Life-long intellectual activities mediate the predictive effect of
early education on cognitive impairment in centenarians:
a retrospective study
M. KLIEGEL,1D. ZIMPRICH1& C. ROTT2
1University of Zurich, Switzerland &2German Centre for Research on Ageing, Heidelberg, Germany
The purpose of this study was to examine the hypothesis of whether early education and/or maintaining intellectual activities
over the life-course have the power to protect against cognitive impairment even in extremely old adults. Ninety centenarians
from the population-based Heidelberg Centenarian Study were assessed with a modified version of the Mini Mental State
Exam (MMSE). Data about education, occupational status, and life-long intellectual activities in four selected domains were
obtained. Results demonstrated that 52% of the sample showed mild-to-severe cognitive impairment. Analyzing the
influence of early education, occupational status, and intellectual activities on cognitive status we applied several (logistic)
regression analyses. Results revealed independent, significant and strong influence of both formal school education and
intellectual activities on the cognitive status in very late life, even after controlling for occupational status. However, about
one fourth of the effect of early education on cognitive status was exerted indirectly via the assessed intellectual activities.
In summary, the present study provides first evidence for the conclusion that even with regard to cognitive performance in
very old age, both early education and life-long intellectual activities seem to be of importance.
Human longevity has dramatically increased in most
developed countries over the last 100 years. In the
last 50 years, a deceleration of mortality at older ages
has been observed (Kannisto et al., 1994), which is
especially salient in the proliferation of centenarians
(Jeune & Kannisto, 1997). Old age mortality will
continue to decrease (Kannisto, 1996) and rate of
growth of the population of the oldest old will further
increase (Vaupel, 2000).
One of the core issues of independence or
autonomy in very old age is the cognitive status.
Several authors have demonstrated that cognitive
impairment is the most powerful predictor of survival
in late life (Allard et al., 2000; Arfken et al., 1999;
Bosworth et al. 1999; Maier & Smith, 1999).
Moreover, intact cognitive functioning seems to be
much more than just a survival attribute. In a sample
of 600 older adults, Lawton et al. (1999) found that
cognitive loss was rated as overwhelmingly threaten-
ing. More than two thirds of their participants
indicated that they would not wish to live any
longer under all forms of cognitive impairment.
Thus one of the key challenges for gerontological
research is how to maintain an intact cognitive
system (Gatz et al., 2001; Hultsch et al., 1999;
Katzman, 1993; Orrell & Sahakian, 1995). A most
prominent finding in this context is that longer and
more complex early education correlates with higher
cognitive performance in old age (e.g., Cummings
et al., 1998; Gatz et al., 2001; Nielsen et al., 1999;
Zhang et al., 1998). Analyzing Swedish twin pairs
discordant for dementia, Gatz et al., (2001), for
example, found that low early education (i.e., six
years or less of schooling) was a risk factor for
developing Alzheimer’s disease. Although the exact
mechanisms are still unknown, a widely accepted
explanation for this effect is the proposition of an
initial cognitive reserve effect (Gatz et al., 2001;
Mortimer, 1997; Satz, 1993). In this view, cognitive
impairment is observed clinically after cognitive
reserve is depleted to some threshold. Less initial
cognitive reserve caused by innate differences in
cognitive resources or by less cognitive activity
during brain maturation in childhood would imply
that smaller changes would be required to bring an
individual to this critical threshold.
Correspondence to: Matthias Kliegel, Institute of Psychology, Department of Gerontopsychology, University of Zurich,
Schaffhauserstr. 15, CH-8006, Zu ¨rich, Switzerland. Tel: þ41 1 635 3431. Fax: þ 41 1 635 3421. E-mail: m.kliegel@
Received for publication 21st March 2003. Accepted 7th December 2003.
Aging & Mental Health, September 2004; 8(5): 430–437
ISSN 1360-7863 print/ISSN 1364-6915 online/04/05430-0437 ? Taylor & Francis Ltd
Another line of research states that not early
education per se but life-long intellectual activities
such as learning a foreign language or travelling as
well as the complexity of one’s occupation may act as
a buffer against cognitive impairment (e.g., Schooler
& Mulatu, 2001; Schooler et al., 1999). Hultsch et al.
(1999), for example, demonstrated significant rela-
tionships between changes in intellectually engaging
activities and changes in cognitive functioning
in a sample of 250 middle-aged and young-old
adults (aged 55–86years)
Longitudinal Study that were tested three times
over six years. These results confirm several findings
from cross-sectional studies reporting correlations
between participation in intellectually engaging daily
activities and performance on various cognitive
tasks (e.g., Erber & Szuchman, 1996; Hill et al.,
1995; Hultsch et al., 1993). The most prominent
explanation for these correlations between life-
long intellectual engagement and cognitive impair-
ment in old age is the disuse hypothesis (e.g., Orrell
& Sahakian, 1995; Salthouse, 1991). According to
this hypothesis, mental activity throughout adulthood
can increase cognitive reserve, either through formal
educational activities (e.g., learning new languages)
or through challenging cognitive occupations or
cognitively stimulating leisure activities (e.g., travel-
ling; Katzman, 1993; Schooler & Mulatu, 2001).
Although the literature on the predictive value of
early education, occupational status, and life-long
intellectual activities for cognitive functioning in old
age is growing, there seem to be at least two essential
shortcomings which we aim to address in the present
paper: first, while many studies focus either on early
education, occupation or life-long intellectual activ-
ities, there are few studies directly testing all
hypotheses in one sample (e.g., Hultsch et al.,
1998; Verghese et al., 2003). Therefore, in the
present paper we intend to compare the influence
of these predictors for cognitive impairment in one
population-based sample. We propose two possible
mechanisms of how early education, occupation and
life-long intellectual activities might interplay in their
effect on cognitive impairment: (1) early education,
occupational status, and life-long intellectual activ-
ities may constitute separate influences on cognitive
impairment in old age. That is, innate differences in
cognitive resources and cognitive activity during
brain maturation in childhood might result in a
specific amount of cognitive reserves, which, how-
ever, does not automatically determine the amount
of intellectual engagement during the life-course.
Empirically, this would imply that early education,
occupational status, and intellectual activities con-
tribute more or less independently to the prediction
of cognitive impairment in the elderly. (2) Higher
early education may lead to more complex occupa-
tions and/or life-long intellectual activities, which,
in turn, predict cognitive impairment. That is, more
initial cognitive reserves and more cognitive activity
during brain maturation in childhood might be
one cause of more intellectual engagement during
adulthood, which, in turn, leads to even more
cognitive reserves that might act as a buffer against
imply that early education has a direct effect on
intellectual engagement but only an indirect effect
on cognitive impairment via intellectual engage-
ment. Both mechanisms may be captured empirically
using a mediational model (Baron & Kenny, 1986),
where the effect of early education on cognitive
impairment in old age is mediated by life-long
intellectual engagement (i.e., occupation and/or
activities). Mechanism (1) would require that there
is no mediation of the early education-cognitive
Mechanism (2), in turn, would require that intellec-
tual engagement completely mediates the early
education-cognitive impairment-relation. However,
it seems unrealistic to expect either no mediation
at all or complete mediation empirically. Rather, one
would expect partial mediation, i.e., that the effects
of occupational status and/or intellectual activities
on cognitive impairment attenuate the direct effect
of early education on cognitive impairment. Thus,
one goal of the present study is to determine the
amount of mediation in the education-cognitive
impairment-relation by occupational status and
Secondly, most of the studies cited above have
been conducted in old adults aged 60 to 90.
However, reviewing the literature on cognitive
impairment in the oldest old (90þ), it becomes
obvious that the question regarding factors predict-
ing cognitive impairment is particularly crucial in
centenarians where cognitive impairment becomes
more and more prevalent (e.g., Jorm et al., 1987;
Thomassen et al., 1998) and plasticity is generally
thought to be highly reduced (e.g., Baltes & Smith,
1999). Recently, converging evidence has been
published that about 50–60% of the centenarians
suffer from substantial cognitive impairment. In the
Heidelberg Centenarian Study, the first population-
based study on centenarians in Germany, Kliegel
et al. (2001) revealed that about 50% of the cente-
narians examined were mildly or strongly cognitively
impaired. Consistent with these numbers are reports
from Denmark (Andersen-Ranberg et al., 2001),
findings from New England, USA (Silver et al.,
2001) as well as data from Japan, Sweden, and
Georgia, USA (Hagberg et al., 2001). Interestingly,
despite the summarized relevance, so far, no study
has been conducted to investigate the effects of early
education and/or life-long intellectual engagement
on cognitive functioning in the oldest old. Therefore,
Centenarian Study to provide initial evidence on
the role of early education and life-long intellectual
engagement in predicting cognitive performance in
very old age.
Cognitive impairment in centenarians
detailed information on the sampling procedure see
Rott et al., 2001) is a population-based study within
a geographically defined area approximately 60
kilometres around Heidelberg, Germany, consisting
of 172 communities with 2.6 million inhabitants. For
recruitment, resident registration offices were asked
to provide information on inhabitants born in 1901
or earlier. Records were obtained from all of the 172
communities that had been approached. Potential
participants were all individuals who were at the age
of 100;1 to 100;11 between January 1, 2000 and
April 30, 2001 (n¼179). During the recruitment
phase it turned out that a substantial proportion of
information provided could not be verified as
persons had either deceased, or their personal
data did not correspond to the information provided,
or refused to participate (see below). Therefore,
in addition, all individuals who were at the age of
99;4 to 99;11 between February 15, 2001 and April
30, 2001 (n¼70) were added to the sample. This
procedure—previously used in oldest old research—
is described as dynamic cohort design (Andersen-
Ranberg et al., 2001).
Combining the identified participants from both
age groups resulted in a potential sample of 156
centenarians or persons close to their 100th birthday
who were contacted in person by telephone. For
individuals who were not able to talk on the phone a
close relative was contacted. The group of non-
participants consisted of 65 persons. Sixteen persons
or their proxies did not agree to participate due to
health problems and the same number due to
dementia. Fourteen persons were not interested in
the study, 11 persons feared that the interview would
be too stressful. Eight proxies refused participation of
the centenarians because those were suffering from
psychiatric disorders. Two possible participants were
hospitalized at the time of recruitment. In eight cases
very personal reasons for non-participation were
mentioned, such as ‘I don’t want to get into contact
with strange persons’. Because nine persons gave
more than one reason these numbers do not add up to
65. One subject had to be excluded from this study
because the cognitive status could not be assessed
due to aggressive behaviour. Thus, a total of 90
extremely old people (10 men and 80 women) with a
mean age of 100.21 years (SD¼0.40) at the time of
examination participated in the study, which corre-
sponds to a participation rate of 58% (see Baltes et al.,
1999, for similar figures in the Berlin Aging Study).
There were no significant differences between parti-
cipants and non-participants regarding sex (non-
participants: 80% female) and the proportion living at
home or in an institution (participants: 52% living at
home; non-participants: 48% living at home). Thus,
Centenarian Study(for more
participants did not represent a positive selection in
terms of the ability of independent living (for more
details on the sample see Rott et al., 2001).
Procedure and measures
All centenarians were interviewed at their place of
residence, including institutions. In addition to the
centenarians, for each participant a proxy respondent
was interviewed. Most of the proxies were children of
the participants (62%) or other close relatives
(spouse: 1%; son-/daughter-in-law, grandchildren,
nephew/niece: 7.5% each). For the remaining 14.5%
the closest care-giving person was interviewed. There
were five exceptions, where the centenarian refused
to have a proxy interviewed. The interview was based
on Fillenbaum’s (1988) Multidimensional functional
assessmentof older adults:
(OARS). The interview included socio-demographic
information such as education and occupation as
well as health-related information such as the
assessment of current activities of daily living
(ADL, IADL), and a cognitive assessment (see
Rott et al., 2001, for more details on these data).
The cognitive assessment included a modified
version of the Mini Mental State Exam (MMSE;
Folstein et al., 1975). We shortened the MMSE
according to results reported by Holtsberg et al.
(1995), demonstrating systematic influences of visual
andliteracy deficits on
performance in items that rely on reading and
writing. Thus, we omitted the reading and writing
divisions, which evaluates naming, following spoken
and written commands, sentence generation, and
constructional abilities. The modified version with
a maximum score of 21 and a cut-off score of 10
does not lead to reduced reliability or validity
(see Holtsberg et al., 1995; Kliegel et al., 2001).
In addition, proxies rated whether and when the
centenarians had shown first indications of cognitive
The assessment of lifetime cognitive activities
was derived from procedures developed by Hultsch
et al. (1998, 1999). Due to assessment restrictions
in studies on centenarians with regard to time of
assessment, only those selected items were adminis-
tered that had been found to be most sensitive in a
previous pilot study with centenarians. Specifically,
we asked whether individuals ever in their lives had
participated in novel information processing activ-
ities (balancing a chequebook and learning a foreign
language) and in travelling activities (abroad and
within the country), and if so, up to what age the
activity was executed.
Finally, occupational status was assessed by a
rating scale on the complexity of the centenarian’s
previous occupation according to the type of
M. Kliegel et al.
occupation (e.g., ranging from 1¼unskilled worker
to 12¼academic work).
First, we report the distribution of the MMSE scores
(see Kliegel et al., 2001, for a more detailed analysis).
As summarized in Figure 1, the MMSE cut-
off of 10 divided the centenarians into two halves
of nearly equal size (MMSE<10: 52%; MMSE>10:
48%). Kliegel et al. (2001) have shown that the cut-
off score of 10 leads to a valid classification of a group
with moderate to severe cognitive impairment. Thus,
because of the non-normal distribution of centenar-
ians’ cognitive scores into two sub-groups, for all
subsequent analyses targeting cognitive impairment
the dependent measure was a dichotomous variable,
where 0 indicated ‘moderate to strong cognitive
impairment (MMSE<10)’ and 1 indicated ‘no to
mild cognitive impairment (MMSE>10)’.
Overall, the participants had M¼8.47 years of
formal school education
Overall, the mean participants’ occupational com-
plexity was rated as M¼7.93; (SD¼9.47; min¼1;
To obtain a score for executed and maintained
intellectual activities that was not confounded by the
present cognitive status of the centenarians, we
analyzed how many of the four assessed activities
had been given up before the age of 80. We chose the
retrospective cut-off of 80 years to avoid a frequent
limitation reported by most previous studies that
were not able to rule out the alternative hypothesis
that correlations between intellectual activities and
cognitive impairment may reflect the possibility that
high-ability individuals lead intellectually active lives
until cognitive impairment in old age limits their
activities rather than that intellectual activities act as
buffer against cognitive impairment. Because recent
analyses have shown that the median survival of
Alzheimer’s disease and vascular dementia ranges
between three and six years (Wolfson et al., 2001), in
living centenarians termination of activities before
the age of 80 (i.e., 20 years before assessment)
is unlikely to be due to the onset of cognitive
impairment observed at the age of 100. This
reasoning was supported by the proxy ratings
concerning the onset of symptoms of cognitive
impairment. According to the proxies, no centenar-
ian had shown any indications of cognitive impair-
ment 20 years before the assessment. The longest
time period since the beginning of cognitive impair-
ment reported by the proxies was 10 years (M¼4.4;
SD¼2.5). The descriptive data are summarized
in Table 1. On average, 2.56 (SD¼1.06) of the
four activities were given up by the age of 80.
0 - 2 3 - 56 - 89 1011 12 - 14 15 - 17 18 - 21
FIG. 1. Cognitive status.
centenarians (complete sample; n¼90)
Maintenance of intellectual engagement among
Given up until
the age of 80 or
after of 80
Learning a foreign
Travelling within the
55 (61.1)35 (38.8)
59 (65.6)31 (34.4)
Cognitive impairment in centenarians
Influence of early education, occupational status, and
intellectual activities on cognitive status
Analyzing the influence of early education, occupa-
tional status, and terminated intellectual activities on
cognitive status, we applied several (logistic) regres-
sion analyses according to the mediational model
depicted in Figure 2. With this model we intended to
combine both models discussed in the introduction,
namely, (a) early education, occupational status, and
intellectual activities may constitute separate influ-
ences on cognitive impairment in old age and (b)
higher early education may lead to more complex
occupations, and more intellectual activities across
the life span, which, in turn, predicts cognitive
In a first step, cognitive status (0¼‘moderate to
strong cognitive impairment’, i.e., MMSE<10,
versus 1¼‘no to mild cognitive impairment’, i.e.,
MMSE>10) was regressed on early education, i.e.,
years of schooling. As can be seen from Table 2,
the logistic regression revealed a statistically signifi-
cant and positive effect of early education on
cognitive status (parameter estimate: 0.819, SE
0.295, p<0.05), corresponding to an odds ratio of
2.268 (95% confidence limits: 1.272–4.042). That is,
with every additional year of early education the
probability of being not or only mildly cognitively
impaired as a centenarian increased by a factor of
2.27. As a measure of effect size, one may use
Pseudo-R2(Demaris, 1992, p. 54), which was 16%.
In a second step, cognitive status was regressed on
occupational status, i.e., the occupational complexity
rating (see Table 2). The logistic regression yielded a
statistically non-significant positive effect of occupa-
tional status on cognitive status (parameter estimate:
0.124, SE 0.069, p>0.05), corresponding to an odds
ratio of 1.132 (95% confidence limits: 0.990–1.296).
Hence, occupational status was not significantly
related to cognitive status, which was also reflected
by the small effect size of Pseudo-R2¼4%.
In a third step, cognitive status was regressed on
the number of activities given up by the age of 80 (see
Table 2). The logistic regression showed a statisti-
cally significant and negative effect of the number of
activities given up by the age of 80 and cognitive
status (parameter estimate: ?0.786, SE 0.234,
p<0.05), corresponding to an odds ratio of 0.456
(95% confidence limits: 0.288–0.720). Thus, with
every activity given up by the age of 80, the
probability of being not or only mildly cognitively
impaired as a centenarian decreased by a factor of
0.456. The Pseudo-R2of this effect was 14%.
Next, in a fourth step, occupational status was
regressed on early education (see Table 2). An
ordinary least squares (OLS) regression revealed
that early education had a statistically significant
and positive effect on occupational status (parameter
estimate: 0.699, SE 0.205, p<0.05). Hence, with
every additional year of schooling, on average,
occupational complexity was raised by 0.7 units.
Early education explained 12% of variance in
Afterwards (step 5), the number of activities given
up by the age of 80 was regressed on early education
(see Table 2). An OLS regression showed a
statistically significant and negative effect of early
education on the number of terminated intellectual
activities (parameter estimate: ?0.287, SE 0.065,
p<0.05). Thisresult impliesthatwith every
TABLE 2.Logistic and OLS regressions examining the mediational model
95% CI Explained
variance (in %)
Step 1Cognitive status
Number of activities
2.268 1.272–4.042 16
Step 2 Occupational status0.124 1.1320.990–1.2964
Step 3 Number of activities given up
Step 4 Early education
Step 5 Early education
Step 6 a) Early education
b) Occupational status
c) Number of activities given up
Cognitive status in
FIG. 2.Heuristic model of data analysis.
M. Kliegel et al.
additional year of early education, on average, 0.287
activities less were given up by the age of 80. Early
education explained 18% of variance in the assessed
In a sixth and final step, cognitive status was
regressed on early education, occupational status,
and number of activities given up simultaneously
(see Table 2). The logistic regression resulted in a
statistically significant and positive effect of early
education on cognitive status (parameter estimate:
0.626, SE 0.299, p<0.05), corresponding to an odds
ratio of 1.869 (95% confidence limits: 1.040–3.258)
as well as a statistically significant and negative effect
of number of activities given up by the age of 80
(parameter estimate: ?0.566, SE 0.262, p<0.05),
corresponding to an odds ratio of 0.566 (95%
confidence limits: 0.339–0.949). By contrast, occu-
pational status had no statistically significant effect
on cognitive status, and the resulting parameter
estimate was small and negative (?0.015, SE 0.086).
The Pseudo-R2for the total model was 21%.
These results imply that the assessed intellectual
activities mediated about 0.819?0.626¼0.193 of
the direct effect of early education on cognitive
status. (Note that this equation holds exactly for
multiple regression models, but it holds only
[Baron & Kenny, 1986]). Stated in a different, but
equivalent way: the direct effect of early education on
0.819)¼0.24 or 24%, i.e., about one fourth of the
effect of early education on cognitive status was
exerted indirectly via the number of activities given
up by the age of 80. In order to test for the statistical
significance of the amount of mediation, a formula
recommended by MacKinnon et al. (1995) was
applied, according to which the mediation was
statistically significant (z¼2.05, p<0.05). Thus,
life-long intellectual activities partially mediated
the effect of early education on cognitive status in
centenarians. By contrast, the effect of early educa-
tion on cognitive status was not mediated by
By analyzing data from the Heidelberg Centenarian
Study, it was the aim of the present study to exam-
ine the hypotheses whether
occupational status, and/or intellectual activities
over the life-course have the power to protect against
cognitive impairment in very old adults. Results of
the cognitive status revealed that 52% of the sample
showed mild-to-severe cognitive impairment, which
is in line with recent reports from other cente-
narian studies (e.g., Andersen-Ranberg et al., 2001;
Hagberg et al., 2001) and underlines the compar-
ability of the collected German sample with other
oldest old databases.
With respect to the core issue of the present
paper, in sum, results showed that both early
education and the assessed intellectual activities,
but not occupational status, had a strong and
statistically significant effect on cognitive status in
the sense that higher early education and the greater
number of intellectual activities maintained acted as
a buffer against becoming cognitively impaired.
Most interestingly, these effects were not indepen-
dent from one another
intellectual activities partially mediated the relation
between early education and cognitive status. Thus,
early education was strongly associated with intel-
lectual activities, which, in turn, were associated
with cognitive status. However, the assessed intel-
lectual activities did not completely mediate the
early education-cognitive status-relation. That is,
both intellectual activities and early education, for
the most part, contributed uniquely to the predic-
tion of cognitive status in centenarians. This finding
is in line with previous studies reporting effects of
very early cognitive performance on the develop-
ment of dementia in late life such as the Nun Study
by Snowdon et al. (1996). Here, the authors demon-
strated that the complexity of thought in essays
written in their early 20s predicted which nuns later
developed Alzheimer’s disease.
Overall, the reported findings support and extend
recent reports concerning the early education-
cognitive impairment-relationship in at least two
major points. First, most of these reports were not
able to rule out the alternative or complementary
idea that besides very early cognitive activity, life-
long intellectual activities also could make a differ-
ence (e.g., Gatz et al., 2001; Moceri et al., 2000).
Assuming that not (only) cognitive reserve acquired
in very early life is responsible for the observed buffer
effects, several authors have proposed that mental
activity throughout adulthood can increase cognitive
capacity (e.g., Katzman, 1993). The present results
underline these hypotheses and reveal that both early
school education (as cognitive reserve) and life-long
intellectual activities share a common influence, but
do also have an independent effect on the develop-
ment of cognitive impairment. Secondly, while
several studies report the positive effects of early
school education (e.g., Gatz et al., 2001) as well as
intellectual activities (e.g., Hultsch et al., 1999) on
cognitive performance in general, so far no study
has been conducted to investigate these effects in
the oldest old. In doing so, the present results
provide the first evidence that even in very late life,
where enormous (biological) ‘pressure’ for cognitive
impairment was repeatedly proposed (e.g., Jorm
et al., 1987; Thomassen et al., 1998) and plasticity
is generally thought to be highly reduced (e.g., Baltes
& Smith, 1999), long-term effects of very early
cognitive activity and life-long intellectual activities
on maintaining efficiency of the intellectual system
can still be demonstrated. Moreover, in the present
because the assessed
Cognitive impairment in centenarians
age group, these effects seemed to be independent
from occupational status. Although occupational
status was related to early school education, the
rated complexity of the centenarians’ previous
occupation did not predict cognitive impairment at
the age of 100. This seems to be in contrast to
previous findings (e.g., Schooler et al., 1999).
However, it underlines the distinctiveness of the
investigated population and might indicate the
particular importance of intellectual activities main-
tained after retirement. However, with the present
data, this hypothesis cannot be tested.
That said, one has to acknowledge some further
limitations of the present data set. The study’s
retrospective nature, the reliance on proxy data,
and a medium-sized number of participants com-
pared to most of the larger epidemiological samples
(e.g., Evans et al., 1997). However, even if a larger
sample size might have allowed detecting even very
small effects, the present power was sufficient to
reveal the substantial effects of early education and
intellectual activities. An important limitation is
represented by the limited number of assessed
variables, especially regarding intellectual activities,
compared to previous longitudinal studies on the
effect of a large number of intellectual activities on
cognitive functioning (e.g., Hultsch et al., 1999).
Thus, the restricted available data do neither allow a
sophisticated investigation of the influence of a broad
theoretical construct such as general life-long intel-
lectual engagement nor can they explicitly test the
interdependence of the assessed variables with
additional potential factors of influence such as a
detailed socio-economic status or lifetime physical
abilities. The same holds, in principle, for the
relation of the present data to the dementia
literature. By only relying on MMSE data, no
direct conclusions can be drawn with regard to the
specific nature of the observed cognitive impairment
in the present centenarians. Here, future studies
focusing on this aspect will have to include more
detailed clinical diagnostic procedures.
In sum, however, given the special nature of the
population investigated in the present study (a small
but very frail population, in which about one half
were not able to provide reliable information
themselves), one has to put up with most of these
limitations as most of the centenarian studies have
done (Martin et al., 2000; Silver et al., 2001).
Therefore, despite the present limitations, the results
obtained provide remarkable initial evidence for the
conclusion that both early school education and the
extent of maintaining at least some life-long intellec-
tual activities are related to cognitive impairment in
late life. Particularly, the findings on the effect of
intellectual activities also seem to have practical
implications. Despite the preliminary nature of the
data, results seem to suggest a positive effect of
cognitive stimulation for cognitive functioning even
in very old age.
ALLARD, M., ROBINE, J.-M. & HENON, N. (2000). Medical
predictors of survival among centenarians. In P. MARTIN,
C. ROTT, B. HAGBERG & K. MORGAN (Eds.), Autonomy
versus dependence in the oldest old (pp. 41–59). Paris:
ANDERSEN-RANBERG, K., VASEGAARD, L. & JEUNE, B.
(2001). Dementia is not inevitable: a population-based
study of Danish centenarians. Journal of Gerontology:
Psychological Sciences, 56B, P152–P159.
ARFKEN, C.L., LICHTENBERG, P.A. & TANCER, M.E. (1999).
Cognitive impairment and depression predict mortality
in medially ill older adults. Journal of Gerontology:
Medical Sciences, 54A, M152–M156.
BALTES, P.B., MAYER, KU., HELMCHEN, H. & STEINHAGEN-
THIESSEN, E. (1999). The Berlin Aging Study (BASE):
sample, design, and overview. In P.B. BALTES &
K.U. MAYER (Eds.), The Berlin Aging Study: Aging from
70 to 100 (pp. 15–55). Cambridge, UK: Cambridge
BALTES, P.B. & SMITH, J. (1999). Multilevel and systemic
analyses of old age: theoretical and empirical evidence
for a fourth age. In V.L. BENGTSON & K.W. SCHAIE
(Eds.), Handbook of theories of aging (pp. 153–173).
New York: Springer.
BARON, R.M. & KENNY, D.A. (1986). The moderator-
mediator variable distinction in social psychological
research: conceptual, strategic and statistical considera-
tions. Journal of Personality and Social Psychology, 51,
BOSWORTH, H.B., SCHAIE, K.W. & WILLIS, S.L. (1999).
Cognitive and socio-demographic risk factors for mor-
tality in the Seattle Longitudinal Study. Journal of
Gerontology: Psychological Sciences, 54B, P273–P282.
KHACHATURIAN, Z.S. (1998). Alzheimer’s disease: etiol-
ogies, pathophysiology, cognitive reserve, and treatment
opportunities. Neurology, 51(Suppl. 1), S2–S17.
DEMARIS, A. (1992). Logit modeling: practical applications.
Thousand Oaks, CA: Sage.
ERBER, J.R. & SZUCHMAN, L.T. (1996). Memory perfor-
mance in relation to age, verbal ability, and activity.
Experimental Aging Research, 22, 59–72.
EVANS, D.A., HERBERT, L.E., BECKETT, L.A., SCHERR, P.A.,
ALBERT, M.S., CHOWN, M.J., PILGRIM, D.M. & TAYLOR,
J.O. (1997). Education and other measures of socio-
economic status and risk of incident Alzheimer disease
in a defined population of older persons. Archives of
Neurology, 54, 1399–1405.
FILLENBAUM, G.G. (1988). Multidimensional functional
assessment of older adults: the Duke Older Americans
FOLSTEIN, M.F., FOLSTEIN, S.E. & MCHUGH, P.R. (1975).
Mini-Mental-State: a practical method for grading the
cognitive state of patients for the clinician. Journal of
Psychiatric Research, 12, 189–198.
GATZ, M., SVEDBERG, P., PEDERSEN, N.L., MORTIMER, J.A.,
BERG, S. & JOHANSSON, B. (2001). Education and the risk
of Alzheimer’s disease: findings from the study of
dementia in Swedish Twins. Journal of Gerontology:
Psychological Sciences, 56B, P292–P300.
HAGBERG, B., BAUER ALFREDSON, B., POON L.W. &
HOMMA, A. (2001). Cognitive functioning in centenar-
ians: a coordinated analysis of results from three
countries. Journal of Gerontology: Psychological Sciences,
HILL, R.D., WAHLIN, A., WINBLAD, B. & BA¨CKMAN, L.
(1995). The role of demographic and lifestyle variables
in utilizing cognitive support for episodic remembering
M. Kliegel et al.
among Download full-text
Psychological Sciences, 50B, P219–P227.
HOLTSBERG, P.A., POON, L.W., NOBLE, C.A. & MARTIN, P.
(1995). Mini-mental state exam status of community-
dwelling cognitively intact centenarians. International
Psychogeriatrics, 7, 417–427.
HULTSCH, D.F., HAMMER, M. & SMALL, B. (1993).
Age differences in cognitive performance in later
life: relationships to self-reported health and activity
life style. Journal of Gerontology: Psychological Sciences,
HULTSCH, D.F., HERTZOG, C., DIXON, R.A. & SMALL, B.J.
Cambridge University Press.
HULTSCH, D.F., HERTZOG, C., SMALL, B.J. & DIXON, R.A.
(1999). Use it or lose it: engaged lifestyle as a buffer of
cognitive decline in aging? Psychology and Aging, 14,
JEUNE, B. & KANNISTO, V. (1997). Emergence of centenar-
ians and super-centenarians. In J.M. ROBINE, J.W.
VAUPEL, B. JEUNE, & M. ALLARD (Eds.), Longevity: to
the limits and beyond (pp. 77–89). Berlin: Springer.
JORM, A.F., KORTEN, A.E. & HENDERSON, A.S. (1987).
The prevalence of dementia: a quantitative integration
of the literature. Acta Psychiatrica Scandinavica, 76,
KANNISTO, V. (1996). The advancing frontier of survival: life
tables for old age. Odense Monographs on Population
Aging, 3. Odense, Denmark: Odense University Press.
Electronic ed.: http://www.demogr.mpg.de/Papers/Books/
KANNISTO, V., LAURITSEN, J., THATCHER, A.R. & VAUPEL,
J.W. (1994). Reductions in mortality at advanced ages:
several decades ofevidence
Population and Development Review, 20, 793–810.
KATZMAN, R. (1993). Education and the prevalence of
KLIEGEL, M. ROTT, C., D’HEUREUSE, V., BECKER, G. &
SCHO¨NEMANN, P. (2001). Demenz im ho ¨chsten Alter ist
Hundertja ¨hrigen-Studie. (Dementia in the oldest old is
no necessity: Results from the Heidelberg Centenarian
Study.) Zeitschrift fu ¨r Gerontopsychologie und Psychiatrie,
LAWTON, M.P., MOSS, M., HOFFMAN, C., GRANT, R.,
HAVE, T.T. & KLEBAN, M.H. (1999). Health, valuation
of life, and wish to live. The Gerontologist, 39, 406–416.
MACKINNON, D.P., WARSI, G. & DWYER, J.H. (1995). A
simulation study ofmediated
Multivariate Behavioral Research, 30, 41–62.
MAIER, H. & SMITH, J. (1999). Psychological predictors of
mortality in old age. Journal of Gerontology: Psychological
Sciences, 54B, P44–P54.
MARTIN, P., ROTT, C., HAGBERG, B. & MORGAN, K. (Eds.).
(2000). Autonomy vs. dependence in the oldest old. New
MOCERI, W.M., KUKULL, W.A., EMANUEL, I., VAN BELLE, G.
& LARSON, E.B. (2000). Early life risk factors and
the development of Alzheimer’s disease. Neurology, 54,
in the aged. New York:
MORTIMER, J.A. (1997). Brain reserve and the clinical
expression of Alzheimer’s disease. Geriatrics, 52(Suppl. 2),
NIELSEN, H., LOLK, A., ANDERSEN, K., ANDERSEN, J. &
KRAGH-SORENSEN, P. (1999). Characteristics of elderly
who develop Alzheimer’s disease during the next two
years: a neuropsychological study using CAMCOG: the
Psychiatry, 14, 957–963.
ORRELL, M. & SAHAKIAN, B. (1995). Education and
dementia: research evidence supports the concept of
‘use it or lose it.’ British Medical Journal, 310, 951–952.
ROTT, C., D’HEUREUSE, V., KLIEGEL, M., SCHO¨NEMANN, P. &
BECKER, G. (2001). Die Heidelberger Hundertja ¨hrigen-
Studie: Theoretische und methodische Grundlagen zur
sozialwissenschaftlichen Hochaltrigkeitsforschung. (The
Heidelberg Centenarian Study: theoretical and metho-
dological foundations of psychosocial research in the
oldest old.) Zeitschrift fu ¨r Gerontologie und Geriatrie, 34,
SALTHOUSE, T.A. (1991). Theoretical perspectives on cognitive
aging. Hillsdale, NJ: Erlbaum.
SATZ, P. (1993). Brain reserve capacity on symptom onset
after brain injury: a formulation and review of evidence
for threshold theory. Neuropsychology, 7, 273–295.
SCHOOLER, C. & MULATU, M.S. (2001). The reciprocal
effects of leisure time activities and intellectual function-
ing in older people: a longitudinal analysis. Psychology
and Aging, 16, 466–482.
SCHOOLER, C., MULATU, M.S. & OATES, G. (1999). The
continuing effects of substantively complex work on the
intellectual functioning of older workers. Psychology and
Aging, 14, 483–506.
SILVER, M.H., JILINSKAIA, E. & PERLS, T.T. (2001).
Cognitive functional status of age-confirmed centenar-
ians in a population-based study. Journal of Gerontology:
Psychological Sciences, 56B, P134–P140.
SNOWDON, D.A., KEMPER, S.J., MORTIMER, J.A., GREINER,
L.H., WEKSTEIN, D.R. & MARKESBERY, W.R. (1996).
Linguistic ability in early life and cognitive function and
Alzheimer’s disease in late life: findings from the Nun
Study. Journal of the American Medical Association, 275,
THOMASSEN, R., VAN SCHAICK, H.W. & BLANSJAAR, B.A.
(1998). Prevalence of dementia over age 100. Neurology,
VERGHESE, J., LIPTON, R.B., KATZ, M.J., HALL, C.B.,
DERBY, C.A., KUSLANSKY, G., et al. (2003). Leisure
activities and the risk of dementia in the elderly. The New
England Journal of Medicine, 348, 2508–2516.
VAUPEL, J.W. (2000). Setting the stage: a generation of
centenarians? The Washington Quarterly, 23, 197–200.
WOLFSON, C., WOLFSON, D.B., ASGHARIAN, M., M’LAN,
C.E., OSTBYE, T., ROCKWOOD, K., et al. (2001). A
re-evaluation of the duration of survival after the onset of
dementia. The New England Journal of Medicine, 344,
ZHANG, M., KATZMAN, R., YU, E., LIU, W., XIAO, S.F.
& YAN, H. (1998). A preliminary analysis of incidence
of dementia in Shanghai, China. Psychiatry and Clinical
Neurosciences, 52(Suppl.), S291–S294.
Cognitive impairment in centenarians