SEARCH

SEARCH BY CITATION

Keywords:

  • cognitive impairment;
  • dementia;
  • Mini-Mental State Examination;
  • Montreal Cognitive Assessment;
  • Parkinson's disease

Abstract

  1. Top of page
  2. Abstract
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. Acknowledgments
  8. References

Background and Aim

The Montreal Cognitive Assessment (MoCA) is the most suitable measure for screening cognitive impairment in Parkinson's disease (PD). However, the utility of the MoCA has not been documented sufficiently, especially in Asian populations. The present multicenter study included a large number of Japanese patients, and compared Mini-Mental State Examination (MMSE) and MoCA scores in PD patients.

Methods

We carried out a cross-sectional study. Idiopathic PD patients (n = 304; age 70.6 ± 8.3 years (mean ± SD); disease duration 6.6 ± 5.1 years; Hoehn and Yahr stage 2.7 ± 0.7) were registered at 13 participating hospitals, and their clinical/neurological/cognitive features were examined using Japanese versions of the MMSE and MoCA.

Results

The MMSE and MoCA scores were 26.3 ± 3.6 (range 12–30) and 20.9 ± 5.0 (range 5–30), respectively, and showed a strong correlation (R2 = 0.74, P < 0.001) with each other. A MMSE score of <26 was observed in 35% of the participants. A MoCA score of <21 had 89% sensitivity and 83% specificity, comparable with a MMSE score of <26. The two scores were correlated with age (R2 = 0.12 and 0.20, respectively; P < 0.0001), but not with Hoehn and Yahr stage or disease duration.

Conclusions

One-third of the patients had a MMSE score of <26, a diagnostic criterion of PD with dementia. A MoCA score of <21 seemed comparable with a MMSE score of <26. The two scores were correlated with age, rather than severity of motor symptoms, suggesting that cognitive decline might be independent of motor decline in PD.


Introduction

  1. Top of page
  2. Abstract
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. Acknowledgments
  8. References

Cognitive impairment in patients with Parkinson's disease (PD) is now recognized as not being rare.[1] This is a challenge to the concept of PD, which is simply a disorder of the extrapyramidal system. Cognitive impairment impacts patient's quality of life (QOL) and caregiver burden adversely. Because cholinesterase inhibitors, which were developed for the treatment of Alzheimer's disease (AD), have been shown to be effective for improving cognitive functions in PD with dementia (PD-D), neurologists require an efficient tool to screen the cognitive functions of PD patients in the context of clinical care. The pattern of cognitive deficits in PD is different from that of AD.[2] Therefore, whether existing screening tests that were developed mainly for detecting cognitive impairments in AD patients are appropriate for detecting cognitive impairments in PD patients remains an urgent question that needs to be answered, for both the purposes of research and clinical settings.

The Mini-Mental State Examination (MMSE) is the most frequently used screening test of cognitive impairments of AD,[3] and remains a gold standard for the population-based screening of mild cognitive impairment (MCI) and dementia. Thus, the Movement Disorder Society Task Force proposed a MMSE score of <26 as one of the diagnostic criteria of Parkinson's disease (PD) with dementia.[4] Conversely, the Montreal Cognitive Assessment (MoCA),[5] which is a 10-min screening tool developed to detect MCI and which assesses a broader range of cognitive domains compared with the MMSE, has been recommended by the Parkinson Study Group Cognitive/Psychiatric Working Group as the most suitable measure for screening cognitive impairment in PD.[6] The MoCA covers a range of executive functions that are commonly impaired in patients with early PD. In contrast, the MMSE does not cover them. The MoCA has been reported to be more sensitive than the MMSE for the detection of MCI and mild AD in the general population, and a score of <26 was found to be the optimal cut-off point for a diagnosis of cognitive impairment.[5] The MoCA has good test–retest reliability, interrater reliability and convergent validity with a neuropsychological battery in patients with PD.[7] Furthermore, it was reported to be more sensitive to cognitive changes compared with the MMSE in a cross-sectional analysis of 221 PD patients.[8] However, the utility of the MoCA for patients with PD has not been sufficiently documented, especially in Asian populations. Additionally, ethnic differences reported for achievements of cognitive screening test require evaluation of MMSE and MoCA in Asian patients with PD.[9] In the present multicenter study of a large number of Japanese patients, we aimed to analyze MMSE and MoCA scores in PD patients comparatively.

Methods

  1. Top of page
  2. Abstract
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. Acknowledgments
  8. References

Participants

The participants were patients with idiopathic PD (n = 304) who were followed at one of the 13 participating hospitals and registered during July to September 2011. All patients agreed to participate in this study. The diagnosis of PD was made according to the UK Parkinson's Disease Society Brain Bank criteria.[10] All study participants provided informed consent, and the study design was approved by the ethical committees of Tachikawa Hospital and all the other hospitals.

Assessments of clinical and neurological features

This was a cross-sectional study that examined the clinical and neurological features of the participants, including sex, age at the assessment, disease duration of PD, Hoehn and Yahr (H&Y) stage, medications, and severity of depression using the 15-item Geriatric Depression Scale (GDS).[11] Examinations were carried out while the patients were “on” state of daily dopaminergic medication. Their cognitive function was evaluated by administering Japanese versions of both the MMSE and MoCA[12] on the same occasion. The MMSE measures five cognitive domains on a 30-point scale: orientation (10 points), memory (registration and short-term recall; 6 points), attention (serial 7 s; 5 points), language (naming, 3-stage command, repeating, reading and writing sentences; 8 points) and visuospatial function (copying pentagons; 1 point). The MoCA measures six cognitive domains on a 30-point scale: orientation (6 points), memory (short-term recall; 5 points), attention (serial 7 s, target tapping and digit span; 6 points), language (naming and repeating sentences; 5 points), visuospatial functions (copying cubes and drawing a clock; 4 points) and executive function (trail making, phonemic fluency and verbal abstraction; 4 points). Education correction, such as a 1-point correction for the MoCA score of a participants with an education of ≤12 years, was not carried out either for MoCA or MMSE scores, because the present study was a paired comparison between these two scores for each participant.

Statistical analysis

The jmp software version 8.0 (SAS Institute, Tokyo, Japan) was used for statistical analysis. The level of statistical significance in the present study was defined as 0.05. One-way anova and the Tukey–Kramer honestly significant difference test were used to examine differences among the groups. Correlations between the variables and MMSE or MoCA score were expressed by R2 using the least-squares method.

Results

  1. Top of page
  2. Abstract
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. Acknowledgments
  8. References

The demographics and clinical/neurological features of the PD patients are shown in Table 1. Their MoCA score was 20.9 ± 5.0 (mean ± SD; range 5–30) and the MMSE score was 26.3 ± 3.6 (range 12–30). The MoCA score was significantly lower than the MMSE score (P < 0.0001). The MoCA score showed a greater range (5–30) compared with the MMSE score (12–30).

Table 1. Demographics and clinical/neurological features in patients with Parkinson's disease (number, male/female = 304, 166/138)
 Mean ± SDRange
  1. a

    < 0.0001, paired Student's t test (MoCA vs MMSE).

Age (years)70.6 ± 8.342–90
Education duration (years)12.5 ± 2.61–20
Disease duration (years)6.6 ± 5.10.2–30.1
Hoehn and Yahr stage2.7 ± 0.71–5
Duration of levodopa treatment (years)4.1 ± 4.20.0–30.1
Levodopa daily dose (mg/day)367 ± 2280–1300
Total Levodopa equivalent daily dose (mg/day)458 ± 2660–1850
MoCA score20.9 ± 5.0a5–30
MMSE score26.3 ± 3.6a12–30

A second-order polynomial equation fitted best with a curve between MoCA and MMSE scores yielding a R2 = 0.74 (P < 0.001; Fig. 1). In 98% of the participants, the MoCA score was lower than the MMSE score. A MMSE score of <26, which is one of the diagnostic criteria of PD-D, was observed in 35% of the participants. A receiver operating characteristic analysis showed that a MoCA cut-off of <21 was the optimal screening cut-off (89% sensitivity and 83% specificity), comparable with a MMSE cut-off of <26 (Fig. 2). A MoCA score of <21 was observed in 42% of the participants, whereas 35% of the participants showed a MMSE score of <26. Participants with a MMSE score of 26–30 showed a greater range of MoCA score (15–30) compared with that of the MMSE score.

image

Figure 1. Correlation between Montreal Cognitive Assessment (MoCA) and Mini-Mental State Examination (MMSE) scores in Parkinson's disease patients (n = 304). Dotted lines indicate the 95% confidence interval.

Download figure to PowerPoint

image

Figure 2. Receiver operating characteristic analysis on Montreal Cognitive Assessment (MoCA) score <21 and Mini-Mental State Examination (MMSE) score <26 (n = 304).

Download figure to PowerPoint

The 304 patients were grouped into tertiles according to MoCA score; that is, into low MoCA score (5–18), middle MoCA score (19–23) and high MoCA score (24–30) groups (Table 2). The MMSE score of the high MoCA score group was 28.8 ± 1.3, and the mean + SD exceeded the maximum MMSE score (30), which represented a ceiling effect. In fact, 17.4% of the participants achieved a perfect score (30) on the MMSE compared with just 0.3% (only one participant!) of perfect scores achieved in the MoCA. This groupwise comparison showed that the low MoCA score group exhibited older age at assessment than did the middle MoCA score group (P < 0.01), and that the middle MoCA score group showed older age at assessment than did the high MoCA score group (P < 0.01). The largest R2 obtained in correlation analyses between MoCA score and clinical/neurological variables was observed for age at assessment (0.20, P < 0.0001; Fig. 3, upper row). A small but significant decline of MoCA score, at 0.27 points/year, was observed between the ages of 42 and 90 years, although this was not a longitudinal study. These findings show that lower MoCA scores are associated with older age in PD patients. A similar association with MoCA score was not found for disease duration, education duration, GDS score, Hoehn and Yahr stage, levodopa duration, levodopa daily dose or total levodopa equivalent daily dose.

Table 2. Comparison of clinical/neurological features among three groups by MoCA score
Three groups according to MoCA scoreLow, 5–18 (n = 98)Middle, 19–23 (n = 98)High, 24–30 (n = 108)Correlation with MoCA score (R2) (n = 304)
  1. *< 0.01 vs low, **< 0.01 vs middle, Tukey–Kramer HSD test. ***< 0.05, ****< 0.0001, one-way anova.

  2. Data are the mean ± SD.

MoCA score14.9 ± 3.021.3 ± 1.425.9 ± 1.5
MMSE score22.7 ± 3.627.1 ± 2.2*28.8 ± 1.3*,**0.74****
GDS (15 items)7.1 ± 4.05.5 ± 3.5*5.1 ± 3.7*0.05***
Age (years)74.3 ± 6.672.0 ± 7.766.0 ± 8.1*,**0.20****
Education duration (years)11.4 ± 3.212.3 ± 2.713.2 ± 2.3*0.09****
Disease duration (years)6.7 ± 5.17.0 ± 5.96.1 ± 4.30.00
Hoehn and Yahr stage3.0 ± 0.82.7 ± 0.7*2.5 ± 0.6**0.09****
Levodopa duration (years)4.5 ± 4.54.2 ± 4.53.8 ± 3.60.00
Levodopa daily dose (mg/day)418 ± 239366 ± 241322 ± 196*0.04***
Total levodopa equivalent daily dose (mg/day)487 ± 272437 ± 253451 ± 2710.01
image

Figure 3. Correlations between age or disease duration and Montreal Cognitive Assessment (MoCA) or Mini-Mental State Examination (MMSE) scores in Parkinson's disease patients (n = 304). Dotted lines indicate the 95% confidence interval.

Download figure to PowerPoint

Similarly, the 304 patients were grouped into tertiles according to MMSE score; that is, into low MMSE score (12–25), middle MMSE score (26–28) and high MMSE score (29–30) groups (Table 3). The score range of the high MMSE score group (29–30) was smaller than that of the high MoCA score group (24–30), indicating a ceiling effect of MMSE in the less cognitively challenged participants. Similarly to what was observed for the MoCA score, the groupwise comparison and correlation analyses showed that lower MMSE scores are associated best with older age in PD patients (R2 = 0.12, P < 0.0001; Fig. 3, lower row). The decline of MMSE score was 0.15 points/year. This R2 and the rate of decline for MMSE score were smaller than those observed for the MoCA score. Disease duration, or any of the other clinical/neurological variables, was not associated with MMSE score.

Table 3. Comparison of clinical/neurological features among three groups according to MMSE score
Three groups according to MMSE scoreLow, 12–25 (n = 105)Middle, 26–28 (n = 93)High, 29–30 (n = 106)Correlation with MMSE score (R2) (n = 304)
  1. *< 0.01 vs low, **< 0.01 vs middle, Tukey–Kramer HSD test. ***< 0.05, ****< 0.0001, one-way anova. Data are the mean ± SD.

MoCA score15.9 ± 3.921.8 ± 3.1*25.0 ± 2.5*,**0.74****
MMSE score22.2 ± 3.027.1 ± 0.829.5 ± 0.5
GDS (15 items)6.7 ± 4.05.8 ± 3.55.1 ± 3.8*0.02***
Age (years)73.9 ± 7.171.0 ± 7.4*67.1 ± 8.7*,**0.12****
Education duration (years)11.6 ± 3.212.3 ± 2.713.2 ± 2.30.08****
Disease duration (years)6.5 ± 5.07.5 ± 5.85.8 ± 4.5**0.00
Hoehn and Yahr stage2.9 ± 0.82.7 ± 0.72.6 ± 0.6*0.08****
Levodopa duration (years)4.4 ± 4.74.4 ± 3.93.7 ± 3.90.00
Levodopa daily dose (mg/day)393 ± 262401 ± 216312 ± 192*,**0.02***
Total levodopa equivalent daily dose (mg/day)464 ± 287501 ± 278414 ± 260.01

Discussion

  1. Top of page
  2. Abstract
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. Acknowledgments
  8. References

The primary finding of the present study was that most (98%) of the participants showed a lower MoCA score compared with the MMSE score. Although the MoCA score in PD showed a significant correlation with the MMSE score, the latter exhibited a ceiling effect, whereas the former did not. The highest tertile of MMSE score was 29–30 points; that is, a full point or a full point −1 point. In addition, 17% of the participants showed full points on the MMSE. Conversely, the highest tertile of MoCA score was 24–30 points, which represents a wider range compared with the highest tertile of the MMSE score. Thus, the MoCA score lacked a ceiling effect. Two studies using a smaller number of PD patients (n = 88[13]; n = 132[14]) also observed a ceiling effect for the MMSE, but not for the MoCA.

The present study also showed that a MoCA cut-off of <21 was comparable with a MMSE cut-off of <26, with 89% sensitivity and 83% specificity. Considering these good sensitivity and specificity values, a score of MoCA <21 might be used as a diagnostic criterion for PD-D, instead of a MMSE <26.[4] We could not, however, determine the precise sensitivity and specificity of each cut-off value for clinical diagnosis of PD-D, because we did not carry out clinical diagnosis of dementia for each participant by extensive cognitive examinations. Furthermore, participants with a MMSE score of 26–30 showed a wider range of MoCA scores (i.e. 15–30), suggesting that the patients can be stratified further into PD with MCI (PD-MCI) and PD with normal cognition (PD-N), according to MoCA score. Dalrymple-Alford et al.[15] reported screening cut-offs of MoCA score similar to those observed for PD-D patients diagnosed according to MDS Task Force criteria[4] and PD-MCI patients. The optimal MoCA screening cut-off was <21 for PD-D (sensitivity of 81% and specificity of 95%), which is the same value found in the present study. In addition, the optimal MoCA screening cut-off was <26 for PD-MCI (sensitivity of 90% and specificity of 75%). Those authors also examined PD-focused Scales for Outcomes in PD-Cognition (SCOPA-COG),[16] which is a PD-focused instrument used to assess detailed cognitive functions, and found that the optimal SCOPA-COG screening cut-offs were <19 for PD-D (sensitivity of 80% and specificity of 98%) and <31 for PD-MCI (sensitivity of 80% and specificity of 51%). Thus, MoCA has equivalent, or even better, sensitivity and specificity compared with the SCOPA-COG in terms of screening PD-D and PD-MCI among PD patients. The MoCA is a suitably accurate and brief test when screening MCI and dementia in PD. The possible explanations for why the MoCA is more sensitive than the MMSE in detecting cognitive decline in PD patients are as follows. First, the MoCA assesses a broader range of cognitive domains, including vulnerable domains in PD patients, compared with the MMSE, which mainly tests memory and language domains. Second, the MoCA consists of more difficult tasks than the MMSE. Recently, Murakami et al.[17] reported the MMSE and MoCA examined in 50 Japanese PD patients, and showed that in the patients with a MMSE score of >27 (n = 25), the patients with a MoCA score of <25 (n = 13) showed lower scores in the visuospatial, executive and memory domains of MoCA compared with the patients with a MoCA score of ≥25. This finding indicates that MoCA can detect characteristic cognitive impairments of PD-MCI.

Another finding in the present study was that lower MoCA and MMSE scores were associated with older age in PD patients. This association seemed to be stronger for the MoCA compared with the MMSE. The association between MoCA score and age, disease duration or severity of motor symptoms has been analyzed in a few studies that included smaller numbers of PD patients compared with the present study. A longitudinal study of 98 PD patients found no significant change (−0.257 points/year) on the MoCA score over 3 years in PD patients.[8] Conversely, these same participants showed an annual rate of changes of –0.539 points on the MMSE score, and individuals with longer disease duration (>18 years) showed a particularly significant decline in MMSE score (−1.96 points/year). A cross-sectional study of 114 PD patients revealed that PD-D (73.4 ± 6.7 years) and PD-MCI (71.5 ± 5.4 years) patients were older compared with PD-N (64.5 ± 8.4 years) patients.[15] The difference in average age between PD-D and PD-MCI was 1.9 year, which was not statistically significant, but similar to the 2.3-year difference between the low and the middle MoCA score groups in the present study. Another cross-sectional study including 132 PD patients also reported that PD-D and PD-MCI (68.1 ± 9.2 years) patients were older than PD-N (63.9 ± 9.7 years) patients.[14] In contrast, disease duration and severity of motor symptoms (H&Y stage) increased stepwise from PD-N to PD-MCI to PD-D.[14] A cross-sectional study of 100 PD patients with normal MMSE score showed that predictors of cognitive impairment on the MoCA were male sex, older age, lower educational level and greater disease severity (H&Y stage) on univariate analyses, whereas older age was the only predictor on multivariate analysis.[18] The question of which parameters among age or disease duration or disease severity is the strongest predictor of cognitive impairment in PD patients requires further study. As was found in the present study, the lack of association between MoCA or MMSE score and disease duration, severity of motor symptoms of PD, or dose of dopaminergic drugs shows that cognitive decline in PD patients might progress independently of motor decline, reflecting a different time-course of pathological progression in the cerebral cortexes compared with the extrapyramidal system of PD patients.

The present study was not without limitations. First, we did not take into consideration the effects of fluctuating cognitive functions or concomitant diseases on the performance of patients in the tests. However, regarding the comparison between MoCA and MMSE, differential effects are not likely, because these two tests were administered on the same occasion. MoCA and MMSE scores were not correlated with GDS score, showing that they were not affected by depression, which is commonly found in PD patients. Second, the present study examined participants on screening tools, but not standard measures of cognition, and our results were not intended to reflect detailed or precise changes in cognitive functions in PD patients. Third, this was not a longitudinal study, but a cross-sectional study; therefore, changes in MoCA and MMSE scores related to age or disease duration do not directly indicate the natural time-course of an individual patient. Fourth, the further analysis on subcategories of MMSE and MoCA might provide different characteristics between these two examinations, although it was not described in the present study.

In conclusion, one-third of the Japanese PD patients had a MMSE score of <26, which is a diagnostic criterion of PD-D. A MoCA score of <21 seemed to be comparable with a MMSE score of <26 in PD patients. A MoCA score of <21 might be an alternative diagnostic criterion of PD-D, instead of a MMSE score of <26, and PD patients with a MoCA score of 21–30 might be stratified into MCI and normal cognition more precisely compared with PD patients with a MMSE score of 26–30. MMSE and MoCA scores were correlated with age, rather than severity of motor symptoms, suggesting that cognitive decline might be independent of motor decline in PD.

Acknowledgments

  1. Top of page
  2. Abstract
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. Acknowledgments
  8. References

We acknowledge Dr K Obara, Department of Neurology, Japanese Red Cross Mito Hospital; Dr M Kunimoto, Department of Neurology, Saiseikai Yokohamashi Tobu Hospital; Dr H Nozaki, Department of Neurology, Kawasaki Municipal Kawasaki Hospital; and Dr H Hattori, Department of Neurology, Saitama Municipal Hospital for their support.

This study was not sponsored or funded by any industry, government or institution.

Dr. Ohta received speaker's honoraria from Novartis Pharma K.K., Janssen Pharmaceutical K.K., Sanofi-Aventis K.K., Eisai Co., Ltd., Otsuka Pharmaceutical Co., Ltd., Bayer Yakuhin Ltd., Daiichi Sankyo Co., Ltd., Takeda Pharmaceutical Co., Ltd. and Ono Pharmaceutical Co., Ltd. Dr. Osada received speaker's honoraria from Novartis Pharma K.K., and royalties from Nankodo Ltd., Tokyo Igakusha Ltd. and Lifemedicom Ltd. Dr Shinohara serves on consultancy for Schering Plough and Pfizer; received speaker's honoraria from Sanofi-Aventis, Mitsubishi Tanabe Pharma, Otsuka Pharmaceutical, Bayer-Health Care and Daiichi Sankyo; and serves as an Associate Editor of Cerebrovascular Diseases and Editor of Journal of Adult Diseases. Dr Suzuki received speaker's honoraria from Sanofi, Eisai and Kowa Pharmaceutical Co. Ltd. Dr Takahashi received speaker's honoraria from Japan Boehringer-Ingelheim Pharma GmbH & Co., GlaxoSmithKline K.K., Novartis Pharma K.K. and Dainippon Sumitomo Pharma Co. Ltd. Dr Seki received speaker's honoraria from Nippon Boehringer Ingelheim Pharma Co., Otsuka Pharmaceutical Co. and Mitsubishi Tanabe Pharma Co. Dr. Nihei received speaker's honoraria from Eisai Co., Ltd., Takeda Pharmaceutical Co., Ltd., Nippon Boehringer Ingelheim Pharma Co., Ltd. and Otsuka Pharmaceutical Co., Ltd. Dr Isozumi received chairperson's honoraria from Japan Boehringer-Ingelheim Pharma GmbH & Co. Dr Kobari received speaker's honoraria from Boehringer-Ingelheim Japan Inc., Otsuka Pharmaceutical Co., Ltd., Sanofi K.K., Eisai Co., Ltd., Kyowa Hakko Kirin Co., Ltd. and Glaxo SmithKline K.K. Dr Shirai received speaker's honoraria from Sanofi, Otsuka Pharmaceutical Co., Ltd. and Janssen Pharmaceutical K.K.

References

  1. Top of page
  2. Abstract
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. Acknowledgments
  8. References