Articles
Leisure-time physical activity at midlife and the risk of
dementia and Alzheimer’s disease
Suvi Rovio, Ingemar Kåreholt, Eeva-Liisa Helkala, Matti Viitanen, Bengt Winblad,Jaakko Tuomilehto, Hilkka Soininen, Aulikki Nissinen,
Miia Kivipelto
Summary
Background Physical activity may help maintain cognitive function and decrease dementia risk, but epidemiological
findings remain controversial. The aim of our study was to investigate the association between leisure-time physical
activity at midlife and the subsequent development of dementia and Alzheimer’s disease (AD).
Methods Participants were randomly selected from the survivors of a population-based cohort previously surveyed in
1972, 1977, 1982, or 1987. 1449 persons (72.5%) age 65–79 years participated in the re-examination in 1998 (mean
follow-up, 21 years). 117 persons had dementia and 76 had AD. Multiple logistic regression methods were used to
analyse the association between leisure-time physical activity and dementia or AD.
Findings Leisure-time physical activity at midlife at least twice a week was associated with a reduced risk of dementia
and AD (odds ratio [OR] 0·48 [95% CI 0·25–0·91] and 0·38 [0·17–0·85], respectively), even after adjustments for
age, sex, education, follow-up time, locomotor disorders, APOE genotype, vascular disorders, smoking, and alcohol
drinking. The associations were more pronounced among the APOE ⑀4 carriers.
Interpretation Leisure-time physical activity at midlife is associated with a decreased risk of dementia and AD later in
life. Regular physical activity may reduce the risk or delay the onset of dementia and AD, especially among
genetically susceptible individuals.
Introduction
Recent studies have shown that a large proportion of the
population undertakes less physical activity than is
necessary to maintain good health.1 At the same time,
the proportion of old people is increasing, and agerelated diseases, such as dementia and Alzheimer’s
disease (AD), are becoming major public health
problems. Interventions that could postpone the onset of
AD even modestly would have a major effect on public
health.2
Current data, epidemiological and experimental,
suggest that physical exercise may promote brain health,
and prevent or slow cognitive decline and development
of dementia; however, results are conflicting.3–7
Prospective epidemiological studies have previously
been done in cohorts of elderly people (baseline age
⬎65 years), but had relatively short follow-up times
(3–7 years), making them prone to biases attributable to
subclinical dementia and other factors.5–12 The aim of our
study was to investigate whether leisure-time physical
activity at midlife is associated with a decreased risk of
dementia and AD later in life. We also investigated
whether sex or the APOE ⑀4 allele modify this
association.
Methods
Participants
The participants of the Cardiovascular risk factors,
Aging and Incidence of Dementia (CAIDE) study were
the survivors of four separate, independent, populationbased random samples examined within the framework
http://neurology.thelancet.com Vol 4 November 2005
of the North Karelia Project and the FINMONICA study.
These surveys assessed the cardiovascular risk factors in
two eastern provinces of Finland: North Karelia and
Kuopio. The study design has been described in detail
elsewhere.13,14
Having being investigated once at midlife (either in
1972, 1977, 1982, or 1987), 2000 randomly selected
individuals, age 65–79 years by the end of 1997, were
invited for a re-examination during 1998 (figure 1).
1449 people (72·5%) participated in the re-examination;
900 (62·1%) were women and 549 (37·9%) were men.
The mean age (SD) at midlife examination was
50·6 (6·0) years (range 39–64), and was 71·6 (4·1) years
(range 65–79) at re-examination. Mean duration of
follow-up was 21 years (SD 4·9). The study was approved
by the local ethics committee, and written informed
consent was obtained from all participants.
Lancet Neurol 2005; 4: 705–11
Published online
October 4, 2005
DOI:10.1016/S1474-4422(05)
70198-8
See Reflection and Reaction
page 690
Aging Research Center, Division
of Geriatric Epidemiology,
Neurotec, Karolinska Institutet,
Stockholm, Sweden
(S Rovio MSc, I Kåreholt PhD,
M Viitanen PhD, B Winblad PhD,
M Kivipelto PhD); Department
of Public Health and General
Practice (E L Helkala PhD),
Department of Neuroscience
and Neurology (H Soininen PhD,
M Kivipelto PhD), and
Department of Neurology,
Kuopio University Hospital,
Kuopio, Finland (H Soininen
PhD), Department of
Epidemiology and Health
Promotion, National Public
Health Institute, Helsinki,
Finland (J Tuomilehto PhD,
A Nissinen PhD), and
Department of Public Health,
University of Helsinki, Helsinki,
and South Ostrobothnia
Central Hospital, Seinäjoki,
Finland (J Tuomilehto PhD)
Correspondence to:
Miia Kivipelto, Aging Research
Center, Division of Geriatric
Epidemiology, Neurotec,
Karolinska Institutet, Box 6401,
113 82, Stockholm, Sweden
Miia.Kivipelto@neurotec.ki.se
Midlife examination
The survey methods used during the baseline (midlife)
visit were carefully standardised and complied with
international recommendations. They followed the
WHO MONICA protocols of 1982 and 1987 and were
similar to the methods used in 1972 and 1977. In brief,
the baseline survey procedures included a selfadministered questionnaire on health behaviour, health
status, and medical history. Participants’ blood
pressure, height, and weight were measured, and bodymass index (BMI) was calculated. A venous blood
sample was taken to determine serum cholesterol
concentrations.13 In addition, the presence of various
705
Articles
Random sample of 2000 CAIDE
study survivors in 1998
1449 (72·5%) participated in 1998 examination
61 (4·2%) had dementia
48 AD
13 other dementias
1388 had no
dementia
551 (27·5%) did not participate in 1998 examination
434 refused
107 refused due to poor health
7 nursing home resident
3 died
56 (10·2%) had dementia*
28 AD
9 other dementias
19 undefined dementia
117 had dementia (5·9% of total population)
76 AD
41 other dementias
495 had no
dementia
1883 had no dementia
Figure 1: Formation of the study population
Percentages of participants and non-participants in the examination in 1998 are indicated. *Dementia diagnoses
from patient records of the local hospitals and primary health-care centres.
locomotor disorders was determined. Participants were
categorised as having a locomotor disorder if at least
one of following was present: rheumatoid arthritis,
arthritis, arthropathy, or arthralgia of the joints of the
extremities or the back.
Active (n=515)
Demographics
Age at midlife (years)
50·8 (6·1)
Age at re-examination (years)
71·5 (4·0)
Follow-up time (years)
20·7 (5·0)
Education (years)
8·7 (3·6)
Men : women
228 (44·3%) : 287 (55·7%)
APOE ⑀4 carriers
171 (33·2%)
Midlife measurements
Systolic blood pressure (mm Hg)
144·0 (19·2)
Diastolic blood pressure (mm Hg)
89·3 (10·3)
2
Body-mass index (kg/m )
26·5 (3·7)
Total serum cholesterol (mmol/L)
6·7 (1·2)
History of locomotor disorders
150 (29·1%)
Re-examination measurements (late-life)
Dementia
15 (2·9%)
Alzheimer’s disease*
10/510 (2·0%)
History of diabetes mellitus
40 (7·8%)
History of stroke
32 (6·2%)
History of myocardial infarction
79 (15·3%)
Smokers
234 (45·4%)
Alcohol drinkers
380 (73·8%)
Sedentary (n=736)
p
49·5 (5·8)
70·9 (3.9)
21·3 (4·7)
8·7 (3·4)
265 (36·0%) : 471 (64·0%)
267 (36·3%)
⬍0·001
0·08
0·02
0·96
0·003
0·26
143·8 (20·3)
89·2 (11·3)
26·4 (3·6)
6·7 (1·2)
215 (29·2%)
0·42
0·93
0·75
0·98
0·97
38 (5·2%)
31/729 (4·3%)
37 (5·0%)
61 (8·3%)
98 (13·3%)
325 (44·2%)
532 (72·3%)
0·05
0·026
0·047
0·17
0·31
0·65
0·56
Data in brackets are means (SD) or percentages (%). The t test was used for means (SD) and the 2 test was used for
percentages (%). The active group comprises people who participated in leisure-time physical activity at least two times
per week at midlife. The sedentary group comprises people who participated in leisure-time physical activity less than two times
per week at midlife. *Only Alzheimer’s disease diagnoses from the re-examination were included (total sample n=1239).
Table 1: Sociodemographic and clinical characteristics of the participants according to the midlife leisure
time physical activity
706
Leisure-time physical activity was assessed on the
questionnaire as follows: “How often do you participate
in leisure-time physical activity that lasts at least
20–30 mins and causes breathlessness and sweating?”
The six response categories were as follows: (1) daily
(n=201), (2) 2–3 times a week (n=580), (3) once a week
(n=421), (4) 2–3 times a month (n=250), (5) a few times a
year (n=404), and (6) not at all (n=79). A trend test for
physical activity including these original categories was
of borderline significance (p=0·086). For the current
analyses, these categories were dichotomised as follows:
“active” people were defined as those who participated in
leisure-time physical activity at least twice a week
(n=781), and “sedentary” people were defined as those
who participated in leisure-time physical activity less
than twice a week (n=1154). The cut-off for this
dichotomisation was based on earlier1,15 and more
conservative16 recommendations for health-promoting
physical activity. Dichotomisation is also used in other
studies on physical activity and dementia, AD, cognitive
decline or memory decline.6,8,10,17 Our preliminary
analyses also showed that the dichotomisation best
distinguished participants who developed dementia
from those who did not. Data on the midlife physical
activity were available for 1935 of 2000 participants who
formed the study population. Only two of those with
missing data on physical activity had dementia, and
none of them had AD.
Re-examination
During the re-examination in 1998, the survey methods
we used were identical to those applied in the previous
surveys. Furthermore, the participants were studied for
their APOE genotypes by use of PCR and Hhal
digestion.18 Cognitive status was determined, and
participants who scored 24 or less on the mini-mental
state examination19 at the screening phase (n=294)
were referred for further examinations, including
thorough neurological and cardiovascular examinations
and a detailed neuropsychological examination.
61 participants were diagnosed as having dementia
according to the Diagnostic and Statistical Manual of
Mental Disorders (fourth edition) criteria,20 of whom 48
fulfilled the diagnostic criteria of AD according to
National Institute of Neurological and Communicative
Disorders and Stroke and the Alzheimer’s Disease and
Related Disorders Association criteria.21 The dementia
diagnoses of non-participants were derived from
patients’ records of the local hospitals and primary
healthcare centres. The total number of dementia cases
increased to 117 (5·9% of the population) when these
diagnoses were also taken into account. The analyses
concerning AD we restricted to AD diagnoses at reexamination (n=48) to ensure diagnostic accuracy, but
we repeated the analyses by including non-participants
and AD cases identified by register linkage only (n=76;
figure 1).
http://neurology.thelancet.com Vol 4 November 2005
Articles
The main analyses were restricted to participants with
no missing data on outcome, physical activity, or any of
the covariates. The total number of participants in the
analyses for dementia was 1251 (61 cases), and 1239 for
AD (48 cases). We also analysed the whole sample
(n=1935, including dementia and AD cases identified by
register linkage) for which the number of individuals
with dementia was 115 (two patients with dementia had
missing data on physical activity), and those with AD
was 76 (no missing data).
way that is comparable to an interaction term, which
makes it easy to compare the results from multiplicative
and additive interactions. The calculation of RERI takes
into account the distributional differences between risk
and odds.23 SEs for the additive interaction were
calculated with the delta method.24 If there is no additive
interaction, then RERI is equal to zero. In this case, if
RERI is greater than zero, then there is a larger
difference among the APOE carriers. The level of
significance was p⬍0·05 in all analyses. The analyses
were done by use of SPSS for Windows, release 12·0.
Statistical analyses
Differences among the participants according to their
midlife leisure-time physical activity categories (active vs
sedentary) were analysed with the 2 test and Student’s
t test as appropriate. The association between midlife
leisure-time physical activity and the subsequent
development of dementia and AD was investigated with
multiple logistic regression analyses, with the sedentary
group as the reference category. First, analyses were
adjusted for potential confounders for the relation
between physical activity and dementia: sociodemographic variables (age at re-examination, sex,
education), follow-up time, and locomotor disorders
(which may limit participation in leisure-time physical
activity; model 1). Second, we adjusted in addition for
vascular risk factors at midlife (BMI, total serum
cholesterol, systolic blood pressure), history of vascular
disorders at re-examination (myocardial infarction,
stroke, diabetes), and APOE ⑀4 carrier status (carriers vs
non-carriers; model 2). Finally, we investigated whether
other lifestyle factors modify the association between
physical activity and dementia, and made additional
adjustments for smoking status (ever smokers vs never
smokers) and alcohol drinking reported at reexamination (yes vs no; model 3).
We also did stratified analyses to assess the effect of
APOE ⑀4 carrier status on the relation between leisuretime physical activity and the risk of dementia and AD.
The putative multiplicative interaction between leisuretime physical activity and APOE ⑀4 carrier status was
then analysed by including an interaction term in the
model. The interaction between leisure-time physical
activity and sex was analysed in a similar manner. In
medical epidemiology, a corroborated view is that risk
factors that act independently have an additive effect.22
This means that when no causal interaction exists, the
total effect of risk factors is equal to the sum of the
effects of the separate risk factors. This implies that
there is an interaction as a departure from additivity,
often called additive interaction, when the total effect of
risk factors are smaller (antagonism) or larger
(synergism) than the sum of the separate effects.22
Therefore, besides multiplicative interaction, we also
calculated additive interactions by using relative excess
risk from interaction (RERI) as a measure.22 We chose
RERI because it shows the size of the interaction in a
http://neurology.thelancet.com Vol 4 November 2005
Role of the funding source
No funding source had a role in the preparation of this
article or the decision to submit it for publication. The
corresponding author had full access to all the data in
the study and final responsability for the decision to
submit for publication.
Results
Individuals who participated in leisure-time physical
activity at least twice a week at midlife (active group)
were somewhat older and had shorter follow-up than
those in the sedentary group (table 1). Other
sociodemographic and clinical characteristics did not
differ significantly between the two activity groups. The
proportions of participants with dementia and AD later
in life was lower in the active group.
The active group had lower odds of dementia later in
life compared with the sedentary group after controlling
for demographic variables, follow-up time, and
locomotor disorders (model 1; table 2). This association
remained significant after further adjustments for
midlife vascular risk factors, history of vascular
disorders at re-examination, and APOE ⑀4 carrier status
(model 2), and for smoking and alcohol drinking
(model 3). In this final model, participants in the active
group had 52% lower odds of dementia compared with
the sedentary group (table 2). The results from the
analyses of the whole sample (including those who did
not participate in the re-examination) were similar to the
results from the main analyses (results not shown).
Odds ratio (95% CI) for active vs sedentary group
Crude model
Model 1
Model 2
Model 3
Dementia (n=1251)
Alzheimer’s disease (n=1239)
0·55 (0·30–1·01)
0·45 (0·24–0·85)
0·46 (0·24–0·88)
0·47 (0·25–0·90)
0·45 (0·22–0·93)
0·34 (0·15–0·74)
0·34 (0·15–0·77)
0·35 (0·16–0·80)
Model 1 was adjusted for age at re-examination, sex, education, follow-up time, and locomotor disorders. Model 2 was
adjusted for the same variables as model 1 with the addition of APOE ⑀4 genotype, midlife body-mass index, systolic blood
pressure, cholesterol, and history of myocardial infarction, stroke, and diabetes mellitus. Model 3 was adjusted for the same
variables as model 2 with the addition of smoking status and alcohol drinking. The analyses included participants with no
missing data on the outcome, physical activity, or any of the covariates (including 61 cases with dementia and 48 with
Alzheimer’s disease).
Table 2: Association between midlife leisure-time physical activity and odds of dementia and
Alzheimer’s disease later in life
707
Articles
Odds ratio (95% CI) for active vs
sedentary group
Dementia
APOE ⑀4 carriers (n=438)
Model 1
Model 2
Model 3
APOE ⑀4 non-carriers (n=813)
Model 1
Model 2
Model 3
Alzheimer’s disease
APOE ⑀4 carriers (n=433)
Model 1
Model 2
Model 3
APOE ⑀4 non-carriers (n=806)
Model 1
Model 2
Model 3
0·39 (0·16–0·98)
0·38 (0·15–0·99)
0·41 (0·16–1·06)
0·55 (0·23–1·32)
0·59 (0·24–1·43)
0·60 (0·25–1·47)
0·23 (0·07–0·73)
0·23 (0·07–0·74)
0·24 (0·07–0·79)
0·54 (0·19–1·50)
0·59 (0·21–1·69)
0·61 (0·21–1·80)
Model 1 was adjusted for age at re-examination, sex, education, follow-up time, and
history of locomotor disorders. Model 2 was adjusted for the same variables as model 1
with the addition of midlife body mass index, systolic blood pressure, cholesterol, and
history of myocardial infarction, stroke, and diabetes mellitus. Model 3 was adjusted for
the same variables as model 2 with the addition of smoking status and alcohol drinking.
Table 3: Association between midlife leisure-time physical activity and
odds of dementia and Alzheimer’s disease by APOE ⑀4 allele status
Socioeconomic factors and depressive symptoms may
affect an individual’s ability or desire to participate in
leisure-time physical activity. Therefore, we did
additional analyses to control for midlife depressive
symptoms (measured with Beck Depression Scale),
income, and marital status; however, these covariates
did not correlate with physical activity and did not
modify the association between physical activity and AD
(results not shown). Physical activity was associated with
an inverse tendency for the odds of dementia in men
(model 3: odds ratio [OR] 0·56, 95% CI 0·22–1·43) and
women (model 3: 0·44, 0·18–1·09), although
confidence intervals were somewhat wider because of
the smaller sample size. There was no evidence for a
significant multiplicative interaction between sex and
physical activity for the odds of dementia (p=0·878).
Physical activity at midlife reduced the risk of AD in all
models (crude model and models 1–3; table 2): active
individuals had approximately 60% lower odds of AD
than sedentary individuals. Results from the analyses in
Age (years)
Education (years)
Systolic blood pressure (mm Hg)
Diastolic blood pressure (mm Hg)
Serum cholesterol (mmol/L)
Body-mass index (kg/m2)
Smokers
Participants (n=1449)
Non-participants (n=551)
p
50·4 (6·0)
8·6 (3·4)
144·5 (20·1)
89·3 (11·0)
6·7 (1·2)
26·6 (3·7)
623 (43·0%)
51·1 (6·0)
7·6 (3·0)
149·8 (21·2)
91·6 (11·1)
7·0 (1·3)
27·2 (4·6)
264 (47·9%)
0·016
⬍0·001
⬍0·001
⬍0·001
⬍0·001
0·002
0·05
Data in brackets are means (SD) or percentages (%). The t test was used for means (SD) and the 2 test was used for
percentages (%).
Table 4: Characteristics of participants and non-participants in the 1998 re-examination
708
which non-participants (diagnoses derived exclusively
from medical records) were also included were similar to
the main analyses that included only AD cases
diagnosed in re-examination (results not shown).
Physical activity seemed to have similar effect both
among men (model 3: OR 0·31, 95% CI 0·09–1·12) and
women (model 3: 0·43, 0·14–1·28), and the
multiplicative interaction term between sex and physical
activity for the odds of AD was not significant (p=0·583).
Among APOE ⑀4 carriers, the inverse association
between midlife physical activity and the subsequent
risk of dementia was significant in models 1 and 2, and
remained of borderline significance after full
adjustments (table 3). Among APOE ⑀4 non-carriers, the
association was not significant. Since the stratified
analyses indicated that the effect of physical activity
might be more pronounced among APOE ⑀4 carriers, we
did further analyses that included all possible groupings
of the physical activity and APOE ⑀4 allele status
interaction term in the logistic regression model.
Compared with sedentary APOE ⑀4 carriers, physically
active ⑀4 carriers had an OR of 0·38 (95% CI 0·15–0·97)
for dementia, sedentary ⑀4 non-carriers had an OR of
0·38 (0·19–0·77), and active ⑀4 non-carriers had an OR
of 0·23 (0·10–0·55) after adjustments. Nevertheless, the
multiplicative interaction term between the APOE ⑀4
allele and midlife leisure-time physical activity was not
significant (p=0·516) after adjustment. The additive
interaction term between physical activity and the APOE
⑀4 allele was of borderline significance (RERI=0·45,
p=0·062), which supported the findings of the stratified
analyses (ie, that APOE ⑀4 carrier status may modify the
association between physical activity and subsequent
development of dementia).
Among the APOE ⑀4 carriers, the association between
physical activity and AD was significant in all models
(table 3), whereas among APOE ⑀4 non-carriers, it was
not significant. All possible groupings of physical
activity and APOE ⑀4 allele status interaction were also
analysed in relation to AD. Compared with the sedentary
APOE ⑀4 carrier group, the active APOE ⑀4 carriers had
an OR of 0·18 (95% CI 0·05–0·67), the sedentary APOE
⑀4 non-carriers had an OR of 0·30 (0·13–0·71), and the
active APOE ⑀4 non-carriers had an OR of 0·22
(0·08–0·60). There was a weak tendency towards a
multiplicative interaction between physical activity and
the APOE ⑀4 allele status for the development of AD
(p=0·12, after adjustments), and the additive interaction
term was significant in the model (RERI=0·73,
p=0·020), which indicated that APOE ⑀4 may modify the
association between physical activity and the odds of AD.
We also investigated whether those who did not
participate in the 1998 re-examination differed from
those who were assessed (table 4). Non-participants were
significantly older, had fewer years of education, and had
higher blood pressure, BMI, and serum cholesterol
concentrations at midlife than participants at rehttp://neurology.thelancet.com Vol 4 November 2005
Articles
examination. Non-participants were also more likely to
be smokers than were participants.
Discussion
This study shows that leisure-time physical activity at
midlife is related to a decreased risk of dementia and
AD. Individuals participating at least twice a week in a
leisure-time physical activity had 50% lower odds of
dementia compared with sedentary persons. The
association was somewhat stronger for AD than for
overall dementia; those in the active group had 60%
lower odds of AD compared to those in sedentary group,
even after adjusting for a wide array of potential
confounding factors. The APOE ⑀4 allele status seemed
to modify the associations between physical activity and
dementia or AD as physical activity had more
pronounced effects against dementia or AD among the
APOE ⑀4 carriers.
Some short-term longitudinal cohort studies have
suggested that an inverse association may exist between
regular and high intensity leisure-time physical
activity,5,6,11 or some specific form of physical activity,
such as dancing10 or walking,7 and the risk of dementia
or AD, whereas others have not found any
association.12,25–28 Most studies that have investigated the
association between physical activity and dementia have
focused on other leisure-time activities and risk factors,
and various methods have been used to assess and group
physical activity. As far as we are aware, our study is one
of the first to investigate the long-term association
between midlife leisure-time physical activity and the
subsequent risk of dementia and AD. Our study
confirms the findings of a retrospective study that
suggested that patients with AD were less active in
midlife compared with non-demented individuals.29
Those findings are also in agreement with some recent
studies reporting that physical activity may be associated
with better cognitive function30 or protect against
cognitive or memory decline,17,31 even after controlling
for baseline cognitive function. A study from Japan
found no association between physical activity and
dementia during a follow-up of 20 years.28 However, that
study combined leisure-time and occupational physical
activity into a single category. Manual work, which
usually includes more physical activity and may be
associated with lower level of education, has been
associated with an increased risk of dementia.32 Thus, the
inclusion of occupational physical activity into the
physical activity definition may partly explain the nonsignificant
results
of
that
study,
but
also other factors, such as genetic, demographic, and
lifestyle characteristics (eg, different basic level of
physical activity) of the populations may account for the
different results.
Our study cohort comprised a large, representative,
and prospective population-based cohort. Midlife risk
factors are of great interest given that AD changes in the
http://neurology.thelancet.com Vol 4 November 2005
brain may start to develop decades before the
manifestation of symptomatic dementia.33 In studies
with shorter follow-up and in which the assessment of
physical activity is only done in old age, subclinical
dementia may have affected the individuals’ observed
physical activity. This kind of bias is unlikely to have
occurred in our study, in which activity assessment was
done at midlife, on average 21 years before the diagnosis
of dementia. However, the possibility of a reverse
causation and residual confounding cannot totally be
ruled out.34
There are several possible pathways through which
physical activity could protect against dementia and AD.3
First, the effect could be mediated through various
vascular risk factors (eg, hypertension, hypercholesterolaemia, diabetes, overweight) that have been
found to contribute to the development of dementia and
AD.35,36 Physical activity is important in promoting overall
and vascular health.37 In our analyses, the association
between physical activity and dementia or AD remained
significant after adjusting for various vascular risk
factors and disorders, indicating that physical activity
has an independent role. However, other vascular
mechanisms, such as subclinical atherosclerosis and
endothelial dysfunction, might be important mediators.
There may also be several neurobiological mechanisms
linking leisure-time physical activity to dementia and
AD. Recent studies have indicated that physical activity
affects several gene transcripts and neurotrophic factors
that are relevant for the maintenance of cognitive
functions,38,39 and that exercise may promote brain
plasticity.40,41 Exercise may even alleviate amyloid burden
in the brain, as suggested by a recent study in a
transgenic mouse model of AD.42 Physical activity has
also been suggested to increase cognitive reserve.43
Finally, participating in leisure-time physical activity
may be associated with other lifestyle and socioeconomic
factors44 associated with the risk of dementia. In our
study, adjustments for several sociodemographic and
vascular factors did not explain the investigated
association, but we still cannot totally exclude some
residual confounding. It would have been interesting to
have data about other socioeconomic factors, general
intellectual ability, and personality traits at midlife to be
able to further elucidate these complex associations and
the potential residual confounding and selection biases
(those with higher cognitive ability might have been
more likely to participate in leisure-time physical
activity, and this ability in general might protect against
dementia). Interestingly, the study by Richards and
coworkers,17 with data about baseline cognition, several
socioeconomic factors, and non-physical types of leisuretime activities, did not indicate that these factors would
explain the association between physical activity and
better memory function.
The effects of physical activity against dementia and
AD were more pronounced among APOE ⑀4 carriers
709
Articles
than non-carriers. A similar interaction has been
reported previously for cognitive decline.45 One
explanation for the possible effect modification of APOE
genotype may be that those individuals carrying the
APOE ⑀4 allele have less effective neural protection and
repair mechanisms, and are thus more dependent on
lifestyle-related factors to protect them against dementia
and AD.46 These findings provide an optimistic outlook
for persons with genetic susceptibility; it may be
possible to reduce the risk of dementia by adopting
positive lifestyle options. Such findings have recently
been reported from the Finnish Diabetes Prevention
Study in which individuals with high-risk genotypes for
type 2 diabetes benefited most from lifestyle interventions.47 The interplay of genes and environment in
the aetiology of AD needs to be further investigated in
other large cohort studies.
Our results may be somewhat compromised by
survival bias. It has been shown previously that physical
inactivity is associated with increased mortality.48 Thus,
if we assume that among the deceased there were more
sedentary persons and that they were also more likely to
have dementia, then our results would represent an
underestimation of the true protective effect associated
with physical activity.
Only individuals who scored 24 or less on the minimental state examination in the screening phase
underwent the exhaustive examinations needed for the
diagnosis of dementia and its subtypes. Some dementia
cases may have been lost because of this cut-off, and this
may have resulted in underestimation of the prevalence
of dementia among the participants. Healthcare-seeking
behaviour and diagnostic bias can lead to somewhat
biased results if the entire sample is included in the
analyses. However, the dementia cases detected in the
study and those ascertained from the registries were
similar in their midlife physical activity as well as in
other characteristics. If the analyses were restricted only
to participants in the re-examination, the results were
similar to the main analyses that also included nonparticipants. Furthermore, when we compared nonparticipants and participants in re-examination, we
found that non-participants were older, less educated,
and had more vascular risk factors than participants at
midlife examination. According to earlier studies,
individuals with cognitive impairment are less likely to
participate in studies of this type.49 It is also known that
medical records often underestimate dementia
diagnoses. Thus, if anything, possible non-participation
bias can be considered to have underestimated the
observed associations rather than the opposite.
A further limitation in most large studies concerning
physical activity is the reliability of physical activity data.
Even though the exact level of physical activity could not
be quantified accurately, the ranking of individuals into
different categories as done in our study is possible.
Having several follow-up measurements would have
710
allowed us to assess changes occurring in physical
activity during follow-up and its relation to dementia.
Future studies are needed to more thoroughly assess
how various types of leisure-time physical activity,
intensities, and frequencies can influence the risk of
dementia and AD.
Our results indicate that regular leisure-time physical
activity at midlife may be protective against dementia
and AD later in life. These findings may have wide
implications for preventive health care; if an individual
adopts an active lifestyle in youth and at midlife, this
may increase their probability of enjoying both
physically and cognitively vital years later in life.
Acknowledgments
The study was supported by EVO 5772720 from Kuopio University
Hospital, grant IIRG-04–1345 from Alzheimer Association, Academy of
Finland grants 103334 and 206951, the Gamla Tjänarinnor Foundation,
and the SADF (Insamligsstiftelsen för Alzheimer och
Demensforskning). We thank Prof Laura Fratiglioni for constructive
comments on the manuscript.
Authors’ contributions
SR and MK were the principal investigators. SR analysed the data and
drafted the paper. MK and IK assisted in analyses and writing. MK and
E-LH did the diagnosing of dementia. MK, E-LH, AN, JT, and HS
contributed to the conception and design of the study. JT and AN were
involved in the baseline surveys for the study. SR, IK, E-LH, MV, BW,
HS, JT, AN, and MK took part in planning the study, interpreting the
data, and commented on the article. MK is the guarantor.
Conflicts of interest
We have no conflicts of interest.
References
1
NIH Consensus Development Panel on Physical Activity and
Cardiovascular Health. Physical activity and cardiovascular health.
JAMA 1996; 276: 241–46.
2
Brookmeyer R, Grey S, Kawas C. Projections of Alzheimer’s
disease in the United States and the public health impact of
delaying disease onset. Am J Public Health 1998; 88: 1337–42.
3
Fratiglioni L, Paillard-Borg S, Winblad B. An active and socially
integrated lifestyle in late life might protect against dementia.
Lancet Neurol 2004; 3: 343–53.
4
Marx J. Preventing Alzheimer’s: a lifelong commitment? Science
2005; 309: 864–66.
5
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.
6
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: 445–53.
7
Abbott RD, White LR, Ross GW, et al. Walking and dementia in
physically capable elderly men. JAMA 2004; 292: 1447–53.
8
Yoshitake T, Kiyohara Y, Kato I, et al. Incidence and risk factors of
vascular dementia and Alzheimer’s disease in a defined elderly
Japanese Hisayama Study. Neurology 1995; 45: 1161–68.
9
Scarmeas N, Levy G, Tang MX, Manly J, Stern Y. Influence of
leisure activity on the incidence of Alzheimer’s disease. Neurology
2001; 57: 2236–42.
10 Verghese J, Lipton RB, Katz MJ, et al. Leisure activities and the risk
of dementia in the elderly. N Engl J Med 2003; 348: 2508–16.
11 Podewills LJ, Guallar E, Kuller LH, et al. Physical activity, APOE
genotype, and dementia risk: findings from the Cardiovascular
Health Cognition Study. Am J Epidemiol 2005; 161: 639–51.
12 Broe GA, Creasey H, Jorm AF, et al. Health habits and risk of
cognitive impairment and dementia in old age: a prospective study
on the effects of exercise, smoking and alcohol consumption.
Aust N Z J Public Health 1998; 22: 621–23.
13 Vartiainen E, Puska P, Jousilahti P, Korhonen HJ, Tuomilehto J,
Nissinen A. Twenty-years trends in coronary risk factors in North-
http://neurology.thelancet.com Vol 4 November 2005
Articles
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
Karelia and in other areas of Finland. Int J Epidemiol 1994;
23: 495–504.
Kivipelto M, Helkala E-L, Hänninen T, et al. Midlife vascular risk
factors and late-life mild cognitive impairment. A population-based
study. Neurology 2001; 56: 1683–89.
Pate R-R, Pratt M, Blair SN et al. Physical activity and public health.
A recommendation from the Center for Disease Control and
Prevention and the American College of Sports Medicine. JAMA
1995; 273: 402–07.
American College of Sports Medicine Position Stand. The
recommended quantity and quality of exercise for developing and
maintaining cardiorespiratory and muscular fitness, and flexibility
in healthy adults. Med Sci Sports Exerc 1998; 30: 975–91.
Richards M, Hardy R, Wadsworth MEJ. Does active leisure protect
cognition? Evidence from a national birth cohort. Soc Sci Med 2003;
56: 785–92.
Tsukamoto K, Watanabe T, Matsushima T, et al. Determination by
PCR-RFLP of ApoE genotype in a Japanese population.
J Lab Clin Med 1993; 121: 598–602.
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: 189–98.
American Psychiatric Association. Diagnostic and statistical
manual of mental disorders (4th edn). Washington, DC: American
Psychiatric Association, 1994.
McKhann G, Drachman D, Folstein M, Katzman R, Price D,
Stadlan EM. Clinical diagnosis of Alzheimer’s disease: report of the
NINCDS-ADRDA Work Group under the auspices of Department
of Health and Human Services Task Force on Alzheimer’s Disease.
Neurology 1984; 34: 939–44.
Chapter 15. In: Modern epidemiology. Rothman KJ. Boston: Little
Brown, 1986: 311–26.
http://www.socarb.su.se/kareholt/kareholt2003.pdf (accessed
Sep 30, 2005).
Assman SF, Hosmer DW, Lemeshow S, Mundt KA. Confidence
intervals for measures of interaction. Epidemiology 1996; 7: 286–90.
Wilson R-J, Bennett D-A, Bienias J-L, et al. Cognitive activity and
incident AD in a population-based sample of older persons.
Neurology 2002; 59: 1910–14.
Fabrigoule C, Letenneur L, Dartigues J-F, et al. Social and leisure
activities and risk of dementia: a prospective longitudinal study.
J Am Geriatr Soc 1995; 43: 485–90.
Wang H-X, Karp A, Winblad B, Fratiglioni L. Late-life engagement
in social and leisure activities is associated with a decreased risk of
dementia: A longitudinal study from the Kungsholmen Project.
Am J Epidemiol 2002; 155: 1081–87.
Yamada M, Kasagi F, Sasaki H, Masunari N, Mimori Y, Suzuki G.
Association between dementia and midlife risk factors: the
Radiation Effects Research Foundation Adult Health Study.
J Am Geriatr Soc 2003; 51: 410–14.
Friedland RP, Fritsch T, Smyth KA, et al. Patients with Alzheimer’s
disease have reduced activities in midlife compared with healthy
control-group members. Proc Natl Acad Sci U S A 2001;
98: 3440–45.
Weuve J, Kang JH, Manson JE, Breteler MMB, Ware JH,
Grodstein F. Physical activity, including walking, and cognitive
function in older women. JAMA 2004; 292: 1454–61.
Yaffe K, Barnes D, Nevitt M, Lui L-Y, Covinsky K. A prospective
http://neurology.thelancet.com Vol 4 November 2005
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
study on physical activity and cognition decline in elderly women.
Arch Intern Med 2001; 161: 1703–08.
Anttila T, Helkala EL, Kivipelto M, et al. Midlife income,
occupation, APOE status, and dementia: a population-based study.
Neurology 2002; 59: 887–93.
Braak E, Griffing K, Arai K, Bohl J, Bratzke H, Braak H.
Neuropathology of Alzheimer’s disease: what is new since
Alzheimer? Eur Arch Psychiatry Clin Neurosci 1999;
249 (suppl 3): 14–22.
Riley KP, Snowdon DA, Desrosiers MF, Markesbery WR. Early life
linguistic ability, late life cognitive function, and neuropathology:
findings from the Nun Study. Neurobiol Aging 2005; 26: 341–47.
Kivipelto M, Helkala EL, Laakso MP, et al. Midlife vascular risk
factors and Alzheimer’s disease in later life. Longitudinal,
population based study. BMJ 2001; 322: 1447–51.
Launer LJ. Demonstrating the case that AD is a vascular disease:
epidemiologic evidence. Ageing Res Rev 2002; 1: 61–77.
Paffenbarger RS Jr, Hyde RT, Wing AL, Lee I-M, Jung DL,
Kampert JB. The association of changes in physical-activity level
and other lifestyle characteristics with mortality among men.
N Engl J Med 1993; 328: 538–45.
Berchtold NC, Kesslak JP, Cotman CW. Hippocampal brainderived neurotrophic factor gene regulation by exercise and the
medial septum. J Neurosci Res 2002; 68: 511–21.
Tong L, Shen H, Perreau VM, Balazs R, Cotman CW. Effects of
exercise on gene-expression profile in the rat hippocampus.
Neurobiol Dis 2001; 8: 1046–56.
Cotman CW, Berchtold NC. Exercise: a behavioural intervention to
enhance brain health and plasticity. Trends Neurosci 2002;
25: 295–301.
Colcombe SJ, Kramer AF, Erickson KI, et al. Cardiovascular fitness,
cortical plasticity, and aging. Proc Natl Ac ad Sci U S A 2004;
101: 3316–21.
Lazarov O, Robinson J, Tang Y-P, et al. Environmental enrichment
reduces A levels and amyloid deposition in transgenic mice. Cell
2005; 120: 701–13.
Kramer A, Hahn S, Cohen NJ, et al. Ageing, fitness and
neurocognitive function. Nature 1999; 400: 418–19.
Laaksonen M, Prattala R, Karisto A. Patterns of unhealthy
behaviour in Finland. Eur J Public Health 2001; 3: 294–300.
Schuit AJ, Feskens EJ, Launer LJ, Kromhout D. Physical activity
and cognitive decline, the role of the apolipoprotein ⑀4 allele.
Med Sci Sports Exerc 2001; 33: 772–77.
Mahley RW, Rall SC Jr. Apolipoprotein E: far more than a lipid
transport protein. Ann Rev Genomics Hum Genet 2000; 1: 507–37.
Siitonen N, Lindström J, Eriksson J, et al. Association between a
deletion/insertion polymorphism in the ␣2B-adrenergic receptor
gene and insulin secretion and type 2 diabetes. The Finnish
Diabetes Prevention Study. Diabetologia 2004; 47: 1416–24.
Hu G, Tuomilehto J, Silventoinen K, Barengo NC,
Peltonen M, Jousilahti P. The effects of physical activity and body
mass index on cardiovascular, cancer and all-cause mortality
among 47212 middle-aged Finnish men and women.
Int J Obes Relat Metab Disord 2005; 8: 894–902.
Launer LJ, Wind WA, Deeg DJH. Nonresponse pattern and bias in
a community-based cross-sectional study of cognitive functioning
among the elderly. Am J Epidemiol 1994; 139: 803–12.
711