correlated with gait improvement were: Naming of 3 animals and sentence repetition in MoCA-J [Table 1] and the task of executing a series of one-, two-, and three-step orally presented commands to use objects in front of the patient in COGNISTAT-J [Table 2]. With regard to brain function, these 2 tasks require sensory integration, executive function and working memory. Sentence comprehension is known to be dependent on executive function [23]. On the other hand, walking is a motor skill that utilizes sensory and conscious inputs and executive activities such as estimation, planning, and real-time adjustment [24]. Thus, performing these language-related tasks, and walking, both require executive input. Recent functional imaging studies show that executive function is linked with the basal ganglia, including the caudate nucleus [25] and striatum [26]. Executive functions, such as working memory, contribute to language functions such as sentence processing and comprehension [21]. On the other hand, both gait and executive impairments involve cortical and subcortical damage [27], and gait requires executive function [24]. Furthermore, executive function is modulated by dopamine [5, 28]. Our result showed that improvement in the similarity subtest in COGNISTAT-J, which examines abstract reasoning, one aspect of executive function, correlated with gait improvement with dopaminergic treatment. Therefore, language and gait functions may share a common, dopaminergic, executive mechanism.
In addition, cognitive assessments whose improvement correlated with gait improvement (language in MoCA-J and language comprehension and similarity in COGNISTAT-J) did not show significant improvement in their mean scores. This is because our study included patients whose scores deteriorated in these tests. Figure 1 shows patients with little or no gait improvement had corresponding deterioration in language and executive function. The language function assessed in these tests depends in part on executive function. On the other hand, executive function can become worse when dopaminergic medication is administered [8]. Further, it is possible that LED of the drugs given to these patients were insufficient to improve gait. We believe change in language function correlates with gait change in drug-naïve PD patients while taking dopaminergic medication whether the changes are positive or negative.
The medication administered to our patients was single use, or combinations of either levodopa/carbidopa, levodopa/benserazide, pramipexole, or ropinirole and selegiline [Table 3]. Correlation between gait and language was common to all medication types. However, the dopaminergic effects on cognitive function are thought to differ among drugs because of differences in receptor subtype affinity [29]. In particular, affinity for the dopamine D1 receptor differs greatly according to drug used. For example, pramipexole and ropinirole do not bind well to the D1 receptor. Affinity for the D2 receptor family does not differ as much as that for D1 [30]. Therefore, the correlated change in both cognitive and motor functions seen in our present study may be mediated mainly by D2 receptors. On the other hand, PD patients are likely to experience a placebo effect [31]. And the effects observed in our patients may reflect this. A possible mechanism may be the dopaminergic reward system as follows; physiological dopamine is released into the striatum when patients expect a reward, such as symptom relief [32]. Changes in motor and cognitive function in our patients may thus include the placebo effect, and may also be influenced not only by the drugs administered but by endogenous physiological dopamine.
Mean score of delayed recall subtest in MoCA-J showed significant improvement with dopaminergic drugs. A previous study using brain functional imaging with [18F]fluorodopa PET showed that a dopaminergic deficit in the caudate nucleus mediated impairment in the delayed recall test [33]. Therefore, ability for delayed recall certainly involves dopaminergic function. However, the score improvement in the delayed recall test in this study did not correlate with improvement in any motor assessment. Our view is that the dopaminergic mechanisms linked to motor function and delayed recall are different. On the other hand, cognitive assessment scales are shown to differ in their sensitivity to detect subtle differences in cognitive function [3, 13]. A similar study using alternative motor or cognitive assessments may show such a correlation. On the other hand, mean score of memory subtest of COGNISTAT-J, which also examines delayed recall abilities, did not show improvement. We think this is because of comparative testing difficulty between McCA-J and COGNISTAT-J. The delayed recall test in MoCA-J requires 5 words and partial points are not calculated even though patients could recall the words after hints were given. The memory test in COGNISTAT-J requires only 4 words and partial points are added when patients could recall the words following given hints. Thus the MoCA-J-delayed recall test is more difficult and sensitive to delayed recall disabilities. Therefore it detects subtle change in this ability.
In our previous study, visuospatial function assessed by MoCA-J, a parietal lobe function, correlated with postural stability [1]. But in this study, visuospatial function did not improve with dopaminergic medication, and no improvement correlated with change in motor assessment. These findings suggest that visuospatial function is not dopamine-dependent.
We recognize that our present study has limitations. There is no control group in this study to evaluate practice effects on cognitive assessments. But we believe our observation, that dopaminergic medication affected both motor and cognitive function, is valid for the following reason. Previous studies established that dopaminergic drugs affect cognitive function [4, 8]. Our previous study included patients with cognitive improvement and with cognitive deterioration while on dopaminergic drugs. This shows the effect of dopaminergic drugs on cognitive function can be either positive or negative as shown in previous reports [8]. In another study examining efficacy of 3 months tango exercise treatment [34], the control group on a stable dose of dopaminergic drugs showed no significant change on MoCA score and the practice effect was not seen. Therefore we believe changes in cognitive assessments in our patients are linked to dopaminergic drugs rather than practice effects. We would like to emphasize changes in gait and language/executive test scores from baseline correlate whether the changes are positive or negative.
CONCLUSION
Dopaminergic drugs can improve cognitive function related to frontal/executive impairment in drug-naïve PD patients. Although mean scores of language-related assessment did not improve with dopaminergic drugs, improvement in these scores correlated with gait improvement. Gait and language dysfunctions may share a common pathophysiology linked to dopamine deficiency. Therefore we should assess both motor and cognitive function while caring for and treating PD patients.
CONFLICT OF INTEREST
The authors have no conflict of interest.
Footnotes
ACKNOWLEDGMENTS
This study was supported by Grant-in-aids for Scientific Research on Innovative Areas, “Face Perception and Recognition” (MEXT, 23119720), and “The Science of Mental Time” (MEXT, 25119006), and a Grant-in-Aid for Scientific Research (MEXT, 23591283). This study was also supported in part by the Showa University Medical Foundation.