See

discussions, stats, and author profiles for this publication at: https://www.researchgate.net/publication/40849758

Cut-off scores in MMSE: A moving target?

Article in European Journal of Neurology · May 2010

Impact Factor: 4.06 · DOI: 10.1111/j.1468-1331.2009.02907.x · Source: PubMed

CITATIONS READS

15 123

5 authors, including:

Manuela Guerreiro Isabel Pavão Martins

University of Lisbon University of Lisbon
60 PUBLICATIONS 1,010 CITATIONS 114 PUBLICATIONS 1,078 CITATIONS

SEE PROFILE SEE PROFILE

Available from: Carolina Maruta

Retrieved on: 01 June 2016
European Journal of Neurology 2009 doi:10.1111/j.1468-1331.2009.02907.x

Cut-off scores in MMSE: a moving target?

J. Morgado, C. S. Rocha, C. Maruta, M. Guerreiro and I. P. Martins
Laboratory of Language Research, Faculty of Medicine of Lisbon, Institute of Molecular Medicine, Hospital de Santa Maria,
Lisboa, Portugal

Keywords:
cognitive ageing, cognitive
neuropsychology in
dementia, neuropsycho-
logical assessment
Received 7 September 2009
Accepted 11 November 2009
Background: Cognitive tests are known to be influenced by language, culture and
education. In addition, there may be an impact of ÔepochÕ in cognition, because there is
secular increase in scores of IQ tests in children. If we assume this is a long lasting
process, then it should persist later in life.
Methods: To test this hypothesis, we compared the performance of two cohorts of
individuals (‡50 years of age), evaluated 20 years apart using the Mini-Mental State
Examination (MMSE).
Results: Study population included 135 participants in 1988 and 411 in 2008. MMSE
scores were higher in 2008 than in 1988 for literacy x age-matched subgroups, the
difference being significant for participants with lower literacy. Score variance was
explained by literacy (b = 0.479, t = 14.598, P = 0.00), epoch (b = 0.34, t = 10.33,
P = 0.00) and age (b = )0.142, t = )4.184, P = 0.00).
Conclusion: The present results are in accordance with a lifelong secular improvement
in cognitive performance. The operational cut-off values may change with time, which
may have clinical impact in the diagnosis of disorders like mild cognitive impairment
or dementia.

Introduction Methods and procedures

Cognitive tests are influenced by language, culture and
formal education. In addition to that, there seems to be
an impact of ÔepochÕ in cognition that is not so well
understood. In children, there is a secular improvement
of general cognitive ability, measured by IQ tests
requiring regular updating of their normative values
[1,2]. This secular change may result from increased
environmental stimulation, earlier contact with sophis-
ticated technology, better general health or more effi-
cient use brain networks. If we assume this is a long
lasting process, then its effect should persist into late
adult life. Accordingly, contemporary elderly subjects
should be expected to perform better on cognitive tasks
than their predecessor and eventually, to have a better
cognitive reserve for the misfortune of brain damage.
To test this hypothesis, we compared the performance
of two similar cohorts of adult populations, evaluated
20 years apart using the Mini-Mental State Examina-
tion (MMSE) [3,4].
Correspondence: I. P. Martins, MD, PhD, Laboratory of Language
Research, Faculty of Medicine of Lisbon, Institute of Molecular
Medicine, Hospital de Santa Maria, Avenida Professor Egas Moniz,
1649-028 Lisboa, Portugal (tel./fax: +351 217934480;
e-mail: labling@fm.ul.pt).
Both samples, 1988 and 2008, were collected to obtain
normative data on the MMSE and other cognitive tests
in individuals above 49 years of age in the same urban
area. Recruitment of participants was made in health
and social security services for the elderly who are part
of the national services. Information concerning the
cognitive status of the individual and autonomy for
daily living activities was obtained either from their
family doctor or from their employees of the community
centres who have a longstanding contact with the indi-
viduals and their families. Consecutive individuals
attending their local centres were invited to participate if
they fulfilled the following criteria: age above 49 years,
autonomy in daily living activities and absence of his-
tory of brain damage (stroke, epilepsy, head trauma,
confirmed or suspected dementia), alcohol or drug
dependence, psychiatric diagnosis of psychosis and of
severe uncontrolled systemic illness (cancer, HIV, liver
or kidney failure). Aims and procedures of the study
were explained by a psychologist, and participants were
asked to give their informed consent. Ethics Commis-
sion of the Faculty of Medicine and a National Com-
mission of Data Protection approved the 2008 study.
MMSE was applied to all participants, followed by a
more extensive cognitive battery aimed to evaluate
memory, language, executive functions, depressive

 2009 The Author(s)

Journal compilation  2009 EFNS 1
2 J. Morgado et al.

Table 1 Characterisation of the two samples
1988 Sample 2008 Sample
N 135 411
Gender (F:M) 88:47 269:142
Age (mean, SD) 72.51 ± 10.06 [50–90] 66.67 ± 8.67 [50–92]
[range]
Literacy (mean, SD) 5.53 ± 5.68 [0–16] 6.48 ± 4.28 [0–17]
[range]
N, number of participants F, female; M, male.
symptoms and constructive abilities. Those tests were
different in the two samples.
Statistical analysis
Analysis was performed through the Statistical Package
for Social Sciences (SPSS) 16.0 software for Windows
(SPSS Inc., Chicago, IL, USA). To avoid the inclusion of
individuals with early undiagnosed dementia, extreme
outliers (under the 2.5th percentile) were excluded from
both samples. Participants were divided into four age
groups (50–59, 60–69, 70–79 and >80 years) and three
literacy groups: no or minimal reading and writing skills
(0–2 years of education), basic school (3–6 years) and
more than 6 years of education subgroups. Mean scores
were compared, by age x literacy matched subgroups,
betweenthetwodifferentsamplesbyindependent samples
tests (Wilcoxon–Mann–Whitney). Repeated regression
analyses were performed where the MMSE score, or
MMSE subtests scores, were the dependent variables and
subjectsÕ gender, age (in years), literacy (in years) and
epoch (1988 or 2008) were the independent variables.
Results
There were 411 participants in 2008 and 135 in 1988
(Table 1). MMSE mean scores were higher in 2008 than
in 1988 for most literacy x age-matched subgroups, but
the difference was only significant for the lower literacy
subgroups (Table 2). On the linear regression analysis,
it was found that literacy was the major determinant of
the MMSE score variation, followed by epoch and age.
Analysing the different subtests of the MMSE, it was
found that epoch explained the variation of three sub-
tests: temporal orientation (five-item score), attention
and calculation, and language (written and oral sub-
tests), but had no effect on the other parts of MMSE
(Table 3).
Discussion
The present study suggests that the cognitive abilities of
our elderly population have changed over the last
20 years. Although awareness of dementia has sub-
stantially increased in the same period of time, because
of the ageing of the population and increased life
expectancy, the present results suggest that nowadays
functionally independent subjects do actually perform
better, at least in this screening cognitive instrument,
than they did 20 years ago. This change may be more
evident in more sensitive specific measures.
The epoch effect, on cognitive performance, was not
observed for all cognitive domains. It was significant
only for temporal orientation, attention and calculation
and language. It may have different explanations that
may be general or specific to the studied populations. It
is possible that an earlier and more intense contact, in
adulthood, with sophisticated technology (television,
mobile phones, computers and domestic electronic de-
vices), the need to change currency and a generalised
access to large amounts of information may have con-
tributed to develop skills that are measured in these
simple tests (working memory, calculation, compre-
hension of sentences). Improved general health might

Table 2 Comparison of Mini-Mental State

1988 Sample 2008 Sample Examination (MMSE) mean scores consid-
ering subgroups of age and literacy
Age Literacy N Mean (SD) N Mean (SD) Mann–Whitney U test

50–59 years 0–2 years 2 18.50 (0.71) 1 29.00 (–) –

3–6 years 6 27.83 (1.72) 36 27.83 (1.96) P = 0.87
>6 years 10 28.90 (1.60) 48 29.44 (0.82) P = 0.68
60–69 years 0–2 years 13 22.85 (3.13) 10 25.40 (1.96) P = 0.03
3–6 years 11 27.09 (1.22) 90 28.02 (1.74) P = 0.02
>6 years 8 29.25 (1.17) 74 29.03 (1.03) P = 0.40
70–79 years 0–2 years 16 20.19 (3.06) 20 24.70 (2.25) P = 0.00
3–6 years 19 26.37 (3.37) 71 27.70 (1.68) P = 0.17
>6 years 11 28.55 (0.93) 27 28.74 (1.38) P = 0.36
‡80 years 0–2 years 15 18.80 (3.01) 7 25.57 (1.90) P = 0.00
3–6 years 9 26.11 (2.37) 17 27.18 (2.04) P = 0.25
>6 years 15 27.60 (1.88) 10 28.20 (1.40) P = 0.50

Bold values denote significant difference between MMSE mean values (P < 0.05). N, number
of participants.

 2009 The Author(s)

Journal compilation  2009 EFNS European Journal of Neurology
 Cut-off scores in MMSE: a moving target? 3

Table 3 Influence of literacy, age and epoch on Mini-Mental State Examination (MMSE) and subtests scores

MMSE R = 0.67 R2 = 0.45 Adjusted R2 = 0.45 SE estimate = 2.23 F = 150.40 P = 0.00

N = 546 B SE B Beta t P-value
Constant 26.72 0.85 31.54 0.00
Literacy 0.31 0.02 0.50 14.60 0.00
Age )0.05 0.01 )0.14 )4.18 0.00
Epoch 2.37 0.23 0.34 10.33 0.00
Temporal orientation R = 0.46 R2 = 0.22 Adjusted R2 = 0.21 SE estimate = 0.67 F = 49.65 P = 0.00
N = 546 B SE B Beta t P-value
Constant 4.38 0.26 – 17.12 0.00
Literacy 0.04 0.01 0.25 6.30 0.00
Age )0.01 0.00 )0.08 )1.92 0.06
Epoch 0.58 0.07 0.33 8.39 0.00
Attention/calculation R = 0.50 R2 = 0.25 Adjusted R2 = 0.25 SE estimate = 1.08 F = 60.09 P = 0.00
N = 546 B SE B Beta t P-value
Constant 3.45 0.41 8.39 0.00
Literacy 0.10 0.01 0.37 9.60 0.00
Age )0.01 0.01 )0.04 )0.97 0.33
Epoch 0.82 0.11 0.28 7.33 0.00
Language R = 0.50 R2 = 0.25 Adjusted R2 = 0.25 SE estimate = 0.62 F = 61.55 P = 0.00
N = 546 B SE B Beta T P-value
Constant 7.11 0.24 30.22 0.00
Literacy 0.06 0.01 0.40 10.52 0.00
Age )0.00 0.00 )0.04 )1.10 0.27
Epoch 0.40 0.06 0.24 6.25 0.00

Bold values denote significant association with MMSE score (P < 0.01). SE, standard error.

also have had an impact on cerebral development or its
protection from several disorders [5]. A better control
of vascular risk factors (hypertension, diabetes, diet,
smoking habits) might have prevented leukoaraiosis or
silent strokes, both associated with cognitive impair-
ment [6–9].
The improved cognitive performance could also be
related to a more widespread access to formal education
and to the overall lower age in the 2008 sample [10].
Yet, although there were more individuals with higher
literacy and lower age in the present sample, which
could account for a higher total sample MMSE score,
the comparison by education x age subgroups shows
that this effect persisted when these variable were con-
trolled, particularly for the lower educated subgroups,
who might have benefited more from ÔepochÕ related
changes.
We acknowledge that the present study has some
limitations. So far, its results only apply to the studied
population. Moreover, this study was performed in a
country that underwent major changes during this 20-
year period, because of its integration in a richer com-
munity, the European Union and suffered a change in
currency. It would be interesting to replicate it in a
community-based sample of the population and in
other communities, with a control of vascular risk
factors and comorbidities, to understand the mecha-
nisms that are underlying this improvement and to
unveil possible protective factors of cognitive decline.
This study stresses the need to update cognitive
normative values, particularly those that are used in
clinical diagnosis or for the selection of subjects for
studies. The present results will have an impact on
dementia or cognitive impairment screening, because
they are now much closer than values obtained in other
countries.
Acknowledgements
This study was sponsored by Fundação Calouste Gul-
benkian (Project 0488). The authors are indebted to
general practitioners who participated in the sample
selection.
References
1. Flynn JR. The mean IQ of Americans: massive gains 1932
to 1978. Psychol Bull 1984; 95: 29–51.
2. Flynn JR, Weiss LG. American IQ gains from 1932 to
2002: the WISC subtests and educational progress. Int J
Test 2007; 7: 209–224.
3. Folstein MF, Folstein SE, McHugh PE. A practical
method for grading the cognitive state of patients for the
clinician. J Psychiatr Res 1975; 12: 189–198.
4. Guerreiro M, Silva AP, Botelho M, Leitão O, Castro-
Caldas A, Garcia C. Adaptação à população portuguesa
da tradução do Mini Mental State Examination (MMSE).
Rev Port Neurol 1994; 1: 9.
5. Rosnick CB, Small BJ, Graves AB, Mortimer JA. The
Association between health and cognitive performance in

 2009 The Author(s)

Journal compilation  2009 EFNS European Journal of Neurology
4 J. Morgado et al.
a population-based study of older adults: the charlotte
county healthy aging study (CCHAS). Aging Neuropsy-
chol Cogn 2004; 11: 89–99.
6. Viana-Baptista M, Bugalho P, Jordão C, et al. Cognitive
function correlates with frontal white matter apparent
diffusion coefficients in patients with leukoaraiosis. J
Neurol 2008; 255(360–366.
7. Benisty S, Gouw AA, Porcher R, et al. Location of
lacunar infarcts correlates with cognition in a sample of
non disabled subjects with age-related white matter
changes: the LADIS study. J Neurol Neurosurg Psychiatry
2009; 80: 478–483.
8. Kearney-Schwartz A, Rossignol P, Bracard S, et al. Vas-
cular structure and function is correlated to cognitive
performance and white matter hyperintensities in older
hypertensive patients with subjective memory complaints.
Stroke 2009; 40: 1229–1236.
9. Hachinski V. Shifts in thinking about dementia. JAMA
2008; 300: 2172–2173.
10. Castro-Caldas A, Petersson KM, Reis A, Stone-Elander
S, Ingvar M. The illiterate brain: learning to read and
write during childhood influences the functional organi-
zation of the adult brain. Brain 1998; 121: 1053–1063.
 2009 The Author(s)
Journal compilation  2009 EFNS European Journal of Neurology