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

net/publication/7450441

Mental, Physical and Social Components in Leisure Activities Equally Contribute

to Decrease Dementia Risk

Article  in  Dementia and Geriatric Cognitive Disorders · February 2006

DOI: 10.1159/000089919 · Source: PubMed

CITATIONS READS

381 2,310

6 authors, including:

Anita Karp Stéphanie Paillard-Borg

Karolinska Institutet Karolinska Institutet
31 PUBLICATIONS   3,430 CITATIONS    7 PUBLICATIONS   1,665 CITATIONS   

SEE PROFILE SEE PROFILE

Hui-Xin Wang Merril Silverstein

Stockholm University Syracuse University
142 PUBLICATIONS   6,963 CITATIONS    108 PUBLICATIONS   8,298 CITATIONS   

SEE PROFILE SEE PROFILE

Some of the authors of this publication are also working on these related projects:

Design and development of molecular chaperone modulators for the treatment of neuro-degenerative disorders. View project

The MOPEAD Project: sEvolving models of patient engagement and access for earlier identification of Alzheimer’s disease View project

All content following this page was uploaded by Hui-Xin Wang on 05 November 2015.

The user has requested enhancement of the downloaded file.

 Original Research Article

Dement Geriatr Cogn Disord 2006;21:65–73 Accepted after revision: September 3, 2005
Published online: November 23, 2005
DOI: 10.1159/000089919

Mental, Physical and Social Components in

Leisure Activities Equally Contribute to Decrease
Dementia Risk

Anita Karp a Stephanie Paillard-Borg a Hui-Xin Wang a Merril Silverstein a, b

Bengt Winblad a Laura Fratiglioni a
a
Aging Research Center, Division of Geriatric Epidemiology, Department of Neurotec, Karolinska Institutet, and
the Stockholm Gerontology Research Center, Stockholm, Sweden; b Andrus Gerontology Center,
University of Southern California, Los Angeles, Calif., USA

Key Words
Dementia risk
Leisure activities, mental, physical and
social components
Elderly population
Abstract
Background: There is accumulating evidence in the lit-
subjects with high scores in all or in two of the compo-
nents (RR of dementia = 0.53; 95% CI: 0.36–0.78). Conclu-
sions: These findings suggest that a broad spectrum of
activities containing more than one of the components
seems to be more beneficial than to be engaged in only
one type of activity.
Copyright © 2006 S. Karger AG, Basel

erature that leisure engagement has a beneficial effect

on dementia. Most studies have grouped activities ac-
cording to whether they were predominantly mental,
physical or social. Since many activities contain more
than one component, we aimed to verify the effect of all
three major components on the dementia risk, as well as
their combined effect. Methods: A mental, social and
physical component score was estimated for each activ-
ity by the researchers and a sample of elderly persons.
The correlation between the ratings of the authors and
the means of the elderly subjects’ ratings was 0.86. The
study population consisted of 776 nondemented sub-
jects, aged 75 years and above, living in Stockholm, Swe-
den, who were still nondemented after 3 years and were
followed for 3 more years to detect incident dementia
cases. Results: Multi-adjusted relative risks (RRs) of de-
mentia for subjects with higher mental, physical and so-
cial component score sums were 0.71 (95% CI: 0.49–1.03),
0.61 (95% CI: 0.42–0.87) and 0.68 (95% CI: 0.47–0.99), re-
spectively. The most beneficial effect was present for
Introduction
After retirement, leisure activities usually constitute a
relatively larger part of the daily life and may take on the
role of providing mental stimulation, social engagement
and physical activity that was earlier offered by school or
employment. Consequently, leisure activities in old age
have been a common focus for studies regarding several
outcomes like mortality [1–3], cognition [4–8], morbidity
[9] and well-being [10–12].
In recent years, prospective studies have found an as-
sociation between leisure activities and decreased risk
of Alzheimer’s disease (AD) and dementia [13]. It has
been suggested that engagement in leisure activities may
result in functionally more efficient cognitive networks,
hereby providing a cognitive reserve that could delay the
onset of clinical manifestations of dementia [14]. Men-
tally stimulating activities, in particular, have been as-

© 2006 S. Karger AG, Basel Anita Karp

1420–8008/06/0212–0065$23.50/0 ARC-Äldrecentrum
Fax +41 61 306 12 34 Box 6401 (Olivecronas väg 4)
E-Mail karger@karger.ch Accessible online at: SE–113 82 Stockholm (Sweden)
www.karger.com www.karger.com/dem Tel. +46 8 6906856, Fax +46 8 6906889, E-Mail anita.karp@neurotec.ki.se
sociated with significantly reduced risk of incident de-
mentia [15–19]. Being engaged in social recreational ac-
tivities [18, 20] and having a rich social network [21]
have also been found to have a protective effect against
dementia. In addition, engaging in regular physical ac-
tivity has been suggested as an important component of
a preventive strategy against AD [22] and, in women,
against cognitive decline at 6–8 years of follow-up [23].
However, other reports did not confirm such an associa-
tion [16, 18].
Most studies of leisure activities and dementia have
grouped activities according to whether they were pre-
dominantly mental, physical or social. Indeed, activities
may simultaneously embody one or more of these three
aspects. As an example, dancing [19] is a physical activ-
ity that contains social interaction, as well as a certain
degree of cognitive involvement, at least if the dance re-
quires learning new steps and turns. Hence, identifying
the activities according to their predominant component
may lead to an underestimation of other components. For
instance, several common leisure activities classified as
social or mental in nature also contain light physical ac-
tivity which, although modest, may be of great impor-
tance for the health of elderly people.
For these reasons, the present study considers activi-
ties as having a multidimensional profile, each manifest-
ing various combinations of mental, social and physical
involvement. Using data from subjects in the Kungshol-
men Project, we hypothesize that all three components
are relevant protective factors against dementia, and that
a combined beneficial effect may be present. Because re-
duced participation in leisure activities may very well be
a consequence and not a cause of cognitive decline in the
preclinical phase of dementia [16, 18, 19], the present
study lags predictors relative to dementia diagnosis by an
average of 6 years through excluding subjects diagnosed
as demented at first follow-up. Furthermore, we excluded
subjects with low baseline cognition, and adjusted for
cognitive function at baseline.
Materials and Methods
Study Population
Subjects were derived from the Kungsholmen Project, a longi-
tudinal population-based study of aging and dementia. All resi-
dents of the Kungsholmen district of Stockholm, Sweden, aged 75
years or older in October 1987 (n = 2,368), were asked to participate
in the initial examination. Subjects who agreed and signed an in-
formed consent were interviewed by nurses, examined by physi-
cians, and tested by psychologists during the baseline (1987–1989),
66 Dement Geriatr Cogn Disord 2006;21:65–73
first follow-up (1991–1993) and second follow-up (1994–1996) ex-
aminations.
Of the 1,810 eligible participants who underwent the baseline
examination, 1,473 were diagnosed as nondemented. The detailed
procedures are available elsewhere [24, 25]. Since institutionaliza-
tion or impaired cognition may limit subjects’ activity [4], 98 sub-
jects whose Mini-Mental State Examination (MMSE) scores [26]
were less or equal to 23 or who were living in an institution were
also excluded from the present study; thus, 1,375 persons remained
for analysis. By the first follow-up examination, 269 subjects had
died, 172 refused participation and 934 participated, of whom 158
were diagnosed as demented. Thus, the population for the present
study was composed of those 776 subjects participating and still
nondemented at the first follow-up examination. Out of these, 732
subjects were followed for another 3 years (second follow-up) to
detect incident dementia cases (44 subjects refused to participate
to the second follow-up).
Diagnosis of Dementia
Incident dementia cases examined in this study were those
subjects who developed dementia during the second follow-up
period. At each follow-up examination, all cohort subjects were
clinically examined following a standardized protocol, which in-
cluded personal and family history collected by nurses, clinical
examination conducted by physicians and psychological tests ad-
ministered by trained personnel. If a participant was not able to
answer, an informant, usually the subject’s next of kin, was inter-
viewed. The study design has been described in detail elsewhere
[24, 25]. Time of dementia onset was assumed to be the midpoint
between the first and second follow-up examinations or the time
of death.
Dementia diagnoses were made according to the Diagnostic and
Statistical Manual of Mental Disorders, Third Edition-Revised cri-
teria [26] following the same three-step procedure used at baseline
examination [24]. First, a preliminary diagnosis was made by the
examining physician. Second, all cases were independently re-
viewed by a specialized clinician and a second diagnosis was made.
If the diagnoses were in agreement, they were accepted as the final
decision. In case of disagreement, a third opinion was obtained and
the concordant diagnosis was accepted. When only one item of the
diagnostic criteria was not fulfilled, the subject was classified as af-
fected by ‘questionable dementia’, which is in contrast with ‘clini-
cally definite dementia’. Both groups were included in the analyses.
For those participants (n = 172) who had died between the first and
second follow-up examinations, a preliminary diagnosis was made
by a physician through consulting medical records and death cer-
tificates, and then reviewed by a senior clinician. When only dis-
charge diagnoses from hospitals or death certificates were available,
the reported diagnosis was accepted.
Assessment of Activity
Information on leisure activities was obtained from the subjects
by means of a personal interview carried out by trained nurses at
baseline (1987–1989) with open questions. Subjects were asked:
whether they regularly engaged in any particular activities, the type
of activities and the frequency of participation. The leisure activi-
ties were grouped into 29 main types of activities. A mental, social
and physical component score was assigned to each of the 29 ac-
tivities. Two authors of this study (A.K. and S.P.-B.) independent-
ly assigned scores to each activity based on their own evaluations
Karp /Paillard-Borg /Wang /Silverstein /
Winblad /Fratiglioni
 Table 1. List of activities, scorings (0 = none; 1 = low; 2 = moderate; 3 = high) and frequency of participation

Type of activity Scores Subjects

mental physical social n %

Reading literature 3 0 0 152 20.8

Handicraft: needlework, weaving, knitting 2 1 1 146 19.9
Doing crossword puzzles 3 0 0 94 12.8
Political or cultural interests 3 1 3 75 10.2
Playing cards or chess 3 0 3 71 9.7
Visiting the summerhouse 2 2 2 70 9.6
Attending courses 3 1 2 59 8.1
Watching TV 2 0 0 55 7.5
Going to theatres or concerts 3 1 2 55 7.5
Doing sport 1 3 2 49 6.7
Going to exhibitions or museums 3 1 1 47 6.4
Meeting friends, participating in groups 2 1 3 47 6.4
Walking 1 3 1 37 5.1
Listening to radio 2 0 0 33 4.5
Travelling 3 2 2 29 4.0
Gardening and flowers 2 2 1 29 4.0
Painting, drawing, photo 3 0 0 26 3.6
Engaging in family or charity 2 2 3 23 3.1
Doing outdoor activities 2 3 2 23 3.1
Collecting stamps or other items 3 0 0 18 2.5
Cooking food 2 1 1 17 2.3
Writing 3 0 1 16 2.2
Housekeeping 1 2 0 15 2.0
Attending church activities 2 1 3 15 2.0
Playing music 2 1 2 14 1.9
Doing solitaire 2 0 0 9 1.2
Following the stock market 3 0 0 8 1.1
Playing bingo 2 1 3 7 1.0
Singing 2 1 2 7 1.0
No activity at all 0 0 0 153 20.9

and then discussed with a third author (L.F.) in order to reach con-
sensus. The grading of the three components was coded as: 0 = none,
1 = low, 2 = moderate, 3 = high (table 1).
To validate the scoring, 13 cognitively intact, elderly raters (7
men, 6 women), aged 75 years or more, but not participants in the
Kungsholmen study, were asked to individually fill in a small ques-
tionnaire containing a list of all 29 activities together with scoring
instructions. The overall correlation between the consensus ratings
of the authors and the means of the elderly subjects’ ratings was
0.86. Reliability analyses were performed. For all three components
taken together, Cronbach’s
was high; 0.90 indicating high inter-
rater consistency. Cronbach’s
values for each of the components
were: 0.89 for the mental component, 0.95 for the physical compo-
nent and 0.82 for the social component. Since the disagreements
between researchers and elderly raters mainly concerned the social
component of activities, four final ratings (handicraft, travels, sing-
ing and collections) were changed to better fit the elderly raters’
opinion. Table 1 shows the final decision regarding the component
scores for the 29 activities and the frequency of each activity.
The three components were highly correlated. The Pearson cor-
relation coefficients were 0.76 between the mental and social com-
ponent score sums, 0.77 between the physical and social and 0.58
between the physical and mental component score sums.
Covariates
We used all covariates with available baseline information from
our database that may have acted as possible confounders: age, sex,
education, cognitive functioning, comorbidity, depressive symp-
toms and physical function at baseline examination. Information
about age and sex was derived from the National Population Reg-
ister. Information about highest level of formal education was col-
lected through personal interview. Cognitive function was mea-
sured by using MMSE [27] since there is a global impairment of
cognitive functioning in the preclinical phase of AD [28]. Depres-
sive symptoms were assessed by two self-reported symptoms: being
in a low mood and/or often feeling lonely. Both of these symptoms
are included in the Center for Epidemiological Studies Depression
Scale [29]. Physical function was defined as disability in at least

Mental, Physical and Social Components Dement Geriatr Cogn Disord 2006;21:65–73 67
in Leisure Activities and Dementia Risk
one of the basic activities of daily living: bathing, dressing, toilet- Table 2. Baseline characteristics of the study population (n = 732)
ing, continence, feeding or transfer [30]. and of the incident dementia cases detected at second follow-up
Data of previous diseases were obtained by reviewing hospital (1994–1996)
discharge diagnoses through the Stockholm Computerized Inpa-
tient Register System. These diagnoses were made according to Characteristic Participants Incident dementia
the International Classification of Disease, 8th edition [31]: coro- (n = 732) cases (n = 123)
nary heart disease (ICD-8: 410–414), cerebrovascular disease
(ICD-8: 430–438), diabetes mellitus (ICD-8: 250), malignancy n n %
(ICD-8: 140–208 and 230–239) and hip fracture (ICD-8: 820).
Comorbidity was defined as subjects who had any of these five Age groups
diseases. 75–79 364 36 9.9
80–84 232 55 23.7
Statistical Analyses 85+ 136 32 23.5
To evaluate the differences in baseline characteristics between Sex
participants and drop-outs, logistic regression was used. Cox pro- Female 543 93 17.1
portional hazards regressions were performed to estimate the rela- Male 189 30 15.9
tive risk (RR) of incident AD and dementia associated with mental,
physical and social components in the different leisure activities. Education1
The mental, physical and social components were analysed in ^7 years 386 70 18.1
the following ways: 68 years 343 53 15.5
(1) We first intended to tap the intensity of involvement in each MMSE score
component by using a threshold to signify moderate/high involve- 24–26 178 38 21.3
ment: for each person and each component (mental, physical and 27–30 554 85 15.3
social), the number of moderately/highly (2–3) scored activities
were summed up separately and then divided into three categories. Comorbidity1
The analyses were performed for each component, contrasting sub- Yes 179 46 25.7
jects with one, and subjects with two or more moderately/highly No 550 77 14.0
scored activities with subjects without any moderately/highly
Depressive symptom1
scored activity.
Yes 201 53 26.4
(2) In order to measure the accumulation of each component
No 528 70 13.3
across the range of activities, the rated scores were added to a sum
of scores for each person and each component. This sum score was Physical dependence2
analysed using (a) the continuous variable, (b) four groups with a Yes 132 30 22.7
similar number of cases and (c) two groups dichotomized according No 592 93 15.7
to the median value.
(3) To capture the combined effect of the three components in Number of activities
reducing the dementia risk, combinations of the mental, physical 0 153 37 24.2
and social components were created using the sum score for each 1–2 400 67 16.8
component after dichotomization according to the median value. 3–7 179 19 10.6
All of the associations studied were first assessed using simple 1
models including age and sex, and then adjustments for all covari- For 3 subjects, information was missing.
2
ates described above were carried out. Age and MMSE score were For 8 subjects, information was missing.
entered into the models as continuous variables. Education was
entered as a categorical variable (112 years, 8–12 years and
!8 years of schooling). Sex (female vs. male), comorbidity (the
presence of coronary heart disease, cerebrovascular disease, dia-
betes mellitus, malignancy or hip fracture vs. the presence of none
of them), depressive symptoms (yes vs. no) and physical function- sex, education, cognitive functioning, depressive symp-
ing (dependent vs. independent) were entered as dichotomous toms and physical functioning.
variables. A total of 123 subjects developed clinical dementia
during the second follow-up. Table 2 shows the baseline
characteristics of the study population and incident de-
Results mentia cases.
The multi-adjusted RRs of dementia associated with
The baseline study population consisted of 776 sub- participation in 1–2 activities, and participation in
jects, but 44 persons refused second follow-up examina- 3–7 activities, versus no participation were 0.77 (95%
tion. While non-participants had less comorbidity than CI: 0.51–1.15) and 0.56 (95% CI: 0.31–0.99), respec-
participants, there were no differences with respect to age, tively.

68 Dement Geriatr Cogn Disord 2006;21:65–73 Karp /Paillard-Borg /Wang /Silverstein /

Winblad /Fratiglioni
 Table 3. RRs of dementia in relation to degree of estimated mental, physical and social component score sums
in baseline leisure activities

n Dementia RR (95% CI)1 RR (95% CI)2

cases

Mental component score sum

0 153 37 1.0 1.0
1–3 207 38 0.75 (0.47–1.18) 0.82 (0.52–1.31)
4–7 251 34 0.61 (0.38–0.98) 0.67 (0.42–1.08)
8 or more 121 14 0.49 (0.26–0.93) 0.56 (0.29–1.07)
Higher score (0–3 vs. 4–18) 372 48 0.67 (0.46–0.98) 0.71 (0.49–1.03)
Physical component score sum
0 285 67 1.0 1.0
1 165 22 0.55 (0.34–0.89) 0.58 (0.36–0.95)
2–3 182 24 0.61 (0.38–0.99) 0.64 (0.40–1.04)
4 or more 100 10 0.51 (0.26–0.99) 0.59 (0.30–1.16)
Higher score (0 vs. 1–12) 447 56 0.57 (0.40–0.81) 0.61 (0.42–0.87)
Social component score sum
0 261 62 1.0 1.0
1 103 13 0.49 (0.27–0.90) 0.53 (0.29–0.97)
2–3 179 23 0.48 (0.29–0.78) 0.51 (0.31–0.82)
4 or more 189 25 0.61 (0.38–0.98) 0.69 (0.43–1.13)
Higher score (0–1 vs. 2–13) 368 48 0.64 (0.44–0.93) 0.68 (0.47–0.99)

For each component (mental, physical and social) two Cox regression models were performed: one analysing
the component score sum as a 4-category indicator variable, and the second as a dichotomous variable.
1
Adjusted for age, sex, education, baseline MMSE score, comorbidity and physical functioning.
2
Adjusted for age, sex, education, baseline MMSE score, comorbidity, physical functioning and depressive
symptoms.

To examine if moderate/high-scored activities were
most beneficial, the number of moderately/highly scored
activities were added up and then divided into three cat-
egories. In comparison to subjects with none or low (0–1)
involvement in the mental component, persons who had
one moderately/highly (2–3) rated activity had a multi-
adjusted RR of 0.90 (95% CI: 0.57–1.41), and persons
with two or more moderately/highly rated activities had
an RR of 0.67 (95% CI: 0.42–1.05). Regarding the phys-
ical component, the multi-adjusted RR of dementia was
0.87 (95% CI: 0.55–1.39) for subjects engaged in one
moderately/highly rated activity and 0.44 (95% CI: 0.14–
1.39) for being engaged in two or more moderately/high-
ly rated activities. For the social component, the multi-
adjusted RR associated with engagement in one moder-
ately/highly rated activity was 0.75 (95% CI: 0.48–1.18)
and the RR associated with two or more moderately/high-
ly rated activities was 0.94 (95% CI: 0.55–1.61).
Next, the total scores of each of the three components
were examined. The mental score sums ranged from 0 to
18, the physical score sums from 0 to 12 and the social
score sums from 0 to 13. The mental component differed
from the physical and social components in that it had no
subjects with a 0 score, aside from the group of subjects
who participated in no activities at all. All of the three
components’ score sum distributions were positively
skewed.
When first analysed as continuous variables, multi-ad-
justed RRs indicated that greater score sums in the men-
tal component (RR = 0.94, 95% CI: 0.87–1.00) and phys-
ical component (RR = 0.89, 95% CI: 0.77–0.99) appeared
to lower the risk of dementia. Higher score sums in the
social component were not significantly associated with
lower risk of dementia when analysed continuously
(RR = 0.95, 95% CI: 0.87–1.04).
All of the components’ score sum distributions were
then categorized into approximate quartiles (table 3).
Having mental component score sums above 3 was sig-
nificantly associated with a decreased risk of dementia,
when controlled for age, sex, education, baseline MMSE

Mental, Physical and Social Components Dement Geriatr Cogn Disord 2006;21:65–73 69
in Leisure Activities and Dementia Risk
 Table 4. RRs of dementia associated with combinations of higher or lower mental, physical and social score
sums
Lower score in all of the mental,
physical and social components
Higher score in one of the mental,
physical or social components
Higher score in two of the mental,
physical or social components
Higher score in all of the mental,
physical and social components
1
Adjusted for age, sex, education, baseline MMSE score, comorbidity and physical functioning.
2
Adjusted for age, sex, education, baseline MMSE score, comorbidity, physical functioning and depressive
symptoms.
score, comorbidity and physical functioning. When the
depressive symptoms (feeling alone and in a low mood)
were added to the model, the association did not reach
the 0.05 level of significance. The mental component
score sum, categorized in four grades, showed a signifi-
cant trend (p = 0.043) in relation to dementia risk. Having
a physical and social component score larger than or
equal to 1 was significantly related to a lower risk of de-
mentia after adjustments for all covariates. There was no
trend with regard to the four score sum categories for the
physical or the social components. The basic adjusted
(only age and sex adjustments) results were similar.
Table 3 also shows the RRs of dementia in relation to
the sum of scores of mental, physical and social compo-
nents using the approximate medians for dichotomiza-
tion. The median for the mental component score was 4,
for the physical component 1 and for the social compo-
nent 2. When analysed one by one, all three components
were protective against dementia, although the mental
score was not entirely significant when adjusted for all
covariates.
We further combined the dichotomized score sums of
the three components into eight groups, with the purpose
of investigating if some specific activity combinations
were more beneficial. The dichotomizations were based
on the median values described above. These results (not
shown) indicated that having high overall scores on all
three components was associated with significantly lower
risks of dementia when all covariates were taken into ac-
count. In two combinations, subjects who scored high on
70 Dement Geriatr Cogn Disord 2006;21:65–73
Subjects Cases RR (95% CI)1 RR (95% CI)2
228 57 1.0 1.0
92 15 0.63 (0.36–1.12) 0.69 (0.39–1.22)
141 16 0.37 (0.21–0.65) 0.39 (0.22–0.69)
271 35 0.57 (0.37–0.89) 0.63 (0.41–0.98)
only two of the three components (mental and physical;
physical and social) were also significantly protected
against dementia. Due to the small number of cases in
some of the combinations, subjects were merged into four
categories: (1) low in all three components, (2) one high,
(3) two high, or (4) three high. The results are shown in
table 4. Having high scores in two or three of the compo-
nents was associated with a significant reduction in risk
of dementia. When those having higher scores on two or
more of the components were combined (category 3 and
4 together), the RR ratio fell to almost half (RR = 0.53,
95% CI: 0.36–0.78).
The following additional analyses were performed to
verify the results: although we had excluded subjects with
MMSE scores ^23, scores above that level could still in-
fluence the studied relation. We therefore performed ad-
ditional analyses for subjects whose MMSE scores were
more than 26 at baseline examination. The results were
largely identical for the mental and physical component
score sums and showed a similar tendency for the social
component score sum. Finally, since there were quite a
few subjects (n = 153) who did not report any activity at
all and might have been qualitatively different from the
other combinations, additional analyses were performed
treating these subjects as a separate category. The results
and overall pattern did not change, and the ‘no activity
group’ had similar RRs as the group who had a lower
score in all three components. We also performed extra
analyses excluding the subjects who did not report any
activities. Similar results were observed.
Karp /Paillard-Borg /Wang /Silverstein /
Winblad /Fratiglioni
 Discussion
This study shows that participation in leisure activities
is associated with reduced risk of dementia and that each
of the mental, physical and social components is of im-
portance. When an intensity threshold was used, compar-
ing the number of activities that rated moderate to high
on the components versus no or low-rated activities, the
results were not statistically significant. However, when
the subjects’ score sums for each component were taken
into account, the RRs of dementia were significantly de-
creased, showing that even small contributions of the
mental, physical or social components mattered when ac-
cumulated across several activities. This was especially
relevant for the physical component, since only few el-
derly had two or more activities with moderate to high
intensity.
The most beneficial effect was, however, present for
subjects with high scores in all or in two of the compo-
nents. In the latter group, nearly half of the subjects had
reported only one activity. Examples of these single ac-
tivities scoring high in more than one component were:
being engaged in political and cultural activities, visiting
the summerhouse, attending courses, going to the theatre
or concerts, being active in sport or outdoor activities,
travelling, being engaged in charity or church activities
or playing music together with others.
Although other researchers [8, 19] have acknowledged
the fact that most leisure activities consist of several over-
lapping components, nobody, to our knowledge, has tried
to separate the different components from each other.
This study used a new approach, where three of the au-
thors (a neurologist, a psychologist and a sociologist), and
13 elderly raters scored the mental, physical and social
components in each of the 29 common leisure activities
reported by the participants. In addition, participation in
the leisure activities was assessed on average 6 years be-
fore dementia diagnosis in an attempt to reduce the influ-
ence of preclinical dementia. None of the covariates (age,
sex, education, baseline MMSE score, comorbidity and
physical function) could explain the reported associa-
tions. ‘Depressive symptoms’ was the only control vari-
able that marginally changed the RRs of dementia – de-
spite the crude estimation of depressive symptomatology
in the present study.
There are a number of different hypotheses concerning
the possible effect of each of the mental, physical and so-
cial dimensions of common leisure activities on dementia
development. Activation of the nerve cells might prolong
their optimal function throughout the life-span [32].
Mental, Physical and Social Components
in Leisure Activities and Dementia Risk
Mentally stimulating activities may buffer or delay path-
ological development. Persistent engagement in effortful
mental activities may even promote plastic changes in the
brain that can circumvent the underlying dementia pa-
thology [33]. Physical activity may increase cerebral oxy-
genation, leading to improved neurotransmitter metabo-
lism. It also reduces the risk of diseases that may be risk
factors for dementia such as hypertension, diabetes and
cardiovascular disease [8]. Predominantly social leisure
activities may affect the immune system [34], which in
turn could influence inflammatory processes in the brain,
which may be involved in dementia.
When interpreting these findings, one needs to take
into account some considerations. First, open-ended
questions were used to obtain information of engagement
in leisure activities. Although these types of questions can
be of great value because they capture a wide variety of
possible and sometimes unexpected answers [35], there
is always the risk that those who did not report any ac-
tivities were subjects who differed systematically in some
important respect. We performed analyses excluding the
subjects who did not report any activities at all (as well as
treating them as a separate group), and the results showed
a similar tendency as compared to the whole group. Fur-
thermore, we could not determine if any changes in lei-
sure engagement had occurred prior to baseline. Findings
from cross-sectional studies indicate that leisure engage-
ment may decrease with age; however, longitudinal, ret-
rospective studies have reported continuity in overall en-
gagement [36].
Secondly, we do not have access to any ‘objective’ mea-
surements of the three components, although our assess-
ment method had good interrater reliability. It is also
difficult to compare the importance of each component
isolated from the others, since they often occur simulta-
neously in the same activities. The fact that each activity
was used to rate all three components meant that the com-
ponents were not independent, and therefore it was prob-
lematic to definitively adjust for the influence of one com-
ponent on the other. Still, we believe that there is an ad-
vantage in this method reducing all the 29 activities to
three main components, instead of merely grouping the
activities in an exclusive manner, not allowing overlap.
In spite of the imprecision of the assessments and catego-
rizations, we were able to detect risk differences.
Thirdly, there is always the possibility that cognitive
disturbances in the preclinical stages of dementia might
affect the subject’s choice of and engagement in leisure
activities in general and in activities with a high degree
of mental components in particular. Although initial cog-
Dement Geriatr Cogn Disord 2006;21:65–73 71
nitive functioning was adjusted for, and activity data
were collected on average 6 years before dementia diag-
nosis and a minimum of 3 years before dementia onset,
there still exists an uncertainty about how early in the
preclinical phase of dementia relevant cognitive deficits
appear. It is therefore premature to interpret these results
as causal or even as delaying the disease onset, since less-
er leisure engagement may be a consequence of early
symptoms. However, a prospective study as extended as
21 years has recently reported results similar to ours [19].
Furthermore, Bennett et al. [37] investigated post-mor-
tem data in the Religious Order Study and found that the
relation between senile plaques and level of cognitive
function differed by level of education, suggesting that
differences in lifestyle may affect cognitive reserve by
partially mediating the relationship between brain dam-
age and the clinical manifestation of AD. Additionally,
in the MRC National Survey of Health and Develop-
ment, Richards et al. [8] had the opportunity to control
for the influence of IQ at 15 years of age, when studying
active leisure and cognition. This research group drew the
conclusion that it was unlikely that the finding of an as-
sociation between active leisure and cognition was a re-
flection of reverse causality.
This study confirms earlier results stating that engage-
ment in leisure activities with mental, social and physical
content in late life is associated with decreased risk of
dementia. Furthermore, it was found that small contribu-
tions of the mental, physical or social components mat-
tered when accumulated. Because few elderly engaged in
more vigorous exercise, the benefit of light physical com-
ponents in activities that are not primarily physical is
especially noteworthy.
The results also indicate that engaging in activities that
cover more than one of the mental, physical and social
components seems to be more beneficial than to be en-
gaged in only one type of activity. Implications for public
health and community may very well be that making dif-
ferent types of activities more accessible to elderly per-
sons could reduce the risk of developing dementia.
Acknowledgements
We thank all members of the Kungsholmen Project study group
for their cooperation in data collection and management. This re-
search was supported by a grant from the Swedish Council for
Working Life and Social Research (No. 2003-0386) and the Amer-
ican Alzheimer foundation.

References

1 Glass TA, de Leon CM, Marotti RA, Berkman
LF: Population-based study of social and pro-
ductive activities as predictors of survival
among elderly Americans. BMJ 1999; 319:
478–483.
2 Musick MA, Herzog AR, House JS: Volunteer-
ing and mortality among older adults: findings
from a national sample. J Gerontol B Psychol
Sci Soc Sci 1999;54:S173–S180.
3 Lennartsson C, Silverstein M: Does engage-
ment with life enhance survival of elderly peo-
ple in Sweden? The role of social and leisure
activities. J Gerontol B Psychol Sci Soc Sci
2001;25:S335–S342.
4 Hultsch DF, Hammer M, Small BJ: Age differ-
ences in cognitive performance in later life: re-
lationships to self-reported health and activity
life style. J Gerontol B Psychol Sci Soc Sci
1993;48:P1–P11.
5 Hill RD, Wahlin A, Winblad B, Backman L:
The role of demographic and life style variables
in utilizing cognitive support for episodic re-
membering among very old adults. J Gerontol
B Psychol Sci Soc Sci 1995;50:P219–P227.
6 Hultsch DF, Hertzog C, Small BJ, Dixon RA:
Use it or lose it: engaged lifestyle as a buffer of
cognitive decline in aging? Psychol Aging
1999;14:245–263.
7 Laurin D, Verreault R, Lindsay J, MacPherson
K, Rockwood K: Physical activity and risk of
cognitive impairment and dementia in elderly
persons. Arch Neurol 2001;58:498–504.
8 Richards M, Hardy R, Wadsworth ME: Does
active leisure protect cognition? Evidence from
a national birth cohort. Soc Sci Med 2003;56:
785–792.
9 Feskanich D, Willett W, Colditz G: Walking
and leisure-time activity and risk of hip frac-
ture in postmenopausal women. JAMA 2002;
288:300–306.
10 Baltes MM, Carstensen LL: The process of suc-
cessful aging. Ageing Soc 1996;16:397–422.
11 McAuley E, Blissmer B, Marquez DX, Jerome
GJ, Kramer AF, Katula J: Social relations,
physical activity, and well-being in older
adults. Prev Med 2000;31:608–617.
12 Silverstein M, Parker MG: Leisure activities
and quality of life among the oldest old in Swe-
den. Res Aging 2002;24:528–547.
13 Fratiglioni L, Paillard-Borg S, Winblad B: An
active and socially integrated lifestyle in late
life might protect against dementia. Lancet
Neurol 2004;3:343–353.
14 Scarmeas N, Stern Y: Cognitive reserve and
lifestyle. J Clin Exp Neuropsychol 2003; 25:
625–633.
15 Scarmeas N, Levy G, Tang MX, Manly J, Stern
Y: Influence of leisure activity on the incidence
of Alzheimer’s disease. Neurology 2001; 57:
2236–2242.
16 Wilson RS, Mendes de Leon CF, Barnes LL, et
al: Participation in cognitively stimulating ac-
tivities and risk of incident Alzheimer disease.
JAMA 2002;287:742–748.
17 Wilson RS, Bennett DA, Bienias JL, et al: Cog-
nitive activity and incident AD in a popula-
tion-based sample of older persons. Neurology
2002;59:1910–1914.
18 Wang HX, Karp A, Winblad B, Fratiglioni L:
Late-life engagement in social and leisure ac-
tivities is associated with a decreased risk of
dementia: a longitudinal study from the
Kungsholmen Project. Am J Epidemiol 2002;
155:1081–1087.
19 Verghese J, Lipton RB, Katz MJ, et al: Leisure
activities and the risk of dementia in the el-
derly. N Engl J Med 2003;348:2508–2516.
20 Fabrigoule C, Letenneur L, Dartigues JF, Zar-
rouk M, Commenges D, Barberger-Gateau P:
Social and leisure activities and risk of demen-
tia: a prospective longitudinal study. J Am
Geriatr Soc 1995;43:485–490.

72 Dement Geriatr Cogn Disord 2006;21:65–73 Karp /Paillard-Borg /Wang /Silverstein /

Winblad /Fratiglioni
21 Fratiglioni L, Wang HX, Ericsson K, Maytan
M, Winblad B: Influence of social network on
occurrence of dementia: a community-based
longitudinal study. Lancet 2000; 355: 1315–
1319.
22 Lindsay J, Laurin D, Verreault R, et al: Risk
factors for Alzheimer’s disease: a prospective
analysis from the Canadian Study of Health
and Aging. Am J Epidemiol 2002; 156: 445–
453.
23 Yaffe K, Barnes D, Nevitt M, Lui LY, Covin-
sky K: A prospective study of physical activity
and cognitive decline in elderly women: wom-
en who walk. Arch Intern Med 2001; 161:
1703–1708.
24 Fratiglioni L, Grut M, Forsell Y, et al: Preva-
lence of Alzheimer’s disease and other demen-
tias in an elderly urban population: relation-
ship with age, sex, and education. Neurology
1991;41:1886–1892.
25 Fratiglioni L, Viitanen M, von Strauss E, Ton-
todonati V, Herlitz A, Winblad B: Very old
women at highest risk of dementia and Alz-
heimer’s disease: incidence data from the
Kungsholmen Project, Stockholm. Neurology
1997;48:132–138.
Mental, Physical and Social Components
in Leisure Activities and Dementia Risk
View publication stats
26 American Psychiatric Association: Diagnostic
and Statistical Manual of Mental Disorders, ed
3 Revised. Washington, American Psychiatric
Association, 1987.
27 Folstein MF, Folstein SE, McHugh PR: ‘Mini-
Mental State’: a practical method for grading
the cognitive state of patients for the clinician.
J Psychiatr Res 1975;12:189–198.
28 Backman L, Jones S, Berger AK, Laukka EJ,
Small BJ: Cognitive impairment in preclinical
Alzheimer’s disease: a meta-analysis. Neuro-
psychology 2005;19:520–531.
29 Radloff LS, Teri L: Use of the Center for Epi-
demiological Studies Depression Scale with
older adults. Clin Gerontol 1986;5:119–136.
30 Katz S, Ford AB, Moskowitz RW, Jackson BA,
Jaffe MW: The index of ADL: a standardized
measure of biological and psychosocial func-
tion. JAMA 1963;185:914–919.
31 World Health Organization: International
Classification of Diseases. Manual of the Inter-
national Statistical Classification of Diseases,
Injuries, and Causes of Death, ed 8. Geneva,
World Health Organization, 1967, vol 1.
Dement Geriatr Cogn Disord 2006;21:65–73
32 Swaab DF: Brain aging and Alzheimer’s dis-
ease, ‘wear and tear’ versus ‘use it or lose it’.
Neurobiol Aging 1991;12:317–324.
33 Coyle JT: Use it or lose it – Do effortful mental
activities protect against dementia? N Engl J
Med 2003;348:2489–2490.
34 Seeman TE: Social ties and health: the benefits
of social integration. Ann Epidemiol 1996; 6:
442–451.
35 Kerlinger FN: Foundations of Behavioral Re-
search, ed 2. New York, Holt, Rinehart and
Winston, 1981, p 484.
36 Cutler J, Hendricks J: Leisure and time use
across the life course; in Binstock RH, George
LK (eds): Handbook of Aging and the Social
Sciences, ed 3. New York, Academic Press,
1990, pp 169–185.
37 Bennett DA, Wilson RS, Schneider JA, Evans
DA, Mendes de Leon CF, Arnold SE, Barnes
LL, Bienias JL: Education modifies the rela-
tion of AD pathology to level of cognitive func-
tion in older persons. Neurology 2003; 60:
1909–1915.
73