Mini-Mental State Examination in Elderly Chinese: A Population-Based Normative Study

Abstract

Background: Chinese nationwide norms of the Mini-Mental State Examination (MMSE) have not been established despite its wide use.

Objective: To obtain norms for the MMSE based on age, gender, education, and rural or urban residences and to determine the optimal cut-off points of the MMSE in elderly Chinese.

Methods: A cross-sectional study was conducted in Chinese community residents aged 65 years or over selected by cluster random sampling. The MMSE was administered to 9,629 subjects (7,110 cognitively normal, 2,024 with mild cognitive impairment, and 495 with dementia). The demographic influences on MMSE scores were investigated and the norms were established considering those factors. Receiver operating characteristic (ROC) analysis was used to determine the optimal cut-off points.

Results: Years of education (standardized β= 0.399), rural residence (standardized β= –0.261), age (standardized β= –0.198), and being female (standardized β= –0.101) had significant effects on MMSE scores (p < 0.001). Accordingly, we presented the demographic-stratified normative data for the MMSE. The optimal cut-off points for dementia screening were 16/17 for illiterate (sensitivity 87.6% and specificity 80.8%), 19/20 for individuals with 1–6 years of education (sensitivity 93.6% and specificity 92.7%), and 23/24 for individuals with 7 or more years of education (sensitivity 94.3% and specificity 94.3%).

Conclusion: We provide the age-, gender-, education-, and residence-specific reference norms for the MMSE derived from an investigation of a large-scale, multicenter, nationwide representative Chinese elderly population. It could be of great improvement for the use of the MMSE in dementia screening in Chinese elderly population.

Keywords

Dementia elderly mild cognitive impairment Mini-Mental State Examination

INTRODUCTION

Since its introduction in 1975, the Mini-Mental State Examination (MMSE) has been modified and translated into several languages [1]. Previous studies have discovered that MMSE scores would be influenced by demographic variables, such as age, gender, and education [2 –4]. Additionally, residence (rural or urban) may also act as an influence factor for the MMSE performance, especially in developing countries where a majority of population lives in rural areas [5 –7]. To alleviate these influences, demographic-stratified reference values have been developed in several countries [2 , 8–10].

The MMSE was also translated into a Chinese version. Despite its wide use, few efforts have been made to investigate the normative data of the MMSE in elderly Chinese. Furthermore, the few reported norms could not be generalized to all populations nationwide because they were not representative: In two studies [5, 11], the norms were derived from more developed regions of Beijing and Shanghai where more residences were better educated than the general population; in another study using a modified version of the MMSE, norms were derived from neurology outpatients rather than the general population, and the sample size was relatively inadequate (370 subjects) [6]. These unrepresentative norms severely limit the clinical value of the MMSE. Consequently, powerful norms for the MMSE should derive from nationwide populations of different regions with a wide range of education levels and ages.

Therefore, we conducted this large-scale, multicenter, cross-sectional study aimed to estimate the influence of the demographic variables on MMSE performance and the demographic-stratified normative data of the MMSE, and to determine the optimal cut-off points of the MMSE for elderly Chinese.

SUBJECTS AND METHODS

Study design and samples

This study is a part of China Cognition and Aging Study (China Coast) that is a national study on the mild cognitive impairment (MCI) and dementia in hospital and community populations. A multistage, stratified, cluster sampling design was used in our study. At the first stage, five regional centers were chosen across China. China is a vast country with significant disparity. Northern China has marked differences in geography, climate, diet, and culture from the south. Eastern China is more developed economically than the west of China. Accordingly, we chose five representative regional centers across China (Changchun for Northeast China, Beijing for North China, Zhengzhou for Central China, Guangzhou for South China, and Guiyang for Southwest China). At the second stage, using random-number tables, 10 urban districts and 12 rural counties were randomly selected from these centers. At the third stage, 30 urban communities and 45 rural villages were randomly sampled within the selected districts and counties.

This study was conducted among elderly Han Chinese 65 years of age or older. Subjects were required to have lived in the target community for at least 1 year prior to the survey date. Subjects who were aged less than 65 years, refused to participate, provided incomplete data, or had inadequate hearing or vision, were excluded. In this study, 893 urban and 690 rural residents from Changchun, 1,419 urban and 702 rural residents from Beijing, 1,423 urban and 1,141 rural residents from Zhengzhou, 970 urban and 806 rural residents from Guangzhou, and 813 urban and 772 rural residents from Guiyang were included. A total of 9,629 participants were included, among which 7,110 (73.8%) participants were cognitively normal (CN), 2,024 (21.0%) were diagnosed with MCI, and 495 (5.1%) were diagnosed with dementia. The protocol was approved by the ethics committee at each center. Informed consent was obtained from each subject either directly or from his or her guardian. The study participants are shown in the Fig. 1.

Assessment and diagnosis procedure

The investigators included interviewers and experts. Interviewers worked as pairs consisting of one junior neurologist and one senior neurology graduate student. An expert group typically included two neurologists and two neuropsychologists with special expertise in cognitive impairment in each region. All interviewers and experts received the same weeklong training course on neuropsychological assessment and diagnosis, and also participated in a retraining course every 3 months. The inter-rater reliability for videotapes of cognitive tests and diagnostic procedures was required to exceed 0.90. The formal education years were stratified into three groups: Illiterate, 1–6 years of formal education, and 7 or more years of formal education. Age of participants was stratified as 65–69, 70–74, 75–79, 80–84, and≥85 years.

A standard diagnostic procedure was used. First, each interviewer pair conducted a 2-hour semistructured interview with participants and their close informants at their residence. Detailed data on sociodemographic characteristics, medical history, current medications, and family history were collected. Then one of the interviewers performed a systematic neuropsychological battery, which included: (1) memory: The World Health Organization-University of California-Los Angeles Auditory Verbal Learning Test (WHO-UCLA AVLT), including immediate recall, short-delay free recall (3 minutes), long-delay free recall (30 minutes), and long-delay recognition [12]; (2) executive function: Trail Making Test B [13]; (3) language: Semantic verbal fluency test (category, animals) [14]; and (4) visuoconstructive skill: Clock-drawing test (CDT) [15]. In addition, the Chinese version of the MMSE, which was first translated and modified by Katzman and colleges [11] and Montreal Cognitive Assessment (MoCA) [16] were performed to assess global cognition. Functional Activities Questionnaire (FAQ) [17] was administered for social functioning, the Center for Epidemiologic Studies Depression Scale (CESD) [18] was adopted for mood, and modified Hachinski Ischemic Score (MHIS) [19, 20] was used to differentiate between degenerative and vascular etiologies. After one interviewer completed these tests, the other determined participants’ Clinical Dementia Rating (CDR) score [21, 22] independently and took a detailed history of all cognitive impairments. The CDR rater was not aware of the other neuropsychological tests’ results. Last, standardized general and neurological examinations were performed. Neither interviewers nor participants were aware of the purpose of the study.

Diagnoses were made at the end of each workday when the expert panel and interviewers reviewed all information collected. When consensus was not reached, an expert returned to the residence the following day to reexamine the participant for a further evaluation and final diagnosis. Neuroimaging like computed tomography or magnetic resonance imaging was performed whenever the diagnosis was difficult to make.

Participants were divided into three categories according to cognitive level. Participants who scored 0 on all six domains evaluated by CDR were considered as CN. Criteria for MCI in our study were established on published criteria [23, 24] and included all the following elements: Cognitive impairment in one or more domains (scored at least 1.5 standard deviations below the norm in memory, executive function, language, or visuoconstructive skill); global CDR score = 0.5 or 0; essentially preserved daily activities and social functions (A total FAQ score≤5); no dementia. The diagnosis with dementia was made according to Diagnostic and Statistical Manual of Mental Disorders (Fourth Edition, DSM-IV) [25].

Statistical analysis

To detect significant difference of the age, years of formal education, and MMSE scores among CN, MCI, and dementia groups, ANOVA and post hoc test was used. Percentage differences of gender and residence were compared using Pearson’s χ² test or Fisher’s exact test. Stepwise multiple linear regression analyses were performed to assess the contributions of the demographic variables on the MMSE scores. The norms of the MMSE were stratified by the demographic variables that had significant effects on the MMSE. Finally, ROC analysis was used to determine the optimal cut-off scores of the MMSE in cognitive impairment screening. Statistical analysis was performed using the SPSS statistical package version 20.0 (SPSS Inc.; Chicago, Illinois, USA).

RESULTS

Demographic factors influencing MMSE performance

The demographic features and the MMSE scores of the subjects in CN, MCI, and dementia groups are provided in Table 1. For age and percentage of female, CN < MCI < dementia (p < 0.001); for years of formal education and mean MMSE scores, CN > MCI > dementia (p < 0.001); for percentage of the rural residence, CN < MCI and CN < dementia (p < 0.001), while no significant difference was found between the MCI and dementia groups (p = 0.097).

Stepwise multiple linear regression analyses were performed to assess the contributions of the demographic variables including age, gender, education, and residence on MMSE scores in CN group. The examined demographic variables, taken together, explained 39.4% of the MMSE scores variance. High level of education was significantly and positively associated with higher MMSE scores and accounted for the greatest proportion of the score variance (31.6%), whereas rural residence, age, and being female were significantly and inversely associated with the MMSE scores and accounted for 3.7%, 3.2%, and 0.9% of the variance, respectively (Table 2).

Normative data

According to the above demographic variables’ effects, age, gender, education, and residence were included in the calculation of normative values of the MMSE. The MMSE score distribution is shown as mean, standard deviation (SD), 10th, 25th, 50th (median), and 75th percentile of the sample (Table 3).

Optimal cut-off points of the MMSE

The optimal cut-off points were determined according to education level since education was the strongest demographic factor influencing the MMSE scores (Table 4). ROC analysis revealed that the area under the curve (AUC) (95% CI) of the illiterate, 1–6 years of education, and 7 or more years of education was 0.92 (0.91–0.94), 0.97 (0.96–0.99), and 0.98 (0.97–0.99), respectively (Fig. 2). The optimal cut-off points were 16/17 for illiterate individuals (sensitivity 87.6% and specificity 80.8%), 19/20 for individuals with 1–6 years of education (sensitivity 93.6% and specificity 92.7%), and 23/24 for individuals with 7 or more years of education (sensitivity 94.3% and specificity 94.3%). Nevertheless, low sensitivity and specificity were detected when we tried to identify MCI against CN using MMSE. The MMSE showed acceptable sensitivity and specificity for identifying MCI against dementia; however, the positive likelihood ratio (PLR) and negative likelihood ratio (NLR) were poor.

DISSCUSSION

The Chinese version of the MMSE was initially developed by Katzman and colleges [11] and later widely used in clinical practice and epidemiological studies. However, the lack of standard reference norms severely limits its clinical value. To address this issue, we conducted this large-scale, multicenter, population-based investigation. The samples recruited in this study represent a substantially broader spectrum of Chinese people than previous studies, and the national norms established in this study represent a great improvement for the use of the MMSE in Chinese elderly population.

Consistent with previous studies, our findings indicated that demographic variables significantly influence the MMSE scores. Longer education correlated with better performance on the MMSE. Urban residents scored higher than rural ones in general. The MMSE scores declined with the increase of age. Furthermore, male scored better than female. In the present study, education had the most potent effect on the MMSE performance. Low levels of education have long been recognized as one of the primary risk factors of dementia, which may influence individuals’ capability of comprehension and execution, or may reflect the innate intelligent status that inhibits the subjects to complete higher levels of education [2]. In addition, aging has always been regarded as an important risk factor for dementia, which has been verified in many epidemiological studies [26 –28]. Living in rural areas and being female are also risk factors for dementia. The residence and gender related variance on the MMSE scores might in part result from selective environmental exposures such as different socio-economic status. Therefore, all the four factors should be taken into consideration when establishing normative data for the MMSE in Chinese elderly.

In our study, besides the mean and SD scores, the normative data of the MMSE was also presented for percentile because of its non-Gaussian distribution. To compare an individual’s score to that expected in the general population, demographic stratified percentile distributions, rather than mean and SD scores, were more appropriate. The percentile MMSE scores in our norms were considerably lower than that in Western populations [2, 8]. One explanation of the lower MMSE scores might be the demographic features of Chinese elderly population. China has a markedly higher proportion of rural and undereducated population, which could significantly lower the MMSE scores. Moreover, increased prevalence of physical diseases detected in less educated individuals who were usually with worse socio-economic background, could also influence the MMSE performance and cause lower MMSE scores [2].

The cut-off points of the MMSE should be carefully chosen since they would directly influence the detection of dementia. Raising cut-off points would give rise to higher sensitivities, yet simultaneously cause higher rates of false positives. Accordingly, the optimal cut-off points should be of high sensitivity and specificity at the same time. Previous studies usually use 2.0 SDs of the mean or the 10th percentile scores to define the cut-off points. However, these definitions have apparent disadvantages. Firstly, the distributions of the MMSE scores are not normal, thus the mean and SD are inappropriate to act as cut-off points. Secondly, the use of the 10th percentile points is based on the assumed or estimated prevalence of dementia, rather than relying on the actual diagnosis, which would decrease the validity of the instrument. In our study, the optimal cut-off points were determined according to the clinical diagnosis of MCI and dementia for samples recruited, which could guarantee the more reliable reference norms of the MMSE.

The most consistently used MMSE cut-off point in dementia screening is 23/24 in Western countries [29]. However, the substantial educational, ethnic, and cultural differences existed in developing countries make these unstratified cut-off points inappropriate for dementia screening. Therefore, education-stratified cut-off points were presented in our study (16/17 for illiterate, 19/20 for those with 1–6 years of education, and 23/24 for those with 7 or more years of education). Before our investigation, only three regional sampling studies presented their norms and cut-off points of the MMSE in Mainland China. In the Beijing study whose sample size was 5,367, the cut-off points were 19/20 for illiterate, 22/23 for individuals with 1–6 years of education, and 26/27 for individuals with 7 or more years of education [5]. While in the study conducted in Shanghai which recruited 5,055 samples, the cut-off points were 17/18, 20/21, and 24/25 for each educational level [11]. Additionally, another study conducted on 370 neurology outpatients suggested that 20/21 for illiterate and 22/23 for literate were the optimal cut-off points for a modified Chinese version of MMSE [6]. The discrepancy of the cut-off points yielded in different studies emphasizes that the norms and cut-off points derived from regional sampling studies could not reflect the distribution of the MMSE scores in general population nationwide.

Regarding use of the MMSE in identifying MCI, limited evidence has been found in previous studies [29]. Our results suggest that the MMSE has limited ability to help identify MCI against CN individuals, or MCI against dementia. To better illustrate this issue, we suggest the need for further validation of the MMSE in MCI patients worldwide.

One limitation of this study is the lack of longitudinal assessment, which would provide further validation of specific relationships between demographic variables and MMSE performance. In addition, some of the strata had small numbers and the findings from these strata might be less reliable. The elderly Chinese residents tend to be less educated, especially for rural ones. Thus for the elderly aged 85 years and over, few subjects were with 7 or more years of education. However, in spite of such limitation, the subjects in most subgroups were sufficient. In general, the sampling population was representative.

In conclusion, our results provided the nationwide MMSE reference values, which could be useful in screening or evaluating dementia in Chinese elderly population.

Footnotes

ACKNOWLEDGMENTS

This study was supported by CHINA-CANADA Joint Initiative on Alzheimer’s Disease and Related Disorders (81261120571), Scientific Promoting Project of Beijing Institute for Brain Disorders (BIBDPXM2014_014226_000016), Seed Grant of International Alliance of Translational Neuroscience (PXM2014_014226_000006), Key Medical Professional Development Plan of Beijing Municipal Administration of Hospitals (ZYLX201301), the National Science and Technology Major Projects for “Major New Drug Innovation and Development” of the Twelfth 5-year Plan Period (2011ZX09307-001-03), the Major Project of the Science and Technology Plan of the Beijing Municipal Science & Technology Commission (D111107003111009), the National Key Technology R&D Program in the Eleventh Five-year Plan Period (2006BAI02B01), the Key Project of the National Natural Science Foundation of China (30830045). We gratefully acknowledge the support of our colleagues and collaborators.

Authors’ disclosures available online ().

References

Folstein

, Folstein

, McHugh

(1975) Mini-mental state: A practical method for grading the cognitive state of patients for the clinician. J Psychiatr Res 12, 189–198.

Crum

, Anthony

, Bassett

, Folstein

(1993) Population-based norms for the Mini-Mental State Examination by age and educational level. JAMA 269, 2386–2391.

Rosselli

, Ardila

, Pradilla

, Morillo

, Bautista

, Rey

, Camacho

(2000) The Mini-Mental State Examination as a selected diagnostic test for dementia: A Colombian population study. GENECO. Rev Neurol 30, 428–432.

Keskinoglu

, Ucku

, Yener

, Yaka

, Kurt

, Tunca

(2009) Reliability and validity of revised Turkish version of Mini Mental State Examination (rMMSE-T) in community-dwelling educated and uneducated elderly. Int J Geriatr Psychiatry 24, 1242–1250.

Zhang

, Hong

, LI

, Zhao

, Huang

, Wei

, Wang

, Li

, Yang

, Wu

, Ji

, Wang

(1999) The mini-mental state examination in the Chinese residents population aged 55 years and over in the urban and rural areas of Beijing. Chin J Neurol 32, 1–7.

, Meyer

, Huang

, Du

, Chowdhury

, Quach

(2003) Adapting mini-mental state examination for dementia screening among illiterate or minimally educated elderly Chinese. Int J Geriatr Psychiatry 18, 609–616.

Nunes

, Silva

, Cruz

, Roriz

, Pais

, Silva

(2010) Prevalence and pattern of cognitive impairment in rural and urban populations from Northern Portugal. BMC Neurol 10, 42.

Magni

, Binetti

, Bianchetti

, Rozzini

, Trabucchi

(1996) Mini-Mental State Examination: A normative study in Italian elderly population. Eur J Neurol 3, 198–202.

Han

, Jo

, Kim

, Park

, Kang

(2008) An adaptation of the Korean mini-mental state examination (K-MMSE) in elderly Koreans: Demographic influence and population-based norms (the AGE study). Arch Gerontol Geriatr 47, 302–310.

10.

De Silva

, Disanayaka

, De Zoysa

, Sanjeewanie

, Somaratne

, Foster

, Srikanth

, Kathriarachchi

, Martins

(2009) Norms for the Mini-Mental State Examination from a sample of Sri Lankan older people. Int J Geriatr Psychiatry 24, 666–670.

11.

Katzman

, Zhang

, Ouang Ya

, Wang

, Liu

, Yu

, Wong

, Salmon

, Grant

(1988) A Chinese version of the Mini-Mental State Examination; impact of illiteracy in a Shanghai dementia survey. J Clin Epidemiol 41, 971–978.

12.

Maj

, Satz

, Janssen

, Zaudig

, Starace

, D’Elia

, Sughondhabirom

, Mussa

, Naber

, Ndetei

, Schulte

, Sartorius

(1994) WHO Neuropsychiatric AIDS study, cross-sectional phase II. Neuropsychological and neurological findings. Arch Gen Psychiatry 51, 51–61.

13.

Reitan

(1958) Validity of the Trail Making Test as an indicator of organic brain damage. Percept Motor Skills 8, 271–276.

14.

Lezak

(1995), Neuropsychological assessment. 3rd ed, Oxford University Press, New York .

15.

Wolf-Klein

, Silverstone

, Levy

, Brod

(1989) Screening for Alzheimer’s disease by clock drawing. J Am Geriatr Soc 37, 730–734.

16.

Nasreddine

, Phillips

, Bedirian

, Charbonneau

, Whitehead

, Collin

, Cummings

, Chertkow

(2005) The Montreal Cognitive Assessment, MoCA: A brief screening tool for mild cognitive impairment. J Am Geriatr Soc 53, 695–699.

17.

Pfeffer

, Kurosaki

, Harrah

Jr , Chance

, Filos

(1982) Measurement of functional activities in older adults in the community. J Gerontol 37, 323–329.

18.

Lewinsohn

, Seeley

, Roberts

, Allen

(1997) Center for Epidemiologic Studies Depression Scale (CES-D) as a screening instrument for depression among community-residing older adults. Psychol Aging 12, 277–287.

19.

Moroney

, Bagiella

, Desmond

, Hachinski

, Molsa

, Gustafson

, Brun

, Fischer

, Erkinjuntti

, Rosen

, Paik

, Tatemichi

(1997) Meta-analysis of the Hachinski Ischemic Score in pathologically verified dementias. Neurology 49, 1096–1105.

20.

Rosen

, Terry

, Fuld

, Katzman

, Peck

(1980) Pathological verification of ischemic score in differentiation of dementias. Ann Neurol 7, 486–488.

21.

Morris

(1993) The Clinical Dementia Rating (CDR): Current version and scoring rules. Neurology 43, 2412–2414.

22.

Hughes

, Berg

, Danziger

, Coben

, Martin

(1982) A new clinical scale for the staging of dementia. Br J Psychiatry 140, 566–572.

23.

Petersen

(2004) Mild cognitive impairment as a diagnostic entity. J Intern Med 256, 183–194.

24.

Winblad

, Palmer

, Kivipelto

, Jelic

, Fratiglioni

, Wahlund

, Nordberg

, Backman

, Albert

, Almkvist

, Arai

, Basun

, Blennow

, de Leon

, DeCarli

, Erkinjuntti

, Giacobini

, Graff

, Hardy

, Jack

, Jorm

, Ritchie

, van Duijn

, Visser

, Petersen

(2004) Mild cognitive impairment–beyond controversies, towards a consensus: Report of the International Working Group on Mild Cognitive Impairment. J Intern Med 256, 240–246.

25.

American Psychiatric Association (1994) Diagnostic and statistical manual of mental disorders: DSM-IV. American Psychiatric Association, Washington, DC.

26.

Folstein

, Bassett

, Anthony

, Romanoski

, Nestadt

(1991) Dementia: Case ascertainment in a community survey. J Gerontol 46, M132–M138.

27.

Jorm

, Jolley

(1998) The incidence of dementia: A meta-analysis. Neurology 51, 728–733.

28.

Kawas

, Gray

, Brookmeyer

, Fozard

, Zonderman

(2000) Age-specific incidence rates of Alzheimer’s disease: The Baltimore Longitudinal Study of Aging. Neurology 54, 2072–2077.

29.

Mitchell

(2009) A meta-analysis of the accuracy of the mini-mental state examination in the detection of dementia and mild cognitive impairment. J Psychiatr Res 43, 411–431.